Tese-Phd 2004 Tese PH.D

Universidade Técnica de Lisboa
Instituto Superior de Agronomia
Francisca Constança Frutuoso de Aguiar
Vegetação ripícola em sistemas fluviais mediterrânicos.

Influência dos ecossistemas envolventes.
Doutoramento em Engenharia Florestal
ORIENTADOR
Doutora Maria Teresa Marques Ferreira da Cunha Cardoso
CO-ORIENTADOR
Doutor Ilídio Rosário dos Santos Moreira
JÚRI
Presidente Reitor da Universidade Técnica de Lisboa

Vogais
Doutor Ilídio Rosário dos Santos Moreira
Professor Catedrático aposentado do Instituto Superior de Agronomia
da Universidade Técnica de Lisboa
Doutor Max Wade
Senior Professor da RPS Ecology
Reino Unido, na qualidade de Especialista
Doutor Joe Caffrey
Senior Research Officer da Central Fisheries Board, Irlanda
Doutora Helena Maria de Oliveira Freitas
Professora Catedrática da Faculdade de Ciências e Tecnologia
da Universidade de Coimbra;
Doutor João Manuel Dias dos Santos Pereira
Professor Catedrático do Instituto Superior de Agronomia
Doutor Mário Fernandes Lousã
Professor Catedrático do Instituto Superior de Agronomia
Doutora Maria Teresa Marques Ferreira da Cunha Cardoso
Professora Associada do Instituto Superior de Agronomia
Dissertação apresentada neste Instituto para obtenção do grau de Doutor

Lisboa 2004
A meu pai e meu filho
To my father and to my son
… roots and branches of myself
ao contrário do que se julga e pôs a correr foi um

botânico o autor da célebre frase, Uma rosa é
uma rosa é uma rosa,
um poeta teria dito apenas, Uma rosa, o resto
caberia no silêncio de contemplá-la
José Saramago (1922- )
e eu teria acrescentado
e já sabem porque está aqui esta Rosa?
(e deixá-los-ia pensar...)
Indice| i
Índice
Contents
Pág.
Page
Indice . Contents i
Lista de Figuras . List of Figures ix
Lista de Quadros . List of Tables xiii
Lista de Anexos . List of Appendices xv
Agradecimentos . Acknowledgments xvii
Resumo xix
Abstract xx
Parte I Introdução 1
Part I Introduction
Capítulo 1 Enquadramento 3
Chapter 1 Thesis framework
1.1 Enquadramento 5
1.2 Thesis Framework 9
Capítulo 2 Ecologia da vegetação ripícola mediterrânica: uma síntese 13

Chapter 2 Ecology of Mediterranean riparian vegetation: a synthesis
2.1 Introdução 15
Introduction
2.2 Conceitos e terminologia 15
Concepts and terms
2.2.1 Área e objecto de estudo 16
Study area, sepcies and vegetation
2.2.2 A flora e as classificações 18

Plant classification systems
2.2.3 Perspectivas do ecossistema 22

Ecosystem perspectives
2.2.4 Invasão ecológica 26

Ecological invasions
ii | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Pág.
Page
2.3 A vegetação ripícola dos sistemas fluviais mediterrânicos: particularidades e respostas 28
ecológicas à perturbação
Riparian vegetation of mediterranean fluvial systems: particularities and ecological
response to human disturbance
2.3.1 O clima mediterrânico e o regime hidrológico 28

The mediterranean climate and the hydrological regime
2.3.2 Composição e estrutura das galerias ripícolas 29

Composition and structure of riparian galleries
2.3.3 Atributos e funções das galerias ribeirinhas e das zonas ripícolas 32

Functions and attributes of riparian zones and of riparian galleries
2.3.4 As galerias ribeirinhas e as actividades humanas 35

Riparian galleries and human activities
2.3.5 A vegetação ripícola como indicador de integridade ecológica 42

Riparian vegetation as indicator of ecological integrity
2.4 Proposta de estudo 44

Study proposal
2.5 Referências bibliográficas 46
References
Parte II As galerias ribeirinhas em sistemas fluviais mediterrânicos: 57

influência dos factores abióticos e da perturbação antropogénica
Part II Riparian woody vegetation in Mediterranean fluvial systems:
influence of abiotic factors and human disturbance
Capítulo 3 Tipologia de galerias ribeirinhas numa bacia Mediterrânica 59
Chapter 3 Riparian types in a Mediterranean basin
3.1 Abstract 61
3.2 Introduction 61
3.3 Material and methods 62
3.3.1 Study area 62
3.3.2 Survey methods 63
3.3.3 Data analysis 64
3.4 Results 64
3.4.1 Riparian composition 64
3.4.2 Site and species grouping 65
3.4.3 Environmental gradients 66
3.4.4 Reference riparian types 67
3.4.5 Riparian structure 69
Indice| iii
Pág.
Page
3.5 Discussion 72
3.6 Acknowledgements 73
3.7 References 74
Capítulo 4 Influência antrópica na composição e integridade das galerias ribeirinhas na 77

bacia do rio Tejo
Chapter 4 Human-disturbed landscapes: effects on composition and integrity of riparian
woody vegetation in the Tagus River basin, Portugal
4.1 Abstract 79
4.2 Introduction 79
4.3 Material and methods 81
4.3.1 Study area 81
4.3.2 Woody plant sampling and analysis 83
4.3.3 Integrity of riparian galleries and analysis 84
4.3.4 Combined effects of land-use and environment 86
4.4 Results 87
4.4.1 Riparian composition 87
4.4.2 Riparian integrity 89
4.4.3 Land-use and environmental features 92
4.4.4 Combined influence of environment and land-use 95
4.5 Discussion 98
4.6 Acknowledgments 100
4.7 Appendices 101
4.8 References 102
Capítulo 5 Alterações espaço-temporais das galerias ribeirinhas: influência dos factores 107
ambientais e do uso do solo
Chapter 5 Changes in riparian woods over space and time: influence of environment and
land use
5.1 Abstract 109

5.2 Introduction 109
5.3 Methods 111
5.3.1 Study area 111
5.3.2 Riparian structure, land-use and environmental data 113
5.3.3 Data treatment 115
5.4 Results 115
5.4.1 Riparian structure and temporal changes therein 115
iv | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Pág.
Page
5.4.2 Land use and temporal changes therein 120
5.4.3 Combined influence of environment and land use 122
5.5 Discussion 125
5.6 Acknowledgments 127
5.7 Appendices 128
5.8 References 129
Parte III Factores ambientais determinantes da vegetação ripícola e 133

resposta à perturbação das comunidades florísticas em bacias de
características semi-áridas
Part III Environmental drivers and response to disturbance of river plant
assemblages from semi-arid river basins
Capítulo 6 Influência dos factores abióticos e das galerias ribeirinhas nas comunidades de 135
macroinvertebrados numa bacia Ibérica
Chapter 6 Relative influence of abiotic factors and riparian features on macroinvertebrate
assemblages from an Iberian basin
6.1 Abstract 137

6.3 Study area 138
6.4 Methods 139
6.4.1 Riparian vegetation 139
6.4.2 Abiotic variables 140
6.4.3 Macroinvertebrates 141
6.4.4 Data analyses 141
6.5 Results 142
6.5.1 Riparian characteristics 142
6.5.2 Abiotic characteristics 144
6.5.3 Macroinvertebrate assemblages 145
6.5.4 Canonical analyses 145
6.6 Discussion 148
6.6.1 Environmental factors 148
6.6.2 Riparian influence 149
6.6.3 Trophic groups 150
6.8 Literature cited 151
Indice| v
Pág.
Page
Capítulo 7 Variações sazonais e anuais de macrófitos em sistemas fluviais semi-áridos 155
Chapter 7 Seasonal and yearly variations of macrophytes in a southern Iberian River

7.2 Study area 158
7.3 Methods 158
7.4 Results 159
7.5 Discussion 163
7.7 References 165
Capítulo 8 Selecção de locais de referência e avaliação da qualidade ecológica da 167

vegetação ribeirinha por métodos de análise multivariada
Chapter 8 Assessing reference sites and ecological quality of river plant assemblages from
an Iberian basin using a multivariate approach
8.1 Abstract 169

8.3 Study area 172
8.4 Materials and methods 174
8.4.1 Plant survey 174
8.4.2 Environmental and anthropogenic variables 174
8.5 Results 176
8.5.1 Plant composition 176
8.5.2 River plant ecotypes and related abiotic determinants 179
8.5.3 Reference sites and floral deviation 184
8.6 Discussion 189
8.7 References 193
Parte IV Biodiversidade da vegetação ripícola e invasão por plantas 197

exóticas em corredores fluviais Mediterrânicos
Part IV Riparian biodiversity and exotic invasions in Mediterranean fluvial
corridors
vi | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Pág.
Page
Capítulo 9 Estabelecimento da vegetação exótica e nativa num rios canalizado 199
mediterrânico
Chapter 9 Exotic and native vegetation establishment following channelization of a western
Iberian river
9.1 Abstract 201

9.3 Study area 203
9.4 Methods 205
9.4.1 Sampling procedures 205
9.5 Results 207
9.5.1 Vegetation patterns at channelized and ‘natural’ sites 207
9.5.2 Channelization effects on species richness and species cover 209
9.5.3 Environmental variables related to species distribution 211
9.5.4 Site clustering and indicator species 214
9.6 Discussion 216
9.8 Appendices 222
9.9 References 225
Capítulo 10 Padrões de distribuição de riqueza e cobertura de plantas exóticas e nativas 231

ao longo de um rio Ibérico semi-árido e do seu leito de cheia
Chapter 10 Patterns of exotic and native plant richness and cover along a semi-arid Iberian
river and across its floodplain
10.1 Abstract 233

10.3.1 Study area 234
10.3.2 Sampling procedures 236
10.4 Results 240
10.4.1 Overview 240
10.4.2 Vegetation patterns along the river basin and across the floodplain 240
10.4.3 Patterns of species richness and species cover 243
10.4.4 Predicting invasibility 249
10.5 Discussion 252
10.5.1 Spatial patterns of plant species richness and cover 252
10.5.2 Predicting invasibility in Iberian riparian floodplains 253
Indice| vii
Pág.
Page
10.7 Appendices 256
10.8 References 257
Capítulo 11 Padrões de invasão do graminhão (Paspalum distichum L.) em habitats 261

ripícolas portugueses
Chapter 11 Invasibility patterns of knotgrass (Paspalum distichum) in Portuguese riparian
habitats
11.1 Abstract 263

11.4 Results and discussion 268
11.6 References 275
Capítulo 12 Flora exótica e endémica em locais de referência e de não-referência de 277

corredores fluviais do Centro e Sul de Portugal
Chapter 12 Exotic and endemic flora on reference and non-reference sites from Iberian
fluvial systems
12.1 Aims 279

12.2 Methods 279
12.3 Results and discussion 279
12.4 Final considerations 282
12.6 Appendices 284
12.7 References 287
Parte V Conclusões e considerações finais 289

Part V Conclusions and final remarks
Capítulo 13 Discussão de resultados e conclusões 291

Chapter 13 Discussion and conclusions
13.1 Introdução 293

Introduction
13.2 Composição e integridade das galerias ribeirinhas em sistemas fluviais mediterrânicos 293
Riparian composition and integrity in mediterranean-type rivers
viii | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Pág.
Page
13.2.1 Padrões de riqueza e composição 293
Patterns of riparian richness and composition
13.2.2 Padrões de continuidade e integridade 295

Patterns of riparian fragmentation and integrity
13.2.3 Padrões de alteração espaço-temporais 299

Patterns of spatial and temporal alteration
13.3 Qualidade ecológica e potencial preditivo da vegetação ripícola em sistemas fluviais 301
mediterrânicos
Ecological quality and predictive potential of riparian vegetation in mediterranean
rivers
13.3.1 Contribuição das galerias para a sustentabilidade de comunidades biológicas em sistemas 301
mediterrânicos
Role of riparian galleries on the support of biotic assemblages in mediterranean rivers
13.3.2 Potencial preditivo da vegetação macrofítica mediterrânica 304
Predictive potential of riparian vegetation
13.3.3 Os macrófitos como bioindicadores da qualidade ecológica de sistemas fluviais 306

Macrophytes as bioindicators of ecological quality in fluvial systems
13.4 Invasibilidade, causalidade abiótica e perturbação antrópica em sistemas fluviais medi- 308
terrânicos
Invasibility, abiotic drivers and human disturbance in mediterraean fluvial systems
13.4.1 Padrões de estabelecimento da vegetação exótica e nativa num rio mediterrânico 308
canalizado
Patterns of exotic vegetation establishment in a channelized river
13.4.2 Padrões de riqueza e cobertura de espécies exóticas em sistemas fluviais de 313
Patterns of exotic richness and cover in semi-arid fluvial systems
13.4.3 Padrões de invasibilidade de ecossistemas mediterrânicos pelo graminhão (Paspalum 317
distichum L.)
Patterns of invasibility of Mediterranean ecosystems by the knotgrass (Paspalum
distichum L.)
13.4.4 Padrões do estabelecimento de espécies exóticas e endémicas em locais de referência e 319
não-referência
Patterns of exotic and endemic vegetation establishment in reference and non-
reference sites
References
Capítulo 14 Considerações finais 331

Chapter 14 Final remarks
14.1 Nota final 333

Final comments
14.2 Outros trabalhos: a continuação e o desafio 335

Further studies: the continuity and the challenge

References
Lista de Figuras| ix
Lista de Figuras
List of Figures
Fig. Pág.
Fig. Page
Parte I
2.1 Modelo conceptual a 4 dimensões do ecossistema lótico. Conceptual model of the four-di- 23
mensional nature of lotic ecosystems.
2.2 Organização hierárquica dos ecossistemas fluviais. Hierarchical view of the fluvial ecosys- 25
tem spatial structure.
2.3 Tipologia transversal de um sistema fluvial e definição dos limites laterais do troço de 26
amostragem. Schematic transversal diagram of the fluvial system, showing the lateral
dimension of the floristic surveys.
2.4 Sistema fluvial do Sul de Portugal e vegetação ripícola característica de climas semi-áridos 30
dominada por tamujo (Rio Guadiana, Estrela 1998). Fluvial system from a semi-arid lands-
cape of south Portugal, with the dominance of the dyer's buckthorn [Flueggea tinctoria
(L.) G.L. Webster] (River Guadiana, Estrela 1998).
2.5 Galeria ribeirinha dominada por amieiros num sistema fluvial mediterrânico do Centro de 31
Portugal (Rio Nabão, Agroal 2004). Riparian galleries dominated by alders [Alnus glutinosa
Mill.] on a fluvial corridor of the Central Portugal (River Nabão, Agroal 2004).
2.6 Etapas da invasão segundo With (2002) e influência da estrutura dos ecossistemas fluviais e 42
envolventes no processo de invasão. Stages of invasion according to With (2002) and
effects of the fluvial ecosystems and of the floodplain structure on the process of in-
vasive spread.
2.7 Enquadramento e síntese dos principais pontos de partida e discussão e da proposta de 45

estudo. Framework and synthesis of the main discussion points and of the key research
questions.
Part II
3.1 Axes I and II of the canonical correspondence analysis, showing the spatial distribution of 65
sampling sites superimposed by the groups obtained from weighted pair-group average
clustering using Bray-Curtis similarity coefficient.
3.2 Axes I and Axis II of the canonical correspondence analysis, showing the spatial distribution 67
of the woody species and the environmental variables.
3.3 Axes I and II of the canonical correspondence analysis of reference sites showing the spatial 69
distribution of the main woody species and of the environmental variables.
3.4 Axes I, II and II of the canonical correspondence analysis showing the spatial distribution of 70
reference sites superimposed by the groups obtained from the weighted pair-group
average clustering method using Bray-Curtis similarity coefficient.
4.1 Iberian Peninsula, showing the Tagus basin and the location of the eight studied basins. 82
4.2 Correspondence Analysis ordination of riparian species, showing the biplot scores of the 88
survey units and the species with the highest scores on the axes (Figure 4.2a), and
Principal Components Analysis ordination of vegetational variables, showing the plot
scores of the survey units (Figure 4.2b).
4.3 Spatial variation of riparian width for the studied corridors. 91

x | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Fig. Pág.
Fig. Page
Parte II
4.4 Global percentage values of the land-use types in the studied basins. 92
4.5 Principal components analysis (PCA) ordination of environmental variables (Figure 4.5a) 94
and of land-use variables (Figure 4.5b), showing the plot scores of the survey units.
4.6 Canonical correspondence analysis (CCA) biplots for the selected environmental (a) and 96
land-use variables (c), showing the field survey scores. Redundancy analysis (RDA) biplots
for the selected environmental (b) and land-use variables (d), showing the survey unit
scores.
5.1 Iberian Peninsula, showing the Tagus basin and the location of the studied basins. 112
5.2 Spatial variation of canopy cover classes (% of river length) on Rivers Nabão and Trancão 118
on different flyover dates.
5.3 Spatial variation of canopy cover classes (% of river length) on River Canha on different 119
flyover dates.
5.4 Changes in land-use classes in the studied basins (given as %). Percentage values were 120
averaged from all transepts.
5.5 Redundancy Analysis (RDA) biplot showing the survey unit scores for the Rivers Nabão and 123
Trancão on the studied flyover dates, and the forward selected variables. The ordination
diagram of riparian structure variables is displayed in the upper right corner.
5.6 Redundancy Analysis (RDA) biplot for the selected variables showing the survey unit scores 124
for the River Canha in 1951 and 1995 and the forward selected variables
Part III
6.1 Iberian Peninsula, showing the Sado basin and the location of the sampling sites. 140
6.2 Canonical correspondence analysis ordination of macroinvertebrate and riparian variables 146
showing the species with the highest scores in the axes and the biplot scores of riparian
variables.
6.3 Canonical correspondence analysis ordination of macroinvertebrate and abiotic variables, 147
showing the species with the highest scores in the axes, and the biplot scores (arrows) of
abiotic variables.
7.1 Position of the surveys from the main river and tributaries, and from all dates, on axes 1 162
and 2 of the correspondence analysis.
8.1 Iberian Peninsula with the indication of the Guadiana basin. Portuguese part of the 173
Guadiana basin with location of the sampling points.
8.2 Hierarchical grouping of all the studied sites from the unweighted pair-group average 181
clustering method using Bray-Curtis similarity coefficient.
8.3 Canonical correspondence analysis biplot for the entire biological data set, considering the 182
forward selected environmental variables.
8.4 Canonical correspondence analysis biplot for the entire biological data set, considering the 183
variables related to human disturbance and the use of the drainage basin.
8.5a Location of the sites from plant ecotype I (northern siliceous) on axis one and two of the 185
canonical correspondence analysis and position and magnitude of the combined variable
anthropogenic disturbance.
Lista de Figuras| xi
Fig. Pág.
Fig. Page
Parte IV
8.5b Hierarchical grouping of ecotype I (northern siliceous) sites from the unweighted pair- 185
group average clustering method using Bray-Curtis similarity coefficient.
8.6a Location of the sites from plant ecotype II (calcareous) sites on axis one and two of the 186
canonical correspondence analysis and position and magnitude of the combined variable
anthropogenic disturbance.
8.6b Hierarchical grouping of sites of plant ecotype II (calcareous) sites from the unweighted 186
pair-group average clustering method using Bray-Curtis similarity coefficient.
8.7a Location of the sites from plant ecotype III (southern siliceous) sites on axis one and two 187
of the canonical correspondence analysis and position and magnitude of the combined
variable anthropogenic disturbance.
8.7b Hierarchical grouping of sites from plant ecotype III (southern siliceous) sites from the 187
unweighted pair-group average clustering method using Bray-Curtis similarity
coefficient.
8.8 Linear regression of Bray-Curtis dissimilarity against the anthropogenic disturbance. 188
Part IV
9.1 The Mondego river basin, showing the floodplain and the location of the nine studied sites. 204
9.2 Axis one and two of the correspondence analysis using the floral composition of the 208
transects from the nine channelized and ‘natural’ sites studied on the River Mondego
9.3 Mean and mean standard error using pooled variance of dominant exotic species cover and 211
total exotic species cover on the plots from the nine studied sites.
9.4 Axis one and two of the canonical correspondence analysis biplot using the floral 212
composition of the 5 plots of each transect from the six sites studied on the channelized
Mondego segment.
9.5 Location of the species on axis one and two of the canonical correspondence analysis biplot 214
using the floral composition of the 5 plots of each transect from the six sites studied on
the channelized Mondego segment.
9.6 Hierarchical grouping of sites from the unweighted pair-group average clustering method 215
using Bray-Curtis similarity coefficient and selected indicator species (maximum
IndVal>45%, p< 0.01) for each level of the typology.
10.1 Iberian Peninsula with indication of the Guadiana basin. Portuguese part of the Guadiana 235
basin with the location of the sampling sites.
10.2 Hierarchical grouping of sites per valley location from the complete clustering method 243
using Bray-Curtis similarity coefficient and selected indicator species (maximum IndVal ≥
50%, P≤ 0.01) for each level of the typology.
10.3 Comparison of species richness and species cover for exotic and native species at the 245
three valley locations.
10.4 Medians, 25/75th quartiles, minimum and maximum values of native and exotic species 246
richness and species cover at the three valley locations from the six studied sites.
xii | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Fig. Pág.
Fig. Page
Parte IV
10.5 Relationship of native species richness to exotic species richness, and of native species 250
cover to exotic species cover for each valley location.
10.6 Kriging map of exotic species cover and fine substrate at the Juromenha sampling site. 255
11.1 Median, 25/75th quartiles, minimum and maximum values for knotgrass cover in the 265
studied river basins. The location of the Iberian Peninsula, Portugal and the studied
basins are represented on the right side of the figure.
11.2 Comparison of the frequency of occurrence (FO) of knotgrass and of giant reed per basin 271
and at all sites.
12.1 Comparison of endemic and exotic species richness and cover for reference and non- 281
reference site-groups.
Lista de Quadros e Anexos| xiii
Lista de Quadros
List of Tables
Pág.
Quadro
Table Page
Parte I
2.1 Alguns conceitos em estudos de invasão e respectivos factores determinantes. Concepts used 27
in biological invasion’ studies and related determinant factors.
2.2 Atributos ecológicos da zona ripícola e respectivas funções e atributos específicos das 33
galerias ribeirinhas. Ecological attributes of riparian areas, related characteristics and
functions of the riparian galleries.
2.3 Actividades humanas e principais efeitos físicos e consequências potenciais para a vegetação 37
ripícola. Human activities, main physical effects and potential consequences on the
riparian vegetation.
Parte II
3.1 Summary statistics for the canonical correspondence analysis including correlation of 68
canonical axes with the environmental variables.
3.2 Ratio of occurrence of main woody species per cluster and total frequency for the basin. 71
4.1 River length (km), extent of river length occupied by dams (%) and drainage area (km2). 84
Minimum and maximum values for the altitude (m), valley slope (‰), fluvial corridor
width (m), Strahler’s order number, annual average rainfall (mm), annual average
temperature (ºC) and annual average humidity (%) of the studied segments.
4.2 Number of survey units. Minimum and maximum values, average ± SD (in parentheses) for 89
the riparian woody cover classes, class changeover and riparian width of the studied
streams.
4.3 Selected variables and results of the partitioning of the riparian floristic and riparian cover 97
variance.
5.1 Characterization of the river basins under study: maximum river length (km), drainage area 113
(km2), maximum Strahler’s order number, dominant river orientation, minimum and
maximum values for altitude (m), fluvial corridor width (m), valley slope (‰), annual
average temperature (ºC), annual average rainfall (mm) and annual average humidity (%).
5.2 Flyover dates and scale of aerial photographs of the river segments expressed as distance 114
from source.
5.3 Minimum and maximum values (mean ± SD in parentheses) for riparian canopy closure 117
classes, canopy class changeover and average riparian width for the observed flyover
dates.
5.4 Minimum and maximum values (mean ± SD in parentheses) for the land-use types for the 121
observed flyover dates.
Parte III
6.1 Mean, SD, and minimum (min) and maximum (max) values for selected environmental 143
variables and macroinvertebrate assemblages at studied sites.
6.2 Summary statistics for the canonical correspondence analysis (CCA) relating 146
macroinvertebrate assemblages to abiotic and riparian variables.
xiv | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Quadro Pág.
Table Page
Part III
7.1 Summary of descriptive floristic parameters for the three surveys, and for the main river 160
and tributaries, including the total number of river species, the average number of
species per site, the species occurring in less than 20% of the sites (number and %), the
species with abundance 1 (number and %) and the total number of terrestrial species
found.
7.2 Summary statistics of the correspondence analysis performed, including the number of 160
active species, the sum of all unconstrained eigenvalues, the eigenvalues of the first two
axes and the cumulative percentage variance explained by species data, in each date and
for all dates.
7.3 Main species contributors to the first and second axis of the correspondence analysis 161
performed for the three dates.
8.1 Species with frequency of occurrence (FO) in the three ecotypes higher than in the 177
Guadiana basin, and their scores on the first and second axes of the global CCA.
8.2 List of species with the highest scores on the first axis of each partial CCA performed and 179
their frequency of occurrence (FO) on the ecotype and on the selected group of reference
sites.
Parte IV
9.1 Comparison of native and exotic species richness and cover in the transects from natural 209
and channelized sites.
10.1 Altitude (m), distance to source (km), annual average rainfall (mm), annual average 237
temperature (ºC), average minimum (min.) and average maximum (max.) annual
temperatures (ºC), valley floor width (m), wetted channel width (m), and dominant
geological background of the sampling sites. Average proportion of particle-size classes of
superficial substrate (%) and valley location limits (m) for each sampling site.
10.2 Origin, lifespan, usual habitat, Raunkjaer life form and frequency of occurrence of the 242
exotic species per site at each valley location.
10.3 Results of ANOVA of native and exotic species richness and cover (means ± SD) at the 244
valley locations of the sampling sites.
10.4 Frequency of occurrence (%) and percentage cover (%) of selected dominant native and 248
exotic plants at the valley locations of the sampling sites.
10.5 Results of forward stepwise multiple regressions of exotic species richness and cover, and 251
habitat characteristics as predictors of exotic species richness and cover at the three
valley locations.
11.1 Characterization of regional, environmental and anthropogenic variables, and respective 266
scale level.
11.2 Origin, Raunkiaer life form and life span of the exotic species recorded on the sampling 269
sites.
11.3 Total frequency of occurrence (FO) and frequency of occurrence of the exotic species with 270
cover above 30% (FO>30% ) at least in one sampling site.
11.4 Summary statistics for the correlations relating the presence and cover of knotgrass to 273
regional, environmental and anthropogenic variables.
12.1 Summary statistics for the correlations relating the richness and cover of exotic and 282
endemic species to regional, environmental and anthropogenic variables. * = p < 0.05;
**= p < 0.01; ns= non-significative, p=0,05.
Lista de Quadros e Anexos| xv
Lista de Anexos
List of Appendices
Anexo Pág.
Appendix Page
Parte II
4.1 Acronyms and source of data of vegetational variables, land-use variables and environ- 101
mental variables.
5.1 Variables considered for vegetation and land use, source of the data and acronyms used in 128
the ordination diagrams and tables.
Parte IV
9.1 Indicator value of the species in the sample groups of the last level of the typology and the 222
abundance/presence pairs for each species on the overall groups.
10.1 Raunkjaer life form, lifespan, propagation mode and dispersal strategies, invasive status, 256
and ecology of the dominant species.
12.1 Type of endemism, usual habitat and life span of the endemic species recorded in the 284
sampling sites.
12.2 Origin, Raunkjaer life form and life span of the exotic species recorded in the sampling 285
sites.
xvi | Vegetação Ripícola de Sistemas Fluviais Mediterrânicos
Agradecimentos
É com satisfação que aproveito esta oportunidade para agradecer a entidades e pessoas que
contribuíram para a realização trabalho, e em particular:
À Professora Teresa Ferreira, orientadora desta dissertação, pelo estreito acompanha-

mento, amizade e disponibilidade sem reservas, demonstrado em todas as fases da
dissertação e pelas ideias e sugestões transmitidas com contagiante entusiasmo. Quero
ainda expressar a minha gratidão por me ter facultado todo o seu saber nesta área e pela
sua orientação criativa e regida sob elevados padrões de método e rigor científico.
Ao Professor Ilídio Moreira, co-orientador desta dissertação, por facultar experiência e

ensinamentos importantes, além de pertinentes sugestões que muito contribuíram para o
enriquecimento deste trabalho. Agradeço, ainda a pormenorizada revisão dos textos de
introdução e conclusões.
Ao Salvador e ao meu filho, pela sã convivência com os meus papéis e “problemas

científicos”, e por retemperarem as minhas forças com um simples olhar ou sorriso. Esta
dissertação também é vossa.
Uma palavra especial à minha mãe, que sempre acreditou e viveu este projecto e que muito
me ajudou a levar, como diz, esta “carta a Garcia” pelo melhor caminho.
Ao meu irmão, Ricardo Aguiar, pela preciosa colaboração dada no processamento e

formatação do manuscripto no contra-relógio desta fase final desta dissertação, e pelas
ajudas no tratamento de dados e actualização do software, interesse e incentivo que
demonstrou ao longo destes seis anos.
Á minha cunhada, Rute Mendes, agradeço todo o interesse demonstrado e em especial o

tratamento e a selecção de imagens e a concepção da capa e separadores de capítulos.
Ao Engenheiro António Albuquerque, pela sua amizade e pelo acompanhamento nas

inesquecíveis saídas de campo, bem como nas frutuosas e empolgadas discussões sobre os
resultados e tratamentos efectuados.
A todos os que pertencem ao Water Lobby, equipa de investigação sob coordenação da

Professora Teresa Ferreira, que acompanharam de perto este trabalho, em especial ao João
Oliveira, José Maria Santos, Luis Lopes, Pedro Vieira, Patricia González, Francisco Godinho,
Paulo Pinheiro e Ivan Bernez pela amizade, pelas críticas, sugestões e demais preciosas
ajudas.
xviii | Agradecimentos
Aos revisores anónimos e editores de revistas internacionais dos artigos publicados no

âmbito desta dissertação, cujas sugestões e críticas em muito enriqueceram este trabalho.
Agradeço em particular, as revisões e sugestões do Dr. Robert C. Bailey, David M.
Rosenberg, Dr. Kevin Murphy, Professor Rui Cortes, Dr. Hiram Li e Dr. Bruce McIntosh.
Ao Professor Paulo Pinto, pela interessante trabalho de cooperação no estudo da influência

das galerias ribeirinhas nas comunidades de macroinvertebrados.
À Carla Nogueira, pela colaboração dada na análise espaço-temporal das galerias ribeirinhas
e pela carinhosa preocupação pela minha “sanidade mental” nas fases mais difíceis.
À Doutora Cristina Duarte, do Instituto de Ciências Tropicais, pelo esclarecimento de

dúvidas sobre ecologia de plantas tropicais e pesquisa de informações na área da Botânica.
Ao Instituto Superior de Agronomia, e em especial ao Departamento de Engenharia Florestal

que proporcionaram bens materiais e um local de trabalho agradável e acolhedor, e
incentivaram a execução deste trabalho. Agradeço ainda, ao Departamento de Protecção
das Plantas e Fitoecologia, e aos Professores João Amaral Franco, Lousã, José Carlos Costa,
Ana Monteiro, Investigadora Dalila Espírito-Santo e Engenheira Teresa Vasconcelos, de quem
sempre recebi pronta ajuda na identificação de plantas e em esclarecimentos vários na área
da Botânica.
À Fundação para a Ciência e a Tecnologia pela ajuda financeira dada pela Bolsa de
Doutoramento (PRAXIS XXI/BD/11173/97), sem a qual seria difícil a realização desta
dissertação. Agradeço ainda, a possibilidade de participação em vários projectos, em
especial no Projecto Praxis POCTI/MGS/42584/2001, que proporcionaram os meios
financeiros para a concretização do plano de estudos proposto e o contacto enriquecedor
com a comunidade científica internacional.
Ao João Luís, Cláudia, Nuno e Rogério pela amizade e interesse, e pela ajuda na resolução
dos inúmeros problemas informáticos que foram surgindo. Ao Salvador, por me ter ajudado
a elaborar as extensas matrizes de dados, a organizar as fotografias aéreas, a ditar os
“estranhíssimos” nomes de plantas, e as muitas ajudas nas revisões dos manuscriptos. À
Anabela Moura e filhas, pela amizade e revisão de textos em inglês e pela ajuda na pesquisa
de citações.
Finalmente, uma palavra de gratidão a todos os amigos e familiares, que de algum modo
contribuíram para a realização deste trabalho.
Resumo
Estudaram-se as relações entre a perturbação antropogénica dos corredores fluviais

mediterrânicos e os padrões de distribuição espaço-temporais da vegetação ripícola. Em
particular, foram analisados os factores que contribuem para a alteração da integridade
florística das galerias ribeirinhas. Os padrões de invasibilidade por plantas exóticas foram
também analisados. Constatou-se uma baixa integridade e riqueza florística, e uma
fragmentação generalizada das galerias ribeirinhas que revelaram também uma relativa
homogeneidade florística, descontinuada apenas por assimetrias climáticas e
hidrogeomórficas. Os corredores fluviais revelaram uma baixa riqueza em flora exótica, mas
uma cobertura elevada, quer em corredores fluviais muito intervencionados, quer em
ecossistemas ripícolas seminaturais. Confirmou-se o potencial preditivo da vegetação
ripícola na avaliação de alterações exógenas e a contribuição das galerias para a qualidade
biológica dos ecossistemas fluviais. Os resultados obtidos têm aplicações práticas,
designadamente nas formas de recuperação de galerias ribeirinhas, na monitorização e
bioavaliação do estado ecológico dos sistemas fluviais e na concepção de modelos de gestão
de ecossistemas lóticos.
Palavras-chave: galeria ribeirinha, perturbação humana, integridade ecológica, Portugal,

invasibilidade, espécie exótica.
Riparian Vegetation of Iberian fluvial systems. The influence of surrounding
ecosystems.
Abstract
The spatial and temporal patterns of the riparian vegetation were studied together with the
human disturbances in the fluvial corridors and in the surrounding landscape. In particular,
the influences of the environmental setting and of the land-use in the river floodplain were
related to the floristic and ecological integrity of riparian galleries.
This study revealed low species richness, low riparian width, and a widespread
fragmentation of the riparian galleries, along with an upstream-downstream variability due
both to the environmental gradient, and to the human-induced disturbances in the fluvial
corridor and in the river valley. The relatively floristic homogeneity of riparian galleries
was only interrupted by large climatic and geomorphological differences. Besides the low
richness in exotic flora, the fluvial corridors presented a high invasibility both in highly
intervened systems and on semi-natural riparian ecosystems. The role of the riparian
vegetation on the sustainability of fluvial ecosystems and its predictive potential in relation
to exogenous disturbances was confirmed.
The results obtained supplied valuable information and scientific support for the restoration
and management of riparian corridors, and for the monitoring and bioassessment of
ecological quality of fluvial systems, as well as for the development of models of invasion in
mediterranean fluvial systems.
Key words: riparian gallery, human disturbance, ecological integrity, Portugal, invasibility,
exotic species.
Parte I Introdução
Part I Introduction
For the cause that lacks

assistance,
The wrong that needs resistance,
For the future in the distance,
And the good that I can do.
George L. Banks (1821-1881)

Capítulo 1
Enquadramento
Chapter 1 Thesis framework

4 | Parte I - Introdução
Capítulo 1 - Enquadramento | 5
1.1 Enquadramento
Esta dissertação inserida na área da ecologia aquática, pretende clarificar a relação entre
as perturbações de origem antrópica nos sistemas fluviais mediterrânicos e a distribuição,
estrutura, integridade e composição florística das galerias ribeirinhas, numa perspectiva de
gestão e sustentabilidade dos recursos hídricos e dos corredores fluviais. As relações causais
e funcionais entre o meio abiótico e a vegetação são abordadas numa óptica limnológica,
i.e. de forma holística e atendendo às dimensões temporais e espaciais do ecossistema e ao
carácter dinâmico e evolutivo da vegetação.
A dissertação é estruturada em cinco partes. Uma primeira parte introdutória, redigida em

Língua Portuguesa, visa demonstrar o enquadramento da dissertação no estado-de-arte
actual e aborda a importância dos trabalhos desenvolvidos para a evolução dos
conhecimentos em sistemas fluviais ibéricos. Apresenta uma caracterização dos sistemas
fluviais mediterrânicos, com ênfase para as suas particularidades a nível florístico,
estrutural, geomorfológico e climático. Os termos técnicos e científicos anglo-saxónicos em
ecologia lótica, flora e vegetação ripícola e em invasão ecológica são revistos, propondo-se
uma correspondente terminologia portuguesa. No final desta parte é apresentada a
proposta de estudo desta dissertação.
O cerne da dissertação (partes II a IV) desenvolve-se em capítulos interdependentes,

apresentados na forma de artigo científico e redigidos em Língua inglesa. Algumas variações
na estrutura destes capítulos devem-se a obrigatoriedade editorial das revistas em causa.
Estes artigos foram objecto de publicação internacional, ou encontram-se submetidos ou
aceites para publicação.
O conteúdo dos artigos e respectivos quadros, figuras e demais anexos não foi modificado,
apenas submetido a formatação para uniformização da apresentação global da dissertação.
No início de cada capítulo apresenta-se em rodapé, a referência de publicação do trabalho
original, e acompanhando o capítulo, colocou-se no cabeçalho o respectivo ‘título corrido’
publicado ou submetido. Permitiu-se apenas alterações a nível ortográfico, de modo a
corrigir pequenos erros identificados após a publicação. Esta fidelidade aos trabalhos
publicados e a publicar implica ligeiras diferenças de nomenclatura taxonómica, ou uso de
vocábulos diferentes consoante se trata de revistas americanas ou europeias, além de
alguma sobreposição de certos conteúdos, como por exemplo na caracterização da área de
estudo.
Na Parte V, redigida em Língua Portuguesa, articulam-se as conclusões parciais dos

trabalhos realizados, fundamentadas pela discussão de resultados e são abordadas as suas
implicações práticas na gestão de sistemas fluviais mediterrânicos. Apresenta-se uma
análise crítica das metodologias de amostragem e de tratamento de dados utilizados e
propõem-se estudos futuros.
Esta modalidade de dissertação, amplamente utilizada a nível internacional, permite uma

maior facilidade de divulgação do trabalho científico realizado, dada a tendência crescente
para a utilização da Língua Inglesa na comunidade científica internacional. Por outro lado,
a estruturação por artigos científicos publicados, permite um acréscimo de qualidade,
dadas as sugestões de revisores especialistas nos temas versados e as subsequentes
alterações ao trabalho inicial. Acresce, ainda, uma maior flexibilização na leitura,
permitida pelo uso de capítulos encadeados, todavia auto-suficientes.
A segunda parte desta dissertação proporciona uma caracterização da vegetação rípicola

arbórea e arbustiva de bacias mediterrânicas, em termos de riqueza, composição, estrutura
e distribuição, com base em amostragens florísticas (capítulos 3 e 4), e em detecção
remota (capítulos 4 e 5) e fundamentada em métodos de análise multivariada (ordenação e
classificação). Perspectiva a contribuição relativa dos factores abióticos e antrópicos na
integridade das galerias ribeirinhas e na composição florística. O primeiro capítulo desta
parte procura definir tipos florísticos e a respectiva afiliação geográfica numa bacia
mediterrânica. Apresenta a distribuição das espécies ripícolas arbóreas e arbustivas e a
estrutura da vegetação ripícola de referência. Discutem-se os efeitos das interferências
antrópicas nos sistemas fluviais mediterrânicos e a importância da selecção de locais de
referência. Destacam-se os factores ambientais (geomorfológicos, topográficos e climáticos)
que contribuem para a distribuição e composição florística ripícola. Os capítulos 4 e 5
estudam a influência relativa da ocupação do solo no vale de cheia na composição e
integridade das formações ripícolas, tendo como base a interpretação estereoscópica de
fotografia aérea e a inventariação florística em sub-bacias do rio Tejo. No capítulo 4,
apresenta-se o estudo da composição e cobertura florística face à heterogeneidade
geográfica, ambiental e de ocupação do solo das várias bacias, e caracteriza a
complexidade intrabacia em termos de integridade longitudinal e de uso dos ecossistemas
envolventes dos corredores fluviais. O capítulo 5 aborda a evolução temporal das formações
ripícolas em sub-bacias com o uso do solo predominantemente urbano (bacias dos rios
Nabão e Trancão) e agro-florestal (bacia da ribeira de Canha), e avalia a importância
relativa da ocupação do solo e dos factores ambientais na fragmentação/integridade das
formações ripícolas.
A terceira parte centra-se no estudo dos padrões de distribuição de grupos florísticos e

respectiva causalidade abiótica, e potencial preditivo da vegetação face a perturbações
antropogénicas e a alterações ambientais. Neste contexto, o capítulo 6 perspectiva o papel
da vegetação ripícola em sistemas com fortes intervenções antrópicas e com regime
intermitente de caudais. Questiona em particular a relação da integridade longitudinal e
composição do coberto ripícola e as comunidades de macroinvertebrados bentónicos, bem
como a influência relativa dos factores abióticos do ecossistema. O capítulo 7 estuda as
variações florísticas derivadas de flutuações climáticas (sobretudo hidrológicas) interanuais
e sazonais, em termos de riqueza e composição florística, pretendendo explorar a
capacidade preditiva da vegetação macrofítica mediterrânica. O capítulo 8 avalia a
resposta de comunidades florísticas à perturbação humana, procurando definir ecótipos
fluviais e condições ecológicas de referência. Estes trabalhos têm como base amostragens
de carácter qualitativo e limnológico em bacias de características mediterrânicas marcadas,
nomeadamente nas bacias dos rios Sado e Guadiana.
A quarta parte visa a análise da expressão quantitativa e qualitativa de plantas exóticas e

os efeitos da sua presença e abundância no decréscimo da biodiversidade ripícola. Pretende
clarificar a relação entre as perturbações de origem antrópica e as componentes riqueza e
cobertura em plantas exóticas nas dimensões longitudinal e transversal dos corredores
fluviais. A importância dos factores e circunstâncias ambientais que conferem uma maior ou
menor invasibilidade dos corredores lóticos é também avaliada. Os efeitos do projecto de
canalização do Baixo Mondego na alteração na composição florística e nos padrões de
distribuição das comunidades vegetais dos corredores ripícolas são apresentados no capítulo
9, como contributo para o conhecimento dos padrões de invasibilidade de ecossistemas
ripícolas portugueses com elevada perturbação humana. Complementarmente, um segundo
trabalho realizado ao longo do curso principal de um sistema fluvial regularizado mas sem
canalização - rio Guadiana (capítulo 10) – visa relacionar a cobertura e riqueza em espécies
exóticas, bem como a exploração do potencial preditivo dos factores ambientais e do uso
da terra no vale de cheia para a presença e invasão por espécies exóticas. O contributo dos
atributos biológicos específicos para o sucesso das invasões foi igualmente objecto de
estudo. Os últimos dois capítulos desta parte, são o resultado de uma abordagem regional
em bacias hidrográficas do Centro e Sul do país. O primeiro (capítulo 11) centra-se no
estudo da invasibilidade de uma das espécies exóticas mais frequentes e abundantes em
corredores ripícolas – Paspalum distichum L. - e sua relação com os factores ambientais e
de perturbação humana. Finalmente, o capítulo 12, respeitante a um artigo em fase de
preparação, tem um enfoque de índole conservacionista, e pretende investigar a relação
causal entre as características habitacionais e a riqueza e cobertura de espécies endémicas
e espécies exóticas em locais de referência e em sistemas muito perturbados.
A presente dissertação compreende um conjunto de trabalhos sobre composição,

integridade e dinâmica da vegetação em ecossistemas aquáticos e ribeirinhos, cujo fio
condutor se centra na perturbação humana como factor modelador dos ecossistemas fluviais
mediterrânicos. São utilizadas abordagens metodológicas distintas, quer a nível de
delineamento experimental, quer em relação ao tratamento de dados, bem como a
utilização de dados qualitativos e quantitativos directos (inventariação) e indirectos
(fotografia aérea) para análise da vegetação. O sistema fluvial é perspectivado nas
dimensões longitudinal, lateral, vertical e temporal, tendo sido escolhidas bacias e locais
de amostragem com padrões geomorfológicos e climáticos variados e com um gradiente de
perturbação humana. Foram ainda abordadas várias escalas espaciais e temporais do
sistemas fluvial. Os trabalhos que compõem esta tese apoiaram-se nos resultados e
conclusões dos estudos entretanto finalizados, com benefício do plano inicial proposto. Por
outro lado, houve a possibilidade de explorar algumas questões que emergiram de trabalhos
precedentes, como é evidente, por exemplo, nos capítulos da segunda parte desta
dissertação.
Espera-se que este estudo contribua para uma gestão sustentada dos sistemas fluviais
portugueses, nomeadamente como base científica de apoio à recuperação, conservação e
intervenção em corredores ripícolas.
1.2 Thesis framework
This thesis intends to clarify the effects of anthropogenic disturbances on the distribution,
structure, composition and integrity of riparian species assemblages, in the context of the
sustainability and management of Mediterranean streamside forests. A limnological
approach was used throughout this study to explore the causal and functional relationships
between the environment and the plant species assemblages.
The thesis is organized into five parts. A first part, written in Portuguese language, aims to
provide a context for the global research work by exposing the state-of-art and the relevant
studies that have been done in the Mediterranean fluvial systems of the Iberian Peninsula
and in other climate-type areas. As some concepts on freshwater ecology in English-
language literature are not clear and are sometimes used erroneously, we discuss the most
important concepts and its possible applications and synonyms, as well as a proposal for the
correspondent Portuguese terms. It also presents a general characterization of the
Mediterranean fluvial systems, and its particularities of geomorphology, climate and
historical features. The study proposal of the dissertation and the key research questions
are also presented in this part.
The core of the thesis (Part II to Part IV) comprises interdependent chapters written in
English language. All the chapters exhibit the standard sequence of a scientific article and
were published, were submitted or are waiting for publication in international journals.
Some variations on the structure between chapters were due to editorial obligations of the
journals where the paper was submitted or published.
The articles were organized into chapters with similar format and references were
homogenised, but none of the text, figures, tables or appendices were modified. Only
orthographic errors were corrected. At the beginning of each chapter, the title of the
published or submitted work is given at the footer, and the header corresponds to the
running head of the published or submitted article. This fidelity involves a certain
incoherence concerning, for instance, the utilization of some terms that are in American
Language or in English Language, depending on the journals, and also on the taxonomic
nomenclature. It also comprises an overlap of information, for example on the description
of the study areas.
The second part presents a characterization of riparian woody vegetation on Mediterranean

fluvial systems, concerning the richness, composition, structure and distribution of species
assemblages, and focuses on the relative contribution of environment and anthropogenic

disturbance as determinants to the integrity and composition of riparian vegetation. The
floristic data were achieved both by phyto-ecological field surveys and by interpretation of
aerial photographs. Univariate methods and indirect ordination techniques were used to
describe the overall spatial patterns, and multivariate analyses (ordination and
classification) were applied to explore the contribution of environmental and land-use
variables on the distribution patterns of species assemblages.
The third part approaches the role of riparian vegetation and the use of river plant
assemblages as indicators of the ecological integrity of Mediterranean fluvial systems and
on the definition of ecotypes. In this context, the functions of riparian woody assemblages
and their contribution to the dynamics of the ecosystem were explored, as well as the floral
variations with inter and intra-annual hydrological fluctuations in a typical semi-arid fluvial
system.
Finally, the fourth part intends to clarify the effects of human disturbance on the
biodiversity of riparian habitats and on the ecosystem susceptibility to exotic plant
invasions. The qualitative and quantitative expression of exotic plant species was related
with environmental and anthropogenic variables, both at the landscape and at the habitat
level. The exotic and native vegetation establishment was also studied along the river
gradient (mouth to source), and across the floodplain (instream-upland) in channelized
corridors and on semi-natural ones. Reference and non-reference conditions were used to
compare richness and cover of species groups (exotic, native and endemic species) between
humanized landscapes and near-natural habitats.
Part V, written in Portuguese language, gathers the main conclusions and considers the
implications of this work for future research and for practical application on the
management of riparian and aquatic habitats. A critical analysis of the sampling methods
and data treatment are presented, as well as a proposal for future works in this area.
Human disturbance is considered all through the present study as a key factor shaping the
ecosystem structure and the habitational features, and promoting changes on the
vegetation patterns. The environmental factors affecting the integrity and structure of
riparian corridors and the human impacts were studied together. For these purposes, river
basins (and sampling locations) with diverse environmental setting, and various human
alterations were studied, considering the fluvial system on its four-dimensions: longitudinal,
lateral, vertical and temporal. In addition, a broad range of spatial scales was considered.
Several methodological approaches were used on data treatment, and the sampling designs
were set according to the research aims of each study. Vegetation data was both collected
by direct (field surveys) and indirect methods (remote sensing). Some questions arose from
the initial works of this thesis, and thus they provided an enriching basis to develop the
following ones.
A desirable target of this work is to supply valuable information and a scientific support for
the conservation, restoration and management of riparian corridors.
Capítulo 2
Ecologia da vegetação ripícola
mediterrânica: uma síntese
Chapter 2 Ecology of Mediterranean riparian

vegetation: a synthesis
Capítulo 2 - Ecologia da Vegetação Ripícola Mediterrânica: Uma Síntese | 15
2.1 Introdução
O meio aquático e ribeirinho suporta os mais diversos e complexos ecossistemas existentes

(Naiman et al. 1993; Jungwirth et al. 2002). Esta diversidade ecológica deriva essencial-
mente da variabilidade geomorfológica e dos regimes hidrológicos, dos gradientes de
altitude e de humidade, e da influência da bacia de drenagem nos corredor fluvial (Naiman
e Décamps 1997). Estas características são particularmente evidentes em sistemas fluviais
mediterrânicos, como os do Centro e Sul da Península Ibérica, onde os corredores ripários
são habitats de excepção (“oásis lineares”) com especificidades próprias distintas das
vegetação xerófita característica dos ecossistemas envolventes (González-Bernáldez et al.
1989).
O reconhecimento das particularidades do meio fluvial e da complexa teia de relações

bióticas e abióticas, levaram ao desenvolvimento de ramos individualizados da ecologia
aquática, como a limnologia e a ecologia lótica, cujos fundamentos são relativamente
recentes. Por outro lado, o cariz de dinamismo e de permanente evolução em face da
multiplicidade de usos dos recursos associados ao sistema fluvial, impulsionou os estudos
em corredores fluviais, nomeadamente na avaliação da influência de actividades humanas e
na fundamentação de técnicas de gestão ecológica e de medidas de conservação e
recuperação em sistemas fluviais, bem como no ordenamento do território numa óptica
integrada de ecologia da paisagem. Nas últimas três décadas tem sido devotada especial
atenção aos processos ecológicos e biogeoquímicos dos ecótonos fluviais, ou seja, das zonas
de interface entre os sistemas ecológicos terrestre e aquático de características espaço-
temporais únicas, cujo dinamismo resulta das interrelações estabelecidas entre ambos os
sistemas (Holland 1988). Este conceito contacta com a noção de zona ripária e de ecótono
fluvial (Gregory et al. 1991; Naiman e Décamps 1997), que compreende a vegetação ripícola
e as galerias ribeirinhas, elementos estruturantes dos ecossistemas lóticos mediterrânicos e
ponto focal deste trabalho.
2.2 Conceitos e terminologia
Em trabalhos científicos redigidos em Língua portuguesa, alguns termos têm sido

erroneamente utilizados como sinónimos, ou em sentidos diversos, ou ainda transpostos
inadequadamente dos termos anglo-saxónicos. Nesta parte introdutória, julga-se vantajoso
apresentar algumas designações utilizadas neste trabalho, bem como as definições que se
tomam como preferíveis, como contributo na clarificação da terminologia limnológica em

Língua portuguesa. Sempre que necessário são apresentados os termos anglo-saxónicos
correspondentes mais utilizados (entre parêntesis rectos ou aspas).
2.2.1 Área e objecto de estudo
Um dos principais aspectos prende-se com a utilização dos termos ripícola e ripário
[riparian]. A raiz etimológica de ambos os vocábulos é latina; o primeiro deriva de ripa
(ribeira) com o sufixo -cola, do latim colere, habitar, ou seja aquele(a) que estabelece o
seu habitat nos cursos de água superficiais, enquanto que ripário deriva de riparius (da
margem) (Torrinha 1939; Silva 1948; Font Quer 1982). Estes termos podem ser assimilados
como sinónimos, adjectivando por exemplo espécie, flora, vegetação, ecossistema,
ecótono. De notar, que é frequente, a nível internacional, o uso de ‘riparian’ em relação a
lagos e outros ecossistemas lênticos, exceptuando albufeiras e outras massas de água
artificiais. O vocábulo ribeirinho foi utilizado por Vasconcellos (1970) na obra pioneira
“Plantas (Angiospérmicas), Aquáticas, Anfíbias e Ribeirinhas”, e mantido recentemente em
publicações no âmbito de projectos de investigação (Aguiar et al. 1999), e no contexto do
Plano Nacional da Água (Moreira et al. 2002a), podendo sem prejuízo substituir ‘ripícola’ e
‘ripário’.
As expressões galeria ribeirinha, galeria ripícola, floresta ribeirinha, bosque ribeirinho,

mata ripária, mata ripícola têm sido utilizadas para designar as comunidades vegetais
dominadas por espécies arbóreas e arbustivas associadas ao meio lótico. O termo mata (do
latim matta), encontra-se definido em vários dicionários e enciclopédias [e.g. Dicionário de
Botânica de Font Quer (1982)], como porção de terreno ocupada por um número elevado de
árvores da mesma espécie. No entanto, um grande número de publicações utilizam-no no
sentido de floresta plantada ou de floresta de produção. Pensa-se que a utilização deste
termo deve ser evitada para designar as formações lenhosas de margens de rios, tal como
floresta, que está conotada com uma grande área coberta de árvores, perdendo-se a noção
da importância da sua natureza linear.
As expressões bosque ribeirinho, bosque ripícola, bosque higrófilo são frequentemente

empregues em estudos fitossociológicos para nomear associações ou comunidades com
dominância de espécies ripícolas lenhosas, como em Costa et al. (1998), Alves et al. (1998),
Aguiar et al. (1999), Moreira et al. (2002a), Marcos et al. (2002). Esta terminologia deve
reservar-se para denominar formações ribeirinhas de composição mais ou menos
homogénea, com uma extensão e densidade considerável. No entanto, na maioria das
situações em regiões de clima mediterrânico, e em particular em sistemas fluviais

portugueses, as expressões julgadas mais adequadas são as de galeria ribeirinha ou galeria
ripícola. Nesta expressão está implícita a disposição preferencialmente longitudinal (em
faixa) ao longo de um curso de água (Font Quer 1982), e permite incluir formações
dominadas por estratos arbóreos e arbustivos mais ou menos fragmentadas. As expressões
anglo-saxónicas ‘riparian forests’, ‘streamside forests’, ‘riverine forests’, ‘riparian
woodlands’ são utilizadas indiferenciadamente para nomear quer as formações de regiões
temperadas húmidas, quer as mediterrânicas ou de climas áridos ou semi-áridos. A
expressão ‘floodplain forests’ aplica-se à vegetação ribeirinha de rios de planície, em vales
abertos. ‘Riparian galleries’, tradução exacta da expressão proposta para sistemas fluviais
mediterrânicos, é pouco utilizada, enquanto que ‘gallery forest’ (floresta-galeria) deve ser
reservada para nomear florestas ribeirinhas de regiões equatoriais, com elevada densidade
e extensão lateral (Font Quer 1982, Moreira et al. 2002a). A expressão vegetação ripícola
[riparian vegetation] é mais abrangente, podendo ser aplicada em relação à vegetação
ribeirinha herbácea ou a galerias ribeirinhas menos estruturadas. Em estudos de ecologia
fluvial de sistemas mediterrânicos é frequentemente utilizada em detrimento de ‘riparian
galleries’, e.g. Stromberg (1993), Ferreira (1994), Gasith e Resh (1999), Salinas et al.
(2000). As expressões ‘riparian woody vegetation’, ‘riparian woods’, ‘riparian woodlands’
aplicam-se também com frequência em relação ao estrato arbóreo e arbustivo da vegetação
ribeirinha.
Sobre o conceito de corredor ripário ou corredor ripícola [riparian corridor], Malanson

(1993) refere que os corredores ripários são faixas abrangendo a área entre as cotas
máximas e mínimas do curso de água, englobando ainda o sistema terrestre desde a cota
superior do meio aquático ao extremo da encosta onde a vegetação pode ser influenciada
por cheias ou por condições hidrológicas, tais como variações decorrentes das alterações do
nível freático associadas com os escorrimentos superficiais. Esta definição é geralmente
aceite na maioria das publicações (Ferreira e Cortes 1998). Neste contexto, a largura do
corredor ripário [riparian width] indica a extensão lateral da galeria ribeirinha. A expressão
corredor fluvial [fluvial corridor] é por vezes aplicada como sinónimo (e.g. Saraiva 1999).
Paralelamente ao conceito de corredor, o conceito de ecótono fluvial [fluvial ecotone;

aquatic-terrestrial ecotone] (Naiman e Décamps 1990), ou de zona ripária [riparian zone]
(Gregory et al. 1991; Naiman e Décamps 1997) é indispensável em estudos de vegetação
ripícola. Indicia entidades funcionais de interface entre o meio aquático e terrestre, de
carácter dinâmico, cujas características e interacções dependem dos ecossistemas
adjacentes, definidas no espaço e no tempo (Holland 1988).
O termo aquático em limnologia é muitas vezes aplicado, por uma questão de simplificação,
no sentido restritivo de dulçaquícola [freshwater], ou seja “de água doce”, relativo a águas
superficiais epicontinentais ou interiores. Interessa ainda notar que por plantas aquáticas
entendem-se as que têm o seu ciclo de vida ligado à água ou a ambientes encharcados ou
húmidos, desde que não sejam frequentes noutros locais (Dawson 1988). Adopta-se assim o
sentido abrangente de plantas aquáticas e ribeirinhas [fluvial plants, river plants],
enquanto que planta ribeirinha [riverine plant] apenas inclui as plantas das margens de
cursos de água.
O termo macrófitos [macrophytes] (do grego macro - grande e de phyton - planta) é

frequentemente utilizado em limnologia para designar as formas macroscópicas de plantas
aquáticas, incluindo macroalgas, líquens, briófitos, pteridófitos e plantas superiores (Denny
1985). Opõe-se a micrófitos, que inclui as plantas microscópicas que constituem o
fitoplâncton e o perifiton que não são objecto do presente estudo. Embora em rigor,
macrófito se refira à generalidade das plantas macroscópicas, é utilizado quase
exclusivamente para as plantas aquáticas, razão pela qual não se aplica o termo macrófito
aquático (Catarino et al. 2001).
2.2.2 A flora e as classificações
Numerosas classificações são adoptadas para distinguir grupos de macrófitos, como por
exemplo segundo o ciclo de vida, habitat usual e tipo fisionómico. Esta classificação das
plantas em grupos visa relacionar características ou atributos biológicos com as condições
de perturbação antrópica (Nilsson et al. 1989; Nilsson et al. 1994), ou ainda com os factores
ambientais, perturbações naturais, nomeadamente no regime hidrológico e com fenómenos
de competição interespecífica (Tabacchi et al. 1996). Uma extensa revisão e crítica sobre
sistemas de classificação de macrófitos pode ser encontrada em Ferreira (1995-1996), pelo
que se abordam unicamente as utilizadas no presente trabalho e as mais frequentes em
estudos limnológicos em Portugal.
A classificação segundo as formas biológicas elementares [life span], muito utilizada a nível
internacional, reduz o universo de plantas recenseadas a duas categorias principais: anuais
e bienais [annuals; biennuals] e vivazes e perenes [perennials].
Outra classificação utilizada neste trabalho segue as designações vulgares: plantas

herbáceas [herbaceous plants], plantas anuais ou vivazes de consistência herbácea ou sub-
herbácea, e plantas lenhosas arbóreas e arbustivas [woody flora; woody plants]
(Vasconcellos 1969). Neste último grupo distinguem-se as árvores [trees], plantas lenhosas
perenes de altura superior a 5 metros na maturidade, com caule distinto e livre de ramos
na parte inferior, arbustos [shrubs], plantas lenhosas perenes de altura geralmente igual ou
inferior a 5 metros na maturidade, e caule geralmente ramificado desde a base, se bem que
não obrigatoriamente, e lianas [woody climbers], plantas trepadoras sarmentosas, com
caule por vezes muito alongado (Fabião 1996). Estas classificações permitem-nos, pela sua
simplicidade, obter um padrão geral da distribuição das plantas.
Para caracterização das formas biológicas existem ainda, outras classificações aceites em
trabalhos da especialidade como a de Raunkjaer (1934), também designada por
classificação fisionómica. Surgiu da necessidade de exprimir melhor a variabilidade de
formas de vida existentes. Define os seguintes tipos biológicos, baseados no número,
posição e protecção das gemas de renovo:
• terófitos [therophytes], ervas propagadas por semente, e cuja vida dura menos de
um ano, tendo um único ciclo vegetativo;
• criptófitos [cryptophytes], ervas vivazes, cujas gemas de renovo se situam abaixo

da superfície do solo (geófitos [geophytes]), abaixo do nível de água (hidrófitos
[hydrophytes]) ou indiferentemente abaixo do nível do solo ou da água (helófitos
[helophytes]);
• hemicriptófitos [hemicryptophytes], plantas vivazes ou bienais com gemas de

renovo situadas ao nível do solo;
• caméfitos [chamaephytes], plantas vivazes com gemas de renovo situadas a menos

de 25 cm do solo;
• fanerófitos [phanerophytes], plantas perenes com gemas de renovo situadas a mais

de 25 cm do solo. Estes últimos são subdivididos em nano-, micro-, meso- e
macrofanerófitos, consoante são providos de gemas de renovo, respectivamente
compreendidas entre 0,25 e 2 metros, 2 e 8 metros, 8 e 30 metros, superiores a 30
metros, a ainda em fanerófitos escandentes [woody climbers; vines; woody lianas],
que correspondem às lianas.
Este sistema mais vulgarizado em ecologia terrestre comporta algumas dificuldades na

adaptação à especificidade ecológica das plantas lóticas (Ferreira 1995-1996), excluindo por
exemplo da denominação de hidrófito as plantas aquáticas anuais, além de ser frequente
uma planta apresentar diferentes formas de vida consoante a situação.
Um dos sistemas de classificação de macrófitos mais frequentemente utilizado é o de

Sculthorpe (1971), seguida em vários trabalhos redigidos em Língua portuguesa. Considera
quatro grandes grupos de macrófitos, a saber:
• emergentes, enraizados sob a superfície da água ou nas margens, com a maior parte
das folhas e orgãos reprodutivos aéreos. Constituem um grupo tolerante à
submersão por períodos mais ou menos longos;
• enraizados com folhas flutuantes, enraizados em solos submersos ou ancoradas ao

substrato, com a maior parte das folhas flutuando à superfície da água. São
característicos de águas pouco profundas;
• totalmente flutuantes, flutuantes à superfície ou submersos, por vezes ancorados

em obstáculos;
• submersos, enraizados, ancorados, com os tecidos vegetativos submersos e orgãos

reprodutores submersos, flutuantes ou aéreos. Inclui as macroalgas.
Vulgarmente, os macrófitos emergentes tomam a denominação de helófitos [helophyte]

(Bartley e Spence 1987), e os macrófitos submersos e macrófitos totalmente flutuantes a
designação de euhidrófitos [eu-hydrophyte] (Dawson 1988). O conjunto dos últimos três
grupos constituem os hidrófitos sensu lato [hydrophyte] (Wetzel 1983). O conceito de
higrófito [hygrophyte] é referenciado como plantas associadas ao meio aquático, ou seja as
que frequentemente se encontram nas proximidades dos ecossistemas aquáticos. Não é
utilizado geralmente pela comunidade científica internacional, dada a inadequação a
condições de regime hidrológico permanente, preferindo-se a distinção entre espécies
ripícolas, aquáticas e terrestres, de carácter mais objectivo.
Os higrófitos e helófitos correspondem à designação vernácula de Vasconcellos (1970) -

anfíbias e ribeirinhas - e representam no seu conjunto as espécies com estruturas
radiculares tolerando ou necessitando de solo húmido, encharcado ou submerso,
eventualmente com o caule total ou parcialmente submerso, tolerando por períodos
variáveis e de forma variável situações de secura e de submersão (Ferreira 1995-1996).
Duarte e Moreira (2002) distinguem ainda entre higrófitos e sub-higrófitos. Estes últimos
ocorrendo em zonas com menores teores de humidade edáfica que os higrófitos,
correspondendo maioritariamente às que têm como habitat preferencial locais de grande
humidade mas onde as inundações são esporádicas ou inexistentes, nomeadamente prados
higrofílicos, bosques ou matos húmidos, zonas sombrias e leitos de cheias.
Os atributos comunitários, medidas indirectas das relações das plantas com o meio, podem
ser usados para distinguir grupos de plantas. Neste contexto, destacam-se os termos
nitrófilo [nithrophilous], que determinam a preferência por solos ricos em azoto, acidófilo
[acidophilous] por solos ácidos, em geral siliciosos, e psamófilo [psamophilous] por solos
arenosos. Na distinção pelo habitat preferencial [location] em sistemas fluviais
mediterrânicos consideram-se geralmente três grupos: plantas de habitat terrestre
[terrestrial], ribeirinho [riparian] e aquático [aquatic], no entanto em rios com regime
permanente de caudais, as duas primeiras categorias são analisadas em oposição às
espécies aquáticas.
A distinção das plantas segundo a sua naturalidade [naturalness] (Nilsson et al. 1989; 1994),
ou seja em espécies ruderais [ruderals] e não ruderais [non-ruderals; naturals] é
controversa e suscita algumas dúvidas quando da colocação das espécies num ou noutro
grupo, uma vez que depende da definição de habitat ruderal e das fontes de informação
utilizadas, recorrendo-se geralmente a pequenos apontamentos nas Floras da região.
Entenda-se como espécie ruderal aquela que é frequente em habitats artificializados, como
sejam bermas de caminhos, taludes de vias de comunicação, imediações de lixeiras e
entulheiras, campos abandonados sujeitos a interferência humana, zonas de acumulação de
águas ricas em nutrientes ou matéria orgânica (Alves et al. 1998). Alguns trabalhos de
padrões de distribuição espacial de vegetação lótica (e.g. Nilsson et al. 1989) incluem as
espécies presentes em terras cultivadas (que corresponde às plantas geralmente
denominadas de adventícias), e reconhecem que os ecossistemas ribeirinhos possam ter sido
o habitat original de algumas destas espécies. Esta classificação é utilizada em vários
trabalhos de referência, como Nilsson et al. (1989; 1994), Tabacchi et al. (1996), e é
geralmente aceite como um bom indicador de perturbação antrópica.
Por último, as definições de planta exótica, naturalizada e invasora necessitam de

clarificação, bem como alguns conceitos associados a estudos de invasão ecológica.
Richardson et al. (2000), após uma apurada revisão bibliográfica e discussão da
terminologia mais adequada recomendou os seguintes termos:
• plantas exóticas [alien plants, exotic plants, non-native plants, non-indigenous
plants] - plantas cuja presença foi devida a introdução acidental ou intencional, em
resultado de actividades humanas;
• plantas naturalizadas [naturalized plants] - plantas exóticas que se reproduzem de

forma consistente e que suportam populações ao longo de vários ciclos sem directa
intervenção humana, produzem gerações livremente, e não invadem
necessariamente ecossistemas naturais, seminaturais ou artificiais;
• plantas invasoras [invasive plants] - plantas naturalizadas que produzem populações

férteis, geralmente com elevado número de indivíduos, e a consideráveis distâncias
dos progenitores, e possuem uma potencial elevado de expansão numa área

considerável de um dado ecossistema;
• infestantes ambientais ou infestantes ecológicas [environmental weeds] - plantas

exóticas que invadem o habitat da vegetação espontânea nativa, afectando
negativamente a biodiversidade nativa e/ou o funcionamento do ecossistema;
• transformadoras [transformers] - grupo de plantas invasoras, que modificam as

características, condições forma e natureza dos ecossistemas numa área de
dimensão considerável em relação à área total desse ecossistema. Exemplos desta
categoria são as plantas que utilizam de forma excessiva determinados recursos,
como a luz e a água (e.g. Arundo donax L.), luz e oxigénio [Eichhornia crassipes
(Mart.) Solms-Laub.] ou que respondem facilmente a perturbações como o fogo (e.g.
Acacia spp.), ou a condições adversas, como elevada salinidade (e.g. Tamarix spp.).
Outras classificações complementares e de uso frequente reportam-se à data de

introdução. Por exemplo, na Europa Central os termos arqueófitos e neófitos referem-se a
taxa introduzidos antes ou depois de 1492, respectivamente (Mandák e Pyšek 1998).
2.2.3 Perspectivas do ecossistema
Os conceitos e perspectivas adoptadas em ecologia lótica têm sofrido uma evolução

significativa nos últimos anos, principalmente no que respeita aos processos e dinâmica dos
sistemas fluviais. Ward (1989) agregou as perspectivas mais focalizadas nas dimensões
individuais do ecossistema como o ‘Conceito do Contínuo Lótico’ [River Continuum Concept]
postulado por Vanote et al. (1980), a ‘Hipótese da Perturbação Intermédia’ [Intermediate
Disturbance Hypothesis] (Connell 1978), a dimensão vertical (Hynes 1983) ou transversal do
sistema (Newbold et al. 1981). Este modelo conceptual holístico do sistema lótico baseia-se
no reconhecimento da heterogeneidade e dinamismo espaço-temporal dos corredores
fluviais nas dimensões longitudinal (montante - jusante), vertical (meio superficial - meio
hiporreico) e lateral (meio aquático / ripário - meio terrestre). A utilização desta
perspectiva tem sido consensual e seguida ou adaptada por vários autores; por exemplo
Gregory et al. (1991) reportam-se à ‘perspectiva do ecossistema‘, Petts e Amoros (1996)
referem a ‘perspectiva do hidrossistema fluvial’ e Pinay et al. (1990) adaptaram esta
concepção aos ecótonos ripícolas (Figura 2.1).
LATERAL LONGITUDINAL
m
ei me
o io
nt t e
-
a q te
s a an
e
uá rre
-ju ont
ti c s t
m
o/ re
rip
ár
io
evolução/alteração natural
DIMENSÃO TEMPORAL
perturbação
meio superficial-
-meio hiporrheico
VERTICAL
Figura 2.1 Modelo conceptual a 4 dimensões do ecossistema lótico (modificado a partir de Ward 1989; Gregory
et al. 1991; Petts e Amoros 1996).
Figure 2.1 Conceptual model of the four-dimensional nature of lotic ecosystems (modified from Ward 1989;
Gregory et al. 1991; Petts and Amoros 1996).
Mais recentemente e em ecologia de rios regularizados tem sido aplicado o ‘conceito da

descontinuidade serial’ [Serial Discontinuity Concept] originalmente desenvolvido por Ward
e Stanford (1983) e revisto por Ward e Stanford (1995) e Stanford e Ward (2001). Este
modelo teórico pretende prever a resposta biofísica dos ecossistemas fluviais às
descontinuidades provocadas por aproveitamentos hidraúlicos (e.g. represamentos,
albufeiras) e outras formas de regularização. São modelos previsivos teóricos que integram
(nas versões mais recentes) a noção de dinamismo e evolução temporal segundo a
perspectiva tridimensional do ecossistema lótico.
As várias perspectivas do sistema fluvial e as interligações espaço-temporais entre

conceitos distintos como o ‘conceito de river continuum’, o ‘conceito de descontinuidade
serial’, o ‘conceito de corredor hiporrheico’ [Hyporheic Corridor Concept] (Stanford e Ward
1993) e o ‘conceito de pulso de cheia‘ [Flood Pulse Concept] (Junk et al. 1989; Tockner et
al. 2000) são discutidas e brilhantemente ilustradas em Poole (2002). Este autor considera a
perspectiva do ‘mosaico dinâmico de habitats’ [Habitat Patch Dynamic], descrito em Wu e
Loucks (1995), como geradora de um enquadramento conceptual robusto e flexível em
ecologia fluvial, uma vez que considera os ecossistemas lóticos como um discontinuum
único de componentes hierárquicos imbricados e interactivos. Esta abordagem permite
integrar os vários conceitos existentes em ecologia lótica e pode ser visualizada como uma
hierarquia de elementos organizados num mosaico estrutural, conectada por processos
físicos e biológicos trans-espaciais, suportando a ideia de singularidade no descontínuo
habitacional lótico.
Estas concepções do ecossistema lótico serviram de base à estruturação dos trabalhos

realizados e à discussão fundamentada dos resultados. Reconhece-se que associada à
organização hierárquica do sistema, a escala espaço-temporal de análise é determinante na
investigação em ecologia lótica. Deste modo, o sistema fluvial é perspectivado como um
entrosamento de níveis espaciais de organização dinâmicos e de crescente dimensão: o
micro-habitat e o habitat, o troço, o segmento e a bacia de drenagem, cujas características
são apresentadas em Frissel et al. (1986). No presente estudo, cada capítulo abrangeu
várias escalas de análise, de acordo com os objectivos propostos e respectivos
delineamento e metodologia (Figura 2.2).
Neste estudo, os inventários fitoecológicos são realizados no leito menor ou leito aparente,
ou seja no canal e na zona afectada pelas cheias anuais, que compreende geralmente a
totalidade da galeria ripícola e da vegetação associada ao meio lótico. A Figura 2.3
apresenta esquematicamente a tipologia transversal do sistema lótico e a largura do troço
de amostragem adoptada neste estudo. O canal de estiagem corresponde à massa de água
associada ao leito durante a Primavera e princípio de Verão, enquanto que o leito maior ou
leito de cheia inclui a área submersa por cheias de maior periodicidade, muito actuada pela
agricultura e possuindo elencos florísticos de feição terrestre (Ferreira 1992). Por
imperativos metodológicos, no capítulo 10, o perfil transversal do troço foi subdividido em
três zonas. Adoptou-se para sistemas com perfil transversal pouco alterado, a terminologia
proposta por Harris (1999) para zonas ripícolas de regiões semi-áridas: 1) canal de
estiagem, ou seja o meio aquático e a zona marginal geralmente inundada [inundated and
frequently flooded banks]; 2) parte do leito menor, correspondente à zona marginal
inundada pelas cheias anuais [infrequently flooded valley floors] e 3) leito maior
correspondente à zona do vale não inundada ou raramente inundada [unflooded or rarely
flooded higher floodplains].
Capítulo 5: Altera- Bacia hidrográfica (106-105) Capítulo 3: Tipologia

ções espaço-tempo- das galerias ribeirinhas
rais das galerias ri- numa bacia mediterrânica
beirinhas: influência
dos factores ambien-
tais e do uso do solo Capítulo 7: Variações
sazonais e anuais de
Capítulo 4: Influência Segmento fluvial (104-103 anos) macrófitos em sistemas
antrópica na compo- fluviais semi-áridos
sição e integridade das
galerias ribeirinhas na
bacia do RioTejo
Capítulo 8: Selecção
de locais de referência
e avaliação da qualidade
ecológica da vegetação
ribeirinha por métodos
de análise multivariada
Troço (102-101 anos)
Capítulo 11: Padrões

de invasão do graminhão
(Paspalum distichum L.)
em habitats ripícolas
portugueses
Capítulo 12: Flora exó-

Habitat (101-100 anos) tica e endémica em
Capítulo 6: Influência
locais de referência e
dos factores abióticos
de não-referência de
e das galerias ribeiri-
corredores fluviais do
nhas nas comunidades
Centro e Sul de
de macroinvertebrados
Portugal
numa bacia Ibérica
Capítulo 10: Padrões de

distribuição da riqueza e
cobertura de plantas exó-
ticas e nativas ao longo
de um rio Ibérico semi- Micro-habitat (100-101 anos)
-árido e do seu leito de Capítulo 9: Estabeleci-
cheia mento da vegetação
exótica e nativa num rio
canalizado mediterrânico
Figura 2.2 Organização hierárquica dos ecossistemas fluviais (adaptada de Frissell et al. 1986), com indicação
das escalas de análise adoptadas em cada um dos casos de estudo do presente trabalho. As barras a preto
indicam a extensão das dimensões espaciais de cada trabalho, e a barra tracejada indica interrupção na escala
de análise.
Figure 2.2 Hierarchical view of the fluvial ecosystem spatial structure (adapted from Frissell et al. 1986).
The extension of black columns refers to the spatial dimension of each work that compose this thesis, and the
doted column notes an interruption on the spatial scale.
O capítulo 9 assenta num estudo comparativo de segmentos do rio Mondego canalizado e

não-canalizado; optou-se igualmente por amostragens tri-facetadas no 1) meio aquático
(corresponde ao canal de estiagem) e nos 1º e 2º terraços, correspondendo ao talude
interior [inner-bank] e talude exterior [outer-bank] respectivamente, no rio ‘seminatural’.
Leito Leito
maior Canal de estiagem maior
Leito menor (= largura do troço de amostragem)
Figura 2.3 Tipologia transversal de um sistema fluvial e definição dos limites laterais do troço de amostragem.
A tracejado está indicado o nível usual das cheias anuais, no período de Outono-Inverno (ilustração adaptada
de Aguiar et al.1999).
Figure 2.3 Schematic transversal diagram of the fluvial system, showing the lateral dimension of the floristic
surveys. The dashed line represents the standard annual flood level. This figure derived from illustrations
presented on Aguiar et al. (1999).
2.2.4 Invasão ecológica
Uma parte importante deste trabalho é dedicada ao estudo da distribuição, frequência e

abundância relativa de espécies exóticas em sistemas fluviais e zonas ripícolas, referindo-se
brevemente alguns conceitos relativos ao ecossistema e utilizados em estudos ecológicos de
invasão. Neste contexto e socorrendo-nos de Richardson et al. (2000), o termo introdução
[introduction] significa que a planta (ou o seu propágulo) foi transportada através de
barreiras geográficas importantes, por exemplo de um continente a outro. A naturalização
[naturalization] inicia-se quando as fronteiras bióticas e abióticas à sobrevivência das
plantas são ultrapassadas, e as barreiras à reprodução regular são igualmente superadas. A
invasão [invasion] implica que as plantas introduzidas se reproduzam e constituam
populações em áreas distantes do local de introdução. Este termo pode ou não ter
implicações negativas de natureza económica ou ambiental. Geralmente considera-se o
processo de invasão dividido nas seguintes etapas: 1) introdução (acidental ou propositada),

2) colonização [colonization], 3) estabelecimento (sobrevivência e reprodução) [establish-
ment], 4) dispersão [dispersal] (para novos locais), o que pode levar a 5) populações
espacialmente distribuídas, e à 6) expansão da invasão [invasive spread] (With 2002).
Num interessante trabalho sobre padrões de invasibilidade, Lonsdale (1999) refere que os
estudos de invasão se desenvolvem em três linhas de estudo distintas, centrando-se (1) nas
propriedades do ecossistema, que pode incluir a resistência à invasão e o grau de perturba-
ção; na (2) pressão de propágulos e importância de taxas de dispersão e (3) nos atributos
das espécies exóticas, como o potencial de invasão e nos atributos das espécies nativas. O
Quadro 2.1 sistematiza os principais conceitos utilizados em estudos de invasão e factores
implicados.
Quadro 2.1 Alguns conceitos em estudos de invasão e respectivos factores determinantes. [adaptado de W.M.
Lonsdale (1999)]. PE –propriedades do ecossistema; AEE – atributos das espécies exóticas; AEN – atributos das
espécies nativas; PP – pressão de propágulos.
Table 2.1 Concepts used in biological invasion’ studies and related determinant factors [adapted from W.M.
Lonsdale (1999)]. PE – ecosystem properties; AEE – biological traits of exotic species; AEN – biological traits of
native species; PP –propagule pressure.
Termo utilizado em estudos de Definição convencional Factores

invasão ecológica determinantes
Perturbação [disturbance] remoção da vegetação competidora e PE, AEE

alteração dos ecossistemas
Resistência das espécies nativas capacidade competitiva das espécies AEN

[native species resistance] nativas
Resistência à perturbação capacidade de recuperação do ecossistema PE, AEN

[resistance to disturbance] ou das espécies nativas após a perturbação
Resistência do ecossistema resistência intrínseca do ecossistema à PE
[ecosystem resistance] invasão através da estrutura das comunidades
Invasibilidade susceptibilidade global do ecossistema à PE; AEN
[invasibility; invadability] invasão
Potencial de invasão1 capacidade intrínseca das espécies para a AEE
[invasion potential] invasão
Pressão de propágulos número de propágulos introduzidos num PP
[propagule pressure] determinado local
1
Prieur-Richard e Lavorel (2000) utilizam o termo ‘invasiveness’ como sinónimo de ‘invasion potential’.
2.3 A vegetação ripícola dos sistemas fluviais mediterrânicos: particularidades e

respostas ecológicas à perturbação
A vegetação ripícola, sobretudo se dominada por espécies arbóreas e arbustivas, constitui

um elemento marcante na paisagem mediterrânica (Aguiar et al. 1999). Os sistemas fluviais
mediterrânicos caracterizam-se pela ocorrência de espécies adaptadas a regimes torrenciais
e intermitentes de caudais, cuja distribuição e ocorrência depende, para além das
características climáticas e hidrológicas, da geomorfologia e da perturbação humana do
sistema fluvial.
2.3.1 O clima mediterrânico e o regime hidrológico
O clima temperado com Verão seco da classificação bioclimática de Köpen (1931),

vulgarmente referido como clima mediterrânico abrange a quase totalidade do território
continental português (Azevedo 1980). A sua expressão mundial é reduzida, ocupando
menos de 1% a 4% das terras emersas do globo, consoante os autores e está confinado a 5
regiões do Globo: Bacia do Mediterrâneo, parte oeste da América Norte (Califórnia), zona
central do Chile, sul e parte ocidental da Austrália e África do Sul (província do Cabo). O
território continental português alberga variações climáticas do clima mediterrânico,
modeladas principalmente pela orografia, latitude e continentalidade. De facto, Portugal
com uma extensão latitudinal de apenas 5º, tem variação considerável de altitude, com os
maiores valores situados geralmente nas faixas dos 1000-1500 m nas regiões a norte do Rio
Tejo, e com extensas planícies a sul com altitudes médias entre os 50 e os 400 m de
altitude. As regiões de Portugal mais afastadas do Oceano Atlântico distam apenas 220 km,
no entanto é suficiente para acentuar diferenças climáticas. Tomando o caso da bacia do
Tejo, a média das precipitações anuais varia desde os 2600 mm na parte nordeste para
cerca de 500 mm na parte sudoeste. Outros parâmetros climáticos apresentam também
elevada variabilidade, como a temperatura média anual, que varia desde 6 a 22ºC a
noroeste e 13 a 27 ºC na zona sudeste da bacia, a radiação solar global (7 MJ/m2 a 27
MJ/m2) ou a humidade relativa (60% a 80%). Deste modo, as variantes “mais ricas” do clima
mediterrânico, onde há maior concentração de precipitação na estação fria e maiores
temperaturas no Verão, associam-se às regiões a Sul do rio Tejo (Azevedo 1980), e
abrangem, parte das bacias do Sado e Mira, a parte sudeste da bacia do rio Tejo, a bacia do
Guadiana e as bacias das Ribeiras do Algarve. São vulgarmente designadas por regiões semi-
áridas.
Em consequência da acentuada sazonalidade no regime pluviométrico (pelo menos 65% a

80% da precipitação ocorre nos meses de Novembro a Abril), os sistemas fluviais
mediterrânicos exibem um regime intermitente de caudais e uma dependência da água
subterrânea nos meses quentes e secos de Verão (Gasith e Resh 1999). Embora o regime
hidrológico tenha uma elevada predictabilidade sazonal, a distribuição interanual da
precipitação é irregular, tendo sido detectados ciclos mais ou menos longos de anos secos e
anos de cheias, embora com padrão inconsistente quanto à sua frequência e duração. Nas
regiões mediterrânicas mais húmidas, os cursos de água conseguem manter um caudal
permanente ou pelo menos ter água subterrânea ao longo de todo o ano. Os cursos de água
das regiões mediterrânicas mais áridas, particularmente no Sul da Península Ibérica,
caracterizam-se por caudal nulo ou muito baixo durante a estiagem, com grandes extensões
do leito ocupada por pegos. O sistema hídrico apresenta-se como uma sucessão temporal e
espacial de ambientes lóticos e lênticos, separada frequentemente por extensões de leito
seco (Alves e Bernardo 1998). A magnitude, frequência, duração, predictabilidade dos
caudais e a taxa de alteração das condições hidrológicas são apontados por Poff et al.
(1997) como componentes críticos do regime hidrológico na regulação dos processos
ecológicos em sistemas lóticos, contribuindo, directa ou indirectamente, para a integridade
ecológica do ecossistema. Por outro lado, em termos de produtividade a disponibilidade em
água é o factor mais limitativo em regiões semi-áridas (Stromberg 1993). A par das
flutuações climáticas, as alterações artificiais no regime hidrológico são uma constante em
regiões naturalmente condicionadas pela irregularidade na disponibilidade hídrica ao longo
do ano. Deste modo, e em particular em Portugal, os aproveitamentos hidraúlicos têm
surgido ao longo dos principais cursos de água, no sentido de produção de energia eléctrica,
rega, abastecimento doméstico e industrial, e ainda para controle de cheias, com
consequente alteração na dinâmica dos ecossistemas ribeirinhos. Outras perturbações
humanas influenciam directamente o regime hidrológico, como por exemplo as obras de
canalização, aprofundamentos do leito, atravessamentos e implantação de esporões para
diminuição da velocidade de escoamento. No entanto, em grande número de bacias
hidrográficas, o uso das terras envolventes, incluindo a exploração florestal e agrícola, o
pastoreio, e a urbanização são mais importantes na modificação do regime hidrológico que
a regularização propriamente dita (Poff et al. 1997).
2.3.2 Composição e estrutura das galerias ripícolas
Em oposição às florestas ripícolas de sistemas temperados húmidos, as galerias ribeirinhas

distinguem-se facilmente das formações vegetais adjacentes pela sua estrutura, morfologia
e composição, e são vulgarmente descritas como estreitos corredores verdes que
atravessam a paisagem mediterrânica (Ferreira e Moreira 1999; Salinas et al. 2000). Este
contraste não é apenas cromático ou paisagístico, uma vez que a zona ripícola é
particularmente diversa em habitats e comunidades (Malanson 1993), e é geralmente
considerada entre os ecossistemas mais produtivos, de maior riqueza e maior diversidade
(Naiman et al. 1993). Ademais, em sistemas fluviais mediterrânicos, é frequente a
observação de uma grande variabilidade longitudinal (cabeceiras→foz) em relação ao
número de espécies, composição e densidade da vegetação ripícola (Ward e Stanford 1995;
Tabacchi et al. 1996).
As comunidades ripícolas típicas dos cursos de água temporários, de regime torrencial e

sujeitos a acentuada secura estival são dominadas por espécies perenifólias, como o
loendro (Nerium oleander L.), ou esclerofílicas, como o tamujo [Flueggea tinctoria (L.) G.L.
Webster] e a tamargueira (Tamarix africana Poiret), ou ainda por salgueirais arbustivos,
comunidades pioneiras adaptadas a flutuações do nível de água que ocorrem sobretudo em
solos aluvionares, geralmente siliciosos (Moreira e Duarte 2002) (Figura 2.4).
Figura 2.4 Sistema fluvial do Sul de Portugal e vegetação ripícola característica de climas semi-áridos
dominada por tamujo (Rio Guadiana, Estrela 1998).
Figure 2.4 Fluvial system from a semi-arid landscape of south Portugal, with the dominance of the dyer's
buckthorn [Flueggea tinctoria (L.) G.L. Webster] (River Guadiana, Estrela 1998).
Em linhas de água de carácter permanente, ou torrencial, mas com menor estiagem, estas
espécies são substituídas por bosques caducifólios, como os salgueirais arbóreo-arbustivos,
freixiais, amiais e olmedos (Aguiar et al. 1999). Nestes bosques ribeirinhos é frequente
encontrar uma zonação transversal dependente do gradiente de humidade, com espécies
adaptadas às condições geomorfológicas e edafo-climáticas, e reflectindo em maior ou
menor grau a alteração do sistema, ou seja o seu desvio em relação às condições de
referência (Figura 2.5).
Figura 2.5 Galeria ribeirinha dominada por amieiros num sistema fluvial mediterrânico do Centro de Portugal
(Rio Nabão, Agroal 2004).
Figure 2.5 Riparian galleries dominated by alders [Alnus glutinosa Mill.] on a fluvial corridor of
the Central Portugal (River Nabão, Agroal 2004).
A distribuição das espécies ribeirinhas ao longo do perfil longitudinal dos sistemas fluviais
resulta numa primeira análise de variações nos factores abióticos regionais ou geográficos
(e.g. clima, altitude, geologia) e locais (e.g. largura do canal, substrato). Estes parâmetros
encontram-se muitas vezes associados, e o seu papel individual é difícil de avaliar (Ferreira
1995-1996). As zonas de cabeceira típicas, de vales encaixados e substratos rochosos,
suportam geralmente comunidades de fraca complexidade e de estrutura arbustiva ou

herbácea. Os troços a jusante, com leitos de cheia frequentemente aluvionares e de solos
mais profundos e enriquecidos em nutrientes, permitem o estabelecimento de comunidades
de estrutura e composição mais complexa. No entanto, esta zonação longitudinal está
dependente de múltiplos factores, dos quais se destaca a perturbação humana. Assim, é
frequente encontrar zonas de cabeceira com frondosa vegetação arbórea (caso das
plantações de Betula celtiberica Rothm. & Vasc. perto da nascente do Rio Zêzere), ou
grandes extensões dominadas por Rubus spp. ou Arundo donax nas planícies aluvionares. De
qualquer modo, alerta-se para a formação de juízos de valor sobre a integridade ecológica
da vegetação ripícola de um dado local, troço ou segmento, dever atender em primeira
análise aos constrangimentos ambientais existentes.
Estudos de carácter limnológico têm sido realizados nos últimos anos no sentido de
conhecer os padrões de distribuição das espécies ripícolas aquáticas e ribeirinhas e
principais factores ambientais determinantes em sistemas fluviais portugueses, dos quais se
destacam os de Ferreira e Smeding (1990), Ferreira e Lousã (1988), Morais (1997), Ferreira
et al. (1998a; 1998b), Ferreira e Figueiredo (1994); Ferreira e Moreira (1999) e Moreira et
al. (1999). Paralelamente, a distribuição e caracterização das principais comunidades
lenhosas dulçaquícolas presentes em Portugal continental têm sido estudadas em diversas
bacias hidrográficas, e foram recentemente apresentadas em Moreira e Duarte (2002) em
resultado da compilação de vários trabalhos de carácter essencialmente fitossociológico
realizados nas últimas duas décadas. Destacam-se os trabalhos de Alves et al. (1998),
Aguiar et al. (1995; 1999), Costa et al. (1998), Espírito-Santo et al. (1999), Lousã et al.
(1998) e Rivas-Martínez et al. (1990; 2001).
Estes estudos permitiram reconhecer a complexidade dos sistemas fluviais portugueses e

identificar importantes lacunas no conhecimento da ecologia da vegetação em sistemas
fluviais portugueses, bem como a impossibilidade de transposição e aplicação de
conhecimentos “importados” de regiões de clima e geomorfologia diferentes, e com
aspectos históricos, sociais e demográficos distintos.
2.3.3 Atributos e funções das galerias ribeirinhas e das zonas ripícolas
As zonas ripícolas em regiões semi-áridas têm uma importância ecológica desproporcionada

em relação à reduzida área que ocupam (Kondolf et al. 1996). De facto, numerosas funções
físicas, biológicas e ecológicas são atribuídas às zonas ripícolas e consequentemente às
galerias ribeirinhas. No Quadro 2.2 sistematizam-se as principais características da
vegetação ripícola e respectivas funções, com base em Pinay et al. (1990), Gregory et al.
(1991), Aguiar et al. (1999), Gilvear et al. (2000) e nas sínteses de Kondolf et al. (1996) e
Naiman e Décamps (1997). De notar que esta perspectiva é apenas uma das propostas
possíveis, dada a complexidade e sobreposição das relações entre atributos e funções.
Quadro 2.2 Atributos ecológicos da zona ripícola e respectivas funções e atributos específicos das galerias
ribeirinhas.
Table 2.2 Ecological attributes of riparian areas, related characteristics and functions of the riparian
galleries.
Atributos gerais Atributos específicos das galerias Funções das galerias

da zona ripícola ribeirinhas ribeirinhas
Complexidade boa distribuição na ocupação do espaço Suporte ecológico de várias
estrutural e aéreo (estratos de vegetação) e radicular comunidades (e.g.
biológica diversidade de comunidades ripícolas macroinvertebrados,
(raramente monoespecíficas); vertebrados tetrápodes):
estratégias de reprodução, propagação e habitat, repouso, refúgio e
dispersão variadas; alimentação;
renovação sazonal ou desfasada da Promoção da diversidade
folhagem (espécies caducifólias e/ou intrafluvial
perenifólias) e “contínua” incorporação no Melhoria da qualidade
sistema de outros materiais orgânicos; cénico-paisagística
elevada produtividade, elevada Valor recreativo
eficiência no uso da água. Valor económico
Perturbação adaptação morfológica e fisiológica: Renovação periódica:
natural resistência a condições de anoxia, a oportunidade para o
(alterações no substratos instáveis e a flutuações sazonais estabelecimento de espécies
regime dos caudais pioneiras; fraccionamento
hidrológico) das populações existentes
Natureza linear efeito de fronteira entre o meio aquático Corredor ecológico:
e o terrestre migração de espécies no
conectividade e continuidade interior, estabelecimento de
relações entre biótopos;
( Corredores de invasão:
promoção do
estabelecimento, dispersão e
propagação de espécies
exóticas)
Quadro 2.2 (cont.) Atributos ecológicos da zona ripícola e respectivas funções e atributos específicos das
galerias ribeirinhas.
Table 2.2 (cont’d) Ecological attributes of riparian areas, related characteristics and functions of the
riparian galleries.
Atributos gerais Atributos específicos das galerias Funções das galerias

da zona ripícola ribeirinhas ribeirinhas
Ecótono incorporação/retenção de energia, Estabilização do leito e das
aquático- materiais e organismos dos meios aquático margens
terrestre terrestre adjacentes; Retenção de sedimentos
Recepção de visitantes
(e.g. aves não-ripícolas) e
invasão por espécies exóticas
Filtro biológico de
nutrientes e substâncias
poluentes (produtos
fitofarmacêuticos,
correctivos, efluentes)
Amenidade densidade e diversidade estrutural e Regulação biofísica do
microclimática florística meio: efeitos na moderação
do crescimento excessivo de
algas, hidrófitos e outros
macrófitos, protecção das
comunidades e do meio
aquático em relação a
temperatura, vento e
luminosidade excessivas.
Em sistemas fluviais mediterrânicos, com zonas marginais declivosas e naturalmente pobres

em vegetação, são particularmente importantes as funções físicas, como a retenção de
sedimentos provenientes da erosão hídrica e a estabilização das margens (Salinas et al.
2000). Os caudais sólidos em suspensão têm efeitos na diminuição da transparência da água,
além da poluição por nutrientes e poluentes que estejam associados aos sedimentos e
causem assoreamento do leito ou sedimentações a jusante. A actuação das galerias ripícolas
como filtro biológico toma igualmente um lugar de destaque no contexto da paisagem
mediterrânica, com o uso frequente dos solos confinantes às linhas de água para culturas
agrícolas intensivas (Aguiar et al. 1999). Por outro lado, a conservação de habitats e a
regulação biofísica do meio é muito importante em regiões constrangidas por factores

geomorfológicos e climáticos, como é o caso das regiões do Centro e Sul de Portugal.
Há uma tendência na bibliografia científica para enfatuar as capacidades das galerias

ripícolas na manutenção da diversidade biológica e no suporte biológico de comunidades
autóctones e associadas ao meio lótico. No entanto, a sua natureza linear, e a maior ou
menor continuidade e conectividade2 contribuem também para formar ‘corredores de
invasão’ no dizer de Stohlgren et al. (1998), a par da função “gémea” mais nobre de
corredor ecológico. Por outro lado, as galerias ripícolas situadas na interface entre o meio
aquático e terrestre, são importantes focos de atracção biológica (Marks 1993) para
espécies terrestres e ruderais (exóticas ou nativas) pela maior disponibilidade em água que
em zonas não influenciadas pelo sistema lótico (Tabacchi et al. 1996).
2.3.4 As galerias ribeirinhas e as actividades humanas
As zonas ripícolas, a par de sistemas muito fragmentados (e.g. áreas florestais em

exploração), ou sistemas típicos de exportação de matéria (e.g. sistemas lagunares e
marinhos) são considerados ecossistemas extremamente frágeis3, uma vez que importam
variados tipos de materiais de fontes diversas e de áreas muito extensas (Nilsson e Grelsson
1995). A vegetação ripícola encontra-se entre os ecossistemas mais severamente alterados
por actividades humanas, principalmente em bacias hidrográficas de características semi-
áridas (Décamps et al. 1988; Lepart e Debussche 1992; Stromberg 1993; Gallego-Fernández
et al. 1999; Corbacho et al. 2003).
As regiões de clima mediterrânico (di Castri 1990; Naveh e Vernet 1991), e em particular a
Bacia Mediterrânica (Lepart e Debussche 1992; Gasith e Resh 1999), são caracterizadas por
uma alteração milenar no uso do solo adjacente aos cursos de água para produção agrícola
e criação de gado, a par de uma forte competição pela água entre as utilizações humanas e
os ecossistemas naturais (García de Jálon 1987). Em resultado, é frequente a fragmentação
e compartimentação das galerias ribeirinhas com ocorrência de alterações significativas na
sua estrutura, riqueza de espécies, distribuição e composição (e.g Knight et al. 1994;
Nilsson 1996; Planty-Tabacchi et al. 1996; Malanson e Cramer 1999). Acresce que a
dinâmica dos ecossistemas envolventes e processos ecológicos fluviais estão
2
A conectividade refere-se à possibilidade de transferência de energia e material (incluindo organismos)
ao longo da zona ripícola ou do sistema fluvial (Ward 1998).
3
Entende-se por fragilidade do ecossistema o grau de alteração da composição e abundância de
espécies, após a perturbação (Nilsson e Grelsson 1995).
particularmente interligados (Gregory et al. 1991), e a perturbação humana nas áreas

envolventes é uma das causas frequentes de interrupção entre estas conexões e interacções
ecológicas (Large et al. 1994; Roth et al. 1999; McIntyre e Hobbs 1999; Jungwirth et al.
2002). Por outro lado, alterações antropogénicas relativamente mais recentes, como as
modificações importantes no regime hidrológico, para regularização de caudais ou
aproveitamento hidráulico, induzem a alterações na morfologia do canal e das margens, no
microclima, no transporte e sedimentação, bem como na qualidade da água, com
repercussões nas estruturas morfogénicas fluviais e consequentemente nos ecossistemas
aquáticos e ribeirinhos.
Um grande leque de actividades humanas afectam as galerias ripícolas, tanto pela remoção
directa da vegetação, como por efeitos secundários da alteração da dinâmica do sistema
fluvial e dos processos hidrológicos. O Quadro 2.3 sumaria as principais actividades humanas
em sistemas fluviais, os principais efeitos físicos e potenciais consequências ecológicas na
vegetação ripícola, e baseia-se sobretudo na revisão de Kondolf et al. (1996),
complementada com os trabalhos de Décamps et al. (1988), Brookes (1988), Lepart e
Debussche (1992), Bravard e Petts (1996), Wood e Armitage 1999).
Quadro 2.3 Actividades humanas e principais efeitos físicos e consequências potenciais para a vegetação
ripícola (baseado principalmente em Kondolf et al. 1996)
Table 2.3 Human activities, main physical effects and potential consequences on the riparian vegetation
(based mainly in Kondolf et al. 1996).
Actividade humana e potenciais efeitos Consequências potenciais para a vegetação

físicos directos ripícola
Extracção de inertes
• construção de infraestruturas, estradas ¨ ¨remoção das galerias ribeirinhas
• indução da instabilidade do canal com

efeitos na erosão das margens ¨ ¨ queda de árvores e alteração da estrutura
das galerias
• indução à incisão do canal, com efeitos no

abaixamento da toalha freática ¨ ¨aumento da mortalidade das árvores
ripícolas, diminuição da taxa de crescimento
e do volume da canópia das árvores
• aproximação do leito à toalha freática nas

zonas de extracção ¨ ¨ ocupação do leito por vegetação em zonas
anteriormente mais profundas
Barragens, represamentos
• diminuição dos caudais provoca o
assoreamento do canal activo ¨ ¨ ocupação do canal activo com vegetação
ripícola
• diminuição dos caudais induz a menores
taxas de meandrização do canal ¨ ¨redução da diversidade de habitats ripários
• descargas artificiais bruscas e alteração da

época da ocorrência de caudais extremos ¨ ¨ submersão da vegetação
¨ estrutura da vegetação ripícola
severamente alterada com mortalidade de
espécies menos tolerantes a flutuações do
nível de água
• efeitos da interposição de um
aproveitamento hidraúlico ¨ ¨ submersão da vegetação existente na área
da albufeira
¨ continuidade longitudinal das galerias
ribeirinhas interrompida
Quadro 2.3 (cont.) Actividades humanas e principais efeitos físicos e consequências potenciais para a
vegetação ripícola (baseado principalmente em Kondolf et al. 1996)
Table 2.3 (cont’d) Human activities, main physical effects and potential consequences on the riparian
vegetation (based mainly in Kondolf et al. 1996).
Actividade humana e potenciais efeitos físicos Consequências potenciais para a vegetação

directos ripícola
Produção de energia hidroeléctrica

• represamento dos cursos de água e desvio
da água para canais ¨ ¨ stresse hídrico
• infraestruturas e canais associados

construídos sobre a zona ripícola ¨ ¨ galerias removidas e substituídas por
estradas e outras infra-estruturas
• flutuações súbitas do nível da água ¨ ¨ erosão das margens e efeitos na vegetação

ripícola
Irrigação
• desvio da água dos sistemas fluviais ¨ ¨stresse hídrico devido à extracção de água
• construção de canais ¨ ¨ estabelecimento de vegetação ripícola nas

valas e canais de irrigação
Captações da água subterrânea
• rebaixamento da toalha freática, perda da
conectividade vertical ¨ ¨ stresse hídrico por diminuição da água na
zona radicular, aumento da mortalidade de
árvores
Drenagem
• rebaixamento da toalha freática ¨ ¨ morte e dessecamento da vegetação ripícola
Uso agrícola e florestal da zona ripícola
• mobilização do solo para uso agrícola ou

florestal e instalação da cultura ¨ ¨ remoção ou fragmentação da galeria ripícola
e substituição pela cultura
¨ entrada de espécies ruderais e terrestres

directos ripícola
Linearização, aprofundamentos,
canalização e outras obras de engenharia
hidráulica (esporões, terraços,...)
• redução do comprimento, complexidade e
dinâmica do sistema fluvial ¨ ¨ redução da diversidade de habitats ripários
¨ remoção/fragmentação da galeria ripícola
Exploração florestal da zona ripícola
• redução da complexidade da zona ripícola,
diminuição da incorporação de materiais no
sistema ¨ ¨ remoção de árvores com valor económico
¨ alteração da estrutura, do microclima e
interrupção da continuidade da galeria
Produção animal
• Compactação das margens ¨ ¨ diminui o potencial de estabelecimento da
vegetação ripícola
• Pastoreio na zona ripícola ¨ ¨ reduz o estabelecimento e renovação da

galeria por alimentação em plântulas de
árvores e arbustos
• Acesso à água e efeitos na alteração do ¨ entrada de sementes e propágulos exógenos
perfil do sistema ¨ por zoocoria
¨ interrupções na galeria, destruição e

diminuição da complexidade de habitats

directos ripícola
Urbanização
• criação de infraestruturas, estradas ¨ ¨ habitat ripário e vegetação associada
substituída por infraestruturas urbanas
¨ alterações indirectas na vegetação induzidas
por alteração na dinâmica e estrutura do
sistema fluvial
• criação de áreas impermeabilizadas ¨ ¨ stresse hídrico
• drenagem para permitir o estabelecimento

do meio urbano ¨ ¨ dessecação da vegetação ripícola
• realocação do canal e canalização para

controle de cheias ¨ ¨ dificuldade do restabelecimento da
vegetação ripícola e alteração da estrutura e
composição da galeria ribbeirinha
• aumento do número de visitantes na zona ¨ver efeitos da recreação
ripícola ¨
Recreação
• Trilhos construídos ao longo dos rios
(pedonais, para cavalos, ciclismo,,...) ¨ ¨ remoção e danificação da vegetação ripícola
¨ interrupção da continuidade por
atravessamentos e acessos ao curso de água
¨ entrada de sementes e propágulos exógenos
por antropocoria
• Compactação da zona ripícola e alteração ¨ diminui o estabelecimento da vegetação

do perfil tranversal¨ ripícola, criação de clareiras
¨ consequências para a vegetação do aumento
da erosão
Estudos recentes sobre o papel ou efeitos da perturbação humana nas componentes

biológicas dos sistemas lóticos centram-se na concepção de modelos explicativos da
modificação dos processos funcionais do ecossistemas lóticos (Ward et al.1999), na
caracterização das alterações de atributos específicos e populacionais ao longo de escalas
espaço-temporais diversas (Ward e Tockner 2001) ou na previsão dos efeitos das actividades
humanas. Neste contexto, são frequentes estudos sobre os efeitos da perturbação na
estrutura comunitária (e.g. Tabacchi et al. 1996), na riqueza em espécies (e.g. McIntyre e
Lavorel 1994; Nilsson e Jansson 1995), na biodiversidade (Ward e Tockner 2001) e na
ocorrência de invasões ecológicas (e.g. Stohlgren et al. 1998).
Particularmente importantes na conservação e gestão dos ecossistemas aquáticos

ribeirinhos são os aspectos relacionados com a diversidade ecológica e com as invasões por
espécies exóticas. Neste âmbito várias vertentes têm sido estudadas, incluindo os efeitos
negativos da sua introdução e expansão (toxicológicos, ecológicos e económicos), os
factores que influenciam as várias etapas do processo de invasão e os atributos dos
ecossistemas e das espécies exóticas em causa (Byers et al. 2002). Geralmente
reconhecem-se relações positivas entre a perturbação antropogénica em zonas ripícolas e a
invasão por espécies exóticas (Hood e Naiman 2000). O tipo, tempo de actuação e
magnitude da perturbação em conjunto com as características do ecossistema interferem
nas várias etapas do processo de invasão e determinam respostas diferentes no sucesso do
processo de invasão (With 2002; Prieur-Richard e Lavorel 2000) (Figura 2.6). Perante um
cenário de crescimento excessivo de taxa invasores ou da possibilidade de infestação,
desenham-se três perspectivas fundamentais de gestão: i) a prevenção/exclusão; ii) a
detecção precoce/rápida avaliação e o iii) combate/contenção/erradicação (Rejmánek
2000).
Em Portugal estão referenciadas 139 taxa exóticos aquáticos ou associados ao sistema

lótico, dos quais apenas 8 (6%) são hidrófitos, 2 são helófitos e os restantes incluem-se no
grupo dos higrófitos e sub-higrófitos (Duarte et al. 2004). As introduções são
maioritariamente acidentais e decorrentes do valor ornamental das espécies exóticas.
Reconhecem-se cerca de uma dezena e meia espécies invasoras com carácter de infestantes
em sistemas fluviais portugueses, entre as quais se destacam taxa lenhosos do género
Acacia spp., Ailanthus altissima (Mill.) Swingle e Arundo donax, os higrófitos Eryngium
pandanifolium Cham. e Schlecht. e Paspalum distichum L. e os hidrófitos Eichhornia
crassipes (Mart.) Solms-Laub., Azolla filiculoides Lam. e Myriophyllum aquaticum (Vell.)
Verdc. (Moreira et al. 2002b; Aguiar et al. 1996; Aguiar et al. 1997). Quanto à possibilidade
de novas introduções, algumas plantas que têm provocado infestações graves em países
europeus de clima mediterrânico foram identificadas e accionados mecanismos de vigilância
às importações de plantas [e.g. Hydrilla verticilata (L.f.) Royle] e aos ecossistemas mais
propícios à sua ocorrência. Por outro lado, a detecção precoce e avaliação de potenciais
invasoras tem sido posta em prática em sistemas fluviais portugueses, como é exemplo o
caso da presumível ocorrência de Salvinia auriculata Aubl. em ribeiras algarvias (Moreira et
al. 2002b).
ETAPAS INFLUÊNCIA DO ECOSSISTEMA REFERÊNCIAS

FLUVIAL E DOS ECOSSISTEMAS
DE INVASÃO ENVOLVENTES
O uso do solo e outras características dos Ferreira e Moreira (1995);
ecossistemas envolventes influenciam King e Buckney (2000);
INTRODUÇÃO a frequência das introduções e o tipo Sax (2001); McKinney (2002);
de espécies exóticas introduzidas
A distribuição espacial e extensão de Lepart e Debussche (1991);

áreas fragmentadas, a complexidade do Pyšek e Prach (1994);
COLONIZAÇÃO ecossistema e a presença dos recursos Stohlgren et al. (1998);
necessários influencia a colonização Stohlgren et al. (1999)
ESTABELECIMENTO A configuração espacial do habitat pode Tabacchi et al. (1996);

DE POPULAÇÕES promover a sobrevivência e a reprodução Deutschewitz et al. (2003);
de espécies exóticas (efeitos na demografia) Lavoie et al. (2003)
(aumento da população)
A dispersão e os factores que influenciam a Johansson et al. (1996);

dipersão estão relacionados com a estrutura Ward et al. (1999);
DISPERSÃO e continuidade do habitat ripário e com as Buchan e Padilla (1999);
perturbações naturais (regime hidrológico, Andersson et al. (2000);
fogo,...) e artificiais (pastoreio, recreação,...) Levine (2001);
POPULAÇÕES A interacção das etapas anteriores Décamps et al. (1995);

com as características dos ecossistemas Planty-Tabacchi et al. (1996)
DISTRIBUÍDAS pode ocasionar aumento na distribuição
espacial das populações
ESPACIALMENTE
Uma perspectiva ecológica global e integrada Wade (1990); With (2000);

é geralmente necessária para a previsão da Réjmánek (2000);Hood e Naiman (2000);
expansão das invasões e gestão dos ecossis- Sax e Brown (2000);
EXPANSÃO temas aquáticos e ribeirinhos, atendendo: Prieur-Richard e Lavorel (2000)
1) à invasividade (susceptibilidade à invasão) Byers (2002); Byers et al. (2002);
DA INVASÃO 2) resistência do ecossistema à invasão
e à perturbação
3) potencial de invasão e atributos biológicos
das espécies invasoras
Figura 2.6 Etapas da invasão segundo With (2002) e influência da estrutura dos ecossistemas fluviais e
envolventes no processo de invasão.
Figure 2.6 Stages of invasion according to With (2002) and effects of the fluvial ecosystems and of the
floodplain structure on the process of invasive spread.
2.3.5 A vegetação ripícola como indicador de integridade ecológica
A determinação da integridade ecológica de sistemas fluviais é um processo complexo e

difícil, envolvendo uma multiplicidade de factores bióticos e abióticos cuja importância não
deve ser minimizada em abordagens reducionistas (Cairns 1995). Para além das dificuldades
de aplicação prática, o conceito de integridade ecológica tem sido amplamente discutido, e

numerosas definições têm sido propostas.
A definição de integridade ecológica está ligada ao de integridade biológica, entendida

como a capacidade de suporte e manutenção de um sistema biológico integrado, adaptado
e equilibrado, com uma multiplicidade de elementos (genes, espécies, comunidades) e de
processos (mutações, interacções bióticas, processos metapopulacionais, ciclos de
nutrientes e processos energéticos) esperados para uma dada região ou habitat ‘naturais’
(Karr´s, comunicação pessoal em Cairns 1995). A integridade biológica de um dado local
será então, uma medida das condições bióticas e dos organismos desse local em relação ao
que seria de esperar na ausência de perturbação humana. No entanto, as actividades
humanas, têm alterado de um modo generalizado a naturalidade dos ecossistemas, facto
que é especialmente evidente nos ecossistemas fluviais da Europa (Verdonschot 2000), e
em particular na Bacia Mediterrânica. A noção de sustentabilidade e co-evolução foi
integrada no conceito de integridade ecológica, possibilitando a sua aplicabilidade a
ecossistemas inevitavelmente perturbados. Assim, e aceitando a definição de Angermeier e
Karr (1994), a integridade ecológica corresponde à manutenção dos processos endógenos e
exógenos e dos atributos que interagem com o ambiente de modo a que a comunidade
biológica corresponda ao estado natural de um habitat tipológico específico, de acordo com
o princípio de auto-regulação, resiliência e resistência.
A preservação e o restabelecimento da diversidade funcional no âmbito da integridade ou

da sustentabilidade ecológica é uma das prioridades actuais na gestão dos recursos hídricos
(Karr 1998; Ward e Tockner 2001). Neste sentido, na recém-aprovada Directiva Europeia-
Quadro da Água, o conceito de qualidade ecológica (‘ecological quality’) é utilizado como
base para a determinação e monitorização da qualidade da água (Pollard e Huxhman 1998;
Boon 2000).
Avaliações funcionais dos ecossistemas estão direccionadas para os processos que os

caracterizam (e.g. hidrologia, geomorfologia, sucessão de comunidades florísticas,
manutenção da biodiversidade) e na avaliação do desvios da integridade ecológica de um
dado local relativamente às condições de referência (locais menos perturbados) (Karr e Chu
2000; Jungwirth et al. 2002). As condições de referência correspondem a atributos
biológicos presentes num grupo de locais com o mínimo de perturbação possível (locais de
referência) que são representativos das condições naturais dominantes numa dada região
(Reynoldson et al. 1997; Reynoldson e Wright 2000).
Avaliações da integridade biológica têm sido usualmente baseados em indicadores bióticos

(fitoplâncton, macrófitos, macroinvertebrados, peixes) e parâmetros fisico-químicos e
hidromorfológicos. Como anteriormente referido, as galerias ribeirinhas estão associadas à
geomorfologia e ao regime hidrológico, pela sua composição, interface lateral e integridade
longitudinal (Harris 1988; 1999; Tabacchi et al. 1996; Jungwirth et al. 2002) e reflectem as
condições ambientais do ecossistema, bem como a perturbação. Deste modo, quer as
espécies, quer os padrões comunitários das plantas fluviais podem ser utilizados para
avaliar a integridade ecológica de uma dada tipologia dos sistemas fluviais (Lange e Van Zon
1981; Holmes 1989; Haury 1996). Enquanto que as espécies ou indivíduos respondem a
condições morfológicas e hidrológicas locais, as comunidades de macrófitos tendem a
reflectir os processos morfogenéticos que ocorrem numa escala espaço-temporal superior
(Habersack 2000), um aspecto de extrema importância quando se pretende utilizar
bioindicadores de qualidade ecológica em sistemas fluviais.
2.4 Proposta de estudo
Os aspectos abordados nos subcapítulos anteriores permitem reconhecer a especificidade da

vegetação ripícola e das galerias ribeirinhas em sistemas fluviais portugueses, face aos
condicionalismos ambientais e à omnipresente perturbação humana nos sistemas fluviais,
zonas ripícolas associadas e ecossistemas envolventes.
Os últimos vinte anos têm sido frutuosos para o conhecimento da dinâmica e estrutura da
vegetação dos ecossistemas fluviais em Portugal. No entanto, identificam-se importantes
lacunas, decorrentes de dois aspectos fundamentais, por um lado, as escalas espaço-
temporais utilizadas não são de um modo geral, abrangentes, ou seja, são analisados casos
de estudo demasiado pontuais ou com carácter sazonal ou fixos numa determinada escala
temporal, inadequados à concepção de modelos funcionais e determinísticos dos sistemas
fluviais portugueses. Por outro lado, a ainda fraca consolidação do edifício teórico basilar
sobre sistemas fluviais mediterrânicos, não permite integrar de forma adequada os esforços
de investigação realizados e suas aplicações práticas.
A Figura 2.7 realiza o enquadramento dos principais dos pontos de partida e de discussão
propostos neste estudo, como contributo para o avanço dos conhecimentos nesta área.
O QUE TEMOS
WHAT WE HAVE
8000 anos de perturbação ecossistemas abertos
antropogénica nos corredores open ecossystems
fluviais e nas zonas envolven-
écotonos ripícolas (i.e. inter-
tes
faces entre os ecossistemas
8000 years of anthropogenic terrestre e aquático)
disturbance in the fluvial corridors
and in the surrounding areas riparian ecotones (i.e. interfaces
between terrestrial and aquatic
clima mediterrânico, com systems)
marcada variabilidade inter-
anual e sazonal mosaico estrutural de habi-
tats e da paisagem envolven-
mediterranean climate, with te
marked inter-annual and sazonal
variability patchy structural pattern of the
aquatic and riparian habitats and of
the surrounding landscape
vegetação ripícola resiliente
resilient riparian vegetation
interligação entre a dinâmica
dinamismo e heterogeneidade dos ecossistemas envolventes
espaço-temporal e complexida- e os processos e interacções
de estrutural e biológica das ga- ecológicas fluviais
lerias ripícolas linkage between the dynamic of the
spatio-temporal heterogeneity and riverine and terrestrial landscapes and
structural and biological complexity of the ecological processes and interactions
riparian galleries of the fluvial corridors
PRINCIPAIS MAIN
PONTOS DE KEY RESEARCH
PARTIDA E QUESTIONS AND
DE DISCUSSÃO DISCUSSION POINTS
Quais os padrões de distribuição espaciais das gale- What are the spatial and temporal patterns of the ripa-
rias ribeirinhas, a respectiva causalidade abiótica e o rian galleries and the related human disturbance?
grau de perturbação associado?
Given the specific features of riparian galleries i n se-
Dadas as características peculiares das galerias ribei- mi-arid regions, what is its contribution to the overall eco-
rinhas em regiões semi-áridas, qual a sua contribuição logical integrity?
para a integridade ecológica dos ecossistemas fluviais ? Do different land-use changes over time in different
Qual a influência das alterações espaço-temporais do regions lead to different responses in riparian integrity?
uso do solo nos padrões de integridade e composição
das galerias ribeirinhas? Do environmental factors play a larger role in shaping
riparian formations than the human activities in the river
Os factores ambientais têm uma maior influência nos valley?
ecossistemas fluviais e ripícolas que as actividades hu-
manas? Que dimensão espacial é mais relevante? Which abiotic specific-features account significantly for
the explanation of the riparian variability? Landscape or
Qual o potencial preditivo da vegetação face a pertur- habitat level: is the scale-factor relevant?
bações antropogénicas e endógenas ? What is the preditive potential of riparian vegetation in
relation to natural and human disturbances?
Quais os padrões de distribuição da flora exótica ao
longo dos gradientes transversal e longitudinal dos sis- Which are the invasive patterns of exotic flora along
temas fluviais? mediterranean-type rivers and across its floodplain?
Qual a relação entre a riqueza e o perfil florístico dos
ecossistemas e a sua invasibilidade? Does the species richness and the floristic profile of
the riparian ecosystems influence its invasibility?
Que factores ambientais e antropogénicos contribuem
para a invasibilidade dos ecossistemas e que atributos What environmental and disturbance factors are more
das espécies exóticas (e nativas) concorrem para o su- strongly associated with the ecosystems' invasibility,and
cesso das invasões? what are the invaders' biological attributes?
Figura 2.7 Enquadramento e síntese dos principais pontos de partida e discussão e da proposta de estudo.
Figure 2.7 Framework and synthesis of the main discussion points and of the key research questions.
2.5 Referências bibliográficas
Aguiar C.F., Capelo J., Costa J.C., Espírito-Santo M.D. e Lousã M. 1995. Tipologia das geoséries
mediterrânicas de Portugal Continental. Actas 3º Congresso Nacional de Conservação da
Natureza, Lisboa. pp. 25-32.
Aguiar F., Costa J.C., Duarte M.C., Fabião A.D., Ferreira M.T., Ramos I.L., Lousã M. e Monteiro F.P.
Moreira I. e Saraiva G. (Coordenadores). 1999. As galerias ribeirinhas na paisagem
mediterrânica: reconhecimento na bacia hidrográfica do rio Sado. ISA Press. Lisboa.
Aguiar F., Moreira I. e Ferreira T. 1996. A Percepção da Vegetação Aquática Infestante pelas
Entidades Gestoras dos Recursos Hídricos. Revista de Ciências Agrárias 19(4): 35-56.
Aguiar F., Moreira I. e Ferreira T. 1997. Estimativa do risco de infestações por plantas aquáticas.
Actas IV Encontro Nacional de Protecção Integrada. Importância da Protecção Integrada
numa Agricultura Sustentável, Angra do Heroísmo. pp. 463-476.
Alves J.M.S., Espírito-Santo, M. D., Costa, J.C., Gonçalves, J.H.C. e Lousã M. 1998. Habitats naturais
e seminaturais de Portugal continental. Tipos de habitats mais significativos e
agrupamentos vegetais característicos. Instituto de Conservação da Natureza, Lisboa.
Alves M.H. e Bernardo J.M. 1998. Novas perspectivas para a determinação do caudal ecológico em
regiões semi-áridas. Seminário Barragens e Ambiente 14 pp.
Andersson E., Nilsson C. e Johansson M.E. 2000. Plant dispersal in boreal rivers and its relation to the
diversity of riparian flora. Journal of Biogeography 27: 1095-1106.
Angermeier P.L. e Karr J.R. 1994. Biological integrity versus biological diversity as policy directives.
BioScience 44: 690-697.
Assessments and scientific basis for management options. Davis: University of California, Centers for
Water and Wildland Resources: 1009-1030.
Azevedo A.L. 1980. An attempt to characterize the mediterranean climate with a special reference
to the case of continental Portugal. Pot. Acta Biol.(A) 16(1-4): 11-18.
Bartley M.R. e Spence D.H. 1987. Dormancy and propagation in helophytes and hydrophytes. Arch.
Hydrobiol. Beih. Ergbn. Limnol. 27: 139-155.
Boon P.J. 2000. The development of integrated methods for assessing river conservation value.
Hydrobiology 422-423: 413-428.
Bravard J.-P. e Petts G.E. 1996. Human impacts on fluvial hydrosystems. Em: Petts G.E. e Amoros C.
(eds), Fluvial Hydrosystems, Chapman & Hall, London. pp. 242-262.
Brookes A. 1988. Channelized Rivers. Perspectives for Environmental Management. John Wiley & Sons
Ltd, Chichester, England.
Buchan L.A.J. e Padilla D.K. 1999. Estimating the probability of long-distance overland dispersal of
invading aquatic species. Ecological Applications 9(1): 254-265.
Byers J. E., Reichard S., Randall J.M., Parker I.M., Smith C.S., Lonsdale W.M., Atkinson I.A.E.,
Seastedt T.R., Williamson M., Chornesky E. e Hayes D. 2002. Directing research to reduce
impacts of nonindigenous species. Conservation Biology 16(3): 630-640.
Byers J.E. 2002. Physical habitat attribute mediates biotic resistance to non-indigenous species
invasion. Oecologia 130: 146-156.
Cairns J.Jr. 1995. Ecological Integrity of aquatic systems. Regulated Rivers: Research and
Management 11: 313-323.
Catarino L., Moreira I., Ferreira T. e Duarte M.C. 2001. Plantas aquáticas infestantes de valas e
canais. ISAPress, Lisboa.
Connell J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302-1310.
Corbacho C., Sánchez J.M. e Costillo E. 2003. Patterns of structural complexity and human
disturbance of riparian vegetation in agricultural landscapes of a Mediterranean area.
Agriculture, Ecosystems and Environment 95: 495-507.
Costa, J.C. Aguiar C., Capelo, J.H. Lousã, M. e Neto C. 1998. Biogeografia de Portugal continental.
Quercetea 0: 5-56.
Dawson F.H. 1988. Water flow and the vegetation of running waters. Em: Wright S.E., Buckley M. e
Symoens J.J. (eds.), Vegetation in Inland Waters. Handbook of Vegetation Science 15(1).
Kluwer Academic Publ., Dordrecht: 283-310.
Décamps H., Fortune M., Gazelle F. e Patou G. 1988. Historical influence of man on the riparian
dynamics of a fluvial landscape. Landscape Ecology 1: 163-173.
Décamps H., Planty-Tabacchi A.-M. e Tabacchi E. 1995. Changes in the hydrological regime and
invasions by plant species along riparian systems of the Adour River, France. Regulated
Rivers: Research and Management 11: 23-33.
Denny, P. 1985. Wetland vegetation and associated life forms. Em: Denny P. (ed.), The Ecology and
Management of African Wetland Vegetation. Dr. Junk Publ. Dordrecht. pp. 1-18.
Deutschewitz K., Lausch A., Kühn I. e Klotz S. 2003. Native and alien plant species richness in relation
to spatial heterogeneity on a regional scale in Germany. Global Ecology & Biogeography 12:
299-311.
di Castri F. 1991. An ecological overview of the five regions of the world with a Mediterranean
climate. Em: Groves R.H. e di Castri F. (eds), Biogeography of Mediterranean Invasions.
Cambridge University Press, Cambridge. pp. 3-16.
di Castri F. 1990. On invading species and invaded ecosystems: the interplay of historical chance and
biological necessity. Em: di Castri F., Hansen A.J. e Debussche M. (eds.), Biological
Invasions in Europe and the Mediterranean Basin. Kluwer Academic Publishers, Dordrecht,
The Netherlands. pp. 3-16
Duarte M.C. e Moreira I. 2002. Suplemento 1. Composição florística e agrupamentos fitossociológicos

dulçaquícolas. Em: Moreira I., Ferreira M.T., Cortes R., Pinto P. e Almeida P.R. (eds.),
Ecossistemas Aquáticos e Ribeirinhos. Ecologia, Gestão e Conservação. Instituto da Água,
MCOTA. pp. S1.1-S1.50.
Duarte M.C., Moreira I. e Ferreira M.T. 2004. Flora de ecossistemas aquáticos e ribeirinhos
portugueses: delimitação taxonómica, tipológica e espacial. Recursos Hídricos 25: 67-94.
Espírito-Santo M.D., Costa J.C., Capelo J. e Arsénio P. 1999. Vegetação das margens das ribeiras do
Algarve. Revista de Biologia 17: 73-87.
Fabião A.M.D. 1996. Árvores e florestas, Colecção EuroAgro nº15 , 2ª ed., Publicações Europa-
América, Mem-Martins.
Ferreira M.T. 1994. Aquatic and marginal vegetation of the river Divor and its relation with land use.
Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte
Limnologie 25: 2309-2316.
Ferreira M.T. 1995-1996. Ecologia de plantas fluviais: Bases teóricas para o seu controle, gestão e
utilização. Anais Instituto Superior de Agronomia 45: 21-83.
Ferreira M.T. e Cortes R.M.V. 1998. Funções dos ecótonos ripários e o ordenamento de bacias
hidrográficas. Actas do Workshop Funções e Ordenamento de Zonas Ripárias, Vila Real.
Ferreira M.T. e Figueiredo J.C. 1994. Spatial evolution of river formations on a partially regulated
river. Limnetica 10(1): 53-57.
Ferreira M.T. e Lousã M. 1988. Fitocenoses ripícolas em Portugal. Actas do 1º Congresso Florestal
Nacional: 264-268.
Ferreira M.T. e Moreira I.S. 1995. The invasive component of river flora under the influence of
Mediterranean agricultural systems. Em: Pysek P., Prach K., Rejmanek M. e Wade M.
(eds.), Plant Invasions - General Aspects and Special Problems, SPB Academic Publishing,
Amsterdam, pp.117-127.
Ferreira M.T. e Moreira I. 1999. River plants from an Iberian basin and environmental factors
influencing their distribution. Hydrobiologia 415: 101-107.
Ferreira M.T. e Moreira I. 2002. Capítulo 1. Ecologia e gestão de ecosistemas dulçaquícolas. Em:
Moreira I., Ferreira M.T., Cortes R., Pinto P. e Almeida P.R. (eds.), Ecossistemas Aquáticos
e Ribeirinhos. Ecologia, Gestão e Conservação. Instituto da Água, MCOTA. pp. C1.1-C1.13.
Ferreira M.T. e Smeding F. 1990. Distribution patterns of aquatic and riparian vegetation in Erra river,
central Portugal. Scientia Gerundensis 16(1): 99-106
Ferreira M.T., Catarino L. e Moreira I. 1998a. Aquatic weed assemblages in an Iberian drainage
channel system and related environmental factors. Weed Research 38:291-300.
Ferreira M.T., Godinho F.N. e Cortes R.M. 1998b. Macrophytes in a southern Iberian river. Vehr.
Internat. Verein. Limnol. 26: 1835-1841.
Ferreira M.T.M.C.C. 1992. Estrutura e dinâmica das comunidades de macrófitos lóticos da bacia
hidrográfica do Rio Sorraia. Interferências dos ecossistemas agrários envolventes.
Dissertação de Doutoramento em Engenharia Florestal. Instituto Superior de Agronomia,
Lisboa.
Font Quer, P. (ed.). 1982. Diccionario de Botánica. Editorial Labor, S.A., Barcelona.
Frissel C.A., Liss W.J., Warren C.E. e Hurley M.D. 1986. A hierarchical framework for stream
classification: viewing streams in a watershed context. Environmental management 10:
199-214.
Gallego-Fernández J.B., García-Mora M.R. e García-Novo F. 1999. Small wetlands lost: a biological
conservation hazard in Mediterranean landscapes. Environmental Conservation 26(3): 190-
199.
Garcia-de-Jalon D. 1987. River regulation in Spain. Regulated Rivers: Research and Management 1: 34-
38.
Gasith A. e Resh V.H. 1999. Streams in Mediterranean climate regions: abiotic influences and biotic
responses to predictable seasonal events. Annual Review of Ecology and Systematics 30:
51-81.
Gilvear D.J., Cecil J. e Parsons H. 2000. Channel change and vegetation diversity on a low-angle
alluvial fan, River Feshie, Scotland. Aquatic Conservation: Marine and Freshwater
Ecosystems 10(1): 53-71.
González-Bernáldez F., Levassor C. e Peco B. 1989. Landscape ecology of uncultivated lowlands in

central Spain. Landscape Ecology 3:3-18.
Gregory S.V., Swanson F.J., McKee W.A. e Cummins K.W. 1991. An ecosystem perspective of riparian
zones. Focus on liks between land and water. Bioscience 41(8): 540-551.
Habersack H.M. 2000. The river scaling concept (RSC): a basis for ecological assessments.
Harris R.R. 1988. Associations between stream valley geomorphology and riparian vegetation as a
basis for landscape analysis in the eastern Sierra Nevada, California, USA. Environmental
Harris R.R. 1999. Defining reference conditions for restoration of riparian plant communities:
examples from California, USA. Environmental Management 24(1): 55-63.
Haury, J. 1996. Assessing functional typology involving water quality, physical features and
macrophytes in a Normandy river. Hydrobiologia 340: 43-49.
Holland M.M. 1988. SCOPE/MAB Technical Consultations on Landscape Boundaries. Report on a

SCOPE/MAB Workshop Ecotones. Biology International, Special Issue, 17: 47-106.
Holmes N. 1985. Typing British Rivers according to their flora, Nature Conservancy Council, nº4.
Huntington.
Hood W.G. e Naiman R.J. 2000. Vulnerability of riparian zones to invasion by exotic vascular plants.
Plant Ecology 148: 105-114.
Hynes H.B.N. 1983. Ground water ans stream ecology. Hydrobiologia 100: 93-99.
Jansson R.1994. A comparison of species richness and traits of riparian plants between a main river
channel and its tributaries. Journal of Ecology 82: 281-295.
Johansson M.E., Nilsson C. e Nilsson E. 1996. Do rivers function as corridors for plant dispersal?
Journal of Vegetation Science 7: 593-598.
Jungwirth M., Muhar S. e Schmutz S. 2002. Re-establishing and assessing ecological integrity in
riverine landscapes. Freshwater Biology 47: 867-887.
Junk W.J., Barley P.B. e Sparks R.E. 1989. The flood pulse concept in river-floodplain system.
Canadian Special Publication of Fisheries and Aquatic Sciences 106: 110-127.
Karr J.R. 1998. Rivers as sentinels: using the biology of rivers to guide landscape management. Em:
Naiman R.J. e Bilby R.E. (eds.), River Ecology and Management: Lessons from the Pacific
Coastal Region. Springer-Verlag. New-York. pp. 502-528.
Karr J.R. and Chu E.W. 2000. Sustaining living rivers. Hydrobiology 422-423: 1-14.
King S.A. e Buckney R.T. 2000. Urbanization and exotic plants in northern Sydney streams. Austral
Ecology 25: 455-461.
Knight C.L., Briggs J.M. e Nellis M.D. 1994. Expansion of gallery forest on Konza Prairie Research
Natural Area, Kansas. Landscape Ecology 9: 117-125.
Kondolf G. M., Kattelmann R., Embury M. e Erman D.C. 1996. Status of riparian habitat. Sierra Nevada
Ecosystem Project: Final Report to Congress, vol. II,
Köpen W. 1931. Der klimate der erde, grundiss der klimakunde, Berlin.
Lange L. e Vanzon J.C. 1981. A system for the evaluation of aquatic biotopes based on the
composition of the macrophytic vegetation. Biological Conservation 25: 273-284.
Large A.R.G., Prach K., Bickerton M.A., Wade M. 1994. Alteration of patch boundaries on the
floodplain of the regulated river trent, UK. Regulated Rivers: Research and Management.
9:71-78.
Lavoie C., Jean M., Delisle F. e Létourneau G. 2003. Exotic plant species of the St Lawrence River
wetlands: a spatial and historical analysis. Journal of Biogeography 30: 537-549.
Lepart J. e Debussche M. 1991. Invasion processes as related with succession and disturbance. Em:
Groves R.H. e Di Castri F. (eds.), Biogeography of Mediterranean Invasions. Cambridge
University Press, Cambridge. pp. 159-177.
Lepart J. e Debussche M. 1992. Human impact on landscape patterning. Em: Hansen A.J. e di Castri F.
(eds.), Landscape Boundaries. Consequences for Biotic Diversity and Ecological Flows,
Ecological Studies 92,. Springer-Verlag, New-York, USA. pp. 76-106.
Levine J.M. 2001. Local interactions, dispersal, and native and exotic plant diversity along a
California stream. OIKOS 95: 397-408.
Lonsdale W.M. 1999. Concepts and Synthesis: Global patterns of plant invasions, and the concept of
invasibility. Ecology 80: 1522-1536.
Lousã M., Lopes M.C., Espírito-Santo M.D. e Costa J.C. 1998.O rio Nabão e afluentes: sua conservação
e valorização. Revista de Biologia 16: 157-171.
Malanson G.P. 1993. Riparian landscapes. Cambridge University Press, Cambridge.
Malanson G.P. e Cramer B.E. 1999. Landscape heterogeneity, connectivity and critical landscapes for
conservation. Diversity and Distributions 5:27-39.
Mandák B. e Pyšek P.1998. History of the spread and habitat preferences of Atriplex sagittata
(Chenopodiaceae) in the Czech Republic. Em: Starfinger U., Edwards K., Kowarik I. e
Williamson M. (eds.), Plant invasions: Ecological mechanisms and human responses,
Backhuys Publishers, Leiden. pp. 209-224.
Marcos N.R.R., Sequeira M.M. e de Koe T. 2002. Cartografia e caracterização da flora e da vegetação
do terço médio do Douro Internacional. Quercetea 3: 47-56.
Marks P.L. 1983. On the origin of the field plants of the northeastern United States. American
Naturalist 122: 210-228.
McIntyre S. e Hobbs R. 1999. A framework conceptualizing human effects on landscapes and its
relevance to management and research models. Conservation Biology 1282-1292.
McIntyre S. e Lavorel S. 1994. Predicting richness of native, rare and exotic plants in response to
habitat and disturbance variables across a variegated landscape. Conservation Biology 8(2):
521-531.
McKinney M.L. 2002. Do human activities raise species richness? Contrasting patterns in United States
plants and fishes. Global Ecology & Biogeography 11: 343-348.
Morais S. 1997. Corredores fluviais da bacia hidrográfica de Alcáçovas. Proposta de ordenamento.

Trabalho Final do Curso de Arquitectura Paisagística. Instituto Superior de Agronomia,
Lisboa.
Moreira I. e Duarte 2002. Capítulo 3. Comunidades vegetais aquáticas e ribeirinhas. Em: I. Moreira,
M.T. Ferreira, R. Cortes, P. Pinto e P.R. Almeida (eds), Ecossistemas Aquáticos e
Ribeirinhos. Ecologia, Gestão e Conservação. Instituto da Água, MCOTA. pp. 3.1-3.30.
Moreira I., Ferreira M.T., Aguiar F. e Duarte M.C. 2002b. Capítulo 4. Plantas infestantes e invasoras
de ecossistemas dulçaquícolas. Em: I. Moreira, M.T. Ferreira, R. Cortes, P. Pinto e P.R.
Almeida (eds), Ecossistemas Aquáticos e Ribeirinhos. Ecologia, Gestão e Conservação.
Instituto da Água, MCOTA. pp. 4.1-4.17.
Moreira I., Ferreira M.T., Cortes R., Pinto P. e Almeida P.R. (eds). 2002a. Ecossistemas Aquáticos e
Ribeirinhos. Ecologia, Gestão e Conservação. Ministério das Cidades, Ordenamento do
Território e Ambiente, Lisboa.
Moreira I., Ferreira M.T., Monteiro A., Catarino L.F. e Vasconcelos T. 1999. Aquatic weeds and their
management in Portugal: insights and the international context. Hydrobiologia 415: 229-
234.
Naiman R.J. e Décamps H. (eds.). 1990. The Ecology and Management of Aquatic-Terrestrial
Ecotones, Man and the Biosphere Series, 4, UNESCO, Paris.
Naiman R.J. e Décamps H. 1997. The ecology of interfaces: riparian zones. Annu. Ver. Ecol. Syst. 28:
621-658.
Naiman R.J., Décamps H. e Pollock M. 1993. The role of riparian corridors in maintaining regional
biodiversity. Ecological Applications 3: 209-212.
Naveh Z. e Vernet J.L. 1991. The palaeohistory of the Mediterranean biota. Em: Di Castri F. e Groves
R.H. (eds), Biogeography of Mediterranean Invasions. Cambridge University Press,
Cambridge. pp. 19-33.
Newbold J.D., Elwood J.W., O’Neill R.V. e Van Winkle W. 1981. Measuring nutrient spiraling in
streams. Canadian Journal of Fisheries and Aquatic Sciences 38: 860-863.
Nilsson C. 1996. Remediating river margin vegetation along fragmented and regulated rivers in the
north: what is possible? Regulated Rivers: Research and Management 12: 415-431.
Nilsson C. e Grelsson G. 1995. The fragility of ecosystems: a review. Journal of Applied Ecology 32:
677-692.
Nilsson C. e Jansson R. 1995. Floristic differences between riparian corridors of regulated and free-
flowing boreal rivers. Regulated Rivers: Research and Management 11: 55-66.
Nilsson C., Ekblad A., Dynesius M., Backe S., Gardfjell M., Carlberg B., Hellqvist S. e
Nilsson C., Grelsson G., Johansson M. e Sperens U. 1989. Patterns of plant species richness along
riverbanks. Ecology 70: 77-84.
Petts G.E. e Amoros C. (eds.) 1996. Fluvial Hydrosystems. Chapman & Hall, London, 322 p.
Pinay G., Décamps H. e Fustec E. 1990. Functions of ecotones in fluvial systems. Em: Naiman R.J. e
Décamps H. (eds.), The Ecology and Management of Aquatic-Terrestrial Ecotones, Man and
the Biosphere Series, 4, UNESCO, Paris. Pp. 141-170.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., De Ferrari C. e Décamps H. 1996. Invasibility of
species-rich community in riparian zones. Conservation Biology 10: 598-607.
Poff N.L., Allan J.D., Bain, M.B., Karr J.R., Prestegaard K.L., Richter B.D., Sparks R.E. e Stromberg
J.C. 1997. The natural flow regime. A paradigm of river conservation and restoration.
Bioscience 47(11): 769-784.
Pollard P. e Huxham M. 1998. The European Water Framework Directive: a new era in the
management of aquatic ecosystem health? Aquatic Conservation: Marine and Freshwater
Ecosystems 8: 773-792.
Poole G.C. Fluvial landscape ecology : addressing uniqueness within the river discontinuum.
Freshwater Biology 47: 641-660.
Prieur-Richard A.-H. e Lavorel S. 2000. Invasions: the perspective of diverse plant communities.
Austral Ecology 25: 1-7.
Pyšek P. e Prach K. 1994. How important are rivers for supporting plant invasions? Em: de Waal L.,
Child L., Wade P.M. and Brock J. (eds.), Ecology and Management of Invasive Riverside
Plants. John Wiley & Sons Ltd., Chichester. pp. 19-26.
Raunkiaer C. 1934. The Life Forms of Plants and Statistical Geography. Oxford Clarendon Press,
Oxford.
Rejmánek M. 2000. Invasive plants: approaches and predictions. Austral Ecology 25: 497-506.
Reynoldson T.B. e Wright J.E. 2000. The reference condition: problems and solutions. Em: Wright
J.F., Sutcliffe D.W. e Furse M.T. (eds.), Assessing Biological Quality of Freshwaters,
RIVPACS and other techniques. The Freshwater Biological Association, Cumbria. pp. 293-
304.
Reynoldson T.B., Norris R.H., Resh V.H., Day K.E. e Rosenberg D.M. 1997. The reference condition: a
comparison of multimetric and multivariate approaches to assess water quality-impairment
using benthic macroinvertebrates. Journal of the North American Benthological Society 16:
833-852.
Richardson D.M., Pyšek P., Rejmanek M., Barbour M.G., Panetta F.D. and West C.J. 2000.
Naturalization and invasion of alien plants: concepts and definitions. Diversity and
Distributions 6: 93-107.
Rivas-Martínez S., Fernández-González F., Izco J., Loidi J., Lousã M. e Penas A. 2001. Syntaxonomical
checklist of vascular plant communities of Spain and Portugal to association level. Itenera
Geobotanica 14:5-541.
Rivas-Martínez S., Lousã M., Diaz T.E., Fernandez-Gonzalez F. e Costa J.C. 1990. La vegetación del
sur de Portugal (Sado, Alentejo Y Algarve). Itenera Geobotanica 3: 5-126.
Roth N.E., Allan J.D. e Erickson D.L. 1996. Landscape influences on stream, biotic integrity assessed
at multiple spatial scales. Landscape Ecology 11(3): 141-156.
Salinas M.J., Blanca G. e Romero T. 2000. Evaluating riparian vegetation in semi-arid Mediterranean
watercourses in the south-eastern Iberian Peninsula. Environmental Conservation 27(1): 24-
35.
Saraiva M.G.A.N. O rio como paisagem. Gestão de corredores fluviais no quadro do ordenamento do
território. Textos Universitários de Ciências Sociais e Humanas. Fundação Calouste
Gulbenkian, Lisboa.
Sax D.F. 2001. Latitudinal gradients and geographic ranges of exotic species: implications for
biogeography. Journal of Biogeography 28:139-150.
Sax D.F. e Brown J.H. 2000. The paradox of invasion. Global Ecology & Biogeography 9: 363-371.
Sculthorpe C.D. 1971. The Biology of Aquatic Vascular Plants. Edward Arnold Publ., London.
Silva A.M. 1948. Grande Dicionário de Língua Portuguesa, 10ª edição, Editora Confluência, Lisboa.
Stanford J.A. e Ward J.V. 1993. An ecosystem perspective of alluvial rivers: connectivity and the
hyporheic zone. Journal of the North American Benthological Society 12: 48-60.
Stanford J.A. e Ward J.V. 2001. Revisiting the serial discontinuity concept. Regulated Rivers:
Research and Management 17: 303-310.
Stohlgren T.J., Binkley D., Chong G.W., Kalkhan M.A., Schell L.D., Bull K.A., Otsuki Y., Newman G.,
Bashkin M. e Son Y. 1999. Exotic plants invade hot spots of native plant diversity.
Ecological Monographs 69(1): 25-46.
Stohlgren T.J., Bull K.A., Otsuki, Y., Villa C.A. e Lee M. 1998. Riparian zones as havens for exotic
plants species in the central grasslands. Plant Ecology 138: 113-125.
Stromberg J.C. 1993. Instream flow models for deciduous riparian vegetation within a semi-arid
region. Regulated Rivers: Research and Management 8: 225-335.
Tabacchi E., Planty-Tabacchi A.-M., Salinas M.J. e Décamps H. 1996. Landscape structure and
diversity in riparian plant communities: a longitudinal comparative study. Regulated
Tockner K., Malard F. e Ward J.V. 2000. An extension of the flood pulse concept. Hydrological
Processes 14: 2861-2883.
Torrinha F. 1939. Dicionário Português-Latino, 2ª edição-E. Editorial Domingos Barreira, Porto.
Vannote R.L., Minshall G.W., Cummins K.W., Sedell J.R. e Colbert C.E. 1980. The river continuum
concept. Canadian Journal of Fisheries and Aquatic Sciences 37: 130-137
Vasconcellos J.C. 1969. Noções sobre a morfologia externa das plantas superiores, 3ª ed. Ministério da
Economia, Direcção-Geral dos Serviços Agrícolas, Lisboa.
Vasconcellos J.C. 1970. Plantas (Angiospérmicas) Aquáticas e Ribeirinhas. Direcção-Geral dos Serviços
Florestais e Aquícolas, Lisboa.
Verdonschot P.F.M. 2000. Integrated ecological assessment methods as a basis for sustainable
catchment management. Hydrobiologia 422/423: 389-412
Wade P.M. 1990. General biology and ecology of aquatic weeds. Em: Pieterse A.H. e Murphy K.J.
(eds.), Aquatic Weeds. The Ecology and Management of Nuisance Aquatic Vegetation,
Oxford University Press, Oxford. pp. 17-30.
Ward J.V e Tockner K. 2001. Biodiversity: towards a unifying theme for river ecology. Freshwater
Biology 46: 807-819.
Ward J.V, Tockner K. e Schiemer F. 1999. Biodiversity of floodplain river ecosystems: ecotones and
connectivity. Regulated Rivers: Research and Management 15: 125-139.
Ward J.V. 1989. The four-dimensional nature of lotic ecosystems. Journal of North American
Benthological Society 8: 2-8.
Ward J.V. 1998. A running water perspective of ecotones, boundaries, and connectivity.
Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte
Ward J.V. e Stanford J.A. 1983. The serial discountinuity concept of lotic ecosystems. Em: Fontaine
T.D. e Bartell S.M. (eds.), Dynamics of Lotic Ecosystems. Ann Arbor Scientific Publishers:
Ann Arbor, MI 347-356.
Ward J.V. e Stanford J.A. 1995. The serial discontinuity concept: extending the model to floodplain
rivers. Regulated Rivers: Research and Management 10: 159-168.
Wetzel R. 1983. Limnology, 2nd ed., Saunders, New York.
With K.A. 2002. The landscape ecology of invasive spread. Conservation Biology 16(5): 1192-1203.
Wood P.J. e Armitage P.D. 1999. Sediment deposition in a small lowland stream – management
implications. Regulated Rivers: Research and Management 15: 199-210.
Wu J. e Loucks O. L. 1995. From balance of nature to hierarchical patch dynamics: a paradigm shift in
ecology. Quarterly Review of Biology 70: 439-466.
Parte II
As galerias ribeirinhas em
sistemas fluviais Mediterrânicos:
influência dos factores abióticos
e da perturbação antropogénica
Part II Riparian woody vegetation

In Mediterranean fluvial systems:
influence of abiotic factors and human
disturbance
Do not go where the path may lead,

go instead where there is no path and leave a trail.
Ralph Waldo Emerson (1803-1882)

Capítulo 3
Tipologia de galerias ribeirinhas
numa bacia mediterrânica
Chapter 3 Riparian types in a Mediterranean basin*
*
PUBLISHED AS: Aguiar F.C., Ferreira M.T., Moreira I.S. and Albuquerque A. 2000. Riparian types in a
Mediterranean basin. Aspects of Applied Biology 58: 221-232.
60 | Parte II - Influência Ambiental e Humana nas Galerias Ribeirinhas
Capítulo 3 – Riparian Types in a Mediterranean Basin | 61
3.1 Abstract
Riparian species assemblages from the Portuguese part of the Tagus basin in the Iberian
Peninsula were studied at 97 sites. Riparian vegetation assemblages were dominated by ash
(Fraxinus angustifolia), alder (Alnus glutinosa), black poplar (Populus nigra) and willows
(Salix atrocinerea, S. alba and S. salviifolia), frequently surrounded by edges of bramble-
thicket (Rubus ulmifolius). Other important species were the hawthorn (Crataegus
monogyna) and Erica arborea at about 20% frequency of occurrence in the studied area.
Five species groups were obtained using weighted pair-group arithmetic average clustering.
The distribution of the 17 major woody species was related to the prevailing environmental
gradients by means of correspondence analysis. Precipitation, temperature, altitude, order
number of the river, river bed substrate type, channel width and physical alteration of the
river banks were significantly related to species distribution. A canonical correspondence
analysis was also performed on 34 non-disturbed sites to obtain reference riparian types.
Reference species distribution was significantly related to precipitation, temperature,
altitude, percentage of bedrock on the riverbed and channel width. The structure of the
riparian groups was analysed using the ratio of occurrence of the main woody species per
cluster.
Key words: riparian species, woody riparian types, multivariate analysis, Tagus basin,
Iberian rivers, Portugal.
3.2 Introduction
Little is known about the ecology of Portuguese riparian formations. Past studies have
included some syntaxonomic approaches (Aguiar et al. 1995; Capelo 1996), broad latitudinal
and altitudinal variations in species composition (Ferreira and Lousã 1988) and vegetation
recovery following mechanical harvest and dredging (Ferreira and Moreira 1990). However,
there are no previous studies on the riparian structure and continuity along river corridors
and their related environmental factors. Furthermore, reference riparian community types
for each ecological region have not been defined, resulting in difficult assessment of the
ecological consequences of human disturbance and analysis of deviation from pristine
ecological conditions.
Riparian habitats in Mediterranean rivers have been subject to cutting and clearing for
centuries. Riparian species must have adapted to these more or less continuous man-made
disturbances that increase the level of high disturbance, typically characterising fluvial
ecotones (Pinay et al. 1990). Meta-stability of these communities (sensu Wissmar and
Swanson 1990) must have evolved in connection with this long lasting interference. The
present assemblage composition will also reflect environmental gradients longitudinally
displayed by the lotic system, such as altitude and channel width, as well as regional
variables such as annual rainfall and temperature.
Currently, the major threats to Portuguese riparian formations, and indeed to the whole
river ecosystems, are the alteration of longitudinal and cross-section river profiles and
instream clear-cutting and dredging. These activities are undertaken for flood control
improvement, for land reclamation for agricultural and urban purposes and for the creation
of access to the river channel for cattle, recreation and angling. Furthermore, flow
regulation prevents the winter flushing of the materials deposited on the riverbed,
promoting invasive riparian species establishment (Johnson 1994), whilst extensive water
abstraction for irrigation decreases the groundwater level with the consequent loss of
vertical connectivity (Kondolff 1995).
3.3 Material and methods
3.3.1 Study area
This study was conducted on the Portuguese part of the Tagus basin, the third largest
hydrographic basin of the Iberian Peninsula. The Tagus basin covers 80 500 km2 and one
third of this area is Portuguese. At the Spanish border, the river presents a E-W orientation
and the main river and right margin tributaries run through a series of deep valleys of
ancient pré-hercinic formations, with gradients of about 0.3m km-1. Then, the orientation
becomes NE-SW and the river and all its tributaries spread over a series of large platforms
of tertiary sediments, limited on the northern edge by calcareous mesozoic formations.
Finally, it enters the quaternary alluvial plain before reaching the estuarine area.
The Tagus basin is located on the central part of the Iberian Peninsula and presents a
typical Mediterranean climate. Yearly average rainfall of the Portuguese part is 810 mm and
average annual temperature is 15ºC, however, regional differences within the basin are
very large, especially between northern and southern parts. The driest southern parts near
the border can experience average annual rainfall as low as 500mm and average annual
temperatures of 16.5ºC. Precipitation is highly irregular with 70% of the annual rain falling
between October and March.
The upper and middle part of the basin (Portuguese part) is occupied by Mediterranean oak
forests and Mediterranean shrubland, though areas of olive yards, Pinus and Eucalyptus also
occur frequently. The floodplain is used for rice fields, irrigated crops and cattle pastures.
The resident population in the basin (Portuguese part) is 2.9 million people, about 32% of
the Portuguese population, but they are concentrated in the littoral and estuarine areas. In
most of the basin, the population density is low, being generally less than 5 inhabitants
km2.
3.3.2 Survey methods
Sampling was conducted at 97 sites during late spring of 1997-98. Sites were defined as
250m long stretches of river and they were located to include all the main tributaries.
Riparian species were identified on site and given a relative cover class, from 0-absence to
5-nearly mono-specific, including trees, shrubs and woody climbers. Lateral limits
considered for the sampling area were the approximate maximum annual wetted width of
the river corridor.
The river corridor width and the wetted channel width were averaged from 10 measured
transepts. Average depth was determined using a graduated dip-net pole within each
sampling site in a number of points (mean = 25) selected by a random point of the finger
index (eyes closed). The percent area of different substrates was visually estimated at
these points, around a 0.5m radius circle and averaged. Substrate classes considered were
bedrock, large substrate (> 64 mm, including cobbles and boulders), medium substrate
(between 2-64 mm, gravel and pebble) and sand (< 2 mm) and fine particulate organic
matter. Distance from source (km), order number and altitude (m) were obtained for each
site from 1:50 000 topographical maps. Mean annual rainfall and mean annual temperature
were obtained by using data from the nearest weather station. The dominant geological
background was separated into four categories and used as dummy variables (0-1):
calcareous materials of jurassic origin, granitic rocks, metasedimentary rocks and
tertiary/quaternary deposits. Bank modification was visually evaluated, on a 1 to 5 scale,
including sand and gravel pits, resectioning, straightening and other physical modifications.
3.3.3 Data analysis
Canonical correspondence analysis was performed by the programme CANOCO version 4.0.
This was used to study the distribution trends of riparian vegetation and their relation to
environmental data (ter Braak and Smilauer 1998). Species occurring in less than 4 sites
were not considered in data analysis. Environmental variables were log transformed
(logx+1) and standardised by row centering (Ludwig and Reynolds 1988). From the 16
variables initially considered, only the subset of the 7 best predictors was retained for the
analysis, a forward selection procedure similar to a step-wise regression, available in
CANOCO (ter Braak 1986). A cut-off point of P<0.10 was used in this routine. A Monte Carlo
test was used to test the significance of both the first axis eigenvalue and the trace, the
sum of all eigenvalues (ter Braak 1986). A second CCA data analysis was performed for the
assessment of reference riparian types using only the 34 sites having low to no bank
modifications.
NTSYS-Pc version 2.0 (Rohlf 1993) was used on the vegetation data matrix, using the Bray-
Curtis coefficient of dissimilarity and the ‘weighted pair-group arithmetic average’
clustering method (WPGMA).
3.4 Results
3.4.1 Riparian composition
A total of 54 woody species were recorded in fluvial corridors of the Tagus basin and 89% of
these were trees and shrubs. The maximum richness per site was 14 and the mean species
richness was 5.6 (± 2.3). One site had no woody species while one third of the species were
found in a single site and half of the species occurred in only one or two sites. A large part
of these low frequency occurrences resulted from plantations in the riparian zone, mainly
of species from the genera Acer, Platanus, Eucalyptus, Pinus, Prunus and Castanea. Analysis
also showed that some terrestrial shrubs typical of the Mediterranean landscape colonise
riverbanks, mainly Ulex minor, Pistacia lenticus, Myrtus communis, Rhamnus alaternus,
Erica sp. and Cistus sp., but again with a low frequency of occurrence. Alien woody flora
represented 14% of the species found and only Acacia dealbata occurred at a significant
frequency (more than 30% of sites).
The riparian formations were dominated by ash (Fraxinus angustifolia), alder (Alnus
glutinosa), black poplar (Populus nigra) and willows (Salix atrocinerea, S. alba and S.
salviifolia), frequently surrounded by edges of bramble-thicket (Rubus ulmifolius). Other
important species were the hawthorn (Crataegus monogyna) and Erica arborea at about 20%
frequency of occurrence in the studied area.
3.4.2 Site and species grouping
Five groups of sites were obtained with the cluster analysis, using woody species cover. Two
of them occurred in areas with higher rainfall, altitude and percentage bedrock (black
symbols in Fig. 3.1). Dominant species in these two groups include Erica arborea, Sambucus
nigra, Frangula alnus and Salix atrocinerea and Alnus glutinosa in one of them.
3.0
1.0
CCA AXIS II
-1.0
-2
-3.0
-4
-1.5 -0.5 0.5 1.5 2.5

-2 -1 0 1 2
CCA AXIS I
Figure 3.1 Axes I and II of the canonical correspondence analysis, showing the spatial distribution of sampling
sites superimposed by the groups obtained from weighted pair-group average clustering using Bray-Curtis
similarity coefficient.
The other three groups are more related to lower altitude but higher annual temperature,
order number, channel width and bank modifications, hence, the river valleys and the
driest parts of the basin (white symbols in Fig. 3.1). Dominant species in these three groups
include Salix salviifolia, Tamarix africana and Salix alba; Ulmus minor, Securinega tinctoria
and Fraxinus angustifolia; Populus nigra, Salix babylonica and Tamarix africana. The group
partition, however, was made at 80% of similarity, which means the distinction between
groups was rather tenuous. Though the first two groups (black symbols) can be related to
the upper areas of the basin whilst the other two mainly occurred in the southern and lower
parts of the basin, the geographical affiliation of the site grouping is weak. In fact, group
separation is mostly dependent of the relative abundance of the tree species and the
presence of typical woody shrubs.
3.4.3 Environmental gradients
Results of the canonical correspondence analysis are shown in Fig. 3.1 (site position on the
first two axes superimposed by the group affiliation) and Fig. 3.2 (17 riparian woody species
and the environmental variables selected by the forward selection procedure). The first two
CCA axes explained 11.5% of the total biological variability and 60.3% of the species-
environment relationship (Table 3.1).
Temperature, altitude and rainfall were highly correlated (r2>0.75, P>0.01) with the first
two removed from the analysis to avoid co-linearity. From the 16 variables originally
considered, only 7 were retained by the forward selection procedure and were significantly
related to the species distribution using the Monte Carlo permutation test (Fig. 3.2). The
influence of environmental variables on riparian composition was significant, P>0.005, for
both the first axis eigenvalue and the trace (Monte Carlo simulation test with 1000
permutations). Consequently, it can be concluded that the riparian composition and the
variables extracted were related. These variables accounted for 19.1% of the species
distribution.
The first axis displayed a gradient of precipitation (and also altitude and temperature) and
was positively related to alder and black willow (Salix atrocinerea) assemblage and
associated riparian shrubs – Frangula alnus, Erica arborea and Sambucus nigra (Fig. 3.2).
Typical rivers on the positive side of this axis are small sized and permanent, running
through V-shaped valleys with considerable gradients. Rivers on the negative side tend to
be warmer and have wider channels, higher order number and softer bottoms and typically
include Tamarix africana and S. alba ssp. vitellina.
2.5
2
Salix babylonica
Sambucus nigra
1.5
Salix alba ssp. alba
BANKMOD JURASS
1
Hedera helix
0.5
Rubus ulmifolius
CCA AXIS II
GRAVEL PREC
Salix atrocinerea
Populus nigra
0 Ulmus minor
Tamarix africana Frangula alnus
Salix salviifolia Alnus glutinosa
Salix alba ssp. vitellina Fraxinus angustifolia
-0.5
Erica arborea
ORDER
Crataegus monogyna
-1
BEDROCK
-1.5
CWIDTH
-2
Securinega tinctoria
-2.5
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
CCA AXIS I
Figure 3.2 Axes I and Axis II of the canonical correspondence analysis, showing the spatial distribution of the
woody species and the environmental variables. BANKMOD- anthropogenic bank modification; BEDROCK-
bedrock; GRAVEL- gravel and pebble; JURASS – Jurassic geological elements; ORDER- number of order; PREC -
average annual precipitation; CWIDTH- average channel width.
The second axis is positively related to riverbank alterations as well as the presence of
Jurassic geological elements, a near-by urban landscape on the lower valleys leading to the
presence of S. alba ssp. alba and S. babylonica. A group of species is distributed throughout
the Tagus basin, namely ashes, Salix salviifolia ssp. salviifolia and the liana bramble-
thicket.
3.4.4 Reference riparian types
A CCA of 34 reference sites (those presenting vestigial physical modifications of the

riverbanks) was performed to extract the human influence and define woody vegetation
types in less disturbed habitats (Fig. 3.3 and Table 3.1). About half of the reference sites
occurred in the northern non-agricultural and less disturbed part of the basin.
Table 3.1 Summary statistics for the canonical correspondence analysis including correlation of canonical axes
with the environmental variables. See Figures 3.2 and 3.3 for acronyms of the variables.
CCA (all sites) CCA (reference sites)
Environmental variables Axis I Axis II Axis I Axis II
GRAVEL -0.1027 0.0730

BEDROCK 0.4052 -0.3423 0.4554 0.3883
CWIDTH -0.2400 -0.4099 -0.5968 -0.4456
BMOD -0.3296 0.2883
ORDER -0.2296 -0.1480
PREC 0.7052 0.0418 0.5846 -0.4296
JURASSIC - 0.0364 0.2741
Eigenvalue 0.191 0.082 0.237 0.119
Cumulative percentage variance of 42.2 60.3 54.5 82.0
species-environment relation
Trace 19.1 % 21.5%
The biplot of environmental variables and woody species is displayed in Fig. 3.3. The rela-
tive position of the 17 species in the biplot is rather similar to the one obtained for all sites
in terms of relative position of the species in the axes plane and also vegetation res-ponse
to environmental gradients, though less variables are now significant to the species
distribution.
The highest altitude and precipitation, lowest temperature, bedrock on the streambed and
smaller channel width are still related to alder and black willow (Salix atrocinerea) as-
semblage and associated riparian shrubs – Frangula alnus, Erica arborea and Sambucus
nigra. The opposite values of these variables were especially related to the presence of
Tamarix africana.
The environmental gradient of the second axis is less obvious but can still be related to so-
me human influence through Salix babylonica (mostly planted) and Salix alba ssp. vitellina,
an invasive species on the regulated river corridors of the Tagus basin (Ferreira and
Figueiredo 1990).
Only three variables were retained by the forward selection. These variables accounted for
21.5% of the species distribution and this variability was totally contained on the first three
axes of the ordination. So, Fig. 3.4 shows the spatial distribution of reference sites on the
first three axes of CCA again superimposed by the groups formed by the reference sites
obtained from weighted pair-group average clustering method using Bray-Curtis similarity
coefficient. In this data analysis, 3 groups were obtained from clustering at 80% of
dissimilarity, again leading to the segregation between upper stream sites (black symbols)
and the other sites, warmer and/or drier.
Crataegus monogyna
1.5
Erica arborea
Salix alba ssp. vitellina
1
BEDROCK
Frangula alnus
0.5 Securinega tinctoria
Fraxinus angustifolia Hedera helix
Salix alba ssp. alba
Rubus ulmifolius
0
CCA AXIS II
Tamarix africana PREC

Alnus glutinosa
CWIDTH Salix salviifolia
-0.5 Salix atrocinerea
Ulmus minor
-1 Sambucus nigra
Populus nigra
-1.5
-2
-2.5
Salix babylonica
-3
-1.5 -0.5 0.5 1.5
-2 -1 0 1 2
CCA AXIS I
Figure 3.3 Axes I and II of the canonical correspondence analysis of reference sites showing the spatial
distribution of the main woody species and of the environmental variables. BEDROCK-bedrock; PREC - average
annual precipitation; CWIDTH- average channel width.
3.4.5 Riparian structure
Table 3.2 summarises the global structure of the woody vegetation using the proportion of
occurrence of the main woody riparian species per cluster (all sites and reference sites) and
their total frequency in the studied area.
The sites represented by a black lozenge are located on the heads, receive more
precipitation, present a large percentage of hard materials on the riverbed and are scarcely
disturbed (Fig. 3.4). Alders, black willows and the shrubs Sambucus nigra, Frangula alnus
and Erica arborea are typically present at, or dominate on, these sites. The black circles on
both clusters consisted in only two sites, less species rich, dominated by Erica arborea and
Frangula alnus and without alders (Table 3.2).
Figure 3.4 Axes I, II and II of the canonical correspondence analysis showing the spatial distribution of
reference sites superimposed by the groups obtained from the weighted pair-group average clustering method
using Bray-Curtis similarity coefficient.
The white square group of the reference sites is scattered throughout the rest of the basin
and their riparian assemblages included mainly ashes, poplars, and willows (Salix alba and
S. salviifolia). Typical woody shrubs of this reference group were Crataegus monogyna,
Securinega tinctoria and Tamarix africana, accompanied by woody climbers. The white
symbols for all sites show a much more complex biological image, scattered into three
groups whose relative species composition showed minor differences. All 3 groups included
highly disturbed sites, where the riparian formations were composed in general by ashes,
poplars and Salix alba and S. salviifolia, accompanied in different parts of the basin and in
different degree by Ulmus minor, Salix babylonica and Tamarix africana. Though a large
part of these sites belong to the lower valleys with lower precipitation and with sand,
gravel and pebble as the main substrate. There were large variations in terms of
environmental features.
3.5 Discussion
Riparian habitats represent an important border between terrestrial and aquatic

ecosystems. Many human activities in the landscape surrounding the river corridor, as well
as those using instream habitats, need to cross this biological boundary and have
contributed constantly to its change over the centuries. Today, virtually no pristine riparian
communities may exist in Europe.
Riparian habitats are widely discussed in limnological literature, mostly in studies of their
buffer capacity (Pinay and Décamps 1988), for maintaining regional diversity and ecotone
functioning (Naiman et al. 1993; Naiman and Décamps 1997), as organic matter contributors
(Benfield 1997; Xiong and Nilsson 1997) or as habitat and structural elements for other
communities and the ecosystem (Gregory 1992; O’Laughlin and Belt 1995). However, there
are surprisingly few studies on the ecology of riparian habitats that include data on
compositional types or other biological features, such as average width, longitudinal
continuity, vertical connectivity, size structure and productivity, or association with
regional landscape features (e.g. Mason et al. 1984; Szaro 1990). In Iberian rivers,
management of riparian habitats is frequently carried out irrespective of regional riparian
characteristics and is often aimed at production of a dense canopy rather than the
reconstruction of near natural assemblages with their proper composition, dimension and
structure.
In this study, Iberian riparian assemblages have been shown to be poor in species and very
similar throughout the basin, in spite of considerable environmental variations. The total
number of species found was larger when considering non-riverine species, however, the
frequency of occurrence of these was very small and they could not be used as typical
components of the riparian habitats. Similar features, and even species, have been
reported for other woody plant communities in England (Mason and MacDonald 1990), but
woody flora was richer in South-western France (Tabacchi and Planty-Tabacchi 1990). Other
studies on Mediterranean type climates from North-America indicate much richer riparian
assemblages, though again, many of the species were non-riverine (Szaro 1990). The
apparent paucity of Iberian riparian species assemblages might be in part related with the
more humanised terrestrial landscape, whilst the wilderness of North-American landscapes
might contribute to their richness.
In the Tagus basin, the floristic entanglement of all sites grouping reduced the cohesiveness
of geographic affiliation, although ordination showed a clear gradient between northern
and cooler permanent rivers and the more drier and warmer southern tributaries and river
lowlands. The major factor of human influence is decisive to this entanglement, and the
difficulty of obtaining distinct regional groupings. In this study, an attempt was made to
derive reference riparian assemblages from reference sites where little physical
modifications were observed. This methodology increased species group’s differentiation
and enabled a better distinction between regional riparian composition to be made.
The establishment of reference sites is an essential and fundamental necessity for

understanding riparian ecology, and the selection of these is a critical step in the process.
Direct physical modifications of the bank and river habitats were used in this study, but
other less obvious interference will probably occur at the same time. Among the most
important indirect interferences are the lowering of the groundwater level, soil enrichment
of the banks and profound changes of the flow regime induced by regulation (Nilsson et al.
1993). For example, Nilsson et al. (1991) found that total species-richness, total percentage
cover and the proportion of perennial species-richness were lower in regulated rivers,
mostly for trees and shrubs. Past interferences such as urban development, tree clearance
and mining strongly affects riparian species composition (Décamps et al. 1988). Historical
records are extremely rare in Portugal and further work has to be done on the assessment
of the indirect and historic components of human influence on riparian assemblages.
In this study, canonical correspondence analysis indicated broad geographical variables as

being most important to riparian species distribution (altitude, temperature and
precipitation), followed by smaller scale hierarchical variables at the reach level (channel
width and hard substrate on the riverbed). However, the biological variability explained by
these variables was relatively small, being less than one third of the total species variation.
Elevation was also found to be the most heavily loaded factor along the first canonical axis
for all Arizona and New Mexico riparian communities types, although stream bearing and
gradient were also important (Szaro 1990). The effect of scale will probably also affect site
grouping and related environmental variables, and there may be more cohesive species
groups when considering wider areas and longer environmental gradients as Myklestad and
Birks (1993) have found for the distribution of Salix species in Europe.
3.6 Acknowledgements
This research has been financed by the Fundação para a Ciência e a Tecnologia, Contract
Praxis XXI/3/3.2/FLOR/2310/95 and Contract PBIC/C/AGR/226895. F.C. Aguiar was
supported by a grant from Fundação para a Ciência e Tecnologia.
3.7 References
Aguiar C.F., Capelo J., Costa J.C., Espírito-Santo M.D. and Lousã M. 1995. Tipologia das geoséries
mediterrânicas de Portugal Continental. Proceedings 3rd National Congress on Nature
Conservation: 25-32.
Benfield E.F. 1997. Comparison of litterfall input in streams. Journal of the North American
Capelo J.H. 1996. Esboço da paisagem vegetal da bacia portuguesa do rio Guadiana. Silva Lusitana, 4:
13-64.
Décamps H., Fortune M., Gazelle F. and Patou G. 1988. Historical influence of man on the riparian
Ferreira M.T. and Figueiredo J.C. 1994. Spatial evolution of river formations on a partially regulated
river. Limnetica 10(1): 53-57.
Ferreira M.T. and Moreira I. 1990. Impacte das actividades agrícolas sobre a dinâmica e evolução de
macrófitos aquáticos do Vale do Sorraia. Proceedings 2nd National Conference on
Environmental Quality: 39-48.
Ferreira M.T. and Lousã M. 1988. Fitocenoses ripícolas em Portugal. Proceedings of the 1st National
Forestry Congress: 264-268.
Gregory K.J. 1992. Vegetation and river channel process interactions. In: Boon P.J., Calow P. and
Petts G.E. (eds), River Conservation and Management: John Wiley & Sons, Chichester. pp.
225-270.
Johnson W.C. 1994. Woodland expansion in the Platte River, Nebraska: patterns and causes.
Kondolf G.M. 1995. Geomorphological stream channel classification in aquatic habitat restoration:
uses and limitations. Aquatic Conservation: Marine and Freshwater Ecosystems 5: 127-141.
Mason C.F. and MacDonald S.M. 1990. The riparian woody plant community of regulated rivers in
Eastern England. Regulated Rivers: Research and Management 5: 159-166.
Mason C.F., MacDonald S.M. and Hussey A. 1984. Structure, management and conservation value of
the riparian woody plant community. Biological Conservation 29: 201-216.
Myklestad A. and Birks H.J.B. 1993. A numerical analysis of the distribution patterns of Salix L.
species in Europe. Journal of Biogeography 20:1-32.
Naiman R.J. and Décamps H. 1997. The ecology of interfaces: riparian zones. Annual Review of
Ecology 28: 621-658.
Naiman R.J., Décamps H. and Pollock M. 1993. The role of riparian corridors in maintaining regional
diversity. Ecological Applications 3: 209-212.
Nilsson C., Ekblad A., Gardfjell M. and Carlberg B. 1991. Long-term effects of river regulation on river
margin vegetation. Journal of Applied Ecology 28: 963-987.
Nilsson C., Nilsson E., Johansson M.E., Dynesius M., Grelsson G., Xiong S., Jansson R. and Danvind M.
1993. Processes structuring riparian vegetation. Current Topics in Botanical Research 1:
419-431.
O’Laughlin J. and Belt G.H. 1995. Functional approaches to riparian buffer strip design. Journal of
Forestry February: 29-32.
Pinay G., Décamps H., Chauvet E. and Fustec E. 1990. Functions of ecotones in fluvial systems. In:
Naiman R. and Décamps H. (eds.), The Ecology and Management of Aquatic-Terrestrial
Ecotones, Man and the Biosphere Series, 4, UNESCO, Paris. pp. 141-169.
Pinay G. and Décamps H. 1988. The role of riparian woods in regulating nitrogen fluxes between the
alluvial aquifer and surface water: a conceptual model. Regulated Rivers: Research and
Rohlf F.J. 1993. NTSYS-Pc, Numerical Taxonomy and Multivariate Analysis System, Ver. 1.80. Exeter
Software, Setauket, NY.
Szaro R.C. 1990. Southwestern riparian plant communities: site characteristics, tree species
distributions and size-class structures. Forest, Ecology and Management 33/34: 315-334.
Tabacchi E. and Planty-Tabacchi A. 1990. Evolution longitudinale de la végétation du corridor de

l’Adour. Botanica Pirenaico-Cantabrica 5: 455-468.
ter Braak C.J.F. 1986. Canonical correspondence analysis: a new technique for multivariate direct
gradient analysis. Ecology 67: 1167-1179.
ter Braak C.J.F. and Smilauer P. 1998. CANOCO. Reference manual and user’s guide to Canoco for
Windows: software for Canonical Community Ordination, Ver. 4. Ithaca, NY.
Wissmar R.C. and Swanson F.J. 1990. Landscape disturbances and lotic ecotones. In: Naiman R. and
Décamps H. (eds.), The Ecology and Management of Aquatic-Terrestrial Ecotones Man and
the Biosphere Series, 4, UNESCO, Paris. pp. 65-89.
Xiong S and Nilsson C. 1997. Dynamics of leaf litter accumulation and its effects on riparian
vegetation: a review. The Botanical Review 63(3): 240-264.
Capítulo 4
Influência antrópica na composição e integridade
das galerias ribeirinhas na bacia do rio Tejo
Chapter 4 Human-disturbed landscapes: effects

on composition and integrity of riparian woody
*
vegetation in the Tagus River basin, Portugal
*
SUBMITTED TO: Environmental Conservation, with authors F.C. Aguiar and M.T. Ferreira.
Capítulo 4 - Human Disturbance Effects on Riparian Woods | 79
4.1 Abstract
The spatial distribution of riparian composition and cover was examined in eight river
basins belonging to the Tagus fluvial system, with varying land-use and environmental
settings. Field surveys, interpretation of aerial photographs and geographical information
systems provided information on riparian cover, species distribution, land-use and
environmental features. The main goal was to determine the relative contribution of land-
use and environment to the integrity of riparian woods and their floristic composition.
Related objectives were to identify the main types of land-use and environment that
contribute to biotic variation. This study revealed a patchy pattern of establishment of
riparian woods, along with generally low average width and low species richness, as well as
significant inter-basin differences and upstream-downstream variations in riparian features.
Whereas species distribution was clearly determined by environmental features, the abiotic
environment and the use of land in the river valley explained riparian integrity and other
structural features to a similar, but modest extent.
Key words: riparian forests, Mediterranean streams, environmental factors, urbanization,

agricultural landscapes, aerial photographs.
4.2 Introduction
Mediterranean riparian landscapes are typically shaped by human disturbances and

intensive land-use (Décamps et al. 1988; Lepart and Debussche 1992; Stromberg 1993;
Gallego-Fernandez et al. 1999; Corbacho et al. 2003) and may cause considerable
patchiness in both the longitudinal continuity of riparian formations and the land adjacent
to river margins (Gumiero and Salmoiraghi 1997). In contrast to riparian forests in
temperate regions, the riparian woody formations in the Mediterranean area are clearly
discernible from the surrounding matrix and have been described as narrow strips along
watercourses that function as corridors within the landscapes and are constrained by the
adjacent terrestrial systems (Malanson 1993; Ferreira and Moreira 1999; Salinas et al.
2000). In addition to the determinant role of the riparian vegetation cover and the
importance of the fluvial corridor width to the sustainability of plant, mammal, bird
(Spackman and Hughes 1995), macroinvertebrate (Tait et al. 1994; Aguiar et al. 2002) and
fish (Roth et al. 1996; Jones et al. 1999) communities, the composition and fragmental
patterns of riparian corridors play a fundamental role in the ecological integrity of riparian
ecosystems (Malanson and Cramer 1999; Aguiar et al. 2000; Allen and O’Connor 2000) and
should be analysed together. In general, under semi-arid conditions riparian vegetation
displays a pronounced spatial variability in species richness, composition and density from
headwaters to lowlands, depending on the environmental setting and the land-use in the
river floodplain (Ward and Stanford 1995; Tabacchi et al. 1996).
Aguiar et al. (2000) highlighted a complex entanglement of floristic riparian groups in the
Tagus basin and attributed them mainly to small-scale and long-lasting anthropogenic
disturbances, such as forestation and the agricultural land-use of river valleys. A recent
study in a South-eastern Iberian basin suggested that water availability and geological
background were related to land-use and disturbance and together contributed to
determining plant distribution (Ferreira et al. 2001). In the wetter range of Mediterranean-
type streams, riparian and aquatic vegetation are highly responsive to flooding and
hydrograph dynamics (Stromberg 1993; Décamps et al. 1995; Ferreira and Moreira 1999) and
thus also to adverse human disturbances, such as water diversion and flow regulation.
However, with increasing aridity, typical species assemblages are composed mainly of
sclerophyllous and evergreen shrubs, such as the Ibero-African shrubby spurge Flueggea
tinctoria (L.) G.L. Webster, and are thus more resilient and adapted to water stress and
poor substrates (Gasith and Resh 1999; Aguiar et al. 2002) than winter deciduous species
like Ulmus minor Miller and Fraxinus angustifolia Vahl. Although the floristic patterns in
Tagus riparian corridors have already been addressed, the links between the distribution
and structure of riparian woody species, and the spatial trends of riparian cover are not
known on a broader scale. So far, the majority of the studies in Iberia have been designed
to describe and locate riparian vegetation types and to assess their conservation value.
What is more, little is known about the worldwide longitudinal continuity and fragmental
patterns of riparian corridors (Mouw and Alaback 2003). In addition, most of the
representative research has been conducted in temperate regions and boreal rivers, where
the flora and the climatic and geomorphic setting differ markedly from the Mediterranean
(Salinas et al. 2000). Given the spatial complexity of land-use on the Tagus floodplain,
together with the region’s intricate geomorphology and the specificity of Mediterranean
flora, our intention was to: (1) outline the spatial patterns of riparian formations in terms
of both riparian composition and cover on a landscape scale; (2) assess the relationships
between species composition and the fragmental patterns of riparian galleries; (3) identify
both a set of significant explanatory variables related to specific contributory types of land-
use, and environmental factors; and (4) quantify the relative contributions that land-use
and abiotic factors make to riparian integrity and floristic composition. A comprehensive
understanding of the riverine landscape would improve the management strategies and
mitigation processes in riparian corridors.
4.3 Material and methods
4.3.1 Study area
The Tagus basin (80 500 km2) is located in the central part of the Iberian Peninsula. Around
a third of it is in Portugal, where the river flows in a northeast/ southwest direction,
towards the Atlantic Ocean (Fig.1, upper right). In rough terms, the main course of the
Tagus separates the mountainous hardrock areas in the north from the extensive areas of
limestone and other sedimentary strata (mostly plateaus or plains) in the south. The
Portuguese part of the Tagus basin, which is where this study was conducted, presents a
characteristic Mediterranean climate, albeit modulated by altitude and distance from the
Atlantic Ocean. The rainfall pattern of Mediterranean-type streams exhibits strong seasonal
and interannual variability, as well as a discharge-type regime (Gasith and Resh 1999). High
floods usually occur in autumn or early winter, with a gradual decline in flow and a
subsequent drying out during late spring and summer. Average annual rainfall varies from
over 2600 mm in the north-eastern area to under 500 mm in the south-western part. Typical
ranges for other meteorological parameters include 6 °C to 22 °C for monthly mean
temperatures in the northwest and 13 °C to 27 °C in the southeast; 6 °C to 10 °C for daily
thermal amplitudes; 60% to 80% for monthly mean relative humidity; and 7 MJ/m2 to 27
MJ/m2 for global horizontal solar irradiation. The existence of various reservoirs, both on
the Tagus River itself and on some tributaries, permits a degree of flood and draught
control.
Eight watersheds in the Portuguese part of the Tagus basin were selected according to
general differences in land-use and environmental setting (Fig. 4.1, lower left).
The Zêzere, Pônsul and Meimoa river basins are located above the Hercinic basement. This
complex geological formation is composed of the oldest materials in the Iberian Peninsula,
including granites, schists and quartzites, with a variable degree of metamorphism. The
Pônsul valley, the upper part of Zêzere valley and its 5th order tributary, the Meimoa, are
mainly occupied by Mediterranean scrubland and oak forests, though areas of olive groves
and pine also occur.
Figure 4.1 Iberian Peninsula, showing the Tagus basin and the location of the eight studied basins.
The River Zêzere is the major Portuguese tributary of the River Tagus and its lower course
is interrupted by three sequential reservoirs. The headwaters of the Zêzere cross the
granite and schist escarpments of the Serra da Estrela through V-shaped valleys and thence
develop sinuously onto the floodplain. The River Nabão also presents a narrow gorge on the
headwater stretch, before flowing down into a wide, fertile valley, which is occupied by
small-scale agriculture, urban settlements and industrial areas. This abrupt change marks
the boundary between the hard Mesozoic limestones and the soft clay and marl deposits of
the ceno-anthropozoic basins of the Rivers Tagus and Sado. Similarly, the presence of more
resistant shales and gneiss rocks makes the valley turn back into a deep gorge until it joins
the Zêzere valley.
The Trancão basin develops on a Mesozoic igneous complex, the floodplain of which is
covered by sedimentary and alluvial deposits. Because it is near Lisbon, numerous human
settlements are established along its course. The river is channelized for about 5 kilometres
before entering the Tagus estuarine area.
Three watersheds were chosen in the driest part of the Tagus basin, where the landscape is
dominated by cork-oak (Quercus suber L.) and holm-oak (Quercus rotundifolia Lam.)
woodlands and Mediterranean sclerophyllous shrubs. Most of these areas are composed of
metasedimentary and metavulcanic rocks below alluvial Cenozoic deposits. Table 4.1
summarizes the main geographical and environmental features of the studied basins.
4.3.2 Woody plant sampling and analysis
A floristic survey was undertaken during late spring (May and early June) of 1998. Sampling
sites were defined as 250-m long sections of the riverbank, between the winter-spring high-
flood level and the summer low-water level, and established at 10-km intervals along each
studied fluvial corridor. Riparian woody species were recorded, and percentage cover was
estimated by eye and grouped into the following cover classes: 0: absence; 1: 1-10% canopy
cover; 2: 11-25%; 3 – 26-50%; 4: 51-75%, and 5: >75%. Species that could not be identified in
the field were collected and identified at the herbarium. The fluvial corridor width was
averaged from 10 measured transepts. The proportions of five particle-size classes of
substrate were visually estimated at these points around a 0.5 m radius circle and
averaged. Bank modification was visually evaluated (range 1-5), including sand and gravel
pits, resectioning, straightening and other physical alterations of the transversal or
longitudinal river profile (Appendix 4.1).
Univariate methods and indirect ordination techniques were used to describe the overall
spatial patterns. In order to test for significant inter-basin differences in species richness, a
Kruskall-Wallis one-way analysis of variance (one-way ANOVA) and non-parametric Mann-
Whitney U tests using STATISTICA/w 6.0 (StatSoft Inc. 2001) were performed, and
significance was assessed at p<0.05. So as to evaluate species distribution trends, the
species data set (21 species, 6 cover classes) was subjected to a correspondence analysis
ordination (CA) and a unimodal response model. Prior to analysis, species cover was
transformed using natural logarithms [y = ln (y + 1)]. In order to improve the ordination
trends, species that were present in less than one survey were removed and rare species
were downweighted.
Table 4.1 River length (km), extent of river length occupied by dams (%) and drainage area (km 2 ).
Minimum and maximum values for the altitude (m), valley slope (‰), fluvial corridor width (m),
Strahler’s order number, annual average rainfall (mm), annual average temperature (ºC) and annual
average humidity (%) of the studied segments.
Stream Length Dams (% Drainage Order Altitude Corridor Valley Annual Annual average Annual average
rainfall (mm) humidity (%)
287.7 5 25 - 100.0 19.5 - 41.6 4-30 16.3 - 16.4 632.3 - 708.0 71.4 - 75.7
758.0 6 20 - 332.5 1.5 - 119.3 2-20 15.1 - 16.1 614.1 - 766.0 68.1 - 70.3
628.5 6 70 - 312.5 6.4 - 40.2 2-20 15.8 - 16.3 619.0 - 730.6 67.8 - 70.2
1016.5 5 25 - 210.0 3.0 - 45.0 4-30 14.4 - 16.1 747.8 - 1109.2 68.5 - 73.3
1081.0 6 15 - 340.0 1.9 - 60.0 2-20 16.0 - 16.4 613.1 - 640.7 69.5 - 72.7
Zêzere 225.0 41.0 4995.7 7 15 -1700.0 6.3 - 145.5 3-45 11.2 - 16.1 751.6 - 1437.0 69.0 - 71.8
1486.6 7 80 - 580.0 5.5 - 122.8 4-30 13.6 - 14.9 977.8 - 1186.6 67.4 - 69.1
481.8 5 390 -1010.0 4.5 - 28.5 4-30 11.8 - 12.4 1052.6 - 1390.9 69.9 - 70.5
(‰) temperature
(km) of river area number (m) width (m) slope average
(ºC)
length) (km2)
Meimoa 59.1 11.2

Pônsul 77.5 5.2
74.0 3.5
Trancão 29.9 -
63.5 -
Nabão 65.9 -
Canha 100.0 -
Divor
Tera
4.3.3 Integrity of riparian galleries and analysis
1:15000 scale monochromatic aerial photographs from the Portuguese Institute of

Cartography were used to assess land-use, fluvial corridor width and the longitudinal
integrity of riparian galleries. Flyover dates were between summer of 1989 and summer of
1995, depending on the material available. We used a WILD ST4™ mirror stereoscope with
an 8x binocular magnifier to analyse 580 vertical aerial photographs. There was an
approximately 60% forward overlap between successive photographs in the same flightline
and around 30% of sidelap between adjacent flightlines. To reduce lens distortion, we
conjugated the principal point - i.e. we made an image of the focal point (achieved from
the joining of fiducial marks) on the overlapping photograph of each stereo pair and worked
only on the limited non-distortion area. Focal points for each photograph were transferred
to a Geographic Information System database in order to collect geographical and
environmental data. The main river courses of each basin were divided into 3-km long river
reaches (survey units, SU) with a Scale Master™ curvimeter. Within each SU two transepts
perpendicular to the stream were positioned on each margin in order to assess the land-use
of the river valley. The transepts were drawn up with a maximum length of 1000 m, or until
the hillside ridge. We first categorized the land-use types into nine classes, which we
considered to be representative of existing land-use: urban and industrial areas; rice fields
and irrigation crops; pastures and extensive crops; orchards, vineyards, olive groves;
scrubland, tilled land and fallow ground; oak and cork-oak woodlands; forest plantations;
mixed woodland; gravel and sand mining; and exposed rock. The percentage of land-use
types in each SU was obtained by measuring the transepts and averaged from the four
transepts pertaining to each SU. Within each SU the integrity of the riparian galleries was
assessed by the length of five types of stand canopy closure (C1: absent; C2: scattered or
semi-continuous on one margin and absent on the other; C3: scattered or semi-continuous
on both margins; C4: continuous on one side and absent or scattered on the other; and C5:
continuous on both margins), the number of cover types per SU (range 1-5), and the number
of times the gallery types changed in each SU (both considered a measure of patchiness).
Riparian width, canopy class length, land-use type length and fluvial corridor width were
measured with a Mitutoyo™ digital calliper (0.01 mm resolution), the former being
averaged from 12 randomly-taken measurements along each SU, and from the
measurements on each transept’s intersection with the river. Riparian width was measured
for the respective riverbank, from the stream to the outer edge of natural vegetation, crops
or bare ground adjacent to the stream. To correct for differences in riparian width due to
the longitudinal variation in fluvial corridor width, a transformed value was also used, in
which the (relative) riparian width = measured (absolute) riparian width/log10(corridor
width). So as to ensure the correct identification of the land-use classes and to eliminate
some doubts that resulted from the photo-interpretation, such as the riparian width in V-
shaped valleys, in the summer of 1998 field expeditions were undertaken in the Zêzere,
Meimoa and Canha basins. Climate data was obtained from a geo-statistical programme (P-
Clima 1999). Distance from source, altitude, Strahler’s order number, valley slope,
geological background and the existence of dams 10 km upstream from each SU were
obtained from a GIS (Geographic Information System), using 1:25 000 digital maps
(Appendix 4.1).
We studied the variation of riparian width along the upstream-downstream gradient for
each river. A smoothing curve was fitted according to the distance-weighted least squares
procedure to reveal non-salient patterns in the data and filter out noise, using a stiffness
value of 0.25. The measured riparian width and the transformed widths were strongly
correlated with one another in all the studied basins (Pearson product moment correlation
>|0.93|, P<0.05) and this enabled us to use absolute values. We used PCA to detect
environmental gradients and land-use differences between basins and within SU’s. The
variables were log transformed (logx+1) and standardised by row centring (Ludwig and
Reynolds 1988). PCA assumes a linear model for the relationship between the responses of
each species and the ordination axes (ter Braak and Prentice 1988).
4.3.4 Combined effects of land-use and environment
The relative contribution of land-use and environmental variables to an explanation of the

species composition and the longitudinal integrity was determined using partial canonical
correspondence analysis (pCCA) and partial redundancy analysis (pRDA), respectively. For
the ordination analysis of the riparian composition, three sets of data were drawn up (61
field surveys): the species data set, the environmental data set (18 environmental
variables) and the land-use data set (9 land-use categories). Similarly, three sets of data
were used for the ordination analysis of the riparian integrity (194 SU’s): the vegetational
data set (9 variables); the environmental data set (13 variables); and the land-use data set
(Appendix 4.1). To start with, in order to identify two subsets of the most effective
explanatory variables (those of the environmental data set and the land-use data set), we
carried out two CCA/RDA ordination runs with forward selection of environmental and land-
use variables, and unrestricted Monte Carlo permutation tests for each variable (ter Braak
1990; ter Braak and Verdonschot 1995). CCA and RDA are multivariate gradient analysis
methods that combine multiple regression and ordination techniques (ter Braak 1986; ter
Braak and Verdonschot 1995). CCA displays unimodal relationships between a set of
response variables and a set of explanatory variables, whereas RDA is the canonical form of
principal component analysis (Rao 1964) and selects the linear combination of
environmental variables that gives the smallest total residual sum of squares (Jongman et
al. 1997). A cut-off point of 0.10 (Magnan et al. 1994) was adopted. Variance inflation
factors for the explanatory variables were examined to detect co-linearity between the
latter. Next, two partial CCA/RDA ordinations were performed (ter Braak 1988) in order to
estimate the relative importance of environmental and land-use variables. Total variation
in the species/vegetational data matrices were partitioned into 4 components (ter Braak
1990; Borcard et al. 1992; Rodriguez and Magnan 1995): 1) the selected environmental
variables; 2) the selected land-use variables; 3) the environmental variables following the
removal of the influence of land-use variables; and 4) the land-use variables following the
removal of the influence of environmental variables. Each component of the variation was
obtained by dividing the canonical eigenvalues of a particular CCA/RDA (or partial
CCA/partial RDA) by the total inertia - i.e. the sum of all the eigenvalues of a
correspondence analysis of the macroinvertebrate abundance matrix. Pure environmental
variation was given by step 3, whereas pure land-use variation was given by step 4. The
total explained variation was obtained from the sum of steps 1 + 4 or 2 + 3, and the
variation shared by both environmental and land-use variables was obtained by subtracting
step 3 from step 1 or step 4 from step 2. For each CCA/RDA or partial CCA/partial RDA that
was performed, a Monte Carlo test (999 permutations) on both the 1st axis eigenvalue and
the trace (i.e., the sum of all canonical eigenvalues) was used to evaluate the significance
of the effects under analysis (ter Braak 1990). All ordinations were performed using the
CANOCO 4.0 programme (ter Braak and Smilauer 1998).
4.4 Results
4.4.1 Riparian composition
We recorded 39 woody species, of which 60% were considered riparian - i.e., related to
aquatic, waterlogged or wet habitats. However, approximately 70% of the woody species
recorded were infrequent - i.e. present at less than 5% of the total sampling sites. Ash,
white-willow (Salix salviifolia Brot.), black-willow (S. atrocinerea Brot.), alder [Alnus
glutinosa (L.) Gaertner], black-poplar (Populus nigra L.), and bramble-thicket (Rubus
ulmifolius Schott) were the most common woody species along the fluvial corridors.
Relatively pure alder communities were widespread throughout the studied basins and
colonized mainly 4th-order streams with permanent moist soils of silt and clay, whereas
remnant stands of Ulmus minor were only occasionally reported. Some exotic species, such
as S. alba L. ssp. vitellina (L.) Arcangelli, were amongst the most frequent and abundant
woody riparian species. In general, we observed a longitudinal zonation of the river-margin
vegetation from shrub and herbaceous communities at the headwaters to shrub and tree
species on the mid-course, but with no clear pattern in the lowland areas. The average
number of woody species per site was 5.7 ± 2.0, and maximum species richness was 12.
There were no significant differences in species richness between the studied basins
(Kruskall-Wallis One-Way ANOVA, p>0.05), or between the northern pluvious basins (Pônsul,
Zêzere, Meimoa and Nabão) and the driest area under study (Mann-Whitney Rank Sum test,
p>0.05). However, correspondence analysis (Figure 4.2a; cumulative variance in the first 2
axes = 16%; 28.7%) revealed floristic differences between the more pluvious basins and the
driest (and warmer) ones.
a)
Securinega tinctoria
Erica arborea
Nerium oleander
Crataegus monogyna
Tamarix africana
Salix alba ssp. vitellina Lonicera implexa
Fraxinus angustifolia Frangula alnus
Salix salvifolia
Salix alba ssp. alba Rubus ulmifolius
Populus nigra Rosa canina
Salix babylonica Salix atrocinerea
Solanum dulcamara Alnus glutinosa
Hedera helix
Acacia dealbata
Sambucus nigra
b)
Meimoa Zêzere Pônsul Nabão
Trancão Tera Divor Canha
Figure 4.2 Correspondence Analysis ordination of riparian species, showing the biplot scores of the survey
units and the species with the highest scores on the axes (Figure 4.2a), and Principal Components Analysis
ordination of vegetational variables, showing the plot scores of the survey units (Figure 4.2b). Acronyms are
given in Appendix 4.1.
Stream Survey units Riparian canopy closure classes (%) Canopy classes Riparian width
(number) changeover (m)
Table 4.2
C1 C2 C3 C4 C5
Zêzere 45 0 – 100 0 – 67.5 0 – 34.0 0 – 91.3 0 – 43.9 0 – 19 0 – 14.2
(54.2 ± 30.3) (24.3 ± 19.2) (6.8 ± 8.4) (11.0 ± 17.3) (3.7 ± 9.8) (5.9 ± 3.4) (4.1 ± 3.6)
floristic heterogeneity.
4.4.2 Riparian integrity

Pônsul 24 0 – 100 0 – 46.5 0 – 48.5 0 - 37.1 0 – 100 0 – 11 0 – 23.9
(53.7 ± 37.6) ( 14.8 ±15.2) (9.1 ± 15.8) (9.4 ± 13.2) (12.9 ± 22.6) (3.5 ± 3.1) (7.8 ± 8.3)
Meimoa 17 0 – 100 0 – 53.0 0 – 38.5 0 – 60.0 0 – 32.0 0 – 18 0 – 15.9
metres in the Canha River (Table 4.2).

(33.8 ± 31.3) (25.4 ± 14.8) (11.8 ± 13.5) (23.4 ± 21.3) (5.7 ± 9.8) (7.7 ± 5.1) (5.9 ± 4.7)
Nabão 20 0 – 100 0 – 39.0 0 – 43.9 0 – 54.3 0 – 62.2 0 – 19 0 – 17.1
(34.4 ± 31.2) ( 19.3 ±14.5 ) (13.6 ± 11.9) (18.2 ± 18.0 ) ( 14.5 ± 18.5) (7.7 ± 4.2) (7.1 ± 5.4)
Trancão 10 0 – 100 0 – 51.8 0 – 32.0 0 – 43.3 0 – 38.6 0 – 11 0 – 11.9
(52.7 ± 37.5) (20.9 ± 18.2) (10.3 ± 12.1) (6.6 ± 13.8) (5.7 ± 12.2) (4.9 ± 3.6) (3.4 ± 3.8)
Canha 34 0 – 100 0 – 57.8 0 – 69.1 0 – 57.0 0 – 77.2 0 – 16 0 – 22.2
(22.7 ± 32.1) (15.1 ± 13.6) (22.3 ± 17.1) (18.8 ± 16.4) (20.9 ± 23.9) (9.0 ± 4.4) (9.8 ± 5.8)
Divor 24 0 – 100 0 – 60.5 0 – 71.0 0 – 24.7 0 – 34.8 0 – 13 0 – 15.8
riparian woody cover classes, class changeover and riparian width of the studied streams.
(38.8 ± 29.1) ( 29.5 ± 18.4) ( 18.5 ± 20.9) (8.0 ± 9.3) (5.2 ± 8.5) (7.4 ± 3.1) (4.2 ± 4.4)
Tera 20 10.3 – 100 0 – 71.0 0 – 47.0 0 – 25.0 0 – 4.7 0 – 12 0 – 9.8
(46.3 ± 26.6) ( 38.9 ± 20.4) ( 9.6 ± 12.3) (4.8 ± 7.5) (0.5 ± 1.4) (6.4 ± 3.3) (2.9 ± 2.5)
Number of survey units. Minimum and maximum values, average ± SD (in parentheses) for the
riparian composition of the Tagus basin, whereas the second axis displays the intra-basin
Frangula alnus Miller, Acacia dealbata Link and Sambucus nigra L. for the first group, and
Pônsul), with an average riparian width ranging from 2,9 metres in the Tera River to 9,8
alba ssp. alba) for the second. The first CA axis accounts for the general differences in the
Tamarix africana Poiret, Nerium oleander L., and S. alba L. (S. alba ssp. vitellina and S.
The riparian galleries were consistently narrow, up to a maximum of 24 metres (observed in

These groups differed mainly in the cover of Erica arborea L., Crataegus monogyna Jacq.,
The highest riparian cover values per SU were found in the Canha corridors, while the Tera,
Divor, Trancão and Zêzere watercourses presented the lowest average values for riparian
width, as well as a low proportion of continuous riparian galleries. All the canopy closure
classes were present on all the rivers, though they covered a large range of values. In most
of the studied basins more than one third of the fluvial corridors presented no woody
vegetation. In addition, non-existent or scattered riparian canopies were frequent
throughout the SU’s, especially at the headwaters. Furthermore, the C4 and C5-classes
never exceeded 20% of total river length on the observed streams. In general, the canopy
closure class changeover along the SU’s revealed a patchy, fragmented riparian landscape
and attained 19 class changes per SU along the River Zêzere.
The general patterns of spatial variation in riparian width are shown in Fig. 4.3. The
riparian width differed both along the longitudinal river gradient and between river
corridors. Data from the Canha, Pônsul and Tera basins revealed wider riparian formations
at mid-course than in the lowland and headwater areas, whilst Divor, Meimoa and Nabão
presented an upstream-downstream increase in width. No clear spatial patterns emerged
from the Zêzere data, except for a tendency towards wider riparian areas on the upper
stretches. The Trancão margins also possessed narrower riparian corridors towards the
lowlands.
The results of the PCA applied to the vegetation data set are illustrated in Fig. 4.2b
(cumulative variance in the first 2 axes = 57.6 %; 74.9%). The first axis accounts for some of
the differences between the SU’s along each river, thus leading to a distinction between
SU’s with a relatively wider riparian width and more continuous riparian galleries on the
one hand and those with scattered vegetation and narrower riparian galleries on the other.
Canha was the most homogeneous fluvial corridor, while Zêzere, Pônsul and Nabão returned
the highest variability.
26
22
22 Pônsul River Canha River
18
18
14 14
10 10
6 6
2
2
-2
0 10 20 30 40 50 60 70 80 -2
3
78 hm
Distance from source (km) 0 20 40 60 80 100
10 12
Tera River 10 Trancão River
8
8
6
6
4
4
2
Riparian width (m)
2
0 0
-2 -2
0 10 20 30 40 50 60 0 4 8 12 16 20 24
10
14 Zêzere River 8 Divor River
10 6
4
6
2
2
0
-2 -2
11.9 hm 3
0 40 80 120 160 200 0 10 20 30 40 50 60

49 hm 3
D 719 hm 3 1100 hm 3 Distace from source (km)
18
18
Nabão River Meimoa River
14
14
10
10
6
6
2 2
-2 -2
0 10 20 30 40 50 60 0 10 40.9 hm 3 20 30 40 50
Distance from source (km)
Figure 4.3 Spatial variation of riparian width for the studied corridors. Note for different scales. Curves
were fitted according to the distance-weighted least squares. Location, extension and gross capacity of
reservoirs are given.
4.4.3 Land-use and environmental features
There were large differences in land-use between basins and within each basin along the
watercourses, although it was possible to observe similar global occupation in some basins,
namely Canha, Divor and Tera (Fig. 4.4). Most of the transepts of the Zêzere, Meimoa,
Nabão and Trancão intersect at least three or more land-use types, whereas Pônsul and the
southern basins displayed a more homogeneous landscape that was dominated by larger
land patches of oak woodlands, scrublands, non-irrigation crops and pastures.
100
80
Land use classes (%)
60
40
20
0
Zêzere Pônsul Meimoa Nabão Trancão Canha Divor Tera
River
MINE FORE QUER PINE SCRU ORCH PAST IRRI URBA
Figure 4.4 Global percentage values of the land-use types in the studied basins. Percentage values were
averaged from the transepts of the overall SU.
A noticeable feature in all the basins was the limited presence of ‘natural’ mixed forest
(Mediterranean woodland and scrubland) in the 1000-meter buffer studied. This type of
vegetation occurred mostly as relict patches on the limits of the valleys and on the
hillsides. This cover type was better represented along the upstream reaches of the Meimoa
basin, which is included in a natural reserve.
The Canha, Divor and Tera landscapes were dominated by holm-oak and cork-oak
woodlands, though both the area above the Divor reservoir and the Canha and Meimoa
lowlands were occupied by rice and irrigation crops. The Nabão and Zêzere hillsides were
extensively forested with Pinus pinaster Aiton and Eucalyptus globulus Labill., but their
floodplains were also occupied by irrigation crops, extensive agriculture, orchards and a
patchy mosaic pattern of small human settlements. The dominant features of the Trancão
basin and the lower Nabão were human settlements and industrial areas surrounded by
orchards.
Figure 4.5 shows the results of the principal component analysis (PCA) using the
environmental data set (4.5a) and the land-use data set (4.5b). The ordination diagram in
Fig. 4.5a reveals a considerable separation between the northern basins on the positive part
of the 2nd axis (eigenvalue=0.21) and the southern and central-western basins on the
negative part, which is determined by climatic and geomorphological variables. The 1st axis
(eigenvalue = 0.58) gives the upstream-downstream gradient of the river.
The PCA land-use plot (Figure 4.5b) distinguishes the basins dominated by Quercus L. sp.
(Canha, Tera, Divor and Pônsul) from those with a patchy pattern of land-use (cumulative
percentage variance axes I and II = 74.9%). The 2nd axes reflect some of the intra-basin
variation.
a)
b)
Figure 4.5 Principal components analysis (PCA) ordination of environmental variables (Figure 4.5a) and of
land-use variables (Figure 4.5b), showing the plot scores of the survey units. Acronyms are given in Appendix
4.1.
4.4.4 Combined influence of environment and land-use
The patterns of woody riparian composition and integrity and the relative selected
environmental and land-use variables are displayed in CCA and RDA biplots shown in Fig. 4.6
[CCA: (a) and (c), RDA:(b) and (d), respectively)].
Eight of the eighteen variables that were initially considered in the ‘environmental CCA’
were selected and together accounted for 28.8% of the species distribution [Figure 4.6(a)].
The first two CCA axes (eigenvalues 0.28 and 0.17, respectively) explained 15.7% of the
total biological variability and 54.4% of the species-environment relationship. Rivers on the
positive side of the 1st axis are permanent and run through V-shaped valleys with
considerable gradients. Rivers on the negative side tend to be drier and have higher order
numbers and softer bottoms. The 2nd axis reflects a slight disturbance gradient.
The ‘environmental RDA’ revealed that local variables, such as fluvial corridor width and
geological background, were related to riparian width and longitudinal integrity [Figure
4.6(b)]. However, the RDA eigenvalues for the 4 axes were extremely low (< 0.10), as was
the explained variance of the ‘species’ distribution (11.8%), though this variability was
totally contained on these axes.
Fig. 4.6(c) shows the CCA biplot of land-use variables and field surveys. Only three variables
were retained by the stepwise forward selection: “oak woodlands”, “scrublands and
pastures”, and “orchards, olive groves and vineyards”. These variables accounted for 13.1%
of the species distribution and this variability was totally contained on the first three axes
of the ordination. The 1st axis (eigenvalue = 0.19) grouped together the sites dominated by
Quercus sp. - mostly Pônsul and the southern basins.
a) b)
c) d)
Meimoa Zêzere Pônsul Nabão Trancão Tera Divor Canha
Figure 4.6 Canonical correspondence analysis (CCA) biplots for the selected environmental (a) and land-use
variables (c), showing the field survey scores. Redundancy analysis (RDA) biplots for the selected envi-
ronmental (b) and land-use variables (d), showing the survey unit scores. Acronyms are given in Appendix 4.1.
A similar set of explanatory variables were selected from the forward selection of ‘RDA
land-use’ ordination: “oak woodlands”; “scrublands and pastures”; “orchards, olive groves
and vineyards”; and “urban and industrial areas”. These variables accounted for 12.8% of
the biological variation. The RDA biplot [Figure 4.6(d); eigenvalues of the first two axes =
0.12; 0.006, respectively] illustrates the fact that no general patterns of spatial distribution
of the SU emerged from the analysis.
Table 4.3 Selected variables and results of the partitioning of the riparian floristic and riparian cover
variance. Acronyms of variables are given in Appendix 4.1.
Riparian composition n=61a Riparian integrity n=194b
Selected variables
Environmental variables BROC, BMOD, ORDN, DSOU, ALTI, MSED, CWID, DSOU, ORDN
CALC, PREC, HUMI
Land-use variables ORCH, SCRU, QUER ORCH, SCRU, QUER, URBA
Variation (%)
All factors 33.7 20.5
“Pure environmental” 20.6 7.7
“Pure land-use” 4.9ns 8.7
Shared 8.2 4.1
Unexplained 66.3 79.5
ns= not significant for the trace and for the 1st canonical axis eigenvalue (Monte Carlo test, p>0.01)
a: number of field surveys; b: number of survey units
Two partial CCA/RDA ordinations were performed in order to assess the relative importance
of the environmental and land-use variables. Table 4.3 summarises the results of their
partitioning. Partial ordinations indicated that land-use variables did not significantly
contribute to an explanation of the floristic variation (Monte Carlo permutation test for the
trace, p>0.05), accounting as they did for a mere 4.9% of the total explained variation.
Moreover, only a relatively low proportion of the total variance in riparian composition was
explained by the environmental variables (significant at p< 0.001), or shared by both (land-
use and environment), and 66.3% remained without explanation. The environmental and
land-use variables significantly accounted for the explained variation of the riparian
integrity distribution patterns, making a similar, but rather small contribution. The
variables employed in this study were only able to explain 20.5% of the total variation
(including 4.1% of shared variation).
4.5 Discussion
In general terms this study revealed a low integrity and a widespread fragmentation of
riparian galleries, along with a small average width of streamside vegetation. Besides the
inter-basin variability, the fragmentation patterns of riparian woods vary greatly along
watercourses. Due to human-induced disturbances, but also to natural features of river
geomorphology and water scarcity, Mediterranean-type ecosystems present contrasting
patterns of fragmentation and connectivity (Malanson and Cramer 1999). Some of the
indicators used - namely the number of classes per SU and the percentage of canopy closure
classes - should thus be analysed with caution. In fact, a great part of the treeless reaches
were located on the intermittent stream headwaters and in areas constrained by
geomorphology or climatic harshness. Canopy openings that naturally result from the
irregular flow regime maintain important assemblages of hydrophytes and emergents. For
instance, multispecific woody assemblages in the Divor, Canha and Tera river corridors
were twice as rich in understorey herbaceous species than those observed in monotypic
willow stands (average species richness=16) in the same systems (F.C. Aguiar, unpublished
data 2001). Conversely, downstream from reservoirs (e.g. Zêzere), regulation has resulted
in continuous riparian stands that are dominated by exotic invaders, especially Acacia
dealbata. In a nearby basin there has been a report of a similar regulation-related
vegetation pattern, followed by vegetation encroachment on the downstream reaches of
reservoirs due to siltation-induced channel-narrowing (M. Kondolf, personal communication
1997).
Along Mediterranean streams riparian formations, can usually be observed as narrow strips
along watercourses (Gasith and Resh, 1999), i.e. ‘riparian galleries’. In contrast, hardwood
vegetation along temperate streams (known as ‘riparian forests’, ‘riverine forests’, or
‘floodplain forests’) can potentially occupy extensive areas of the floodplain, with up to
150 metres of lateral development. For example, when Roth et al. (1996) examined an
agricultural and urbanizing landscape in south-eastern Michigan, USA, they found that
median streamside vegetation width ranged from 0 to 130 metres. Though a standard
minimum fluvial corridor width for riparian conservation should be assessed locally and will
be typical for each community and even for individual taxa (Spackman and Hughes 1995),
the average values for both riparian width and also species richness encountered in this
study can be considered low when compared to other Mediterranean-type streams
(Tabacchi et al. 1996). They probably reflect some degradation, the amplitude of which has
yet to be assessed, namely for the purpose of riparian restoration.
Riparian width varied along the upstream-downstream river gradient in three different
spatial patterns: one was characterized by wider riparian formations at mid-course and
included the Canha, Pônsul and Tera basins; a second group displayed an upstream-
downstream increase in width and included the Divor, Meimoa and Nabão basins; and a
third, composed of the Zêzere and Trancão basins, where substantial regulation and
channelization, respectively, seemed to contribute to the irregular pattern of riparian
cover in the lowland areas. Some of the longitudinal patterns that were observed in the
riparian cover can be explained by the composition of riparian vegetation. Naturally narrow
galleries composed of Erica arborea, Sambucus nigra, and Frangula alnus were located at
the headwaters of the northern basins, while upstream southern reaches also presented
narrow riparian strips composed of resilient species such as tamarisk and Nerium oleander,
which are adapted to dryness. At the other extreme, the wider riparian galleries in the
Canha’s shallow-valleys were characterized by continuous stands of willows (Salix alba
group and S. salviifolia), associated with irrigation-crops (Ferreira and Moreira 1999). The
headwaters of the Divor displayed similar land-use on the upstream reaches, and
homogeneous willow communities had also established themselves following silting and
nutrient enrichment of the river system (Ferreira 1994).
Riverine landscapes are currently recognized as four-dimensional systems (Ward 1989), the
function and structure of which are determined mainly by the river and its flow regime,
which integrate a complex web of interrelations between the fluvial corridor and its
surroundings (Jungwirth et al. 2002). These intricate relationships between interacting
factors complicate the task of assessing anthropogenic influences (Allen and O’Connor
2000). There is a lot of information - e.g. Décamps et al. (1988); Jongman (1992); Lepart
and Debussche (1992); Allen and O’Connor (2000); Haslam (2000); Rao and Pant (2001);
Inoue and Nakagoshi (2001); Corbacho et al. (2003); Snyder et al. (2003) - to suggest that
land-use plays a large role in explaining fragmentation and riparian cover patterns, along
with strong cross-influences between land-use and environment. However, in this study,
land-use (8.7%) and environment (7.7%) each offered a similar, small contribution to the
explained variation in riparian integrity, whereas the partitioning performed with the
floristic data set attributed 20.6% of the explained variation to the environment, but no
significant contribution to land-use. Moreover, the shared variation explained
(environment/land-use) was also low for both analyses. The ‘environmental’ CCA and RDA
indicated broad geographical variables (altitude, humidity, precipitation) as the most
important factors in species composition, followed by small-scale variables at the reach
level (channel width, and hard substrate on the riverbed). Riparian integrity was also best
explained by local variables, such as fluvial corridor width, and geological background.
These results sustain the idea of the existence of hydrogeomorphological and climate
constraints on Mediterranean riparian ecosystems and highlight the potential predictive

capacity of reach-level features.
Nonetheless, riparian strips in this study were both narrow and extremely poor in species.
Riparian ecotones of Mediterranean streams are highly dependent on the life-history
features of land-use and on its long-term effects and the structure and fragmental patterns
therefore represent an expression of woody vegetation constrained by a long-lasting human
influence. Chronological studies on recent changes and paleoenvironmental research studies
are prime requisites if we are to understand present-day landscape patterns in this region.
4.6 Acknowledgments
This study received backing from the National Foundation for Science and Technology
(PRAXIS/3/3.2/FLOR/2130/95). F. C. Aguiar was supported by a grant (PRAXIS XXI /BD/
11173/97) from the same institution.
4.7 Appendices
Appendix 4.1 Acronyms and source of data of vegetational variables, land-use variables and environmental
variables.
Acronym Variable Source of data

Vegetation Aerial
photographs
RWID Riparian width (m)
NCLA Number of canopy classes in each survey unit, SU (0-5)
CCHA Number of canopy class changeovers in each SU
C1 Absence of riparian woody vegetation (%)
C2 Woody riparian layer scattered or semi-continuous on one margin and
absent on the other (%)
C3 Woody riparian layer scattered or semi-continuous on both margins (%)
C4 Woody riparian layer continuous on one margin and absent or scattered
on the other (%)
C5 Woody riparian layer continuous on both margins (%)
Land-use Aerial
photographs
URBA Urban and industrial areas (%)
IRRI Rice fields and irrigation crops (%)
PAST Pastures, extensive crops (%)
ORCH Orchards, vineyards, olive groves (%)
SCRU Scrubland, tilled land, fallow ground (%)
PINE Forest plantations (%)
QUER Oak and cork-oak woodlands (%)
FORE Mixed woodland (%)
MINE Gravel and sand mining, and exposed rock (%)
Environment
PREC HUMI Yearly average precipitation (mm), yearly average humidity (%), yearly P- Clima
TEMP average temperature (ºC) Software
SLOP ALTI Valley slope (‰), altitude (m) distance from source (km), Strahler’s Geographical
DSOU ORDN order number, existence of dams 10 km upstream the SU (0-1). Information
DAMS System
CWID Fluvial corridor width (m). Aerial
photographs
CALC META Dominant geological background (0-1): Calcareous and carbonate Geographical
ERUP SEDI rocks, conglomerates, sandstones; metamorphic basement: shales, Information
quartzites, schists, gneisses, slates; igneous rocks; terrace and gravel System
deposits, coarse sands, clays and mudstones, siltstones, alluvia.
BMOD Bank modification (1-5) Field survey
BROC LSUBS River margin substrate (%): bedrock; large substrate (> 64 mm,
MSUBS SAND including cobbles and boulders); medium substrate (2-64 mm, gravel
SILT and pebble); sand (0,2-2 mm); lime, silt, and fine particulate organic
matter (< 0.2 mm).
4.8 References
Aguiar F.C., Ferreira M.T. and Pinto P. 2002. Relative influence of environmental variables on
macroinvertebrate assemblages from an Iberian basin. Journal of the North American
Benthological Society 21(1): 443-453.
Aguiar F.C., Ferreira M.T., Moreira I. and Albuquerque A. 2000. Riparian types in a Mediterranean
basin. Aspects of Applied Biology 58: 221-232.
Allen A.P. and O’Connor R.J. 2000. Interactive effects of land-use and other factors on regional bird
distributions. Journal of Biogeography 21:889-900.
Borcard D., Legendre P. and Drapeau P. 1992. Partialling out the spatial component of spatial
variation. Ecology 73: 1045-1055.
Corbacho C., Sánchez J.M. and Costillo E. 2003. Patterns of structural complexity and human
Décamps H., Planty-Tabacchi A.-M. and Tabacchi E. 1995. Changes in the hydrological regime and
invasions by plant species along riparian systems of the Adour River, France. Regulated
Ferreira M.T. 1994. Aquatic and marginal vegetation of river Divor and its relation with land use.
Verhandlungen der Internationalen Vereinigung für theoretische und angewandte
Ferreira M.T. and Moreira I.S. 1999. River plants from an Iberian basin and environmental factors
Ferreira M.T., Albuquerque A., Aguiar F.C. and Catarino L.F. 2001. Seasonal and yearly variations of
macrophytes in a Southern Iberian River. Verhandlungen der Internationalen Vereinigung
für theoretische und angewandte Limnologie 27 : 3833-3837.
Gallego-Fernández J.B., García-Mora M.R. and García-Novo F. 1999. Small wetlands lost: a biological
199.
Gasith A. and Resh V.H. 1999 Streams in Mediterranean climate regions: abiotic influences and biotic
51-81.
Gumiero B. and Salmoiraghi G. 1997. The relationship between land use, river bank quality and lotic
invertebrates community. In: Haycock N.E., Burt T.E., Goulding G.W.T. and Pinay G.
(eds.), Buffer Zones: their Protection and Potential in Water Protection, Proceedings of
the International Conference on Buffer Zones, Quest Environmental, Hertfordshire, UK. pp.
74-82.
Haslam S.M. 2000. Impact of land use changes on rivers. Aspects of Applied Biology 58: 197-204.
Inoue M. and Nakagoshi M. 2001. The effects of human impact on spatial structure of the riparian
vegetation along the Ashida river, Japan. Landscape and Urban Planning 53: 111-121.
Jones III E.B.D., Helfman G.S., Harper J.O. and Bolstad P.V. 1999. Effects of Riparian Forest Removal
on Fish Assemblages in Southern Appalachian Streams. Conservation Biology 13 (6): 1454-
1465.
Jongman R.H.G. 1992. Vegetation, river management and land use in the Dutch Rhine floodplains.
Regulated Rivers: Research and Management 7: 279-289.
Jongman R.H.G., ter Braak C.J.F. and van Tregen O.F.R. 1997. Data analysis in community and
Landscape Ecology. Cambridge University Press.
Jungwirth M., Muhar S. and Schmutz S. 2002. Re-establishing and assessing ecological integrity in
Lepart J. and Debussche M. 1992. Human impact on landscape patterning. In: Hansen A.J. and di
Castri F. (eds.), Landscape Boundaries. Consequences for Biotic Diversity and Ecological
Flows, Ecological Studies 92,. Springer-Verlag, New-York, USA. pp. 76-106.
Ludwig J.A. and Reynolds J.F. 1988. Statistical Ecology: A Primer on Methods and Computing, John
Wiley & Sons Ltd., New York, US.
Magnan P., Rodriguez M.A., Legendre P. and Lacasse S. 1994. Dietary variation in a freshwater fish
species: relative contributions of biotic interactions, abiotic factors and spatial structure.
Canadian Journal of Fisheries and Aquatic Sciences 51: 2856-2865.
Malanson G.P. and Cramer B.E. 1999. Landscape heterogeneity, connectivity, and critical landscapes
for conservation. Diversity and Distributions 5: 27-39.
Mouw J.E.B. and Alaback P.B. 2003. Putting floodplain hyperdiversity in a regional context: an
assessment of terrestrial-floodplain connectivity in a montane environment. Journal of
Biogeography 30: 87-103.
P-Clima. 1999. Software Provedor de Dados Climatológicos para Portugal. Ver. 1.2. Project ALTENER
XVII/4.1013/Z/96-092, ed. R. Aguiar, INETI / DER, Lisbon.
Rao C.R. 1964. The use and interpretation of principal component analysis in applied research.
Sankhya A 26: 329-358.
Rao K.S. and Pant R. 2001. Land use dynamics and landscape3 change pattern in a typical
microwatershed in the mid elevation zone of central Himalaya, India. Agriculture,
Ecosystems and Environment 86: 113-123.
Rodriguez M.A. and Magnan P. 1995. Applications of multivariate analysis in studies of the
organisation and structure of fish and macroinvertebrate communities. Aquatic Sciences
57: 199-216.
Roth N.E., Allan J.D. and Eriksson D.L. 1996. Landscape influences on stream biotic integrity assessed
Salinas M.J., Banca G. and Romero T. 2000. Evaluating riparian vegetation in semi-arid Mediterranean
35.
Snyder C.D., Young J.A., Villella R. and Lemarié D.P. 2003. Influences of upland and riparian land use
patterns on stream biotic integrity. Landscape Ecology 18: 647-664.
StatSoft Inc. 1995. STATISTICA for Windows [Computer program manual]. Tulsa, OK: StatSoft, Inc.
Spackman S.C. and Hughes J.W. 1995. Assessment of minimum stream corridor width for biological
conservation: species richness and distribution along mid-order streams in Vermont, USA.
Biological Conservation 71: 325-332.
Stromberg J.C. 1993. Instream flow models for mixed deciduous riparian vegetation within a semi-arid
region. Regulated Rivers: Research and Management 8: 225-235.
Tabacchi E., Planty-Tabacchi A.-M., Salinas M.J. and Décamps H. 1996. Landscape structure and
diversity in riparian plant communities: a longitudinal comparative study. Regulated.
Tait C.K., Li J.L., Lamberti G.A., Pearsons T.N. and Li H.W. 1994. Relationships between riparian
cover and the community structure of high desert streams. Journal of the North American
ter Braak C.F.J. 1986. Canonical correspondence analysis: a new eigenvector technique for
multivariate direct gradient analysis. Ecology 67(5): 1167-1179.
ter Braak C.F.J. 1988. Partial canonical correspondence analysis. In: Bock H.H. (ed.), Classification
and related methods of data analysis, North Holland, Amsterdam. pp. 551-558.
ter Braak C.J.F. 1990. CANOCO version 3.1. Update notes. Agricultural Mathematics Group,
Wageningen, The Netherlands.
ter Braak C.F.J. and Prentice I.C. 1988. A theory of gradient analysis. Advances in Ecological
Research 18: 271-317.
ter Braak C.J.F. and Smilauer P. 1998. CANOCO. Reference manual and user´s guide to Canoco for
Windows: software for Canonical Community Ordination (version 4). Microcomputer Power,
Ithaca, NY, USA.
ter Braak C.F.J. and Verdonschot P.F.M. 1995. Canonical correspondence analysis and the related
multivariate methods in aquatic ecology. Aquatic Sciences 57: 254-289.
Ward J.V. 1989. The four-dimensional nature of lotic ecosystems. Journal of the North-American
Ward J.V. and Stanford J.A. 1995. The serial discontinuity concept: extending the model to floodplain
rivers. Regulated Rivers: Research and Management 10: 159-168.
Capítulo 5
Alterações espaço-temporais das galerias
ribeirinhas: influência dos factores
ambientais e do uso do solo
Chapter 5 Changes in riparian woods over space and time:

influence of environment and land use*
*
SUBMITTED to: Forest Ecology and Management, with authors F.C. Aguiar, M. T. Ferreira and C.
Nogueira.
Capítulo 5 - Temporal and Spatial Variation of Riparian Woods | 109
5.1 Abstract
With the objective of relating alterations to differences in the integrity of riparian

galleries, chronological studies (1951-1995) of the changing land use in three river basins
belonging to the Tagus fluvial system were conducted using remote sensing data. The
spatial characteristics of riparian woody vegetation and the temporal changes therein were
quantified using landscape metrics, such as average riparian width, length of canopy
closure types on river segments and patchiness indicators – e.g. the number of changes per
survey unit. We also studied the predictive potential of geomorphology and climate
features for explaining biological variability. We identify significant spatio-temporal land-
use changes on the adjacent floodplains, accompanied by remarkable differences in
riparian integrity. However, neither fragmentation patterns nor the loss or increase of
riparian cover were consistent with similar land-use changes throughout the rest of the
studied basins, particularly in urbanizing landscape patches. The longitudinal patterns of
streamside forests along Mediterranean-type rivers – particularly in Iberia – appear to be
complex and dependent upon the life-history features of land use. Results also suggest that
both the proximity and the magnitude of land-use patch interplay influence the degree of
changes in riparian integrity, which in overall terms displayed appreciable resistance to
alterations and indirect disturbance. The results are broadly discussed with regard to
spatial fine-grained scale changes in both land-use and riparian integrity.
Key words: streamside forests, riparian galleries, land-use change, aerial photographs,
Mediterranean ecosystems, Portugal.
5.2 Introduction
Riparian woodlands act as habitat connectors for the ecological flow of species, energy and
nutrients (Delcourt and Delcourt 1992; Malanson 1993), as players in trophic interactions
within fluvial corridors (Tait et al. 1994; Aguiar et al. 2002) and as promoters of regional
biological diversity (Naiman et al. 1993). The maintenance of riparian vegetation is
recognized as vital to the integrity of river ecosystems (Hancock and Froend 1996; Roth et
al. 1996) and its conservation value has repeatedly been emphasized (Naiman and Décamps
1997; Gilvear et al. 2000). In Mediterranean areas riparian galleries perform an essential
role in structuring and shaping landscapes by supporting a wide range of biotic assemblages
in a seasonally water-stressed environment (Gasith and Resh 1999). Riparian corridors are
ecotones – i.e. dynamic boundaries between the terrestrial and the aquatic ecosystems –
and thus receive a lot of inputs, either from the adjacent floodplain, or from the river, and
are probably among the most fragile of all ecosystems (Nilsson and Grelsson 1995).
A large number of human activities along a river and its valley, including agriculture,
urbanization, surface mining, water abstraction, flow regulation and grazing, are known to
influence riparian vegetation (Malanson 1993; Kondolf et al. 1996) and often result in the
fragmentation and compartmentalization of the riparian galleries, or changes in riparian
vegetation structure and composition (e.g. Décamps et al. 1988; Knight et al. 1994; Nilsson
1996; Planty-Tabacchi et al. 1996). Furthermore, linkages between landscape dynamics and
fluvial ecological processes are especially interwoven (Gregory et al. 1991) and human
interferences often disrupt these ecological bonds and interactions (Jungwirth et al. 2002).
Consequences of recent landscape changes in Mediterranean riparian vegetation have been

studied in Southern France (Lepart and Debussche 1992; Décamps et al. 1988), Malta
(Haslam 2000), Italy (Farina 1991) and semi-arid regions of Spain (Gallego-Fernández et al.
1999; Salinas et al. 2000) and California (Kondolf et al. 1987), and biotic descriptors such as
total cover and richness, number of exotic species and connectivity have been used to
express the degree of riparian degradation (Olson and Harris 1997; Salinas et al. 2000). So
far, studies on the influence that land-use activities have on riparian vegetation in
Portuguese river systems have included changes in plant composition (Ferreira 1994;
Ferreira and Moreira 1999), the introgression of plant assemblages by exotic species
(Ferreira and Moreira 1995; Aguiar et al. 2001) and attempts to establish regional reference
community types (Aguiar et al. 2000; Ferreira et al. 2002). Some of the effects of
Mediterranean hydric harshness are particularly difficult to distinguish from human
interferences, in that both produce naturally sparse and bushy riparian woody strata
(Ferreira et al. 1998; Aguiar et al. 2000; Salinas et al. 2000). What is more, Mediterranean
landscapes are characterized by a long-lasting history of intensive land-use (di Castri 1990;
Naveh and Vernet 1991), which probably interferes with riparian gallery patches.
Understanding both landscape and riparian patterns and processes will therefore benefit
from the analysis of past and present human activities (Lepart and Debussche 1992).
There have been numerous societal changes in Portugal over the last 50 years and they have
led to major modifications in the area and type of forestry, agriculture and urban
development that have logically affected the spatial integrity of riparian landscapes.
However, historical imagery data is scarce and fragmentary in this country. The first aerial
photographs were taken during World War II and the earliest routine flyovers only date from
the nineteen-eighties. Previous flyovers were generally made for specific purposes and
areas – e.g. regional urban planning or mapping of fast-growing forest stands. The data set
used in this study presented limitations of this type and is a ‘best available approach’ to
the aim of studying chronological changes in Mediterranean riparian integrity.
In this study we addressed the following questions: Have riparian corridors experienced a
significant alteration in riparian width and integrity over the 35 to 45 year-period for which
historical data exist? Are these alterations related to land-use changes? Do different land-
use changes over time in different regions lead to different responses in riparian integrity?
Do environmental factors play a larger role in shaping riparian formations than the human
activities in the river valley?
5.3 Methods
5.3.1 Study area
The River Tagus presents a Mediterranean climate, modulated mainly by altitude and
distance from the Atlantic Ocean. The interannual hydrological variability is high. The
woody strata of the riparian vegetation is dominated by ash (Fraxinus angustifolia Vahl ssp.
angustifolia), alder (Alnus glutinosa (L.) Gaertn.), black poplar (Populus nigra L.) and
willows (Salix atrocinerea Brot., S. alba L. and S. salviifolia Brot.), frequently surrounded
by boundaries of bramble-thicket (Rubus ulmifolius Schott). Other important species
include the hawthorn (Crataegus monogyna Jacq.), Flueggea tinctoria (L.) G.L. Webster and
Erica arborea L. The oleander (Nerium oleander L.) and the Tamarix africana Poir. are
frequent in the southern areas, and the exotic Acacia dealbata Link occurs throughout the
basin (Aguiar et al. 2000).
The Tagus basin displays marked differences in geomorphology, rainfall, population density
and land use. This pronounced heterogeneity led us to study three different regions (Fig.
5.1). The Nabão and Trancão basins have a similar mild climate, which is due to the
proximity of the Atlantic Ocean, as well as land-use features that are dominated by human
settlements, small-scale farming and crop areas in both cases; however, they differ in their
geological characteristics. The Canha basin presents semi-arid characteristics and is quite
different from the other two in terms of land-use and geomorphology (Table 5.1).
Figure 5.1 Iberian Peninsula (on the right), showing the Tagus basin (Portuguese part shaded) and the location
of the studied basins (on the left).
The upstream reaches of the River Nabão pass through narrow gorges and then run down
into a wide, fertile valley, which is occupied by agriculture, urban and industrial areas. This
abrupt change in landscape marks the boundary between hard Mesozoic limestone and soft
clay and marl deposits pertaining to the “Tertiary Tagus Sedimentary Basin”. The Trancão
basin develops over a Mesozoic igneous complex, which is covered by sedimentary and
alluvial deposits. Because of its proximity to Lisbon, there are numerous human settlements
along its watercourse. The river is channelized along a 5 km segment, before entering the
Tagus estuary. The Canha basin is located in the driest part of the Tagus basin, with a
landscape that is dominated by cork-oak (Quercus suber L.) and holm-oak (Quercus
rotundifolia Lam.) woodlands and Mediterranean sclerophyllous shrubs. The floodplain lies
on metasedimentary and metavulcanic rocks covered by alluvial Cenozoic deposits.
Table 5.1 Characterization of the river basins under study: maximum river length (km), drainage area (km2),
maximum Strahler’s order number, dominant river orientation, minimum and maximum values for altitude
(m), fluvial corridor width (m), valley slope (‰), annual average temperature (ºC), annual average rainfall
(mm) and annual average humidity (%).
Fluvial system Nabão Trancão Canha
Maximum length (km) 65.9 29.9 100.0
Drainage area (km²) 1016.5 287.7 1081.0
Maximum order number 5 5 6
Orientation NW – SE NW – SE SE – NW
Altitude (m) 25 – 662 3 – 442 2 – 409
Fluvial corridor width (m) 3.0 – 45.0 19.5 – 41.6 1.9 – 60.0
Valley slope (‰) 8 – 25 4 – 25 <4 – 8
Annual average temperature (ºC) 14.4 – 16.1 16.3 – 16.4 16.0 – 16.4
Annual average rainfall (mm) 747.8 – 1109.2 632.3 – 808.0 613.1 – 640.7
Annual average humidity (%) 68.5 – 73.3 71.4 – 75.7 69.5 – 72.7
5.3.2 Riparian structure, land-use and environmental data
Black and white aerial photographs of the Nabão, Canha and Trancão basins were obtained
from the Portuguese Institute of Cartography and analysed with a WILD ST4Ô mirror
stereoscope, using an 8x binocular magnifier.
The flyover dates and the river segments studied depended on the available material
(Table 5.2). The river segments were divided into 3-km-long river reaches (or survey units,
SU) with a Scale Master® curvimeter. Ground control points such as bridges, roads,
settlements and landscape features were used to co-register the photographs for each date.
Within each SU the integrity of the riparian galleries was assessed by the length of five
types of stand canopy closure (C1: absent; C2: scattered or semi-continuous on one margin
and absent on the other; C3: scattered or semi-continuous on both margins; C4: continuous
on one side and absent or scattered on the other; and C5: continuous on both margins), the
richness of cover types per SU (range 1-5) and the number of times the canopy closure
types changed in each SU (“canopy changeover”), both of which were considered to be
measures of riparian patchiness (Appendix 5.1). Riparian width and canopy class length
were measured with a Mitutoyo® digital calliper (0.01 mm resolution) and the former was
averaged from 12 measurements taken randomly along each SU.
Table 5.2 Flyover dates and scale of aerial photographs of the river segments expressed as distance from
source (total river length in parentheses, km).
Basin Flyover date Scale River segment under

study (km)
Nabão 1958 1: 13 500 31.0-55.0
1971 1: 15 000 (65.9)
1993 1: 15 000
Trancão 1965 1: 15 000 10.5-28.5
1989 1: 15 000 (29.9)
Canha 1951 1:13 500 21.4-99.4
1995 1: 15 000 (100.3)
Within each SU two transepts perpendicular to the stream were positioned on each margin
in order to quantify the land use of the river valley. The transepts were drawn up with a
maximum length of 1000 m or until the hillside ridge and made coincident for the different
flyover dates. Ten land-use types were defined to express valley occupation landscape
patterns: urban and industrial areas; rice fields; irrigation crops; pastures and non-irrigated
crops; orchards, vineyards and olive groves; fast-growing forest plantations (mostly Pinus
pinaster Aiton and Eucalyptus globulus Labill.); holm-oak and cork-oak woodlands; mixed
Mediterranean woodland and shrubland; scrubland, tilled land and fallow ground; and
surface mining and exposed rocks.
The percentage of these land-use types in each SU was obtained via their intersection with
the transepts and measured with a Mitutoyo® digital calliper (0.01 mm resolution),
following stereoscopic examination of the aerial photographs, and averaged for the four
transepts pertaining to each SU.
Mean distance to source, altitude, order number, gradient and geological background were
derived from cartographic and geological maps (Appendix 5.1). Climate data was obtained
using P-Clima software®, version 1.2 (1999).
5.3.3 Data treatment
For the Nabão data set (8 SU and 3 dates), a one-way ANOVA was performed to test for
significant differences in riparian cover and land use between dates. The Bartlett test was
used to analyse the homogeneity of the variance and Kolmogorov-Smirnov to test the
normality of the data. A non-parametric Friedman ANOVA was used when data was not
normally distributed. Student’s Newman-Keuls test permitted the identification of pairs
with significant differences (Zar 1996). For the Canha and Trancão data sets (26 and 6 SU’s
respectively, two dates), paired t-tests were performed and the Wilcoxon test was applied
to non-normal data. All statistical analyses were performed using STATISTICA/w 5.0®
(StatSoft Inc. 1995).
A redundancy analysis ordination (RDA) was performed with CANOCO 4.0Ò (ter Braak and
Smilauer 1998) in order to analyse the combined influence of landscape use and
environmental gradients on the riparian structure. RDA is a canonical form of principal
component analysis (Rao 1964) that selects the linear combination of environmental
variables giving the smallest total residual sum of squares (Jongman et al. 1997). Of the 23
variables that were initially considered, only the subset of the best predictors was retained
for analysis, using a forward selection procedure similar to a step-wise regression. A cut-off
point of P<0.05 was used in this routine, which is available in CANOCO. The variables were
log transformed (logx+1) and standardised by row centring (Ludwig and Reynolds 1988).
Geological background categories were included as dummy variables. A Monte Carlo
simulation of both the first axis eigenvalue and the trace (i.e. the sum of all canonical
eigenvalues) was used to test the significance of the effects under analysis (ter Braak
1990). Variance inflation factors for the explanatory variables were examined to detect co-
linearity between the latter.
5.4 Results
5.4.1 Riparian structure and temporal changes therein
Table 5.3 summarises the characteristics of the riparian structure found in the three basins
and the trends displayed by the temporal changes. All the canopy classes considered were
nearly always present within each SU and taken together with a generally high canopy
changeover (up to 19 class changes), this indicates a high degree of riparian patchiness for
all dates. The percentage cover of each canopy class varied widely from one SU to another,
with the higher average values (≥30%) occurring in different classes across various dates (C2
in Nabão 1958, C3 in Nabão 1971, C1 in Trancão all dates, C2 in Canha 1951 and C5 in Canha
1995). All the canopy classes were represented in all the rivers and dates, but covered a
large range of values. In both Nabão and Trancão there was an in-between-dates increase in
the parts of the river without a riparian canopy (significant in the case of the Trancão). On
the contrary, along the Canha there was a significant decrease in river reaches with non-
existent or scattered riparian canopies, together with an equally significant increase in C4
and C5-type canopy cover classes.
Riparian width presented low values on all the studied rivers. Average width per SU ranged
from about 3 to 18 m on the Nabão and from no-cover to 13 m on the Trancão and to 22 m
on the Canha. There were no significant detectable changes between the photo-records in
terms of riparian width on the Nabão and the Trancão, or in the canopy class changeover.
However, in the southernmost Canha basin significant increases were observed between
1951 and 1995, with a move towards a larger, more continuous canopy, along with a higher
changeover of canopy classes within each SU.
Longitudinal patterns were roughly conserved across dates for both the Nabão and the
Trancão (Fig. 5.2). Despite the lack of past-records from the head-reaches on both rivers,
there was a clear downstream increase in riparian width and cover (C4 and C5 galleries)
along the Nabão, whereas the River Trancão presented consistently wider and more integral
canopies along its upper and mid-courses. There was also a certain tendency towards an
increase in C1 galleries across dates, but this was non-significant. In the case of the River
Canha there were no obvious longitudinal trends in riparian canopy cover and width (Fig.
5.3). However, across dates there were important temporal changes throughout its course,
with a significant increase in riparian canopy width and cover (especially C5 galleries),
including some reaches that displayed current encroachment on the active channel by
riparian vegetation.
Riparian width (m) Nabão Trancão

20 20
1958 1965
16 16
1971 1989
1993
12 12
8 8
4 4
0 0
100 100
80 80
60 60
40
C1 40
20 20
0 0
100 100
Canopy classes cover (% of river length)
80 80
60
C2 60
40 40
20 20
0 0
100 100
80 80
60 60
40
C3 40
20 20
0 0
100 100
80 80
60 60
40 C4 40
20 20
0 0
100 100
80 80
60 60
40
C5 40
20 20
0 0
31 34 37 40 43 46 49 52 55 10.5 13.5 16.5 19.5 22.5 25.5 28.5
Figure 5.2 Spatial variation of canopy cover classes (% of river length) on Rivers Nabão and Trancão on
different flyover dates.
Riparian width (m)

25
1951
20
1995
15
10
C1
100
80
60
40
20
C2
100
80
Canopy cover classes (% of river length)
60
40
20
C3
100
80
60
40
20
C4
100
80
60
40
20
C5
100
80
60
40
20
0
21.4 33.4 45.4 57.4 69.4 81.4 93.4
Figure 5.3 Spatial variation of canopy cover classes (% of river length) on River Canha on different flyover
dates.
5.4.2 Land use and temporal changes therein
Land use and the temporal changes in it in the three basins are shown in Fig. 5.4 and Table
5.4. Both the Nabão and especially the Trancão basins presented quite a patchy landscape,
with a major component of urban and industrial areas and related impervious pavement,
which was concentrated along the river corridors as of the middle and lower courses (Fig.
5.4). A comparatively larger proportion of urban areas, scrubland and small orchards
occurred along the River Trancão, due to the proximity of the city of Lisbon. In contrast
with the Trancão and the Nabão, the Canha basin was characterised by a rural and less
fragmented landscape, dominated by holm-oak and cork-oak woodlands, with large farm
units that were mostly devoted to rice, irrigation crops and pastures occupying the more
fertile parts of the lowland valleys. Human settlements and industrial units (most of the
latter related to the processing of forest and crop products or farm animals) were fewer in
number, less concentrated and frequently associated with logging and unpaved roads,
rather than impervious pavements.
Figure 5.4 Changes in land-use classes in the studied basins (given as %). Percentage values were averaged
from all transepts. Acronyms are given in Appendix 5.1.
Forest stands of pine and eucalyptus (the two major fast-growing species used in
Portuguese forestry) dominated the upper reaches of the Nabão, whereas orchards and
pastures were the major land-use features around human settlements downstream (Table
5.4). The areas of holm-oak and cork-oak woodlands have decreased significantly since the
nineteen-fifties, while urban settlements have expanded significantly. In the Trancão basin
a fine land reticulate of pastures, orchards and irrigation crops also surrounded urban and
industrial areas, but a notorious significant increase in human settlements took place from
1965 to 1989 (an increment of 17.7%), at the expense of agricultural activities. In the Canha
basin there has been a significant decrease in cork-oak forestry and rice fields in the 45-
year period under study, both having been progressively substituted by crops that use
irrigation equipment, while urban development has also increased significantly.
5.4.3 Combined influence of environment and land use
Riparian characteristics are likely to be influenced in a multivariate way across space and
time by both environmental (intrinsic to the river system) and land-use (essentially
anthropogenic) factors. RDA was used to analyse the combined effects of these influences
on the riparian structure. In addition, the Canha basin was dominated by agro-forestry
activities and presented large landscape units, while landscapes in the Nabão and Trancão
basins presented small-scale patches of land use, influenced by numerous, scattered human
settlements. Data from these two basins was pooled to increase the ordination’s ability to
display temporal and spatial changes, as well as because of the small number of available
SU’s. The scores of the SU’s on the RDA biplots were superimposed with their temporal
affiliation, thereby depicting major temporal patterns (Figs. 5 and 6). The scores of the
riparian characteristics are displayed in the ordination diagrams in the top corner of both
Figures (see Appendix 5.1 for acronyms).
The eigenvalues of the first two RDA ordination axes with Nabão and Trancão data were
0.40 and 0.16 respectively, and together explained 55.6 % of the total variance in the
riparian structure (Fig. 5.5). Two categories of geological background strongly influenced
the riparian structure: alluvia (correlation with axis I =0.78) and sedimentary deposits from
Tertiary epoch (correlation with axis II =0.71). The first RDA axis was also positively
influenced by urban/industrial areas, which tend to occupy the more fertile alluvium-
dominated downstream valleys and to be associated with higher proportions of C1 river
reaches and less riparian width and patchiness (expressed by less C4 canopy class and
canopy changeover in each SU). The second RDA axis also received an important
contribution from forestry activities (fast-growing species), associated non-existent or small
irregular riparian cover (C1, C2 and C3) and non-urban land-use. No interpretable temporal
patterns emerged from the multivariate data treatment in these two basins.
0.8
TERT
C5
0.4
AWID
0.0
C4
CCHA
-0.4
NCLA
C3 C1
-0.8 C2
-0.8 -0.4 0.0 0.4 0.8
URBA
RDA axis 2
ALLU
Nabão 1958
PINE Nabão 1971
Nabão 1993
Trancão 1965
Trancão 1989
RDA axis 1
Figure 5.5 Redundancy Analysis (RDA) biplot showing the survey unit scores for the Rivers Nabão and Trancão
on the studied flyover dates, and the forward selected variables. The ordination diagram of riparian structure
variables is displayed in the upper right corner. Acronyms are given in Appendix 5.1.
The eigenvalues of the first two RDA ordination axes using Canha data were 0.33 and 0.18
respectively, and together they explained 38.4% of the total variance in the riparian
structure (Fig. 5.6). Unlike the previous analysis, the RDA biplot from Canha revealed a
clear separation between flyover dates. The dominant riparian structure from the nineteen-
fifties was non-existent or fragmentary (associated with C1 and C2 canopy categories) and
occurred together with a patchy landscape of rice fields and small-scale non-irrigated crop
areas, which were present throughout the basin and were associated with Tertiary
sedimentary formations. At present the dominant riparian structure includes a larger
riparian canopy associated with larger land-use patches of irrigated crops and a slowly
decreasing rice-field area, particularly in the more fertile, downstream valleys. Wherever
human settlements have developed, so has a higher riparian patchiness, expressed as C3
canopies, class richness per SU and canopy changeover, all of which are related with the
second ordination axis.
4.5
NCLA
0.6
CCHA
0.4
3.5 C3
0.2
C2
URBA 0.0
C1
AWID
2.5 C4
-0.2
C5
-0.4
-0.7 -0.3 0.1 0.5
1.5
RICE
RDA axis 2
0.5
ORCH
-0.5
META
WIDT
-1.5
MINE
-2.5
-3.5
1951
1995
-4.5
-2.0 0.0 2.0
-3.0 -1.0 1.0 3.0
RDA axis 1
Figure 5.6 Redundancy Analysis (RDA) biplot for the selected variables showing the survey unit scores for the
River Canha in 1951 and 1995 and the forward selected variables. The ordination diagram of the riparian
structure measures is displayed in the upper right corner. Acronyms are given in Appendix 5.1.
5.5 Discussion
Though many small-scale changes in land use have occurred with time and have locally
influenced riparian patchiness (Table 5.3 and Fig. 5.4), a considerable amount of change
seems to have been needed to significantly alter canopy structure (cover, width or
patchiness) in the studied river systems. This was the case with the Trancão, which
displayed an expansion of urban and industrial areas, accompanied by an increase in the C1
canopy class (Table 5.4 and Fig. 5.5), and especially with the River Canha, where
expansions in the area covered by irrigated crops and urban development were respectively
associated with increases in C4 and C5 canopy classes and a higher class changeover (Table
5.4 and Fig. 5.6). However, the consequences of changing land use were less obvious in the
Nabão basin, where there has been an increase in pinewoods at the expense of oak land and
Mediterranean scrubland, while urban development has occurred throughout the valley
floor, but in scattered patterns that therefore potentially have less direct effects on
riparian integrity. The results suggest that both the proximity and magnitude of land-use
patches interplay to influence the degree of changes in riparian structure, which in overall
terms displayed a remarkable resistance to alterations and indirect disturbance.
Both environmental and land-use variables influenced riparian cover (Gilvear et al. 2000),
though the former are frequently overlooked in studies on riparian integrity. In the present
study geological background and land-use variables were selected as the significant
determinants of riparian integrity in all the river basins (Fig. 5.5 and 5.6). In the case of the
Canha, the river hierarchy (expressed as fluvial corridor width) also significantly helped
explain the variability of riparian vegetation trends. In this basin a considerably longer river
axis (about 100 kms) resulted in more developed soils and deeper alluvia downstream,
which in recent decades has in turn favoured an increase in the area occupied by irrigation
crops and a more intensive type of agriculture.
Notwithstanding the environmental and land-use factors that were found to significantly
influence the riparian structure, an important part of the latter’s variability remained
without explanation. This may result from a scale effect whereby land-use activities that
occur spatially closer to the riparian corridor and direct interferences such as tree clearing,
aquatic weed harvesting, grazing, waste dumping, fire and channel re-profiling, would have
as great an influence on the riparian structure as larger landscape and human activities
(McIntyre and Hobbs 1999). In addition, historical geomorphic changes in river structure and
the impacts of series of indirect human-induced disturbances also disrupt lateral, vertical
and longitudinal linkages and can be directly related to alterations in riparian vegetation,
as Brierley et al. (1999) have demonstrated for the Bega catchment in Australia. However,
it may also result from the intrinsic features of Mediterranean-type rivers, including water
and geological harshness (Gasith and Resh 1999). In Southern Iberia riverine floral features
are poorly structured and modestly related to the environmental variables that are usually
considered in river ecology, frequently leaving a major unexplained component of
variability (Ferreira 1994; Ferreira and Moreira 1999; Ferreira et al. 1998; 2002; Aguiar et
al. 2002). A large number of the smaller river segments studied presented no flow during
the summer period, when a major component of exposed bedrock was observed, so water
and substrate constraints are probably responsible for a substantial part of the C1 and C2
canopy cover types. In warm temperate areas water availability has generally decreased
during the last two decades, due to agriculture-driven water abstraction from surface and
subterranean sources, thereby imposing a continually growing stress on fluvial ecosystems
(Wood and Armitage 1999). What is more, in general the small differences across dates
found in the C1 canopy cover classes on headwater reaches (Fig. 5.2 and Fig. 5.3) reflect
those same environmental constraints.
An important component of the riparian structure may also result from the cumulative
effects of millenary disturbances occurring in Mediterranean regions (di Castri 1990; Lepart
and Debussche 1992). In our study the riparian galleries consisted of narrow and laterally
constrained strips along the watercourses (generally up to 10 m wide), with patchy and
discontinuous canopies – a pattern that was displayed across dates and in spite of land-use
and canopy-related changes. However, this study can only be seen as a snapshot image from
a 35-45-year period, while the human interference in this region has been exerted for six
thousand years, since the Neolithic period, with tremendous changes in forest cover.
Riparian forest types based on overstorey dominance are extremely difficult to identify
(Aguiar et al. 2000) and determining the environmental factors that have affected the
establishment of each particular species is equally hard. The same riparian narrowness and
species poorness of Iberian corridors was observed by Salinas (2000). Whether the observed
patterns of compositional types and structure existed in pre-human times or are the result
of a co-evolution with a long-lasting human interference remains to be demonstrated.
A few studies suggest that riparian patches in Mediterranean areas were much larger a
hundred years ago (Gallego-Fernández et al. 1999) and that the existing communities
suffered millenary fragmentation and area reduction, at the same time as a decrease in
water availability from surface and sub-superficial runoff, which has been especially
stressful given their naturally greater dependence on a vertical water connection. Present
communities may well be the remnants of pre-historic riparian formations – the last,
resilient survivors. This hypothesis would be consistent with the low number of woody
species compared to riparian formations in climatically proximate regions (e.g. Szaro 1990),
the low number of riparian type communities and their low spatial discrimination ability
(Aguiar et al. 2000; Salinas et al. 2000), and the riparian vegetation’s apparent resilience in
the face of changes, whether in the case of direct actions such as the reprofiling and
straightening of the river channel (Aguiar et al. 2001), or in that of the present changes in
valley land use.
5.6 Acknowledgments
(PRAXIS/3/3.2/FLOR/2130/95). F. C. Aguiar was supported by a grant (PRAXIS XXI/ BD/
11173/97) from the same institution.
5.7 Appendices
Appendix 5.1 Variables considered for vegetation and land use, source of the data and acronyms used in the
ordination diagrams and tables.
Acronym Variable Source of data
Vegetation Aerial photographs

AWID Riparian width
NCLA Number of canopy classes identified in each survey unit (range 1-5)
CCHA Number of canopy class changeovers in each survey unit
C1 Absence of riparian woody species
C2 Woody riparian layer scattered or semi-continuous on one margin
and absent on the other
C3 Woody riparian layer scattered or semi-continuous on both margins
C4 Woody riparian layer continuous on one margin and absent or
scattered on the other
C5 Woody riparian layer continuous on both margins
Land-use Aerial photographs
URBA Urban and industrial areas
RICE Rice fields
IRRI Irrigated crops
PAST Pastures, non-irrigated crops
ORCH Orchards, vineyards, olive groves
SCRU Scrubland, tilled land, fallow grounds
PINE Fast-growing forest plantations
QUER Holm-oak and cork-oak woodlands
FORE Mixed woodland and Mediterranean scrubland
MINE Gravel and sand surface mining, exposed rock
Environment
Average annual humidity (%) P-Clima Software
Average annual temperature (ºC)
Average annual rainfall (mm)
Distance to source (km) 1:50,000 maps
Strahler’s order number
Valley slope (‰)
Altitude (m)
Geological background categories 1:100,000 maps
CALC Calcareous and carbonate rocks, conglomerates, sandstones
META Metamorphic deposits: shale’s, quartzite’s, schist’s, gneisses
IGNI Igneous rocks
TERT Tertiary sedimentary formations
ALLU Pleistocene and modern alluvia: gravel deposits, coarse and fine
sands, mudstones and siltstones
WIDT Fluvial corridor width (m) Aerial photographs
5.8 References
Aguiar F.C., Ferreira M.T., Moreira I. and Albuquerque A. 2000. Riparian types in a Mediterranean
Aguiar F.C, Ferreira M.T. and Moreira I. 2001. Exotic and native vegetation establishment following
channelization of a western Iberian river. Regulated Rivers: Research and Management 17:
509-526.
Aguiar F.C., Ferreira M.T. and Pinto P. 2002. Relative influence of environmental variables on
macroinvertebrate assemblages from an Iberian basin. Journal of the North American
Brierley G.J., Cohen T., Fryirs K. and Brooks A. 1999. Post-European changes to the fluvial
geomorphology of Bega catchment, Australia: implications for river ecology. Freshwater
Biology 41: 839-848.
Décamps H. and Naiman R.J. 1990. Towards an ecotone perspective. In: Naiman R.J. and Décamps H.
(eds.), Ecology and Management of Aquatic-Terrestrial Ecotones. UNESCO, Paris, and
Parthenon Publishing Group, Carnforth, UK. pp. 1-5.
Delcourt P.A. and Delcourt H.R. 1992. Ecotone dynamics in space and time. In: Hansen A.J. and Di
Flows. Ecological Studies 92. Springer-Verlag, New-York, pp. 19-54.
di Castri F., Hansen A.J. and Debussche M. (eds.) 1990. Biological invasions in Europe and in the
Mediterranean Basin. Kluwer Academic Publ., Dordrecht.
Farina A. 1991. Recent changes of the mosaic patterns in a mountainous landscape (north Italy) and
consequences to the breeding fauna. Options Méditerranéennes 15: 121-134.
Vehr. Internat. Verein. Limnol. 25: 2304-2315.
Ferreira M.T. and Moreira I.S. 1995. The invasive component of river flora under the influence of
mediterranean agricultural systems. In: Pysek P., Prach K., Rejmanek M. and Wade M.
(eds.), Plant Invasions - General Aspects and Special Problems. SPB Academic Publishing,
Amsterdam. pp. 117-127.
Ferreira M.T., Godinho F.N. and Cortes R.M. 1998. Macrophytes in a southern iberian river. Vehr.
Ferreira M.T., Albuquerque A., Aguiar F.C. and Sidorkewicz N. 2002. Assessing reference sites and
ecological quality of river plant assemblages from an Iberian basin using a multivariate
approach. Arch. Hydrobiol. 155(1): 121-145.
Gallego-Fernández J.B., García-Mora M.R. and García-Novo F. 1999. Small wetlands lost: a biological
1999.
Gasith A. and Resh V.H. 1999. Streams in Mediterranean climate regions: abiotic influences and biotic
51-81.
Gilvear D.J., Cecil J. and Parsons H. 2000. Channel change and vegetation diversity on a low-angle
alluvial fan, River Feshie, Scotland. Aquatic Conservation: Marine and Freshwater
Ecosystems 10(1): 53-71.
Gregory S.V., Swanson F.J., McKee W.A. and Cummins K.W. 1991. An ecosystem perspective of
riparian zones. BioScience 41: 540-551.
Hancock C.N. and Froend R.H. 1996. Biodiversity and management of riparian vegetation in Western
Australia. Forest Ecology and Management 85: 239-250.
Jongman R.H.G., ter Braak C.J.F. and van Tregen O.F.R. 1997. Data analysis in community and
Landscape Ecology. Cambridge University Press.
Jungwirth M., Muhar S. and Schmutz S. 2002. Re-establishing and assessing ecological integrity in
Knight C.L., Briggs J.M. and Nellis M.D. 1994. Expansion of gallery forest on Konza Prairie Research
Natural Area, Kansas. Landscape Ecology 9: 117-125.
Kondolf G.M., Webb J.W., Sale M.J. and Felando T. 1987. Basic hydrological studies for assessing
impacts of flow diversions on riparian vegetation: examples from streams of the eastern
Sierra Nevada, California. Environmental Management 11: 757-769.
Kondolf G. M., Kattelmann R., Embury M. and Erman D.C. 1996. Status of riparian habitat. Sierra
Nevada Ecosystem Project: Final Report to Congress, vol. II, Assessments and scientific
basis for management options. Davis: University of California, Centers for Water and
Wildland Resources: 1009-1030.
Lepart J.and Debussche M. 1992. Human impact on landscape patterning. In: Hansen A.J. and Di
Flows. Ecological Studies 92. Springer-Verlag, New-York. pp. 76-106.
Wiley & Sons Ltd., New York.

McIntyre S. and Hobbs R.A. 1999. A framework for conceptualising human effects on landscapes and
its relevance to management and research models. Conservation Biology 13(6): 1282-1292.
Naveh Z. and Vernet J.L. 1991. The palaeohistory of the Mediterranean biota. In: di Castri F. and
Groves R.H. (eds.), Biogeography of Mediterranean Invasions. Cambridge University Press,
Cambridge, pp. 19-33.
Naiman R.J. and Décamps H. 1997. The ecology of interfaces: riparian zones. Annu. Ver. Ecol. Syst.
28: 621-658.
biodiversity. Ecolological Applications 3: 209-212.
Nilsson C. and Grelsson G. 1995. The fragility of ecosystems: a review. Journal of Applied Ecology 32:
677-692.
Nilsson C. 1996. Remediating river margin vegetation along fragmented and regulated rivers in the
north: what is possible? Regulated Rivers: Research and Management 12: 415-431.
Olson C. and Harris R.R. 1997. Applying a two-stage system to prioritize riparian restoration at San
Luis Rey River, San Diego County, CA. Restoration Ecology 5(4SUPPL): 43-55.
P-Clima software®. 1999. Software Provedor de Dados Climatológicos para Portugal. Versão 1.2.
Projecto ALTENER XVII/4.1013/Z/96-092. R. Aguiar (ed.), INETI / DER, Lisboa.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., De Ferrari C. and Décamps H. 1996. Invasibility of
species-rich communities in riparian zones. Conservation Biology 10(2): 598-607.
Rao C.R. 1964. The use and interpretation of principal component analysis in applied research.
Sankhya A 26: 329-358.
Roth N.E., Allan J.D. and Erickson D.L. 1996. Landscape influences on stream biotic integrity assessed
Salinas M.J., Banca G. and Romero T. 2000. Evaluating riparian vegetation in semi-arid Mediterranean
35.
Ithaca, NY, USA.
Wissmar R.C. and Swanson F.J. 1990. Landscape disturbances and lotic ecotones. In: Naiman R.J. and
Décamps H. (eds.), The Ecology and Management of Aquatic-terrestrial Ecotones. Man and
Wood P.J. and Armitage P.D. 1999. Sediment deposition in a small lowland stream – management
implications. Regulated Rivers: Research and Management 15: 199-210.
Zar J.H. 1996. Biostatistical Analysis. Prentice-Hall International, Inc., New Jersey.
Parte III
Factores ambientais
determinantes da vegetação ripícola
e resposta à perturbação das
comunidades florísticas em
bacias de características semi-áridas
Part III Environmental drivers

and response to disturbance of river plant
assemblages from semi-arid river basins
The important thing is not to stop questioning.

Curiosity has its own reason for existing.
Albert Einstein (1875-1955)

Capítulo 6 Influência dos factores abióticos
e das galerias ribeirinhas nas comunidades de
macroinvertebrados numa bacia Ibérica
Chapter 6 Relative influence of abiotic factors and

riparian features on macroinvertebrate assemblages from an
Iberian basin*
*PUBLISHED AS: Aguiar F.C., Ferreira M.T. and Pinto P. 2002. Relative influence of environmental
variables on macroinvertebrate assemblages from an Iberian basin. Journal of North American
136 | Parte III - Resposta das Comunidades Ripícolas ao Ambiente e à Perturbação
Capítulo 6 – Influence of Environmental Variables on Macroinvertebrates | 137
6.1 Abstract
The composition of macroinvertebrate assemblages was studied at 31 sites in the Sado

basin, a lowland area with intermittent rivers located in southern Portugal. Riparian
features of the sites, including bankfull width, composition, cover, shade, and relative
amount of tree debris, were also measured, and local and basin abiotic variables were
determined. Woody vegetation in the Sado basin was species poor and was dominated by
the ash Fraxinus angustifolia, the alder Alnus glutinosa, and the willows Salix alba and S.
salviifolia. The study sites included a broad range of cover and riparian longitudinal
integrity. Macroinvertebrate assemblages revealed low levels of richness (4 - 29 taxa,
average 17.6) and Shannon’s diversity (0.65 - 2.87 bits, average 1.89), and were dominated
by species or groups (especially chironomids) that tolerate organic pollution and habitat
disturbance. Assemblages were composed of collector-gatherers (5.4 - 100%, average 58.1),
and to a lesser extent collector-filterers (0 - 81.3%, average 14.1) and shredders (0 - 58.3%,
average 12.1). Partial canonical correspondence analysis was used to relate the
macroinvertebrate assemblages to the riparian characteristics of a site and to local and
basin abiotic variables. Out of 8 riparian variables, overhanging tree shade, total tree
cover, and abundance of F. angustifolia and A. glutinosa were significant predictors of
assemblage variation. Of the 21 abiotic variables considered, only conductivity, distance
from source, and the average number of flowless summer months were significant
predictors of assemblage variation. Total macroinvertebrate variation was divided into
portions that were: 1) explained solely by abiotic variables (~14%); 2) explained solely by
riparian variables (~18%); 3) explained by both riparian and abiotic variables (~2%); and 4)
unexplained (~66%). Riparian features had greater influence than other environmental
characteristics on the composition of macroinvertebrate assemblages in the Sado basin,
perhaps because riparian features are closely related to food types.
Key words: freshwater macroinvertebrates, riparian features, partial canonical

correspondence analysis, Portugal.
6.2 Introduction
Inputs of coarse particulate organic matter (CPOM) from riparian vegetation are a source of
nourishment for aquatic organisms in streams (Conners and Naiman 1984; Benfield 1997),
and strongly influence the aquatic food chain (Cortes et al. 1995). The timing, amount, and
quality of the CPOM in leaf litter originating from near and upstream locations may have a
major influence on instream assemblages (Cummins et al. 1989). Bank storage of CPOM not
only creates specific aquatic assemblages but also turns the lateral ecotone into a major
transfer pathway through which terrestrial organic matter enters streams (Naiman and
Décamps 1997). Riparian vegetation also promotes physical habitat heterogeneity and
instream patchiness as a result of shade and the accumulation of tree debris (Hall et al.
1994).
A number of studies have examined macroinvertebrate assemblages and their relationships

with riparian vegetation and land use in Iberian rivers. These studies include work on the
spatial structure of macroinvertebrate assemblages and related environmental variables
(Graça et al. 1989; Cortes 1992), leaf processing (Cortes et al. 1995; Graça and Pereira
1995), and the consequences of afforestation on exotic species and deforestation on
harvesting and wildfire (Cortes et al. 1992; Canhoto and Graça 1995), and were conducted
on mountain streams and in central areas of Portugal. However, the link between streams
and their riparian areas has not been extensively studied in very dry southern Iberia. The
more low-lying and intermittent rivers in southern Iberian regions present differences in
riparian composition, such as increased proportions of small evergreen woody species
compared to deciduous trees (Aguiar et al. 2000). Deciduous trees, evergreen shrubs, and
herbaceous plant growth (Ferreira and Moreira 1999) contribute to the detritus pool of
southern Iberian rivers. Macroinvertebrate composition is also different, in that chironomids
are very abundant (Coimbra et al. 1996; Pinto and Morais 1998).
Severe reductions in water flow that result from water removal for irrigation and municipal
supplies (a problem on many dryland river systems, Davies et al. 1994), are the major
human-related impacts on southern Iberian rivers. Loss of natural flashiness and the
summer isolation of river stretches as a result of desiccation likely interfere with seasonal
patterns in the transport and deposition of organic matter along the river system and
negatively influence macroinvertebrate assemblages (Poff et al. 1997). In addition,
wastewater discharges from farming units increase river organic loads and may decrease
the litter dependence observed in studies such as Wallace et al. (1997).
6.3 Study area
The Sado basin (7640 km2) is a typical mediterranean lowland basin located in southwestern
Portugal (Fig. 6.1). The main river runs through a valley with a south-north orientation that
is dominated by fine tertiary deposits and quaternary alluvia. Most of the tributaries are on
soft slopes and some headwaters reach 300 m above sea level. The area has a complex
surficial geology, including granites and quartzodiorites, and metasedimentary rocks and
quartzites, although a large part is also occupied by coarse and fine tertiary deposits.
Rainfall is concentrated in the winter months. The annual average precipitation is 600 to
700 mm, although the northwestern tributaries can have up to 750 mm. The upper and
middle tributaries are intermittent and naturally dry out during the summer months,
thereby isolating parts of the river system. There can be ≥ 5 mo of natural drought in some
tributaries.
Heavy industry is concentrated near the estuary, whereas the rest of the basin contains
cropland (maize, melon, tomato, rice), livestock (cattle, pigs, poultry), and farm-related
industries that process crop products. Most of the tributaries have been regulated since
1996 to support the needs of the irrigated farmland. Water is retained during winter and
spring and is used for irrigation in late spring and summer. Most of the reservoirs have not
been supplied with the hydraulic devices needed to guarantee minimal flow requirements.
As a result, the summer drought period has increased both yearly and over a series of years.
The flow regime thus lacks the erratic interannual fluctuations typical of a Mediterranean
river. In addition, point or diffuse sources of water pollution from farm activities also
currently affect river biotic assemblages (Moreira et al. 1996).
6.4 Methods
6.4.1 Riparian vegetation
Sampling was conducted at 31 sites during the spring of 1996. Each of them comprised a
100-m-long stretch of 2nd- to 5th-order stream (Fig. 6.1). Each site was divided into ten 10-
m units for evaluation, each of which was independently visually assessed. The results of
these assessments were averaged across the site. Riparian species were identified on site
and their relative abundances were determined (0: absence, 1: 1 - 20% canopy cover, 2: 21
- 40%, 3: 41 - 60%, 4: 61 - 80%, and 5: >80%). Dominant tree species included Fraxinus
angustifolia, Alnus glutinosa and Salix spp. (S. salviifolia ssp. australis and S. alba ssp.
vitellina). The longitudinal integrity of the riparian vegetation was also assessed (1:
isolated trees or shrubs; 2, 3, and 4: tree clumps with a total canopy length of >20, >40,
and >60 m respectively; 5: continuous formations). The percentages of total tree cover,
total herbaceous cover (helophytes and hydrophytes, as in Ferreira and Moreira 1999), and
tree shade (overhanging trees) were determined. The relative amount of tree debris
(CPOM), including recognizable leaves, branches, twigs, and logs, was also assessed (0:
absence, 1: 1 - 20% cover of stream substrate, 2: 21 - 40%, 3: 41 - 60%, 4: 61 - 80%, and 5:
>80%). Bankfull width was estimated 10 times on each stretch of stream and averaged.
Figure 6.1 Iberian Peninsula, showing the Sado basin (top). Sado basin and the location of the sampling sites
(bottom).
6.4.2 Abiotic variables
Water temperature, dissolved oxygen (DO), pH, and conductivity were determined on site
using field probes. Average depth was determined using a graduated dipnet pole at a
number of randomly selected points (mean n = 25) within each sampling site. The
percentile area of different substrates was visually estimated for a circle of 0.5 m radius
around these points, and then averaged (>64 mm [large]: cobbles, bedrock, and boulders; 2
- 64 mm [medium]: gravel and pebble; and <2 mm [small]: sand, silt, and fine particulate
organic matter [FPOM]). The wetted width was averaged from 10 measured transects.
Subsurface water velocity was determined in the middle of the sampling location. Distance
from source (km) and height above sea level (m) were obtained for each site from 1:50,000
topographical maps. Water samples were collected and immediately taken to the laboratory
to determine biological oxygen demand (BOD), phosphates, total P (TP), NH4-N, and NH3-N.
Mean annual rainfall, mean annual temperature, and average number of months without
flow were determined using the information from the closest weather station. In situ man-
made site disturbances (e.g., cattle walking and grazing, human walking and use of the
banks [e.g., for angling] and organic waste deposits) were visually evaluated (gradient from
1: no disturbance to 5: heavily disturbed).
6.4.3 Macroinvertebrates
Benthic macroinvertebrates were collected in a zigzag pattern, running from the edge to
the middle of each sampling station, with a fixed time (5 min) kick sample, using a hand
net with a blade scraper (330 x 150 mm frame, mesh aperture 550 µm). Each habitat (e.g.,
riffle, pool, edge) was sampled proportionally to its representation at the site.
Macroinvertebrates were preserved in the field using a 5% formaldehyde solution,
subsequently identified in the laboratory to species level (except Diptera), and counted.
Richness, abundance, Shannon’s diversity, and the Biological Monitoring Working Party
(BMWP) index (Armitage et al. 1983) were calculated. BMWP’ is an Iberian adaptation of the
original BMWP score system (Alba-Torcedor and Sanchez-Ortega 1988); values >100 indicate
good water quality (Zamora-Munoz et al. 1995). Invertebrates were also grouped into 5
functional feeding categories, according to Merritt and Cummins (1984).
6.4.4 Data analyses
Two sets of environmental data were used in multivariate analyses. The 1st set had 8
riparian variables (total tree cover, total herbaceous cover, overhanging tree shade,
riparian longitudinal integrity, tree debris, and abundance of the dominant tree species),
and the 2nd set had 21 abiotic variables (water temperature, pH, conductivity, DO, BOD,
phosphates, TP, NH4-N, NH3-N, subsurface water velocity, average depth, wetted width,
large, medium, and small substrates, altitude, distance from source, mean annual rainfall,
mean annual temperature, number of mo with no flow, and human disturbance).
Data sets were subjected to canonical correspondence analyses (CCA) using CANOCO (ter
Braak 1987; 1990), which is an eigenvalue ordination that was specifically developed to
relate multivariate ecological data matrices. Prior to analysis, macroinvertebrate numbers
were transformed using natural logarithms [y = ln (y + 1)]; species that were present in only
1 sampling unit were removed. The best riparian and abiotic predictors (those retained for
analysis) were identified by a forward selection procedure available in version 3.1 of
CANOCO (ter Braak 1990; ter Braak and Verdonschot 1995), using a cutoff point of 0.10
(Magnan et al. 1994).
Total variation in the macroinvertebrate data matrix was partitioned into 4 components
(ter Braak 1990; Borcard et al. 1992, Rodriguez and Magnan 1995), accounted for by: 1) the
selected riparian variables (obtained using a CCA relating macroinvertebrate composition to
riparian variables), 2) the selected abiotic variables (obtained using a CCA relating
macroinvertebrate composition to abiotic variables), 3) the riparian variables following the
removal of the influence of abiotic variables using partial CCA (ter Braak 1987, 1988), and
4) the abiotic variables following the removal of the influence of riparian variables using
partial CCA. Each component of variation was obtained by dividing the canonical
eigenvalues of a particular CCA (or partial CCA) by the total inertia, i.e., the sum of all the
eigenvalues of a correspondence analysis of the macroinvertebrate abundance matrix. Pure
riparian variation was given by step 3, whereas pure abiotic variation was given by step 4.
The total explained variation was obtained from the sum of steps 1 + 4 or 2 + 3, and the
variation shared by both riparian and abiotic variables was obtained by subtracting step 3
from 1 or 4 from 2. The unexplained proportion of variation was obtained by subtracting the
pure riparian variation, the pure abiotic variation, and the shared variation from the total
variation. For each CCA or partial CCA that was done, a Monte Carlo test (1000 permu-
tations) on both the 1st axis eigenvalue and the trace (i.e., the sum of all canonical eigen-
values) was used to evaluate the significance of the effects under analysis (ter Braak 1990).
6.5 Results
6.5.1 Riparian characteristics
Twelve riparian woody species were found in the river corridors of the Sado basin (Table
6.1). Only F. angustifolia, A. glutinosa and Salix spp. occurred in a large part of the studied
sites (>70%). Other tall woody species included Populus nigra, Ulmus minor, P. alba, and
Acer negundo, but did not occur frequently and were therefore not included as riparian
variables. Smaller trees and shrubs that typically compose the mediterranean riparian
understory were Erica arborea, Tamarix africana, S. atrocinerea, Crataegus monogyna, and
Ruscus aculeatus.
Table 6.1 Mean, SD, and minimum (min) and maximum (max) values for selected environmental variables and
macroinvertebrate assemblages at studied sites (n = 31). BOD = biological oxygen demand. BMWP’ = Biological
Monitoring Working Party score, adapted by Alba-Torcedor and Sanchez-Ortega (1988).
Mean SD Min-max
Riparian variables
Tree richness per site 3.8 2.5 1 - 12
Herbaceous richness per site 40.8 10.3 17 - 58
Total tree cover (%) 47.9 24.1 10 - 90
Total herbaceous cover (%) 49.5 25.1 5 - 95
Overhanging tree shade (%) 32.9 30.2 0 - 95
Bankfull width (m) 10.2 3.3 1.2 - 18.6
Abiotic variables – local

Water temperature (ºC) 14.7 3.4 11.4 - 17.4
Conductivity (µS/cm) 874 590 190 - 3100
Dissolved oxygen (%) 84.5 33.4 11 - 160
pH 7.4 0.7 5.5 - 9.1
BOD (mg/L) 1.4 0.6 0.5 - 4.8
Phosphates (mg/L) 0.980 0.441 0.211 - 0.287
Total P (mg/L) 3.118 2.579 0.010 - 8.370
NH4-N (mg/L) 0.625 0.872 0.250 - 3.856
NH3-N (mg/L) 1.833 1.246 0.543 - 3.906
Wetted width (m) 6.4 3.8 1 - 15
Average depth (m) 0.8 0.4 0.2 - 2
Subsurface water velocity (cm/s) 7.8 3.3 0 - 29
Small substrate (%) 36.8 28.1 0 - 90
Medium substrate (%) 52.3 28 10 - 100
Large substrate (%) 10.9 19.4 0 - 80
Abiotic variables – basin

Altitude (m) 163 43.8 96 - 240
Distance from source (km) 23.7 19.3 3 - 96
Mean annual rainfall (mm) 621.6 24.7 593.9 - 699.7
Mean annual temperature (ºC) 16.6 0.2 16.1 - 16.7
Number of mo with no flow 2.7 1.1 1-5
Table 6.1 (cont.’d) Mean, SD, and minimum (min) and maximum (max) values for selected environmental
variables and macroinvertebrate assemblages at studied sites (n = 31). BOD = biological oxygen demand.
BMWP’ = Biological Monitoring Working Party score, adapted by Alba-Torcedor and Sanchez-Ortega (1988).
Mean SD Min-max
Macroinvertebrates
Richness per site 17.6 6.9 4 - 29
Abundance 523 545 13 - 1635
Shannon’s diversity H’ (bit) 1.89 0.53 0.65 - 2.87
BMWP’ water quality 36.5 27.1 4 - 99
Collector-gatherers (%) 58.1 27.4 5.4 - 100
Collector-filterers (%) 14.1 22.1 0 - 81.3
Predators (%) 10.6 13.5 0 - 51.1
Scrapers (%) 5.0 7.8 0 - 25.5
Shredders (%) 12.1 15.1 0 - 58.3
Total tree cover, longitudinal integrity of the riparian vegetation, and bankfull width varied
greatly both between the studied sites and along given stretches of river (Table 6.1). Shade
provided by overhanging trees was related to total tree cover (r = 0.72, p < 0.05), a
variable that ranged from isolated riparian patches of trees to dense stands. Tree shade
also depended on bank slope and geology of the riverbed and banks, which influenced
colonization of the banks by riparian trees (cf. Aguiar et al. 2000). CPOM, including many
types of tree debris, was abundant at most sites. At most locations, surfaces that were
available for colonization were occupied by herbaceous plants, which were generally
dominated by resilient grasses and graminoids, such as Paspalum paspalodes, Juncus
acutiflorus, Cyperus longus, Scirpus lacustris, or Phragmites australis.
6.5.2 Abiotic characteristics
The sampled sites included generally shallow and slow-moving, small to medium-sized
channels (Table 6.1). Stream water had high conductivity, highly variable pH and DO values,
and generally high levels of N and P, related to the widespread use of the river valleys for
different agricultural purposes. The riverbed was usually composed of hard substrates, but
the amount of small substrates was frequently high.
6.5.3 Macroinvertebrate assemblages
A total of 100 macroinvertebrate taxa was found, but richness, Shannon’s diversity, and
abundance per site were all relatively low (Table 6.1). Chironomidae was one of the most
important groups, and included up to 93.3% of the individuals per site (average 32.9%).
Oligochaeta and Simuliidae also frequently occurred. Brachyptera risi, Nemoura spp.,
Protonemoura sp., Isoperla sp., and Capnia bifrons (organisms intolerant to disturbance and
organic pollution) occurred mainly in northern and western tributaries. Water quality was
generally very poor and the BMWP’ score was always below the threshold of 100 (Table
6.1). The highest water-quality values occurred in the northern part of the basin, but even
then they were still low in absolute terms. Macroinvertebrate assemblages at most sites
were dominated by collector-gatherers, whereas all other functional groups were less
abundant (Table 6.1).
There was no bivariate correlation >0.4 (Pearson or Spearman) between abiotic variables or
between most riparian variables and macroinvertebrate assemblage descriptors. However,
the abundance of both ash and alder was significantly related to shredder abundance (r =
0.51, p < 0.01).
6.5.4 Canonical analyses
Four riparian variables (total tree cover, overhanging tree shade, abundance of A.
glutinosa, and abundance of F. angustifolia) of the 8 considered for inclusion in the CCA
were retained by the forward selection procedure (Table 6.2, Fig. 6.2). This CCA accounted
for ~20% of the total variation in the macroinvertebrate assemblage (p < 0.001). The
variance inflation factors of the 4 selected variables were <20 (ter Braak 1990). The 1st axis
(11.3% of the variation) was related to tree cover and tree shade, whereas the 2nd axis
(9.8% of the variation) was related to F. angustifolia abundance (Fig .6.2). There was a
gradient of macroinvertebrate assemblages from sites with no riparian trees to heavily
forested and shaded sites. However, the macroinvertebrate gradient observed along the 1st
axis was not related to the presence of shredders. Both the tree species present and their
abundances appeared to be important. Ash and alder were selected as significant variables.
Ash was highly related to the 2nd axis (Table 6.2) and increased with distance from source
(MTF and FCA, unpublished data). The stonefly shredder Nemoura fulvicepes was positively
related to the presence of F. angustifolia.
Table 6.2 Summary statistics for the canonical correspondence analysis (CCA) relating macroinvertebrate
assemblages to abiotic and riparian variables. Eigenvalues were 0.25 for the 1st axis and 0.22 for the 2nd in
the riparian CCA, and 0.32 for the 1st axis and 0.16 for the 2nd axis in the abiotic CCA. Also shown are the
acronyms of the variables.* = p < 0.05.
Forward selection of variables Correlation with Canonical

canonical axes coefficients
Extra-fit F-ratio p-value Axis 1 Axis 2 Axis 1 Axis 2
Riparian variables
Fraxinus angustifolia FANG 0.22 1.81 0.01 0.11 -0.90* 0.34 -0.96*
Total tree cover (%) COVER 0.22 1.77 0.06 -0.70* -0.33* -0.66* -0.15
Alnus glutinosa AGLU 0.16 1.33 0.08 -0.17 0.19 -0.16 0.10
Overhanging tree shade (%) SHADE 0.16 1.31 0.10 -0.74* 0.17 -0.62 0.05
Abiotic variables
Conductivity (µS/cm) COND 0.27 2.20 0.01 0.79* -0.05 0.51* -0.33
Number of mo with no flow NFLOW 0.17 1.42 0.03 0.64* -0.44* 0.59* -0.19
Distance from source (km) DSOU 0.17 1.48 0.02 0.39* 0.76* 0.38* 0.97*
300
Baetis buceratus
Baetis sp.
Piona sp.
Adicela reducta
200 Beraidae Notonecta sp.
Nychia sp.
Hidrovatus sp.
100
Hydrochus sp. Ceratopogonidae
AGLU
SHADE
Axis II
0 Boyeria irene
-100 COVER Nemoura fulviceps Physa acuta

Nepa cinerea
-200
FANG
-300
-300 -100 100 300
-400 -200 0 200 400
Axis I
Figure 6.2 Canonical correspondence analysis ordination of macroinvertebrate and riparian variables showing
the species with the highest scores in the axes and the biplot scores (arrows) of riparian variables. Triangles =
taxa scores. SHADE= overhanging tree shade, AGLU = abundance of Alnus glutinosa , COVER = total tree cover,
FANG = abundance of Fraxinus angustifolia.
Three abiotic variables (number of mo with no flow, conductivity, and distance from
source) were retained by the forward selection procedure, and together accounted for ~16
% of the total variation in the macroinvertebrate assemblage (p < 0.005; Table 6.2,
Fig. 6.3).
250
DSOU
200
Platycnemis sp.
150
100
50 Crocotemis servilia
Anax imperator
Brachyptera risi
Axis II
0 Coenagrion scitulum
COND
-50 Hydrophilidae
-100
Athericidae Hydrochus sp. NFLOW
-150 Baetis buceratus Nychia sp.
Ceratopogonidae
Beraidae
-200
Notonecta sp.
Hidrovatus sp.
-250
-150 -50 50 150 250
-200 -100 0 100 200
Axis I
Figure 6.3 Canonical correspondence analysis ordination of macroinvertebrate and abiotic variables, showing
the species with the highest scores in the axes, and the biplot scores (arrows) of abiotic variables. Triangles =
taxa scores. DSOU = distance from source, COND = conductivity, NFLOW = months with no flow.
The variance inflation factors of the 3 selected variables were <20 (ter Braak 1990). The 1st
axis (10.4% of the variation) was related to the more intermittent sites that are subject to
drought for long periods and possess a higher mineral content (mostly because of
agricultural activities). This intermittency results from natural conditions, but also from
river regulation and water withdrawal. Most of these sites occurred in the eastern and
southern part of the basin, in 1st- to 4th- order streams. Lentic-type macroinvertebrate
species, such as Odonata (Crocotemis servilia, Anax imperator, and Coenagrion scitulum),
Coleoptera (Nychia sp., Hydrochus sp., and Hidrovatus sp.), Notonecta sp., and
Ceratopogonidae (all of which presented the highest positive values on the 1st axis) tended
to be associated with these low-flow environments. The more permanent river sites
positioned on the negative side of the 1st axis were associated with rheophilous taxa, such
as Brachyptera risi and Beraidae. These sites included the northern and western tributaries
and the main river. The 2nd axis (8.4% of the variation) was related to distance from source
and, hence, is the major river axis (Table 6.2).
The total explained variation accounted for ~34% of the total variation, whereas ~66%
remained unexplained. The largest proportion of variation was explained by pure riparian
variables (~18%), whereas the pure abiotic component of the variation was lower (~14%),
although both proved to be significant in the Monte Carlo permutation test for the 1st axis
eigenvalue and the trace (p < 0.05). The variation shared by both the riparian and the
abiotic variables accounted for the remaining ~2% of total variation.
6.6 Discussion
6.6.1 Environmental factors
Environmental variables generally exert an important influence on macroinvertebrate

assemblages (Hawkins et al. 1982; Townsend et al. 1983; Ormerod and Edwards 1987;
Collier 1995; Wright 1995). Although 29 environmental variables (riparian plus abiotic) were
considered on this study, only 7 were significantly related to macroinvertebrate assemblage
variation, and these variables accounted for ~34% of the total variability. Studies of other
Portuguese basins located in northern regions and displaying a more regular flow regime
made it possible to identify groups of sites with similar macroinvertebrate assemblages that
could be related to environmental variables (Graça et al. 1989; Cortes 1992). For example,
Graça et al. (1989) significantly related 17 groups of co-occurring taxa to variations in
water chemistry, substrate particle size, and yearly temperature and flow.
However, our results were more similar to observations in the Guadiana basin (located in
the same region as the Sado), where macroinvertebrates were poorly grouped and poorly
related to environmental variables (Cortes et al. 1998). Multivariate analysis in this basin
did not show spatial patterns of macroinvertebrate assemblages related to stream hierarchy
variables (e.g., order number). Trophic groups also did not seem related to the habitat
descriptors, although a similar species composition and the same proportion of feeding
groups occurred throughout low-flow conditions in the Guadiana basin (Cortes et al. 1998).
The natural high disturbance levels that are typical of Mediterranean river systems (Gasith
and Resh 1999) could result in the ubiquitous presence of resilient, opportunistic, and
relatively persistent species assemblages, which are poorly related to the environmental
variables that are generally considered to be important in stream ecosystems (Puig et al.
1991). How this pattern is related to the degree of species sensitivity to present and past
habitat changes remains to be determined.
6.6.2 Riparian influence
Riparian variables significantly accounted for ~18% of the macrobenthos distribution in the
Sado river system. Moreover, tree cover and tree shade were important variables. Tree
cover and tree shade contribute to habitat heterogeneity and within-stream patchiness and,
thus, to the distribution of benthic organisms (Downes et al. 1993; Hall et al. 1994). It is
well known that riparian features influence macroinvertebrate assemblages, and many
studies have addressed the consequences of different riparian composition and disturbance
on macroinvertebrates (Hawkins et al. 1982; Rundle et al. 1992; Tait et al. 1994). Far less
is known about riparian inputs to intermittent streams than about those to temperate
forested ones, but other studies on intermittent streams also suggest that canopy cover and
allochthonous organic matter are important, at least in the wetter range of mediterranean-
type streams (Maamri et al. 1994).
In our study, it was not possible to relate either macroinvertebrate distribution or shredders
to tree debris (CPOM). It is probable that many southern Iberian woody species are of poor
food quality and availability for macroinvertebrate consumption. Some have scale-like
leaves (e.g., Erica arborea and Tamarix africana) whereas others have hard, sclerophyllous
leaves (e.g., Crataegus monogyna) or a strong pubescence (e.g., Salix atrocinerea and S.
salvifolia). These plant materials are quite different from those that are usually preferred
by macroinvertebrates (Anderson and Sedell 1979).
Ash and alder, which were significantly related to the 2nd canonical axis, were the only
species-level variables to be related to macroinvertebrate distribution, and they were also
significantly related to shredder abundance. These tree species are known to have leaves
that decompose easily and provide good quality food for macroinvertebrates (Bärlocher
1985). They were both common at many sites and, in some cases, were dominant in terms
of canopy cover. Alder usually developed dense and continuous riparian stands, whereas ash
frequently occurred in isolation or in small groups mixed with other riparian species.
6.6.3 Trophic groups
There was a high proportion of collector-gatherers and a low proportion of shredders in our
study. The high proportion of collector-gatherers was probably related to the abundant
organic matter, which is supplemented by fine particulate allochthonous inputs of
agricultural origin that usually offer considerable nutritional value (Shepard and Minshall
1981). Luxurious herbaceous growth is also common in southern Iberian rivers (Ferreira and
Moreira 1999), suggesting an abundant food supply that is readily accessible through either
grazing or the detrital pathway. The extended period of low-flow conditions means that
detrital organic matter remains in the riverbed for long periods (Gasith and Resh 1999).
The lentic-type characteristics of these rivers are probably not favorable to shredder
populations, which are usually more closely related to rheophilous, well-oxygenated, and
lotic-type conditions, where retention structures like roots and logs lead to the
accumulation of tree debris (Prochaza et al. 1991; Friberg 1997; Flory and Milner 1999).
However, little attention has been paid to the effects of drying on functional feeding
groups (Vidal-Abarca et al. 1992). Gasith and Resh (1999) proposed that habitat conditions
developed during the drought period of mediterranean-type streams cause shredders to
play a smaller role than they do in more humid regions.
Also, generalist food habits may prevail in many lotic systems (Mihuc 1997). It is frequently
assumed that species belong to the same guild as taxonomically related species, but
misplacements often occur (Hawkins and MacMahon 1989). In fact, accurate autoecological
data for Iberian macroinvertebrates are scarce, and little is known about their use of
multiple food resources.
In conclusion, after accounting for the pure abiotic and pure riparian plus shared variation,
>50% of the variability in macroinvertebrates assemblages still remained to be explained.
The unpredictability of the flow regime of mediterranean river systems combined with a
strong water demand for human purposes creates flow variability and environmental
harshness (Puig et al. 1991; Poff et al. 1997; Gasith and Resh 1999) that is probably
responsible for the poorly predictable macroinvertebrate assemblages in the Sado basin.
The implementation of minimum flow requirements, including the compulsory release of
flushing flows by irrigation dams and the maintenance of a medium to high level of riparian
cover, including the preservation of certain riparian species like ash and alder, are likely to
be the key elements in the conservation and management of these southern Iberian river
systems.
We are indebted to Kevin Murphy, who revised an initial version of the manuscript. The
comments of 2 anonymous reviewers, Robert C. Bailey, and David M. Rosenberg greatly
improved the final version. We thank Antonio Albuquerque and Pedro A. Vieira for field and
laboratory assistance. This study received backing from the Portuguese Ministry of
Agriculture, Rural Development and Fisheries’ PAMAF Project 4059/95. F. C. Aguiar was
supported by the grant PRAXIS XXI/BD/11173/97 from the National Foundation for Science &
Technology.
6.8 Literature Cited
Aguiar F.C., Ferreira M.T., Moreira I.S. and Albuquerque A. 2000. Riparian types in a mediterranean
Alba-Torcedor J. and Sanchez-Ortega A. 1988. A quick and simple method to evaluate biological
quality of running fresh water based on Hellawell (1978). Limnetica 4: 51-56.
Anderson N.H. and Sedell J.R. 1979. Detritus processing by macroinvertebrates in stream ecosystems.
Annual Review of Entomology 24: 351-357.
Armitage P.D., Moss D., Wright J.F. and Furse M.T. 1983. The performance of a new biological water
quality score system based on macroinvertebrates over a wide range of unpolluted running
waters. Water Research 17: 333-347.
Bärlocher F. 1985. The role of fungi in the nutrition of stream invertebrates. Botanical Journal of the
Linnean Society 91: 83-94.
Benfield E.F. 1997. Comparison of litterfall input to streams. Journal of the North American
Borcard D.P. Legendre and P. Drapeau. 1992. Partialling out the spatial component of spatial
variation. Ecology 73:1045-1055.
Canhoto C. and Graça M. 1995. Food value of introduced eucalyptus leaves for a Mediterranean
stream detritivore: Tipula lateralis. Freshwater Biology 34: 209-214.
Coimbra C.N., Graça M.A.S. and Cortes R.M. 1996. The effects of a basic effluent on
macroinvertebrate community structure in a temporary Mediterranean river.
Environmental Pollution 3: 301-307.
Collier K.J. 1995. Environmental factors affecting the taxonomic composition of aquatic
macroinvertebrate communities in lowland waterways of Northland, New Zealand. New
Zealand Journal of Marine and Freshwater Research 4: 453-465.
Conners M.E. and Naiman R.J. 1984. Particulate allochtonous inputs: relationships with stream size in
undisturbed watersheds. Canadian Journal of Aquatic Sciences 41: 1437-1484.
Cortes R.M. 1992. Seasonal patterns of benthic communities along the longitudinal axis of river
systems and the influence of abiotic factors on the spatial structure of those communities.
Archiv für Hydrobiologie 126: 85-103.
Cortes R.M., Graça M.A. and Monzon A. 1992. Replacement of alder by eucalyptus along two streams
with different characteristics: differences in decay rates and consequences to the stream
functioning. Verhandlungen der Internationalen Vereinigung für theoretische und
angewandte Limnologie 25: 2169-2172.
Cortes R.M., Graça M., Vingada J.N. and Oliveira S.V. 1995. Stream typology and dynamics of leaf
processing. Annales de Limnologie 31: 119-131.
Cortes R.V., Ferreira M. T. and Godinho F.N. 1998. Benthic community organization in an intermittent
lotic system. Verhandlungen der Internationalen Vereinigung für theoretische und
Cummins K.W., Wilbach M.A., Gates D.M., Perry J.B. and Taliaferro W.B. 1989. Shredders and riparian
vegetation. BioScience 39: 24-30.
Davies B.R., Thoms M.C., Walker K.F., O’Keefe J.H. and Gore J.A. 1994. Dryland rivers: their ecology,
conservation and management. In: Calow P. and Petts G. (eds), The rivers handbook,
Volume 2. Blackwell Scientific Publications, London, UK. pp. 484-511.
Downes B.J., Lake P.S. and Schreiber E.S.G. 1993. Spatial variation in the distribution of stream
invertebrates: implications of patchiness for models of community organisation.
Ferreira M.T. and Moreira I. 1999. River plants from an Iberian basin and environmental factors
Flory E.A. and Milner A.M. 1999. Influence of riparian vegetation on invertebrate assemblages in a
recently formed stream in Glacier Bay National Park, Alaska. Journal of the North
American Benthological Society 18: 261-273.
Friberg N. 1997. Benthic invertebrate communities in six Danish forest streams: Impact of forest type
on structure and function. Ecography 20: 19-28.
51-81.
Graça M.A.S., Fonseca D.M. and Castro S.T. 1989. The distribution of macroinvertebrate communities
in two Portuguese rivers. Freshwater Biology 22: 297-308.
Graça, M.A.S. and Pereira A.P. 1995. The degradation of pine needles in a Mediterranean stream.
Archiv für Hydrobiologie 134:119-128.
Hall S.J., Rafaelli D. and Thrush S.F. 1994. Patchiness and disturbance in shallow water benthic
assemblages. In: Giller P.S., Hildrew A.G. and Rafaelli D.G. (eds), Aquatic ecology. Scale,
pattern and processes. Blackwell Science, London, UK. pp. 333-375.
Hawkins C.P. and MacMahon J.A. 1989. Guilds: the multiple meanings of a concept. Annual Review of
Entomology 34: 423-451.
Hawkins C.P., Murphy M.L. and Anderson N.H. 1982. Effects of canopy, substrate composition and
gradient on the structure of macroinvertebrate communities in Cascade Range streams of
Oregon. Ecology 63: 1840-1856.
Maamri A., Chergi H. and Pattee E. 1994. Allochthonous input of coarse particulate organic matter to
a Moroccan mountain stream. Acta Oecologica 15: 495-508.
Canadian Journal of Fisheries and Aquatic Sciences 51: 2856-2865.
Merritt R.W. and Cummins K.W. (eds). 1984. An introduction to the aquatic insects of North America.
2nd edition. Kendall/ Hunt, Dubuque, Iowa.
Mihuc T.B. 1997. The functional trophic role of lotic primary consumers: generalist versus specialist
strategies. Freshwater Biology 37: 455-462.
Moreira I., Saraiva M.G. and Pinto P. 1996. Assessing the conservation value of a Mediterranean basin.
In: Boon P.J. and Howell D.L. (eds), Freshwater quality: defining the indefinable. His
Majesty’s Stationary Office, Edinburgh, UK. pp. 292-300.
Naiman R.J. and Décamps H. 1997. The ecology of interfaces: riparian zones. Annual Review of
Ecology and Systematics 28: 621-658.
Ormerod S.J. and Edwards R.W. 1987. The ordination and classification of macroinvertebrate
assemblages in the catchment of the River Wye in relation to environmental factors.
Pinto P. and Morais M. 1998. Colonisation curves of macroinvertebrate fauna in a temporary river
(river Degebe, Guadiana basin, Portugal). Verhandlungen der Internationalen Vereinigung
für theoretische und angewandte Limnologie 26: 113-115.
Poff N.L., Allan J.D., Bain M.B., Karr J.R., Prestegaard K.L., Richter B.D., Sparks R.E., and Stromberg
J.C. 1997. The natural flow regime. A paradigm for river conservation and restoration.
Prochaza K., Stewart B.A. and Davies B.R. 1991. Leaf litter retention and its implications for shredder
distribution in two headwater streams. Archiv für Hydrobiologie 120: 315-325.
Puig M. A., Aboal M. and Soston A. 1991. New approaches to Mediterranean fluvial communities.
Oecologia Aquatica 21: 13-20.
Rodriguez M.A. and Magnan P. 1995. Applications of multivariate analysis in studies of the
organisation and structure of fish and macroinvertebrate communities. Aquatic Sciences
57: 199-216.
Rundle S.D., Lloyd E.C. and Ormerod S.J. 1992. The effects of riparian management and physico-
chemistry on macroinvertebrate feeding guilds and community structure in upland British
streams. Aquatic Conservation: Marine and Freshwater Ecosystems 2: 309-324.
Shepard R.B. and Minshall G.W. 1981. Nutritional value of lotic species compared with allochtonous
materials. Archiv für Hydrobiologie 90: 467-488.
ter Braak C.J.F. 1987. CANOCO, a Fortran program for canonical community ordination by (partial)
(detrended) (canonical) correspondence analysis (version 2.1) ITI-TNO. Institute of Applied
Computer Sciences, Wageningen, The Netherlands.
ter Braak C.J.F. 1988. Partial canonical correspondence analysis. In: Bock H.H. (ed), Classification
and related methods of data analysis. North-Holland, Amsterdam, The Netherlands. pp.
551-558.
ter Braak C.J.F. and Verdonschot P.F.M. 1995. Canonical correspondence analysis and related
multivariate methods in aquatic ecology. Aquatic Sciences 57: 255-289.
Townsend C.R. 1989. The patch dynamics concept of stream community ecology. Journal of the North
Townsend C.R., Hildrew A.G. and Francis J. 1983. Community structure in some southern streams: the
influence of physico-chemical factors. Freshwater Biology 13: 521-544.
Vidal-Abarca M.R., Suarez M.L. and Ramirez-Diaz L. 1992. Ecology of Spanish semi-arid streams.
Limnetica 8: 151-160.
Wallace J.B., Eggert S.L., Meyer J.L. and Webster J.R. 1997. Multiple trophic levels of a forest stream
linked to terrestrial litter inputs. Science 277: 102-104.
Wright J. 1995. Development and use of a system for predicting the macroinvertebrate fauna in
flowing waters. Australian Journal of Ecology 20: 181-197.
Zamora-Munoz C., Sainz-Cantero C., Sanchez-Ortega A. and Alba-Torcedor J. 1995. Are biological
indices BMWP’ and ASPT’ and their significance regarding water quality temperature
dependent? Factors explaining their variations. Water Research 29: 285-290.
Capítulo 7
Variações sazonais e anuais de macrófitos em
sistemas fluviais semi-áridos
Chapter 7 Seasonal and yearly variations of

macrophytes in a southern Iberian River*
*
PUBLISHED AS: Ferreira M.T., Albuquerque A., Aguiar F.C. and Catarino L. 2001. Seasonal and yearly
variations of macrophytes in a southern Iberian River. Verh. Internat. Verein. Limnol. 27: 3833-3837.
Capítulo 7 - Temporal Variations in River Macrophytes | 157
7.1 Introduction
Many authors have attempted to group river plants, and to relate these groups directly or
indirectly to environmental factors (Wiegleb 1981; Holmes 1985; Haslam 1987; Haury and
Peltre 1993), and some have suggested the use of river plant assemblages as indicators of
the environmental quality of the river system (Lange and Vanzon 1981; Holmes 1989; Haury
1996). However, one important issue to analyse in this context would be the tightness and
permanency of the plant groups formed, along the temporal variation of the river systems,
especially when these are subject to particularly large intra-annual and inter-annual water
level fluctuations, such as the ones occurring in Mediterranean rivers.
As one of the southernmost Iberian rivers, the Guadiana (Portuguese part) presents a
natural intermittency in most of its tributaries. During summer, the river can be reduced to
a succession of natural ‘pools’ for a period as long as 4 months. However, some of the
tributaries presently maintain flow throughout the year due to water regulation from
upstream reservoirs (Caia, Lucefecit, Degebe and Beliche). When series of low rainfall years
occurs, the main river tends to present long periods of very low flow, or even no flow at all
in some parts, because of water retention in more than 50 Spanish reservoirs upstream.
Such a period of low rainfall, and low flow occurred during 1990-1994 (total discharge in
the period 1992-1994 was 1.45 x 109 m3), while the following years presented large rainfall
and frequent floods (total discharge in the period 1995-1997 was 10.20 x 109 m3).
The lotic vegetation found in the basin in 1994 (30 sites, Ferreira et al. 1998) was rather
homogenous and dominated by a group of ubiquitous emergents, highly resistant to drought
and disturbance, such as Cyperus longus, Scirpus lacustris, S. holoschoenus, Lythrum
salicaria, Paspalum pasplodes and Polypogon monspeliensis. Frequently more than 50% of
the terrestrial plants found on the river corridor were terrestrial and adventitious. Some
argument was made then over the possibility that the weak distribution trends observed
would be an indicator of highly disturbed and water-stressed systems (Ferreira et al. 1998).
A previous survey in 1985 (six sites, Ferreira and Monteiro, 1986), also pointed to a major
group of species dominating the river flora, however, several years of abundant rainfall
(1983-1985) apparently increased the wet usable area, and allowed the diversification of
habitats, thus resulting in the detection of rarer species like Carex pendula, Ranunculus
peltatus ssp. baudotti and Marsilea batardae.
In this work, we compare the floral characteristics and grouping of river macrophytes from
the Guadiana river, in two years of quite different rainfall and flow characteristics (1994
and 1996), and also from two different seasons within one year (Spring 1996 and Summer
1997). These floral variations are important to assess the efficacy of using plant
assemblages as metrics for environmental quality in Mediterranean river systems.
7.2 Study area
The Guadiana basin covers an area of 66,960 km2, of which 11,700 km2 (17%) is Portuguese.
The Portuguese part of the basin, where this study took place, presents a N-S orientation,
with a total length of 260 km2 of the main river, and an average altitude of 237 m. The
range of the average annual temperature is 12-18ºC, and the average rainfall is 598 mm.
The Guadiana basin is included in the hercinic basement of the Iberian Peninsula, and its
geological background is highly complex. Most of the basin is composed of metasedimentary
and metavulcanic rocks, but some alkaline or acidic igneous areas are also present as well
as Cenozoic deposits. Areas with alkaline igneous rocks, metavulcanites and detritic rocks,
along with some small areas of alluvial deposits, are used for agriculture, which exists
especially on the uppermost part of the basin. The remaining areas are covered with cork-
oak forest and Mediterranean shrubland.
Human settlements are small and scattered, with only a few having more than 2,000
inhabitants. Industries are also few, and most of them related to olive processing and pig
farming.
7.3 Methods
Nineteen sites of 250 m length were studied in August 1994 (very dry year), in August 1996
(very wet year) and in Spring 1997 (very wet year). Five sites occurred on the main river
(G4, G7, G8, G17 and G28) and the rest (Luc, Ca1, Deg2, Odia, Deg3, Alc2, Ard2, Fou, Bel,
Va2, Oei, Va1, Ter and Lim) occurred throughout the basin. For the geological areas and the
location of studied sites, please refer to Fig. 1 of Ferreira et al. (1998, p.1836).
All species found on water, margins and inner banks were identified and given a combined
cover/abundance scale of six degrees (Ferreira 1994). Because of their climatic dryness and
intermittency of the flow, Mediterranean rivers have their banks colonised to various
degrees by terrestrial adventitious plants. Both the identification and the separation
between lotic and terrestrial plants were made according to Franco (1971; 1984), Franco
and Rocha-Afonso (1994), and the latter were discarded from the data treatment. This
procedure avoids the tendency of site grouping being dominated by rarer and non-riverine
species (Ferreira 1994; Ferreira et al. 1998).
All statistical significances for the comparison of site richness were obtained by Mann-
Whitney (U) tests at 95% confidence level. Correspondence analysis, CA (ter Braak and
Smilauer 1998), was used to study the temporal variations of the community composition.
To improve the ordination trends, rare species were downweighted.
7.4 Results
A total of 144 river species was found that could be related to wet, waterlogged or aquatic
environments. When comparing survey dates (Table 7.1), it can be observed that Spring
presented a lower total richness, number of species per site and number of terrestrial
species, than both Summers. The percentage of species with a low frequency of occurrence
(occurring in less than two sites) and low abundance (rated 1 in a scale up to 6) was always
large, sometimes more than 50% of the total number of river species found. The number of
terrestrial species found within the river corridor was higher in both Summers than in
Spring. Marsilea batardae, a rare pteridophyte subject to legal protection, was only found
in May in one site.
When considering a comparison between tributaries and the main river, the number of
species per site is lower in the former, but the total number of species (riverine and
terrestrial) is larger, and so is the proportion of rarer and less abundant species. The
richness of the main river and the tributaries was always significantly different (Table 7.1).
Table 7.1 Summary of descriptive floristic parameters for the three surveys, and for the main river and
tributaries, including the total number of river species, the average number of species per site, the species
occurring in less than 20% of the sites (number and %), the species with abundance 1 (number and %) and the
total number of terrestrial species found. n=5 for the main river; n=14 for the tributaries. *Protection status
(Diário da República Nº 167/81). Small letters indicate a significant difference at P<0.05 using the Mann-
Whitney rank-sum test.
August 94 August 96 May 97
Main Tributaries Main Tributaries Main Tributaries

river river river
Total plant richness 54 84 48 79 41 69
Species/site 27 a 23 a 26 b,d 24 b,e 21 c,d 25 c,e
Species ≤ 20% frequent 17 (31) 35 (42) 13 (27) 39 (49) 14 (34) 30 (43)
Low abundance no. (%) 30 (56) 35 (42) 30 (63) 42 (53) 29 (68) 30 (41)
Terrestrial species (no.) 74 96 71 90 63 78
Marsilea
Threatened species* - - - - -
batardae
The summary of the statistics resulting from the various correspondence analyses
performed showed that, in Spring, there was a lower number of active species, lower values
of the sum of all unconstrained eigenvalues and also a lower percentage variance explained
by the first two axes. Hence, plant grouping and distribution trends were harder to detect
in this season (Table 7.2). The sum of all unconstrained eigenvalues is higher when all data
are considered because then spatial and temporal variations are included in the results.
However, the data variability increases and the percentage variance explained in the two
first axes is much lower. All eigenvalues obtained were medium-low.
Table 7.2 Summary statistics of the correspondence analysis performed, including the number of active
species, the sum of all unconstrained eigenvalues, the eigenvalues of the first two axes and the cumulative
percentage variance explained by species data, in each date and for all dates.
August 94 August 96 May 97 All dates

Axis I Axis II Axis I Axis II Axis I Axis II Axis I Axis II
Nº active species 92 83 80 125
Sum of eigenvalues 1.770 1.749 1.539 2.248

Eigenvalues 0.268 0.226 0.299 0.216 0.211 0.197 0.239 0.179
Cumulative % 15.1 27.9 17.1 29.4 13.7 26.6 10.6 18.6
Table 7.3 indicates the major species contributors (based on species scores) to the axes
extremes of the various analyses performed. There are almost no common major
contributing species when comparing the surveys. The species selected as such were not
the rarer, less frequent (downweighted in the analysis), but generally the ones with a
medium abundance and only occurring on that particularly survey. In fact, another group of
species is evenly distributed and generally site dominant, in all surveys and dates. This
group included the herbaceous species Agrostis stolonifera, Cyperus longus, Eleocharis
palustris, Lythrum salicaria, Paspalum paspalodes, Scirpus holoschoenus, S. lacustris, and
Typha latifolia, and the woody species Fraxinus angustifolia, Nerium oleander and
Securinega tinctoria.
Table 7.3 Main species contributors to the first and second axis of the correspondence analysis performed for
the three dates (positive or negative side of the axis is also indicated).
Axis I Axis II
Eleocharis palustris+
E. multicaulis+ Alisma lanceolata+
Holcus lanatus+ Cyperus eragrostis+
Juncus inflexus + Equisetum telmateia+
Myosotis stolonifera+ Epilobium hirsutum+
August 94 Scrophularia auriculata+ Juncus squarrosus+
Elodea canadensis+ J.subnodulosus+
Euphorbia peplis- Trifolium repens+
Iris pseudacorus- Baldellia ranunculoides-
Ammania coccinea- Gratiola linifolia-
Lemna gibba-
Echinochloa crus-galli+
Epilobium hirsutum +
Fimbristylis bisumbellata+
Glyceria declinata +
Juncus acutiflorus+
Juncus subnodulosus+
Juncus bufonius+
Lycopus europaeus+
Polygonum persicaria+
August 96 Scrophularia auriculata+
Setaria verticillata+
Cyperus eragrostis+
Potamogeton lucens-
Ludwigia palustris-
Najas minor-
Potamogeton crispus-
Gratiola linifolia-
Saponaria officinalis-
Baldellia ranunculoides-
Carex acuta+
Echinochloa crus-galli+
Adianthus capillus-veneris+
Phragmites australis+
Carex otrubae+
Portulaca oleracea+
Elodea canadensis+
May 97 Salix alba+
Lycopus europaeus+
Chara fragilis-
Ceratophyllum demersum-
Marsilea batardae-
Lythrum hyssopifolia-
Poa trivialis-
Potamogeton lucens-
The spatial position of all sites and dates is shown in Fig. 7.1. The main river presented a
clear floristic distinction from the tributaries, and was generally associated with the
negative portion of both axes. However, there was also a clear separation of the Summer
and Spring inventories, the latter positioned close together and related to the positive
portion on the first axis (Fig. 7.1). Thus, the first axis opposed the Spring survey and
especially the tributaries (as most floristically divergent) in the Summer surveys, whilst the
second axis tended to separate the main river and the tributaries of the Summer sampling
dates, regardless of the type of hydrological year.
300
250
200
150
100
50
AXIS II
0
-50
-100
-150
-200
-250
-300
-200 -150 -100 -50 0 50 100 150 200 250
AXIS I
MR94 MR96 MR97 T94 T96 T97
Figure 7.1 Position of the surveys from the main river and tributaries, and from all dates, on axes 1 and 2 of
the correspondence analysis (ter Braak and Smilauer 1998). MR, main river; T, tributaries.
Species especially related to the Spring tributaries were Myosotis debilis, Carex otrubae,
Adiantum capillus-veneris, Juncus acutus, Myosotis lusitanica, Polygonum amphibium and
Prunella vulgaris. The Summer main river is related to two exotic species, Ammania
coccinea and Azolla filiculoides. The negative part of the first axis, the main river in
Summer, received strong contributions from Ammania coccinea, Poa annua, Cyperus fuscus,
C. capitatus, Carex divulsa, Echinochloa crus-galli, Fimbristylis bisumbellata, and
Polygonum persicaria. Baldellia ranunculoides, Callitriche stagnalis, Juncus articulatus,
Eleocharis multicaulis, Ludwigia palustris, Potamogeton lucens and Setaria pumilla were
associated with the Summer tributaries.
7.5 Discussion
In this study, no significant differences were found between Summer surveys in spite of
quite different hydrological conditions. Simple floral characteristics such as richness and
percentage of rare and less abundant species did not change much in different hydrological
years and seasons. A smaller site richness and number of terrestrial plants was generally
found in Spring and the correspondence analysis was less able to explain the biological
variability for the Spring survey data. Also, some early flowering plants could only be found
in this survey, namely the threatened species Marsilea batardae. Thus, the early Summer
period (May and early June) could be the best one for surveying river plant assemblages in
this region, especially when considering single surveys.
Richness was always lower on the main river and these results might be related to the
higher number of plots surveyed on the tributaries. However, previous studies using a much
larger number of surveys (up to 17 in the main river and 43 in tributaries) gave the same
results (Ferreira and Monteiro 1986; Ferreira et al. 1998). Some environmental factors other
than the sampling regime were probably involved, namely the higher habitat diversity found
on the tributaries and the nutrient enrichment occurring within areas of irrigation crops.
Thus, the proportional sampling of the river length (e.g. up to 10% of the river surveyed,
Holmes 1985) seems to be suitable for the study of plant assemblages in these rivers.
The river plant assemblages of the Guadiana remained relatively poor (with low site
richness) when compared with other adjacent basins (Ferreira and Smeding 1990, Ferreira
1994). Also, local assemblages always had a high percentage of non-riverine species, and of
rare and less abundant plants, and were dominated by a small group of ubiquitous species,
regardless of the water availability of the hydrological year.
The same results had already been obtained with the more general survey made in the
Summer 1994 (Ferreira et al. 1998); thus, it must correspond to the prevailing ecological
characteristics of these biological assemblages, and not necessarily to man-made
disturbance such as regulation. River habitats and riverine ecosystems have intrinsic
features of intermediate-to-high disturbance (Wissmar and Swanson 1990), and these
Mediterranean ecosystems also present a geological and evolutionary history of frequent
natural disturbances, together with an early man-related history of exogenous disturbance
(di Castri 1990). Disturbances, such as perturbations of the flow regime, may not be
immediately detected in plant surveys unless persistent over many years.
Rarer but relatively abundant species tended to determine the position of that particular
survey on the CA axes. Such types of species are often discarded when dealing with larger
number of surveys, and if that was the case the ordination trends would be very difficult to
detect, as they were by Ferreira et al. (1998). These results also point to the idea that the
Guadiana basin (lower Portuguese part) is a relatively homogeneous region in terms of river
plant distribution. Besides the time factor, the scale factor (i.e. the magnitude of the
studied area and its intrinsic abiotic diversity) will also be a determinant when considering
the distribution and grouping of river plants.
This study was supported by the project PRAXIS XXI/2/2.1/BIA/113/94 of the Foundation for
Science and Technology. We are in debt to an Anonymous referee whose comments were
highly appreciated.
7.7 References
di Castri F. 1990. On invading species and invading ecosystems: the interplay of historical chance and
biological necessity. In: di Castri F., Hansen A.J. and Debussche M. (eds), Biological
Invasions in Europe and the Mediterranean Basin. Kluwer Academic Publishers. Dordrecht.
pp. 3-16.
Ferreira M.T. and Monteiro A. 1986. Estudo das comunidades de macrófitos aquáticos do rio Guadiana.
Actas do I Colóquio Luso-Espanhol sobre Ecologia de Bacias Hidrográficas e Recursos
Zooloógicos: 129-137.
Ferreira M.T. and Smeding F. 1990. Distribution patterns of aquatic and riparian vegetation in Erra
river, Central Portugal. Sci. Gerund. 16: 99-106.
Ferreira M.T., Godinho F.N. and Cortes R.M. 1998. Macrophytes in a southern Iberian river. Vehr.
Franco J.A. (ed.) 1971. Nova Flora de Portugal (Continente e Açores). 1. Lycopodiaceae-Umbelliferae.
Ed. Autor, Lisboa.
Franco J.A. (ed.) 1984. Nova Flora de Portugal (Continente e Açores). 2. Clethraceae-Compositae. Ed.
Autor, Lisboa.
Franco J.A. and Rocha-Afonso M.L. 1994. Nova Flora de Portugal (Continente e Açores). 3(1)
Alismataceae-Iridaceae. Escolar Editora, Lisboa.
Haslam S. 1987. River Plants of Western Europe. Cambridge University Press. Cambridge.
Haury, J. 1996. Assessing functional typology involving water quality, physical features and
macrophytes in a Normandy river. Hydrobiologia 340: 43-49.
Haury J. and Peltre M.-C. 1993. Interêts et limites des indices macrophytiques pour qualifier la
mésologie et la physico-chimie des cours d’eau: exemples armoricains, picards et lorrains.
Ann. Limnol. 29: 239-253.
Holmes N. 1985. Typing British Rivers according to their flora, Nature Conservancy Council, nº4.
Huntington.
Holmes N. 1989- British rivers: a working classification. Br. Wildlife 1: 20-36.
Lange L. and Vanzon J.C. 1981. A system for the evaluation of aquatic biotopes based on the
Windows: software for Canonical Community Ordination, Ver. 4. Microcomputer Power,
Ithaca, NY.
Wiegleb G. 1981. Application of multiple discriminant analysis on the analysis of correlation between
macrophyte vegetation and water quality in running waters of Central Europe.
Hydrobiologia 79: 91-100.
Wissmar R.C. and Swanson F.J. 1990. Landscape disturbances and lotic ecotones. In: R. Naiman and
H. Décamps (eds), The Ecology and Management of Aquatic-Terrestrial Ecotones, Man and
Capítulo 8 Selecção de locais de referência
e avaliação da qualidade ecológica da vegetação
ribeirinha por métodos de análise multivariada
Chapter 8 Assessing reference sites and ecological

quality of river plant assemblages from an Iberian basin
using a multivariate approach*
*
PUBLISHED AS: Ferreira M.T., Albuquerque A., Aguiar F.C. and Sidorkewicz N. 2002. Assessing
reference sites and ecological quality of river plant assemblages from an Iberian basin using a
multivariate approach. Archiv für Hydrobiologie 155(1): 121-145.
Capítulo 8 – Ecological Quality of River Plants | 169
8.1 Abstract
This study attempts to develop and discuss the applicability of a method for using river
plant assemblages as indicators of ecological quality in a southern Iberian basin, located in
Portugal. Several relevant aspects contribute to make this purpose scientifically
challenging. Two thirds of the basin are located in Spanish territory, a large part of the
river system is intermittent, the historical reference data on river plants in the basin are
non-existent and the distribution patterns of Iberian river plants are scarcely known. A
survey of 40 sites was conducted in late spring 1999, covering all the basin. Vegetation data
were treated by hierarchical classification and by canonical correspondence analysis, using
both environmental (19) and anthropogenic (9) variables. Three plant ecotypes were
identified, having different geological and geographical backgrounds. For the selection of
reference sites of each ecotype, the environmental variables significantly explaining within-
ecotype species distribution were determined by a forward selection procedure and latterly
used as co-variables. A partial canonical correspondence analysis was performed for each
ecotype, using a unique variable for the anthropogenic influence, henceforth, leading to a
sequence of less human disturbed sites being opposed to the more disturbed ones along the
first canonical axis. A cluster analysis for the group of sites of each ecotype contributed to
the final distinction of the reference sites, still presenting some intrinsic natural floral
variability. The Bray-Curtis dissimilarity within the group of sites from each ecotype ranged
between a natural similarity variation and a maximum dissimilarity (a community within the
same ecotype but with a plant composition completely different from the floral reference).
This range was partitioned into four classes of floristic quality corresponding to an
increasing deviation from the pro-natural floristic state. The use of river plants as
indicators of river integrity and the method used are discussed, considering the particular
features of the river system studied and its plant assemblages.
Key words: macrophyte community, intermittent streams; multivariate analysis, ecotypes,

Portugal.
8.2 Introduction
River health, or its ecological condition, though receiving some discussion on its meaning
and forms of application (e.g. Shrader-Frechette 1995; Moulton 1999), has however rapidly
been taken up as a working concept for both river conservation and management purposes
(Norris and Thoms 1999; Karr and Chu 2000). Ecosystem health has been defined in terms of
system organization, vigour and resilience as well as through the absence of anthropogenic
stress (Rapport et al. 1998).
The recently approved European Water Framework Directive uses the now well-established
idea of ecosystem health for determining and monitoring the water quality status of all
surface waters of the European Community (Pollard and Huxhman 1998; Boon 2000). The
ecological status expresses the quality of the structure and functioning of the aquatic
ecosystems and in order to assess the ecological status of any surface water, a series of
ecological quality elements from several biological communities are compared with type-
specific reference conditions.
An important development in water quality monitoring has been the incorporation of the
concept of reference conditions (Hughes 1995; Reynoldson and Wright 2000). The best
available reference condition is represented by information from a number of sites and can
be defined as the ecological condition that is representative of a group of minimally
disturbed sites. The reference condition is used by comparing the biological attributes of
individual test sites with a group of reference sites expected to be similar and to represent
the potential biological conditions in a region (Reynoldson et al. 1997).
However, identification of the different types of biological condition that exist within a
region requires that reference sites should be previously placed into groups with similar
habitat and biological communities, and in the absence of the disturbance existing in that
same region. The identification of reference conditions is thus a critical step for
bioassessment, and reference sites have been grouped differently (Johnson et al. 1993;
Hughes 1995). The multimetric (or a priori) method currently applied in most parts of the
USA (Barbour et al. 1999), uses independent geophysical and chemical data to group
reference sites and to determine reference conditions whilst the multivariate (or a
posteriori) method, currently applied in the UK, Australia and Canada (Wright et al. 2000),
uses the similarity of their species composition based on clustering methods.
To compare the condition of test sites with that of the reference sites (hence, their biologi-
cal and ecological deviation), the multimetric approach uses the quartile distributions of
additive community attributes or metrics, whilst the multivariate approach uses taxa ordi-
nation space or weighted probabilities of taxa occurrence (Reynoldson et al. 1997; Wright
et al. 2000). However, in situations where community structure, function or response, as
well as physico-chemical data, are still poorly known, it is difficult to identify ecologically
meaningful metrics, and it is also questionable if biological attributes already in use for
other regions and communities are transferable to local conditions. This might be the case
in Mediterranean-type stream ecosystems, of which the ecology is still poorly understood

(Gasith and Resh 1999). In this situation, basic community composition may be an
appropriate first approach to determine the reference conditions and the ecological status.
The multivariate approach undertaken in Australia has defined river health as the ability of
the aquatic ecosystem to support and maintain key ecological processes and a community
of organisms with a species composition, diversity and functional organisation as
comparable as possible to the natural habitats within a region (Davies 2000). Until now,
multivariate approaches have exclusively used macroinvertebrate communities (Johnson et
al. 1993). However, river health is a more holistic concept and ecological status includes all
the biological components of the ecosystem, albeit the functional organisation of different
biological communities expressing themselves at different time and space scales (Habersack
2000) and presenting different degrees and patterns of deviation from reference conditions
(Karr and Chu 2000).
Many authors have attempted to group river plants and to relate these groups directly or
indirectly to environmental factors (e.g. Wiegleb 1981; Penuelas and Sabater 1987; Holmes
et al. 1998; Dawson and Szoskiewicz 1999). Some have used river plant assemblages as
indicators of biological quality of the river systems (e.g. Lange and Van zon 1981; Haury et
al. 1996; Holmes et al. 1998). However, these methods for establishing environmental
quality using macrophytes have some limitations because they are not based on the specific
conditions of reference sites, but rather on a multiple of samples within and between sites,
that produce important shifts of the species’ bioindication meaning, therefore introducing
important bias on the final results.
River plants seem to be particularly dependent on the hydro-geophysical characteristics of

a river, such as bed substrate, flow regime and water velocity (Dawson 1988; Nilsson et al.
1993). Their ability to indicate the increase in nutrients has been demonstrated locally
(e.g. Haury and Peltre 1993), but they frequently provide a better indication of nutrient
enrichment and other man-made disturbances at the river segment or basin’ spatial levels
(Nilsson and Jansson 1995; Holmes et al. 1998). In fact, river plants are highly responsive to
the development of the lateral interface between the aquatic and the terrestrial
environments and related in-stream habitat patchiness (Décamps and Tabacchi 1992).
However, the degree of development of this interface changes along the river system and
also according to the land use of the river floodplain (Ward and Stanford 1995; Tabacchi et
al. 1996). Though individuals may respond to local hydrological and morphological
conditions, plant assemblages tend to reflect morphodynamic processes occurring at larger
scales (Habersack 2000), an important point when considering them for bioindication
purposes.
Some particular aspects further contribute to complicate the use of river plant assemblages
as bioindicators in the present study. The region corresponds to only one fifth of the basin,
while the remaining area is in Spanish territory. Intra-basin similarity of river plants was
previously found to be high (Ferreira et al. 1998). Mediterranean river systems present a
high degree of natural disturbance, especially resulting from the natural flow regime, and
most of the rivers in the studied area are intermittent, with an important in-stream
component being of terrestrial flora (Ferreira et al. 1998). Historical and present human-
made disturbance are omnipresent in southern Iberia, whilst historical data on river plants
are nearly non-existent, and the distribution patterns of Iberian river plants are scarcely
known (Ferreira and Moreira 1999).
This study aimed to use river plant assemblages as a bioindicator of ecological quality,
considering type-specific reference conditions as benchmarks. Furthermore, this study
attempted to apply the multivariate approach to river plant communities, in order to derive
groups of sites with similar compositional properties. These sites are used to determine the
degree of deviation from the reference taxonomic composition.
8.3 Study area
The Guadiana basin covers an area of 66,960 km2, of which 11,700 km2 (17%) are Portuguese
and the remainder is Spanish (Fig. 8.1, upper right). The Portuguese part of the basin,
where this study took place, presents a N-S orientation, with a total length of 260 km2 of
the main river and an average altitude of 237 m (Fig. 8.1, bottom left). The Guadiana is an
eight-order river, though the tributaries in its Portuguese part only go up to fourth-order.
The range of the average annual temperature is 12-18 ºC, and average annual rainfall is
598 mm.
The Guadiana basin is included in the hercinic basement of the Iberian Peninsula, and its
geological background is spatially complex. Most of the basin is composed of metasedimen-
tary and metavolcanic rocks, but some calcareous areas are also present as well as some
recent deposits. However, shales are predominant in the southern parts and alkaline
materials in the central part, whilst the northern part of the basin is more heterogeneous,
including acidic igneous and metavolcanic rocks along with cenozoic deposits. The areas
dominated by calcareous bedrock and cenozoic and quaternary deposits are used for agri-
culture, irrigated and non-irrigated crops, olive orchards and pastures (central and northern
part of the basin), while the areas dominated by metasedimentary rocks are generally
occupied by cork-oak forest and Mediterranean shrubland (the most southern part of the
basin). For further details on the mosaic of these geological areas, please refer to Figure 1
of Ferreira et al. (1998).
Sampling sites
Northern siliceous
Calcareous
Southern siliceous
Floristic deviation
Very small
Medium
High
0 10 20 30 km
Figure 8.1 Iberian Peninsula (upper right) with the indication of the Guadiana basin. Portuguese part of the
Guadiana basin (bottom left) with location of the sampling points. The symbols from each sampling point
indicate the ecotype affiliation and the degree of shade indicates the floristic deviation of each river plant
assemblage from the reference plant community within each ecotype.
Human settlements are small (few have more than 2,000 inhabitants) and scattered. The
industries are also scattered and small-sized, most of them related to cork, olive oil
processing and animal farming.
8.4 Materials and methods
8.4.1 Plant survey
Forty sites each 250 m in length, were dispersed throughout the basin (Fig. 8.1) and
surveyed in June 1999. The late spring period (May and early June in these latitudes) had
previously been found to be the best one for surveying river plant assemblages in this
region (Ferreira et al. 2001). Sites located on the main river were not considered in this
study because it was impossible to define reference floral conditions for them within the
basin. All species found in water, on margins and inner banks were identified and given an
overall abundance on a scale of 0-5, where 0-absent to 5-highly abundant. Nomenclature
followed Tutin et al. (1980), Franco (1971; 1984) and Franco and Rocha-Afonso (1994;
1998).
8.4.2 Environmental and anthropogenic variables
A total of 20 environmental variables were considered for each site. Water temperature,
dissolved oxygen, pH and conductivity were determined on site using WTW™ field probes.
Average depth was determined using a graduated dip-net pole within each sampling site for
a number of points (mean number = 25) selected at random. The percent area of different
substrates was visually estimated at these points, around a 0.5m radius circle and averaged
for the 25 points. Substrate classes considered were large substrate (more than 64 mm,
including cobbles, bedrock and boulders), medium substrate (between 2-64 mm, gravel and
pebble) and fine substrate (less than 2 mm, including sand, silt and fine particulate organic
matter). The wetted width was averaged from ten measured transects. Mean water velocity
(at 0.4 m from bottom) was determined using a R.OTT™ instrument. Distance from source
and Strahler’s order number were obtained for each site from 1:50,000 topographical maps.
Other environmental variables considered for each site were altitude, average annual
rainfall, average rainfall for the driest period (June-September), average annual
temperature, average annual flow and average annual run-off, obtained or estimated from
data collected at the nearest weather or gauging station. Furthermore, two basic types of
geological background were defined, calcareous and siliceous materials, from 1:100,000
lithological maps.
Variables related to human disturbance and the use of the drainage basin were also
considered. At the site level, variables included the disturbance of the banks and riverbed
by cattle and people, the condition (or more precisely the level of degradation) of the
riparian woody assemblages (both of which were visually assessed on site) and the number
of pollution sources directly impacting each site. At the drainage area level, variables
included the number of pollution sources, the area of human settlements and the area of
irrigated crops, non-irrigated crops, forestry and natural vegetation. All these variables
were given a scale with four degrees of magnitude, from 0 (absent) to 4 - highly disturbed,
polluted, perturbed or influenced by human activities or landscape vegetation.
The classification of vegetation groups both for the whole basin and for each plant ecotype
was performed with NTSYS-pc 2.0 (Rohlf 1993), using the Bray-Curtis coefficient of
dissimilarity and the ‘unweighted pair-group arithmetic average’ clustering method
(UPGMA).
Canonical correspondence analysis (CCA) performed with CANOCO 4.0 (ter Braak and
Smilauer 1998) was used to study the distribution trends of the river plant assemblages, and
separately analyse their relation with the environmental and human disturbance variables.
To improve the ordination trends, rare species were downweighted. Environmental
variables were log transformed (logx+1) and standardised by row centring (Ludwig and
Reynolds 1988). Basic geological types were used as binary dummy variables. From all the
explanatory variables considered, the best predictors (those retained for analysis) were
selected by a forward selection procedure available in CANOCO (ter Braak 1990; ter Braak
and Verdonschot 1995), with a cut-off point of 0.10 (Magnan et al. 1994).
For the specific purpose of identifying the reference sites from each plant ecotype, the
anthropogenic variables (condition of the river channel and riparian formations, pollution
sources for each site and its drainage area and the proportion of the drainage area used for
human activities) were additively combined (range 0-25) to obtain a unique variable named
anthropogenic disturbance. For each plant ecotype, a partial CCA was performed using the
anthropogenic disturbance as the sole abiotic variable and the environmental variables
(that were forward selected previously) as co-variables, therefore removing the natural
variability of each group of sites.
A Monte Carlo simulation of both the first axis eigenvalue and the trace (i.e. the sum of all
canonical eigenvalues) was used to test the significance of the effects under analysis (ter
Braak 1990).
8.5 Results
8.5.1 Plant composition
A total of 264 plant species were found in water, on margins and banks of the river
corridors surveyed. Only 99 (37.5%) of the species found were riverine, that is, could be
related to aquatic, waterlogged or wet habitats and requiring some degree of relevant
moisture during their lifecycles. Of these riverine species, only 25 (9.5%) were emergents
and true aquatic plants, the most frequent being Azolla filiculoides Lam., Myriophyllum
spicatum L., Ceratophyllum demersum L., Apium nodiflorum (L.) Lag., Scirpus lacustris L.
ssp. lacustris, Typha latifolia L. and Sparganium erectum L.. The remaining species were
hygrophytes, and associated with moist environments, such as Pulicaria paludosa Link,
Chamaemelum fuscatum (Brot.) Vasc. and Ranunculus repens L.. Because of their
intermittency of flow the banks and beds of southern Mediterranean rivers are typically
colonised in various degrees by adventitious terrestrial species common to the surrounding
landscape (Ferreira et al. 1998, Ferreira and Moreira 1999; Ferreira et al. in press).
Only one third of the species found on the river sites were annuals, and only 6.7% of them
were nitrophyllous or related to eutrophic environments. Only 4% of the species found were
exotic, though some of them were local invaders of river habitats, such as Aster squamatus
(Spreng.) Hieron., Conyza bonariensis (L.) Cronq., Cyperus eragrostis Lam., Paspalum
paspalodes (Michx) Scribner and Azolla filiculoides Lam.. Further information of species
composition is included in Table 8.1 and Table 8.2.
Table 8.1 Species with frequency of occurrence (FO) in the three ecotypes higher than in the Guadiana basin,
and their scores on the first and second axes of the global CCA.
Relevant species FO (%) FO (%) AXIS I AXIS II

Ecotype Basin scores scores
Plant ecotype I (northern siliceous)
Chamaemelum fuscatum (Brot.) Vasc. 100 85 -0.16 -0.26
Rubus ulmifolius Schott. 100 75 0.07 0.14
Fraxinus angustifolia Vahl. ssp. angustifolia 88 75 0.36 0.27
Rumex crispus L. 88 75 0.50 -0.33
Sonchus oleraceus L. 88 45 -0.16 0.02
Echium plantagineum L. 77 70 -0.38 -0.23
Ranunculus trilobus Desf. 77 53 0.08 -0.35
Sinapis arvensis L. 77 58 0.16 0.13
Rosa canina L. 55 50 0.15 0.43
Scirpus lacustris L. ssp. lacustris 55 55 -0.40 0.43
Trifolium fragiferum L. 55 18 -0.46 0.04
Callitriche stagnalis Scop. 44 15 0.08 -0.02
Foeniculum vulgare Miller ssp. piperitum (Ucria) Coutinho 44 33 0.33 -0.02
Cyperus longus L. 33 30 0.38 -0.23
Plant ecotype II (calcareous)
Scirpus holoschoenus L. 100 95 0.09 -0.22
Cyperus longus L. 91 75 0.19 -0.23
Nerium oleander L. 91 65 -0.01 0.15
Oenanthe crocata L. 91 88 -0.02 -0.21
Mentha pulegium L. 82 53 0.35 -0.23
Raphanus raphanistrum L. 82 73 0.18 -0.08
Echium plantagineum L. 77 70 0.41 -0.22
Eleocharis palustris (L.) Roemer & Schultes 77 55 0.14 -0.06
Fraxinus angustifolia L. ssp. angustifolia 77 75 -0.27 -0.22
Lythrum junceum Banks & Solander 73 60 0.42 -0.12
Plantago lanceolata L. 73 55 0.32 -0.23
Scirpus lacustris L. ssp. lacustris 73 55 0.10 -0.21
Securinega tinctoria (L.) Rothm. 73 45 0.34 0.36
Coleostephus myconis (L.) Reichenb. fil. 68 53 0.48 -0.01
Paronychia argentea Lam. 68 40 0.02 -0.40
Aristolochia baetica L. 64 50 0.23 -0.31
Briza maxima L. 59 38 -0.37 -0.38
Cerastium glomeratum Thuill. 59 53 0.46 -0.13
Juncus acutus L. 59 35 0.18 -0.28
Table 8.1 (cont.’d) Species with frequency of occurrence (FO) in the three ecotypes higher than in the
Guadiana basin, and their scores on the first and second axes of the global CCA.
Relevant species FO (%) FO (%) AXIS I AXIS II

Ecotype Basin scores scores
Ranunculus trilobus Desf. 59 53 0.44 -0.30
Campanula rapunculus L. 55 30 0.04 0.10
Cistus salvifolius L. 55 30 -0.06 -0.12
Sonchus asper (L.) Hill 55 50 0.48 -0.16
Gladiolus illyricus Kock ssp. reuteri (Boiss.) Coutinho 45 30 0.30 -0.05
Rumex conglomeratus Murray 45 33 -0.49 -0.03
Convolvulus arvensis L. 41 40 0.04 -0.34
Juncus acutiflorus Ehrh. ex Hoffm 41 28 0.14 -0.09
Lycopus europaeus L. 41 33 0.04 -0.11
Trifolium repens L. ssp. repens 36 33 0.33 -0.40
Galium palustre L. 32 28 -0.34 -0.49
Gladiolus italicus Miller 27 15 -0.40 -0.05
Notoscordum gracile (Aiton) Stearn 27 15 -0.40 -0.15
Cistus ladanifer L. 18 10 0.04 0.16
Myriophyllum spicatum L. 18 13 0.33 0.22
Saponaria officinalis L. 18 15 0.04 0.18
Linum bienne Miller 14 13 0.48 -0.25
Sanguisorba minor Scop. ssp. magnolii (Spach) Briq 14 13 0.48 -0.25
Alnus glutinosa (L.) Gaertner 10 9 0.04 0.31
Carex pendula Hudson 10 9 -0.44 -0.31
Rumex bucephalophorus L. 10 9 0.04 0.30
Ecotype III (southern siliceous)
Arundo donax L. 89 55 -0.43 -0.28
Papaver rhoeas L. 89 30 0.12 -0.03
Phalaris minor Retz. 89 25 0.13 -0.36
Picris echioides L. 89 43 0.50 0.42
Plantago major L. ssp. intermedia (DC.) Arcangeli 89 45 0.48 0.25
Rubus ulmifolius Schott. 89 75 -0.06 0.21
Sinapis arvensis L. 89 58 -0.20 0.15
Piptatherum milliaceum (L.) Cosson 78 18 0.30 -0.07
Lythrum junceum Banks & Solander 67 60 0.08 -0.41
Lactuca serriola L. 55 25 -0.28 0.22
Phragmites australis (Cav.) Trin. ex Stendel 44 15 0.12 -0.02
Scrophularia canina L. ssp. canina 33 18 0.30 0.29
Arisarum vulgare Targ-Tozz 22 10 0.02 0.47
8.5.2 River plant ecotypes and related abiotic determinants
The hierarchical classification of the sites using Bray-Curtis dissimilarity enabled the
identification of three major floristic groups (Fig. 8.2). The highest dissimilarity separates a
group of siliceous from a group of calcareous sites, and a further separation occurs in the
siliceous sites, including a larger group from the northern part of the basin and a smaller
group from the southern part of the basin. These three groups of sites had between-group
floral dissimilarity greater than 50% and were considered the floral ecotypes of the study
area (and named ecotype I- northern siliceous; ecotype II- calcareous; ecotype III- southern
siliceous).
Five sites presented floral deviation from each or all of these groups (11, 12, 31, 36 and 45;
Fig. 8.2). This deviation was assumed to result from extreme human disturbance (all these
sites were located immediately downstream of reservoirs). These sites were henceforth
included on the southern siliceous ecotype according to their common geological
background: metasedimentary rocks.
Table 8.2 List of species with the highest scores on the first axis of each partial CCA performed and their
frequency of occurrence (FO) on the ecotype and on the selected group of reference sites.
Relevant species AXIS I FO (%) FO (%)

scores Ecotype Reference sites
Plant ecotype I (northern siliceous)
Arisarum vulgare Targ-Tozz 2.05 11 20
Convolvulus althaeoides L. 2.05 11 20
Urtica dioica L. 1.80 33 60
Saponaria officinalis L. 1.55 22 40
Solanum nigrum L. 1.55 22 40
Fumaria officinalis L. 1.31 44 80
Calystegia sepium (L.) R. Br. ssp. sepium 1.06 22 40
Coleostephus myconis (L.) Reichenb. fil. 1.06 56 80
Crataegus monogyna Jacq. ssp. brevispina 1.06 44 60
(G. Kunze) Franco
Lactuca serriola L. 1.06 33 60
Lythrum junceum Banks & Solander 1.06 22 40
Securinega tinctoria (L.) Rothm. 1.06 22 40
Table 8.2 (cont.’d) List of species with the highest scores on the first axis of each partial CCA performed
and their frequency of occurrence (FO) on the ecotype and on the selected group of reference sites.
Relevant species AXIS I FO (%) FO (%)

scores Ecotype Reference sites
Plant ecotype II (calcareous)
Geranium rotundifolium L. -2.67 5 25
Juncus striatus Schousboe -2.67 5 25
Lamarckia aurea (L.) Moench -2.67 13 50
Picris echioides L. -1.68 18 75
Stachys arvensis L. -1.68 18 75
Cistus crispus L. -1.65 18 50
Chrysanthemum segetum L. -1.35 9 25
Dorycnium rectum (L.) Ser. -1.35 9 25
Fumaria officinalis L. -1.35 9 25
Ranunculus repens L. -1.35 9 25
Vicia sativa L. ssp. sativa -1.35 14 50
Lactuca serriola L. -1.20 9 50
Tolpis barbata (L.) Gaertner -1.16 32 75
Plant ecotype III (southern siliceous)
Alisma plantago-aquatica L. 2.60 11 33
Stachys arvensis L. 1.87 22 67
Populus nigra L. 1.63 33 100
Rumex conglomeratus Murray 1.63 33 100
Carex paniculata L. ssp. lusitanica (Schkuhr) Maire 1.48 22 33
Borago officinalis L. 1.15 22 67
Carex pendula Hudson 1.15 22 100
Chenopodium opulifolium Schrader 1.15 44 100
Cyperus longus L. 1.15 22 67
Fumaria capreolata L. 1.15 44 100
Plantago coronopus L. 1.15 11 33
Silene gallica L. 1.15 22 67
Torilis arvensis (Hudson) Link 1.15 44 100
Vicia sativa L. ssp. sativa 1.15 44 100
Malva L. sp. 1.04 55 100
Fraxinus angustifolia L. ssp. angustifolia 1.00 55 100
Figure 8.2 Hierarchical grouping of all the studied sites from the unweighted pair-group average clustering
method using Bray-Curtis similarity coefficient. The extent of the three ecotypes is delineated: I – northern
siliceous, II – calcareous, III – southern siliceous.
From the 20 environmental variables originally considered, only six were retained by the
forward selection, and were significantly related to the species distribution using the Monte
Carlo permutation test (Fig. 8.3). The first two CCA axes explained 19.1% of the total
biological variability and 60.3% of the species-environment relation. The geological
background (siliceous or calcareous) and the mean annual temperature, and to some extent
the distance to the source and related variables width and depth of the channel, were the
most important environmental factors determining the floristic variability. Two other basin-
scale variables, altitude and annual rainfall, were related to the second axis and associated
to siliceous areas, presumably from the northern part of the basin. The superimposition of
the previously established ecotypes onto the site scores confirmed their reality and
illustrated the major environmental gradients detected in the basin: geological background
and water availability (Fig. 8.3).
1.5
1.0
Altitude
0.5
Average annual
rainfall
CCA AXIS II
Siliceous
0.0
Distance
to source Calcareous
-0.5
Average annual
temperature
-1.0
-1.5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
CCA AXIS I
Figure 8.3 Canonical correspondence analysis biplot for the entire biological data set, considering the
forward selected environmental variables. Symbols indicate the site scores in the first two axes: - ecotype
I, - ecotype II, - ecotype III. The relative position and magnitude of the explanatory variables is
represented by arrows. Eigenvalues for the first and second canonical axes were 0.33 and 0.30, respectively.
Table 8.1 shows the species with a frequency of occurrence on each of the three ecotypes
higher than their frequency of occurrence on the Guadiana basin.
From the nine variables related to human disturbance and the use of the drainage basin
originally considered, five were retained by the forward selection, and were significantly
related to the species distribution using the Monte Carlo permutation test (Fig. 8.4). The
first two CCA axes explained 18.9% of the total biological variability and 55.9% of the
species-environment relation. The first axis opposed highly human disturbed areas (>3 point
effluent sources near the site, an important component of human settlements and irrigation
crops dominating the landscape) to less disturbed areas dominated by less water demanding
crops and forestry-related activities. However, superimposition of the previously
established ecotypes onto the site scores again illustrated the underlying dominant
environmental gradients, as the geological background and the water availability clearly
determine the type of land use, hence, the degree of human disturbance.
The influence of environmental and anthropogenic variables on river plant composition was
significant, P>0.005, for the first axis eigenvalue and the trace of both CCA analyses (Monte
Carlo simulation test with 1,000 permutations). Consequently, it can be concluded that the
floristic composition and the extracted variables on both canonical analyses were related.
1.5
1.0
0.5 Forestry
Urban areas
CCA AXIS II
Non-irrigation crops Irrigation crops
0.0
Site
effluents
-0.5
-1.0
-1.5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
CCA AXIS I
Figure 8.4 Canonical correspondence analysis biplot for the entire biological data set, considering the
variables related to human disturbance and the use of the drainage basin. Symbols indicate the site scores in
the first two axes - ecotype I, - ecotype II, - ecotype III. The relative position and magnitude of the
forward selected anthropogenic variables is represented by arrows. Eigenvalues for the first and second
canonical axes were 0.27 and 0.22, respectively.
8.5.3 Reference sites and floral deviation
For assessing the reference conditions, that is, the floristic composition that can be consi-
dered the ‘pro-natural’ reference for each ecotype, a canonical correspondence analysis
was performed for each of these, using the synthetic additive variable ‘anthropogenic
disturbance’ (Figures 8.5a, 8.6a and 8.7a, respectively for ecotypes I, II and III).
The floral composition of each ecotype is also subjected to the influence of environmental
variables, such as rainfall and distance to the source, the relative importance of which
varies according to the type and range of the site characteristics from each ecotype. For
example, average annual rainfall is a significant variable for determining floral composition
of ecotype I (northern siliceous) but not for ecotype III, where its value is relatively similar
for all of its sites. This within-group environmental variability was partialled out in each
ecotype CCA, by using the environmental variables forward selected previously, as co-
variables in the canonical procedure. As a result, and for each ecotype, the first canonical
axis displayed a gradient from the less disturbed to the more disturbed sites.
However, each group of less disturbed sites from each ecotype, still displayed an intrinsic
variability, which would be expected from biological assemblages. The hierarchical
clustering of the group of sites of each ecotype was used to help selecting the reference
sites on each canonical plane (Figures 8.5b, 8.6b and 8.7b, respectively for ecotypes I, II
and III). As a result, four floral reference sites were selected for ecotype I, three reference
sites for ecotype II and four sites for ecotype III (cut-off points on each cluster are
indicated by black arrows).
Table 8.2 shows the species with the highest scores on the first axis of each partial CCA
performed and their frequency of occurrence in the ecotype and in the selected group of
reference sites. All these species present a higher frequency of occurrence in the group of
reference sites than in the group of sites of the respective ecotype.
1.5
(a)
4
1.0
0.5 16 27
CCA AXIS II 7 8
Anthropogenic
0.0
disturbance
9
-0.5
10
1
2
-1.0
-1.5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
CCA AXIS I
(b)
Figure 8.5a Location of the sites from plant ecotype I (northern siliceous) on axis one and two of the canonical
correspondence analysis and position and magnitude of the combined variable anthropogenic disturbance.
Environmental variables were previously forward selected for this ecotype and used as co-variables. Eigenvalues
for the first and second canonical axes were 0.23 and 0.22, respectively.
Figure 8.5b Hierarchical grouping of ecotype I (northern siliceous) sites from the unweighted pair-group
average clustering method using Bray-Curtis similarity coefficient. The black arrow indicates the root of the
group of sites with floral reference conditions.
1.5
(a)
15
1.0 29
13
0.5
23
CCA AXIS II
6
28
0.0
Anthropogenic
disturbance
-0.5 20
22
24
-1.0
-1.5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
CCA AXIS I
(b)
Figure 8.6a Location of the sites from plant ecotype II (calcareous) sites on axis one and two of the canonical
correspondence analysis and position and magnitude of the combined variable anthropogenic disturbance.
Environmental variables were previously forward selected for this ecotype and used as co-variables.
Eigenvalues for the first and second canonical axes were 0.26 and 0.20, respectively.
Figure 8.6b Hierarchical grouping of sites of plant ecotype II (calcareous) sites from the unweighted pair-
group average clustering method using Bray-Curtis similarity coefficient. The black arrow indicates the root of
the group of sites with floral reference conditions.
(a)
36
4.0
3.0
CCA AXIS II
2.0
1.0
12 43 44
Anthropogenic 39 37
31 disturbance 42 33
19 40
0.0 41
18 17
45 11 14
26
32 3
34 30
-1.0
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

CCA AXIS I
(b)
Figure 8.7a Location of the sites from plant ecotype III (southern siliceous) sites on axis one and two of the
canonical correspondence analysis and position and magnitude of the combined variable anthropogenic
disturbance. Environmental variables were previously forward selected for this ecotype and used as co-
variables. Eigenvalues for the first and second canonical axes were 0.23 and 0.18, respectively.
Figure 8.7b Hierarchical grouping of sites from plant ecotype III (southern siliceous) sites from the
unweighted pair-group average clustering method using Bray-Curtis similarity coefficient. The black arrow
indicates the root of the group of sites with floral reference conditions.
20.0
18.0
ANTHROPOGENIC DISTURBANCE
16.0
14.0
12.0
10.0
8.0
6.0
4.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
BRAY-CURTIS COEFFICIENT
Figure 8.8 Linear regression of Bray-Curtis dissimilarity against the anthropogenic disturbance. Ecotype I ( ),
y=0.5949 x –15.827, R2=0.916, p=0.05 (n=9); Ecotype II ( ), y=0.2359 x + 2.2473, R2=0.647, p=0.05 (n=9);
Ecotype III ( ),y=0.2442 x –2.1574, R2 =0.8040, p=0.05 (n=22). Symbols filled in black indicate the reference
sites for each ecotype. A few symbols represent more than one site with the same dissimilarity value.
Linear regressions were obtained between the dissimilarity values of each site on the global
cluster (Fig. 8.2) and their values of the anthropogenic disturbance, for the group of sites
of each ecotype (Fig. 8.8). The probability values found indicate a strong relationship
between the synthetic variable used and the floral deviation of the sites. The slope of each
equation illustrates the range of floristic variability displayed by the group of sites from
each ecotype (northern siliceous presenting the smallest range). The spatial scattering of
the sites from each ecotype indicates the existing range of disturbance (southern siliceous
presenting the highest range). The spatial tightness of the reference sites from each
ecotype illustrates the magnitude of their intrinsic biological variability (the calcareous
ecotype presented the highest dissimilarity between the selected reference sites).
Based on the dissimilarity Bray-Curtis values obtained (Figures 8.5b, 8.6b and 8.7b), a cut-
off point of 0.50 was considered to separate sites with good floral quality from those
disturbed and presenting deviant plant associations. Five classes of quality status for river
plant assemblages were established, with the following dissimilarity thresholds: <0.45 (pro-
natural, very good river plant quality); 0.46-0.50 (small floristic deviation from pro-
natural); 0.51-0.55 (considerable floristic deviation); 0.56-0.60 (large floristic deviation)
and >0.60 (very large floristic deviation). Each quality status class from each ecotype is
indicated on Fig. 8.1.
8.6 Discussion
In this work, a multivariate approach for bioassessment was successfully applied to river
plant assemblages from southern Iberia, including the definition of the ecotypes existing in
the basin and the selection of their reference sites, both based on hierarchical clustering
and canonical ordination. Furthermore, both physicochemical and anthropogenic variables
were used in the procedure.
The methodological procedure developed included some subjective steps, necessitating

delicate decisions. Amongst these, there was the selection of the field sampling method,
such as the reach length surveyed and the lateral limits of the survey, both previously
determined for southern Iberian rivers (Ferreira and Moreira 1999). Also critical was the
time of the sampling, which is dependent on river flow and within and between year
temperatures. This had also been previously established (Ferreira et al. 2001).
Field procedures and time of sampling have been pointed out as critical to the success of
riparian surveys (e.g. Szaro and King 1990) and also for final results of bioassessment using
the multivariate approach (Bunn and Davies 2000; Wright 2000). Ideally, environmental
attributes should be relatively stable over time and monitored communities should be
persistent and presenting a deterministic underlying structure (Wright 2000). However,
intermittent rivers often show low persistence and biological assemblages are unpredictable
over time (Gasith and Resh 1999; Bunn and Davies 2000) and in some cases multivariate
models have been developed for different seasons of the year (Humphrey and Storey 2000).
In this study, using a specific time of sampling before the dry period enabled the study of
the river plant assemblages at a point of their highest spatial organization (Ferreira et al.
2001). For other communities and systems, further studies would be needed to establish the
best time for sampling.
There is also an important decision over which biological components of plant assemblages
should be considered. Whether or not the adventitious terrestrial component of the river
plant assemblages should be incorporated into the evaluation of the ecological quality, may
well be a critical issue for intermittent southern Iberian rivers. Drying in Mediterranean-
type streams is a continuous, gradual process over the summer period, with important
changes in the composition and life forms of riparian vegetation (Gasith and Resh 1999).
The exclusion of the terrestrial component results in a reduction of river plant richness per
site; however, its inclusion frequently results in much less predictable patterns of river
plant distribution, with terrestrial species as fortuitous plant group indicators (Ferreira
1994; Ferreira and Moreira 1999).
A certain number of exotic species, such as Arundo donax L., Paspalum paspalodes and
Cyperus eragrostis, became well established in southern Iberian regions and developed
persistent, naturalised populations (senso Richardson et al. 2000). The presence of
naturalised exotic species is a common occurrence in most sites and it is likely that its
exclusion from the multivariate data treatment would produce biased results on the
multivariate model. On the other hand, its inclusion potentially leads to their presence on
the reference sites. Because the number and abundance of exotic species is generally low
in these river systems, and they are consistently widespread, none is reported in Table 8.2
as significant for the reference sites. However, dominance of a site by exotic species (alias
site invasion) will be related to an increase of man-made disturbance on the river channel
(Aguiar et al. 2001) and on its floodplain (Ferreira and Moreira 1995), thus contributing to
the position as the disturbed site on the partial CCA. The inclusion of the exotic species on
data treatment can also be considered a side-effect of the multivariate approach, which
deals with the presence and/or the relative abundance of taxa present in each site and not
with biological parameters such as indicator species, taxonomic groups or functional
metrics. In the multimetric approach, for example, the presence or abundance of exotic
species would be usual metrics to consider (Barbour et al. 1999), but then they should be
tested as if significant for the indication of anthropogenic stress and to determine river
condition.
In this study, all species found were incorporated into data treatment, including those
which were not strictly aquatic plants. Aquatic species in mesic fluvial systems are few,
frequently ubiquitous and highly tolerant to adverse abiotic conditions, tending to develop
large mono-specific stands whenever their specific requirements are met by locally
available habitat conditions (Haslam 1987; Ferreira and Moreira 1999). When the terrestrial
species are not incorporated in the methodological procedure, helophytes and hygrophytes
generally determine plant groups, and not hydrophytes. However, the distribution patterns
obtained have been shown to result in a relatively homogeneous composition and few
floristic types at the basin scale (Ferreira 1994; Ferreira et al. 1998), and in such case, the
ecotypes would probably be structured at a supra-basin level and be relatively few across
the ecoregions. For example, the two siliceous ecotypes from the Guadiana basin would
probably coalesce and eventually become linked to other siliceous groups from the near-by
basins, such as Tejo, Sado and Guadalquivir.
Different biological assemblages express themselves at different spatio-temporal scales and

also present different responses to environmental changes at a range of scales (Habersack
2000). In this study, the water availability and the geological background were related to
land use and disturbance, and together determined plant distribution.
The environmental and anthropogenic variables used in data treatment permitted an

understanding of the abiotic factors influencing river vegetation and explaining the
ecotypes. The number of variables considered in the data treatment will depend on data
availability, but it should be as high as possible, unless previous country-wide studies have
already identified the specific group of variables that determine the distribution of each
biological assemblage within each ecotype (e.g. Holmes et al. 1998). Finally, some parts of
the data treatment may include a certain degree of (best) professional judgement, such as
the cut-off points of the various clusters chosen to determine the ecotypes and the
reference sites.
The benefits and difficulties associated with the acquisition and use of reference sites were
discussed recently in Reynoldson and Wright (2000). Traditionally, multivariate approaches
use references sites previously identified as minimally impacted by considering their hydro-
geomorphological and physico-chemical features. For this approach, there is a need for
comprehensive environmental data, which is not easily available for Iberian rivers (Alba-
Torcedor and Pujante 2000). To overcome this limitation, multivariate techniques were also
used to identify reference conditions in this study, but following the identification of the
ecotypes and the environmental features that determine them, using the first and overall
canonical procedure. The incorporation of multivariate analysis for defining reference
conditions and assessing the impairment of test sites, though recently suggested by some
authors (Gasith and Resh 1999) has not been applied until now, but may be the only
possible alternative for regions with low availability of environmental data and unclear
regional ecotypes. In the whole procedure, the basis of the multivariate approach was
maintained, in which biological assemblages are brought together with no previous
subjective assumptions about environmental features that influence species occurrence.
Nonetheless, uneven distribution of data, outliers and the absence of data from minimally
disturbed sites will distort this model.
The Mediterranean basin was one of the first regions of the world to suffer the
environmental impacts of human disturbances, such as forest clearing, grazing pressure and
agriculture (di Castri 1991). Reference sites, albeit being closer to the original
environmental conditions, cannot be considered pristine; indeed it will be very difficult to
find pristine situations in Portugal or anywhere else in Europe (Verdonschot 2000).
However, both reference and disturbed sites were subjected to the same geological and
evolutionary history of endogenous (natural) disturbances together with an early human-
related history of exogenous disturbances. A high level of recent human-made disturbance
throughout the basin will interfere with the initial cluster that identifies the ecotypes and
eventually short circuit the multivariate procedure. At the ecotype level, a high
disturbance throughout the area will make the establishment of the reference conditions
difficult. In this case study, all sites from the calcareous ecotype presented some degree of
disturbance resulting from its type of soils, which are quite favourable for agricultural
activities. Working at a supra-basin scale would probably improve the application of the
multivariate approach by promoting a higher number of sites per ecotype and a wider
gradient of disturbance, from the less to the highest perturbed sites.
Ecotypes made up of a group of sites including low variability, whether disturbed or

preserved, will present the largest difficulty for the determination of reference sites,
because their range of dissimilarity will be small. Also, there is a lower limit to the number
of sites that an ecotype can have, below which there after will be a lack of detectable
vectors of disturbance. Both cases were illustrated in the present study. The northern
siliceous ecotype presented a narrow range of disturbance and therefore its reference
conditions were harder to define. The southern siliceous ecotype had a relatively small
number of sites, but included three sites strongly affected by regulation, which increased
its range of disturbance, allowing a good distinction between less and more disturbed sites.
In summary, therefore, for the application of the multivariate method, there must be a
range of (intermediate) biological variability and perturbation within each ecotype, with a
reasonable number of surveyed sites.
8.7 References
Aguiar F.C., Ferreira M.T. and Moreira I. 2001. Exotic and native vegetation establishment following
channelization of a western Iberian river. Regul. Rivers: Res. Mgmt. 17: 509-526.
Alba-Torcedor J. and Pujante A.M. 2000. Running water biomonitoring in Spain: opportunities for a
predictive model. In: Wright J.F., Sutcliffe D.W. and Furse M.T. (eds), Assessing Biological
Quality of Freshwaters. RIVPACS and other techniques. The Freshwater Biological
Association, Ambleside, Cumbria. pp. 207-217.
Barbour M.T., Gerritsen J., Snyder B.D. and Stribling J.B. (eds). 1999. Rapid bioassessment protocols
for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and Fish.
2nd Ed. EPA/841-B-98-010. - EPA Office of Water, Washington DC.
Boon P.J. 2000. The development of integrated methods for assessing river conservation value.
Bunn S.E. and Davies P.M. 2000. Biological processes in running waters and their implications for the
assessment of ecological integrity. Hydrobiology 422-423: 61-70.
Davies P.E. 2000. Development of a national river bioassessment system (AUSRIVAS) in Australia. In:
Wright J.F., Sutcliffe D.W. and Furse M.T. (eds), Assessing the Biological Quality of
Freshwaters: RIVPACS and other techniques. Freshwater Biological Association, Ambleside,
Cumbria. pp. 113-124.
Dawson F.H. and Szoskiewicz K. 1999. Relationship of some ecological factors with the associations of
vegetation in British rivers. Hydrobiology 415: 117-122.
Dawson F.H. 1988. Water flow and vegetation of running waters. In: Symoens J.J. (ed), Vegetation of
Inland Waters. Kluwer Academic Publ., Dordrecht. pp. 283-310.
Décamps H. and Tabacchi E. 1992. Species richness in vegetation along river margins. In: Giller P.S.,
Hildrew A.G. and Raffaelli D.G. (eds), Aquatic Ecology. Scale, Pattern and Process.
Blackwell Scientific Publications, Oxford. Pp. 1-20.
climate. In: Groves R.H. and di Castri F. (eds), Biogeography of Mediterranean Invasions.
mediterranean agricultural systems. In: Pysek P., Prach K., Rejmanek M. and Wade M.
(eds), Plant Invasions - General Aspects and Special Problems. SPB Academic Publishing,
influencing their distribution. Hydrobiology 415: 101-107.
Ferreira M.T., Godinho F.N. and Cortes R.M. 1998. Macrophytes in a southern iberian river. Vehr.
macrophytes in a southern Iberian river. Vehr. Internat. Verein. Limnol. 27: 3833-3837.
Franco J.A. and Rocha-Afonso M.L. 1994; 1998. Nova Flora de Portugal (Continente e Açores). 3(1)
Alismataceae-Iridaceae. 3(2) Gramineae. Author’s Edition, Lisboa.
Franco J.A. 1971; 1984. Nova Flora de Portugal (Continente e Açores). 1. Lycopodiaceae-
Umbelliferae. 2. Clethraceae-Compositae. Author’s Edition, Lisboa.
51-81.
Haslam S. 1987. River Plants of Western Europe. Cambridge University Press, Cambridge.
Haury J. and Peltre M.-C. 1993. Interêts et limites des indices macrophytiques pour qualifier la
mésologie et la physico-chimie des cours d’eau: exemples armoricains, picards et lorrains.
Annals Limnol. 29: 239-253.
Haury J., Peltre M.C., Muller S., Tremolieres M., Barbe J., Dutartre A. and Guerlesquin M. 1996. Des
indices macrophytiques pour estimer la qualité des cours d’eau français: premieres
propositions. Ecologie 27: 233-244.
Holmes N.T.H., Boon P.J. and Rowell T.A. 1998. A revised classification system for British rivers based
on their aquatic plant communities. Aquatic Conservation: marine and freshwater
ecosystems 8: 555-578.
Hughes R.M. 1995. Defining acceptable biological status by comparing with reference conditions. In:
Davies W.S. and Simon T.P. (eds), Biological Assessment and Criteria. Tools for Water
Resource Planning and Decision Making. Lewis Publishers, Boca Raton, Florida. pp. 31-47.
Humphrey C.L. and Storey A.W. 2000. AUSRIVAS: operator sampling processing errors and temporal
variability - implications for model sensibility. In: Wright J.F., Sutcliffe D.W. and Furse
M.T. Assessing Biological Quality of Freshwaters. RIVPACS and other techniques. The
Freshwater Biological Association, Ambleside, Cumbria. pp. 143-164.
Johnson R.K., Wiederholm T. and Rosenberg D.M. 1993. Freshwater monitoring using individual
organisms populations and species assemblages of benthic macroinvertebrates. In:
Rosenberg D.M. and Resh V.H. (eds), Freshwater Monitoring and Benthic
Macroinvertebrates. Chapman and Hall, New York. pp. 40-158.
Karr J.R. and Chu E.W. 2000. Sustaining living rivers. Hydrobiology 422-423: 1-14.
Lange L. and van Zon J.C. 1981. A system for the evaluation of aquatic biotopes based on the
Ludwig J.A. and Reynolds J.F. 1988. Statistical Ecology: A Primer on Methods and Computing. John
Can. Journ. Fish. Aquatic. Sc. 51: 2856-2865.
Moulton T.P. 1999. Biodiversity and ecosystem functioning in conservation of rivers and streams.
Aquatic conservation: marine and freshwater ecosystems 9: 573-578.
Nilsson C. and Jansson R. 1995. Floristic differences between riparian corridors of regulated and free-
flowing boreal rivers. Regul. Riv.: Res. Mgmt. 11: 55-66.
Nilsson C., Nilsson E., Johansson M.E., Dynessius M., Grelsson G., Xiong S., Jansson R. and Danvind M.
419-431.
Norris R.H. and Thoms M.C. 1999. What is river health? Freshwater Biology 41: 197-209.
Penuelas J. and Sabater F. 1987. Distribution of aquatic plants in relation to environmental factors in
the Ter river, N.E. Spain. Int. Rev. Ges. Hydrobiol. 78: 41-58.
Pollard P. and Huxham M. 1998. The European Water Framework Directive: a new era in the
management of aquatic ecosystem health? Aquatic conservation: marine and freshwater
Rapport D.J., Constanza R. and Mcmichael A.J. 1998. Assessing ecosystem health. Trends Ecol. Evol.
13: 397-402.
Reynoldson T.B. and Wright J.E. 2000. The reference condition: problems and solutions. In: Wright
J.F., Sutcliffe D.W. and Furse M.T. Assessing Biological Quality of Freshwaters. RIVPACS
and other techniques. The Freshwater Biological Association, Cumbria. pp. 293-304.
Reynoldson T.B., Norris R.H., Resh V.H., Day K.E. and Rosenberg D.M. 1997. The reference condition:
a comparison of multimetric and multivariate approaches to assess water quality-
impairment using benthic macroinvertebrates. J. North Amer. Benthol. Soc. 16: 833-852.
Richardson D.M., Pysek P., Rejmanek M., Barbour M.G., Panetta F.D. and West C.J. 2000.
Rohlf F.J. 1993. NTSYS-Pc, Numerical Taxonomy and Multivariate Analysis System, Ver. 1.80. Exeter
Shrader-Frechette K. 1995. Hard ecology, soft ecology and ecosystem integrity. In: Westra L. and
Lemons J. (eds), Perspectives on ecological integrity. Kluwer Academic Publishers,
Dordrecht, The Netherlands. pp. 125-145.
Szaro R.C. and King R.M. 1990. Sampling intensity and species richness: effects of delineating
southwestern riparian plant communities. Forest Ecol. & Mgmt. 33-34: 335-349.
diversity in riparian plant communities: a longitudinal comparative study. Regul. Rivers:
Res. Mgmt. 12: 367-390.
ter Braak C.J.F. and Smilauer P. 1998. CANOCO. Reference manual and user´s guide to CANOCO for
Windows: software for Canonical Community Ordination Ver. 4. Microcomputer Power,
Ithaca, NY.
ter Braak C.J.F 1990. CANOCO Ver. 3.1. Update notes. Agricultural Mathematics Group, Wageningen.
ter Braak C.J.F. and Verdonschot P.F.M. 1995. Canonical correspondence analysis and related
multivariate methods in aquatic ecology. Aquatic Sc. 57: 255-289.
Tutin T.C., Heywood V.H., Burges N.A, Moore D.M., Valentine D.H., Walters S.M. and Webb D.A. 1980.
Flora Europaea. Cambridge University Press, Cambridge.
Verdonschot P.F.M. 2000. Integrated ecological assessment methods as a basis for sustainable
catchment management. Hydrobiology 422-423: 389-412.
Ward J.V. and Stanford J.A. 1995. The serial discontinuity concept: extending the model to floodplain
rivers. Regul. Riv.: Res. Mgmt. 10: 159-168.
Hydrobiology 79: 91-100.
Wright J.F. 2000. An introduction to RIVPACS. In: Wright J.F., Sutcliffe D.W. and Furse M.T. (eds),
Assessing Biological Quality of Freshwaters. RIVPACS and other techniques. The Freshwater
Biological Association, Ambleside, Cumbria. pp. 293-304.
Wright J.F., Sutcliffe D.W. and Furse M.T. (eds) 2000. Assessing Biological Quality of Freshwaters.
RIVPACS and other techniques. The Freshwater Biological Association, Ambleside, Cumbria.
Parte IV
Biodiversidade da vegetação ripícola
e invasão por plantas exóticas em
corredores fluviais Mediterrânicos
Part IV Riparian biodiversity

and exotic invasions in Mediterranean
fluvial corridors
If you can look into the seeds of time,

And say which grain will grow and which will not,
Speak then to me . . .
William Shakespeare (1564-1616)

Capítulo 9 Estabelecimento da vegetação
exótica e nativa num rio canalizado
mediterrânico
Chapter 9 Exotic and native vegetation

establishment following channelization of a western
Iberian river*
*
PUBLISHED AS: Aguiar F.C., Ferreira M.T. and Moreira I. 2001. Exotic and native vegetation
establishment following channelization of a western Iberian river. Regulated Rivers: Research and
200 | Parte IV - Biodiversidade Ripícola e Invasão por Plantas Exóticas
Capítulo 9 - Vegetation Establishment on a Channelized Iberian River | 201
9.1 Abstract
Channelization is often a major cause of human impacts on river systems. It affects both
hydrogeomorphic features and habitat characteristics and potentially impacts riverine flora
and fauna. Human-disturbed fluvial ecosystems also appear to be particularly vulnerable to
exotic plant establishment. Following a twelve-year recovery period, the distribution,
composition and cover of both exotic and native plant species were studied along a
Portuguese lowland river segment, which had been subjected to resectioning, straightening
and two-stage bank reinforcement, and were compared with those of a nearby, less
impacted segment. The species distribution was also related to environmental data. Species
richness and floristic composition in the channelized river segment were found to be similar
to those at the more ‘natural’ river sites. Floral differences were primarily consistent with
the dominance of cover by certain species. However there were significant differences in
exotic and native species richness and cover between the ‘natural’ corridor and the
channelized segment, which was more susceptible to invasion by exotic perennial taxa,
such as Eryngium pandanifolium, Paspalum paspalodes, Tradescantia fluminensis and
Acacia dealbata. Factorial and canonical correspondence analyses revealed considerable
patchiness in the distribution of species assemblages. The latter were associated with small
differences in substrate composition and their own relative position across the banks and
along the river segments in question. Data was also subjected to an unweighted pair-group
arithmetic average clustering, and the Indicator Value methodology was applied to selected
cluster noda in order to obtain significant indicator species.
Key words: river plants, channelized river, species richness, exotic species, indicator value,
Portugal.
9.2 Introduction
Conventional river engineering operations often produce major anthropogenic impacts on

the fluvial ecosystem (Brookes 1988) and channelization affects nearly all the hydrogeo-
morphic characteristics and processes both on and downstream of a channelized reach
(Hupp 1992).
Numerous studies have shown how channelization generally reduces the physical
heterogeneity of river beds and banks, accelerates erosional processes and changes flow
and sediment load patterns, and how consequently, at ecosystem level the river system
experiences a reduction of habitat heterogeneity, niche potential, and frequently
ecological diversity as well (MacCarthy 1985; Swales 1988; Shields Jr. and Hoover 1991;
Jongman 1992; Petersen et al. 1992; Higler 1993). Fewer case studies draw a comparison
between the impact of channelization and other anthropogenic factors that influence rivers
(Boon 1988; Iversen et al. 1993) or look at vegetation response and recovery in this context
(Bravard et al. 1986; Brookes 1986; 1995; Hupp 1992).
Different river systems or reaches possess differing degrees of susceptibility to and

potential to recover from disturbances. Lowland alluvial systems are amongst those with
the highest recovery potential (Petts 2000), probably because juxtaposed land and water
patches occupy relatively stable positions in a landscape where disturbances are less
frequent than the patch adjustment response (Wissmar and Swanson 1990); however,
channelization of lowland rivers often results in the loss of the very riparian wetlands that
play a major role in ensuring this recovery potential (Petersen et al. 1992).
Diverse riparian ecosystems are maintained by an active natural disturbance regime working
over a wide range of spatial and temporal scales (Naiman and Décamps 1997). A river’s
geomorphological dynamics are fundamental to the maintenance of vegetation diversity –
channel change and disturbance flood events are the key controls (Bravard and Petts 1996).
Habitat mosaics are destroyed during floods, thereby creating new patches for colonisation
by pioneer species and the rejuvenation of early successional stages. The disproportionately
high species richness in some riparian zones (Naiman et al. 1993) can be partly explained by
their patchiness or habitat heterogeneity, as well as by high-frequency natural disturbance
regimes (Gregory et al. 1991; Décamps and Tabacchi 1992; Pollock et al. 1998; Ward et al.
1999), all of which are affected by channelization. Bravard et al. (1986) examined
vegetation dynamics along the River Rhône (France) and showed that the building of flood
embankments on the river and a reduction in the extent of the geomorphologically active
river bed led to the loss of both pioneer vegetation communities and high diversity.
In some situations, however, channelization has had only small long-term effects (Duvel et
al. 1976). Management regimes involving repeated disturbance by dredging, plant cutting
and the use of herbicides along man-made drainage channels and navigation canals consis-
tently induced changes, albeit small ones, in vegetation composition (Wade and Edwards
1980; Murphy et al. 1987). Brookes (1986) also reported that following the widening of
some English and Welsh rivers there was an increase in vegetation biomass during channel
and plantform morphological adjustment, but no changes in species composition. In fact,
the rate and nature of vegetation recovery following channelization has been found to be
dependent upon a number of factors – particularly the substrate of the riverbed (Brookes
1987).
Unlike some other European countries such as Denmark and England (Brookes 1988; Iversen
et al. 1993), Portugal only has a few, localised stretches of channelized river, most of
which are set in urban environments. Most channel networks are artificial and related to
crop irrigation (Ferreira et al. 1998a; Ferreira and Moreira 1999). However, there are a few
channelized river segments located on large floodplains, not only for reasons related to
flood control, but also for crop protection, the maximisation of land use and the drainage
of irrigation water. In these cases the riverbed has been lowered, while riparian land is
periodically cleared of trees and riverbank edges are cleaned. However, little attention has
so far been paid to the characteristics of the riparian vegetation that develops extensively
following an engineering intervention. What is more, human-influenced disturbance in
Mediterranean-type aquatic ecosystems has been poorly documented (Gasith and Resh
1999).
Our study looked at the riparian vegetation of a lowland river in Portugal 12 years after
river channelization, and compared it to that of non-channelized reaches. Besides species
composition and the foliar cover provided by both woody and herbaceous species, particular
attention was paid to the relative proportion of exotic and native species in plant
assemblages.
9.3 Study area
The River Mondego possesses a catchment area of 6,671 km2 - among the largest of all
those that are located entirely in Portuguese territory. The river is born at 1,547 m above
sea level in the Serra da Estrela mountains and flows for about 227 kms, running in a
general Northeast-Southwest direction before entering the Atlantic Ocean (Fig. 9.1). The
floodplain on which this study was conducted spreads right across a low-lying area (less
than 300 m amsl). It possesses a complex geological background, comprising a thick layer of
recent fine Tertiary deposits and Quaternary alluvia over a Hercinic basement. For
centuries this floodplain has been primarily devoted to agricultural activities and related
urban development and industries.
Figure 9.1 The Mondego river basin, showing the floodplain and the location of the nine studied sites. The
arrow points to the upper limit of the brackish water influence.
The natural hydrological regime, which undergoes marked seasonal fluctuations, includes a
succession of large floods in late autumn and winter (high flow levels of more than 1,200
m3/s have been documented since the thirteenth century) and extensive periods of low
summer flow. A large-scale river regulation scheme including three upstream reservoirs and
the channelization of the main river and tributaries downstream from the city of Coimbra
was implemented between 1980 and 1986 for flood protection, irrigation and land drainage
purposes (Fig 9.1). The alluvial floodplain, (c. 15,000 hectares) is mostly used to grow rice,
maize and other water-consuming Mediterranean crops, and has a complex system of
irrigation and drainage channels and ditches.
The channelized river segment is about 40 kms long. It was subjected to reprofiling by
dredging, straightening and two-stage bank reinforcement. The width of the channelized
segment varies between 60 and 140 m, while bank height ranges from 2.5 to 6 m. The bank
substrate consists of boulders with local origins covered by cobbles and pebbles. Finer
materials – mainly sand – cover this hard rock matrix. The riverbed is mostly uniform and
cobbles and coarse sand predominate.
Riparian vegetation was nearly eliminated during the resectioning of the river channel. The
reestablishment of bed and bank vegetation proceeded spontaneously thereafter. This study
was conducted on the freshwater part of the channelized river segment down to the limit of
brackish water influence (Fig. 9.1).
9.4 Methods
9.4.1 Sampling procedures
Sampling was carried out at six sites located at regular intervals along the channelized
segment of the Mondego (sites 1 to 6 in Fig. 9.1) in May-June 1999. The first sampling site
was positioned 15 kms upstream of the estuarine area to avoid the effects of salinity. Three
longitudinal transects consisting of five 5 x 25 m plots set in a sequence were laid out at
each studied site. The first transect was positioned in such a way as to include the water
and the inner bank, while the others were placed on each of the two-stage terraces.
As no detailed vegetation studies had been carried out on the main channel of the Mondego
prior to channelization, three sites on the main river upstream of Coimbra were included
for comparative purposes. Their characteristics may be considered to be closer to those of
the ‘natural’ environmental conditions occurring in the river (sites 7 to 9 on Fig. 9.1). At
these sites transects were located in the corresponding zones of the river channel – i.e. the
instream, waterlogged margin, inner-bank and outer-bank points.
The richness and percentage of foliar cover in each plot were visually estimated for each
exotic and native species by averaging the results from 2.5 x 2.5 m sub-plots (n=10 in each
plot). Nomenclature followed Tutin et al. (1968; 1980), Franco (1971; 1984) and Franco and
Rocha-Afonso (1994; 1998). Exotic plants (synonym: alien) consist of recently or remotely
introduced plant taxa that can reproduce consistently and sustain populations over many
life cycles without direct intervention by humans [as per Richardson et al. (2000)].
The dominant substrate, the distance from the water and the height above the water were
measured for each plot, using no less than 10 randomly set assessment points. The
substrate types considered were large substrate (more than 64 mm, including cobbles and
boulders), medium substrate (between 2-64 mm, gravel and pebble) and fine substrate (less
than 2 mm).
The river corridor width and the wetted channel width were also determined at each site.
The average depth of the river channel was determined during plant surveys, using a
graduated dip-net pole at the outer limit of the instream transect and at no less than 10
assessment points. Conductivity (µS/cm) was determined on site using WTW field probes.
Distance to the sea (kms) was obtained for each site from 1:50 000 topographical maps. The
dominant land use of the surrounding terrestrial environment was also taken into account
by considering the following types of vegetation and land use: sedges of Juncus and
Cyperus; shrubland and woodland; irrigation crops and rice fields.
First of all a correspondence analysis (CA) was carried out using the CANOCO 4.0
programme (ter Braak and Smilauer 1998), so as to evaluate channelization effects by
comparing plant species composition and distribution in ‘natural’ and channelized plots. In
order to improve the ordination trends, rare species were also downweighted.
In every test non-parametric Mann-Whitney U tests using STATISTICA/w 5.0 (StatSoft Inc.
1995) were performed and significance was assessed at p<0.05 in order to compare the
number and foliar cover of native and exotic species of channelized and ‘natural’ transects.
Canonical correspondence analysis (CCA) performed with CANOCO was used to study the
distribution trends of exotic and native species and their relationship with the
environmental data within the channelized river corridor. Environmental variables were log
transformed (logx+1) and standardised by row centring (Ludwig and Reynolds 1988).
Terrestrial vegetation types were used as dummy variables. Of the variables that were
initially considered, only the subset of the best predictors was retained for the analysis,
using a forward selection procedure similar to a step-wise regression available in CANOCO.
A cut-off point of P<0.05 was used in this routine. A Monte Carlo simulation of both the first
axis eigenvalue and the trace (i.e. the sum of all canonical eigenvalues) was used to test
the significance of the effects under analysis (ter Braak 1990).
The classification of the vegetation groups that were found within the channelized river
segment was performed with NTSYS-pc 2.0 (Rohlf 1993), using the Bray-Curtis dissimilarity
coefficient and the ‘unweighted pair-group arithmetic average’ clustering method
(UPGMA). The IndVal 2.0 programme (Dufrêne and Legendre 1997) was used to determine
the indicator species and species assemblages that characterised the groups of sites that
had been obtained using the previous typology. Indicator species are defined as those that
are the most characteristic of a given typological group and are most frequently and
abundantly found mostly at the majority of sites belonging thereto. For each species i in
each site group j, the ‘Indicator Value’ (IndValij) is the product of Aij, which is the mean
abundance of a species i in the sites of group j compared to all groups in the study, and Bij,
which is the relative frequency of occurrence of species i in sites of group j (Dufrêne and
Legendre 1997). The index (multiplied by 100) is at its maximum (100%) when a given
species is observed at all the sites of a certain group and at no others. A random procedure
with 499 permutations was used to reallocate sites within each site group and served to test
the significance of the maximum indicator value of a species i, which is the largest value
for IndValij observed in all groups j belonging to the typology in question.
9.5 Results
9.5.1 Vegetation patterns at channelized and ‘natural’ sites
Ecologists commonly use CA ordination to summarise and analyse biological data on species
distribution. The first axes are interpreted as latent variables and are presumed to relate
underlying variables (ter Braak 1985). The results of the correspondence analysis that was
applied to the entire vegetation data set are shown in Fig. 9.2. The positions of the
transepts along the axes were overlain by the site affiliation, thereby enabling the
visualisation of the major features of the vegetation spatial patterns.
The eigenvalues of the first two CA ordination axes were 0.65 and 0.56 respectively, and
together they explain 37.6% of the total variance of the species data. These eigenvalues are
relatively high compared to those commonly obtained in community ecology studies (ter
Braak and Smilauer 1998). The first axis can be interpreted as a transverse gradient from
the water to the outerbank/2nd terrace. However, the inner bank of the ‘natural’ sites
presented a species composition that was closer to the marginal vegetation of the river
channel and the 1st terrace of the channelized bank, whilst the vegetation on the 2 nd
terrace clearly diverges in terms of species composition and is more related to a fully
terrestrial environment, despite being generally closer to the water than the transects on
the outer bank.
Channelized instream
3
Channelized 1st terrace
Channelized 2nd terrace
'Natural' instream
2 'Natural' inner bank

'Natural' outer bank
CA AXIS II
-1
-2
-2 -1 0 1 2 3
CA AXIS I
Figure 9.2 Axis one and two of the correspondence analysis using the floral composition of the transects from
the nine channelized and ‘natural’ sites studied on the River Mondego.
The second axis might be related to the disturbance effects of the channelization, with the
more ‘natural’ transects located in the positive part of the graph and the channelized
transects in the negative part. The environmental gradient associated with the disturbance
effects was more obvious in the instream communities, though the inner-bank/1st terrace
communities also revealed considerable group separation between channelized and
‘natural’ transects. However, there was a lesser degree of separation between the ‘natural’
and channelized environments on the outer bank/2nd terrace.
9.5.2 Channelization effects on species richness and species cover
Vegetation response to channelization was also analysed at species level, using the
observed values of the richness and cover of both exotic and native species. The total
number of species found in channelized and ‘natural’ sites was quite similar, with 150
species present in the ‘natural’ corridor, 160 species on the channelized part of the river
and 98 species common to both. Only 21 of the 212 species recorded were exotic, in
addition to which a relatively low number of exotic species was found per plot (maximum 8
species, average 2.8). However, in many plots exotic species cover was found to be
substantial, especially on the channelized river. Differences in species richness and species
cover between the channelized and the ‘natural’ sites were tested using the Mann-Whitney
U-test. The results of the statistical analysis are summarised in Table 9.1.
Table 9.1 Comparison of native and exotic species richness and cover (means 1 SE in parentheses) in the
transects from natural and channelized sites. p denotes probability (Mann-Whitney U-tests).
Natives Exotics
Richness (nº) Cover (%) Richness (nº) Cover (%)

*
Instream Channelized 14.1 30 4.5 41.5*
(1.46) (3.16) (0.27) (4.43)

‘Natural’ 15.3 46.3 2.2 3.1
(2.56) (9.11) (0.38) (0.69)
p† 0.7451 0.3476 <0.0001 <0.0001

* * * *
1st terrace/ Channelized 25.2 55.3 5.4 62.5
inner bank (1.23) (2.66) (0.31) (4.11)

‘Natural’ 17.0 81.3 2.9 29.7
(2.64) (8.10) (0.47) (9.65)
p† 0.0075 <0.0001 <0.0001 0.0039
2nd terrace/ Channelized 16.4 58.3* 2.4* 38.0
outer bank (1.01) (5.10) (0.26) (6.62)

‘Natural’ 20.2 82.7 3.9 15.2
(1.63) (4.19) (0.46) (3.32)
p† 0.0672 0.0065 0.0155 0.1204
†
- Mann-Whitney U-test, two-tailed tests of probability
*
- significant differences between channelized and natural groups (p<0.05)
In the instream plots there were no significant differences in total and native species
richness between the two site-groups (Mann-Whitney U-test, p>0.05). However, the richness
and cover of exotic species were consistently greater in the channelized plots. A low
floristic diversity was found in the waterlogged and submerged area in both the channelized
and the ‘natural’ plots.
A species assemblage that was dominated by riparian trees and included a variety of trees,
shrubs and woody climbers occupied the first terrace of the channelized embankment. Many
herbaceous hygrophilous species were established under the relatively open riparian canopy
and presented the highest average number of native species per plot. However, as in the
instream plots, the richness and cover of exotic species were significantly greater in the
channelized inner-bank plots than they were in the corresponding ‘natural’ ones.
The second terrace exhibited a significantly lower number of exotic species than the outer
bank of the ‘natural’ sites, but possessed a significantly higher cover, whereas native
species tended to present a significantly higher cover on the ‘natural’ corridor. The total
exotic cover and the average cover of the main exotic species per plot (selected species
possessed an average cover >50% in all the plots in each group of transects) are set out in
Fig. 9.3.
The instream transects of the channelized river revealed a step-by-step longitudinal

substitution of the dominant exotic species, which had begun downstream and had
gradually extended upstream. Eryngium pandanifolium, a swamp-dwelling plant from the
South-American subtropical zone, dominated the instream plots at downstream sites, but
was substituted by Paspalum paspalodes at the upstream ones. E. pandanifolium was also
present at the same first-terrace sites, but in more shady habitats shared the ground
surface with dense stands of Tradescantia fluminensis.
On the second terrace the increase in exotic cover was due mainly to the presence of the
woody Australian species Acacia dealbata.
P. paspalodes, T. fluminensis and A. dealbata were also found at the ‘natural’ sites, but,
with the exception of T. fluminensis on the first terrace, consistently possessed a lower
degree of cover.
Figure 9.3 Mean and mean standard error using pooled variance of dominant exotic species cover and total
exotic species cover (multiplier 10) on the plots from the nine studied sites.
9.5.3 Environmental variables related to species distribution
Seven variables out of the fourteen initially considered were selected by the CANOCO
forward selection procedure and accounted for 36.1% of the species distribution. Distance
to the sea, channel depth and water conductivity were strongly correlated with one another
(Pearson product moment correlation >|0.90|, P<0.05). To avoid multicollinearity, only the
first of these variables was used for the CCA, although it is possible to use channel depth
and conductivity for interpretation purposes (ter Braak 1990). The first two CCA axes
(eigenvalues of 0.64 and 0.37, respectively) explained 21.2% of the total biological
variability and 58.8% of the species-environment relationship.
The first axis clearly represented a transverse moisture gradient (water→terraces) and had
a high correlation with the variables ‘height above water’ (r=0.93) and ‘distance to the
water’ (r=0.95). The finer elements of the substrate composition (sand and silt) were more
closely related to the instream environment (Fig. 9.4).
1
2.5 Instream plots
2
12
1st terrace plots
2 2nd terrace plots
2
1
1.5 111
1
2
1
2 2 1
1 1 2
1
BOULDERS 1 1
CCA AXIS II
1 1 2
22 2 2
2 2
0.5
33
3 3 54 3 3
DISTW
3
0 2 65 64 3 HEIGHTW
3 6 6 3 6 4
SAND.SILT 5 5 4 6 4 4
3 3 4 3
3 6 4 5 6 64 6 4
WIDTH
-0.5 5 IRRIGATION CROPS 5 5
5 56
3
-1
464 DSEA
6
-1.5 4 55 6
4 56
4 5
5
6
-2
-1.5 -0.5 0.5 1.5 2.5
-2 -1 0 1 2 3
CCA AXIS I
Figure 9.4 Axis one and two of the canonical correspondence analysis biplot using the floral composition of
the 5 plots of each transect from the six sites studied on the channelized Mondego segment. Numbers indicate
transects. BOULDERS- average percentage of boulders on the plot substrate; WIDTH- average channel width;
DSEA- distance to the sea; SAND.SILT- average percentage of sand and silt on the plot substrate; DISTW-
distance of the plot to the water; HEIGTHW- height of the plot above water; IRRIGATION CROPS - the
dominant use of the land surrounding the site was irrigated crops and rice fields.
The longitudinal gradient of the river (upstream→downstream) is displayed on the second

axis, which is positively correlated with the percentage of boulders. Conductivity and
channel depth are also set against the distance from the sea, channel width and irrigated
cropland. The second axis gradient also illustrates the longitudinal separation between sites
1 and 2 (less than 20 kms from the estuary) on the one hand and sites 4, 5 and 6 (near
Coimbra) on the other. Also noticeable in Fig. 9.4 is the fact that species composition is
quite similar within each plot group but not between transects. Site 3 seemed to act as a
transitional zone in terms of species composition.
Figure 9.5 indicates each species’ position on the first two axes of the CCA, although only
exotic species names are shown. Many species did not display any clear site allocation, but
were widely distributed along the river segment. This was true of ash (Fraxinus
angustifolia), black-poplar (Populus nigra), willow (Salix alba), bramble-thicket (Rubus
ulmifolius), yellow-iris (Iris pseudacorus), Juncus effusus, Mentha suaveolens, Rumex
conglomeratus and the exotic species Tradescantia fluminensis, boxelder (Acer negundo),
Solanum sublobatum, Aster squamatus and Conyza bonariensis.
Instream transects with deeper and higher conductivity waters were dominated by the
exotic hydrophytes Azolla filiculoides and Myriophyllum aquaticum, while under the
opposite environmental conditions the native species Potamogeton crispus, Ranunculus
peltatus and Callitriche stagnalis predominated. Some amphibious species, namely great
yellow-cress (Rorippa amphibia), creeping buttercup (Ranunculus repens) and the
psammophyllous Pseudognaphalium luteo-album displayed a preference for moist soils with
higher percentages of fine elements in the substrate.
The second terrace at sites 4, 5 and 6 mostly supported ubiquitous plants that are typical of
dry soils (e.g. Dittrichia viscosa, Foeniculum vulgare), crop weeds (e.g. Avena sterilis,
Medicago nigra, Echium plantagineum) and the exotics Acacia dealbata, Oxalis pes-caprae
and Arundo donax. Plots further downstream were composed mainly of native perennials,
such as Lonicera peryclymenum, Galium aparine, Humulus lupulus and Hedera helix ssp.
canariensis.
2.5
2
Eryngium pandanifolium
1.5
Azolla filiculoides
1
CCA AXIS II
Erigeron karvinskianus
0.5
Tradescantia fluminensis
Solanum sublobatum Arundo donax
0 Oxalis pes-caprae
Acer negundo
Aster squamatus
Acacia dealbata
-0.5
Myriophyllum aquaticum
Ailanthus altissima
Galinsoga ciliata
-1 Cyperus eragrostis
Bidens frondosa
-1.5
Paspalum paspalodes
-2
-1.5 -0.5 0.5 1.5 2.5
-2 -1 0 1 2
CCA AXIS I
Figure 9.5 Location of the species on axis one and two of the canonical correspondence analysis biplot using
the floral composition of the 5 plots of each transect from the six sites studied on the channelized Mondego
segment. Black dots indicate the position of exotic species. Full names for the exotic species found are also
given.
9.5.4 Site clustering and indicator species
The dendrogram obtained via the NTSys classification is presented on Fig. 9.6. A 70% simi-
larity level was taken as the cut-off point. The results showed the presence of six groups –
which were determined by aggregating plots from different transects and different sites – in
an orderly longitudinal (upstream-downstream) and transverse (water-terrace) manner.
Relevant indicator species (i.e. maximum IndVal >45%, P≤0.01) for each clustering level are
also shown in Fig. 9.6. Within each group of indicator species, the indicator values
(between brackets) are ranked by decreasing order. The indicator value of the species in
the sample groups of the last level of the typology and the abundance/presence pairs for
each species on the overall clusters obtained are shown in Appendix 9.1.
Coefficient
0.94 0.91 0.88 0.85 0.81 0.78 0.75 0.72 0.69
Eryngium pandanifolium* (82.5)

Azolla filiculoides* (80.0)
GROUP A
Paspalum paspalodes* (90.0) INSTREAM
Polygonum hydropiper (74.3) Sites: 1, 2
Ranunculus baudotii (72.8)
Apium nodiflorum (62.3)
Nasturtium officinale (57.1)
Plantago major (50.8)
GROUP B
Potamogeton crispus (45.3)
Cyperus eragrostis* (70.6) INSTREAM
Tradescantia fluminensis* (67.8) Sites: 4, 5, 6
Lythrum salicaria (62.1)
Pseudognaphalium luteo-album (50.9) Rorippa amphibia (67.7)
Aster squamatus* (50.8) Phragmites australis (62.3)
Iris pseudacorus (46.8)
GROUP C
INSTREAM
1st TERRACE
Site: 3
Salix atrocinerea (73.1)
Salix alba (70.4)
Acer negundo* (59.5) GROUP D
Urtica dioica (47.0)
Fraxinus angustifolia (84.0) 1st TERRACE
Sonchus oleraceus (73.8) Acacia dealbata* (96.7)
Alnus glutinosa (62.6) Sites: 1, 2, 4, 5, 6
Rubus ulmifolius (55.0) Vicia sativa (59.8)
Populus nigra (52.0)
Conyza bonariensis* (48.8) Oxalis pes-caprae* (59.7)
Coleostephus myconis (56.9)
Scirpus holoschoenus (53.6) GROUP E
2nd TERRACE
Sites: 4, 5, 6
Foeniculum vulgare (75.1)

Dittrichia viscosa (60.8)
GROUP F
Geranium purpureum (72.7) 2nd TERRACE
Galium aparine (59.7) Sites: 1, 2, 3
Figure 9.6 Hierarchical grouping of sites from the unweighted pair-group average clustering method using
Bray-Curtis similarity coefficient and selected indicator species (maximum IndVal>45%, p< 0.01) for each level
of the typology.
The highest floristic difference occurred between the instream plots at sites situated less
than 20 km from the sea (Group A) and the remaining ones. The significant indicator species
in Group A were hydrophytes, namely the exotics Azolla filiculoides and Myriophyllum
aquaticum, and the emergent E. pandanifolium, as opposed to the native terrestrial species
that characterise the other clustering node.
The third intermediate level divided the second terrace plots from the first terrace and
remaining instream plots. Adventitious species on the surrounding crop fields, ruderals and
species that are typical of the Mediterranean landscape are characteristic of the
riverbank’s second terrace, whereas hygrophilous and riparian woody species are associated
with transects that lie closer to the water. Most of the first terrace plots were clustered in
Group D and comprised riparian trees such as ash, alder, black-poplar, boxelder, Platanus
hybrida, Salix atrocinerea and Salix alba. The transitional area identified by the canonical
analysis (Group C) is represented by a species assemblage with a small niche breath, with
native species as the main indicators.
The less abundant exotic species are typical of higher order groups and displayed a
decrease in indicator values along the clustering levels. They covered a larger ecological
range that encompassed the entire surveyed area and probably the river valley as a whole,
while the dominant exotics tended to be species with narrower habitat requirements.
9.6 Discussion
The total riparian richness in the channelized river segment was found to be similar to that
at the more ‘natural’ sites and was relatively high for an Iberian Mediterranean river
(Ferreira 1994; Ferreira et al. 1998b; Ferreira and Moreira 1999). Numerous factors,
including fluvial-geomorphic landforms, river size, instream flow volume and biotics, among
others, influence the patterns of plant species richness along fluvial corridors (Tabbachi et
al. 1996; Pollock et al. 1998). In addition, the Mediterranean basin was one of the first
regions of the world to suffer the environmental impacts of ‘recent’ human disturbances,
such as forest clearing, grazing pressure and agriculture (di Castri 1991). Albeit the
‘natural’ sites considered in the Mondego floodplain are closer to the study area’s original
environmental conditions, they still cannot be considered to be pristine. However, both the
‘natural’ and the currently channelized sites have been subjected to the same geological
and evolutionary history of endogenous (natural) disturbances and an early man-related
history of exogenous disturbances.
Floodplain rivers tend to present a high resilience to disturbance and short recovery periods
(Wissmar and Swanson 1990); what is more, species richness in Mediterranean-type rivers
may be expected to be relatively persistent, thereby reflecting the evolutionary adaptation
of the biota (Gasith and Resh 1999). Along a river on a similar Iberian floodplain further to
the south – the Sorraia, which is also channelized and subject to periodic dredging of the
river channel and banks (including underground plant parts) – it took riparian vegetation
only four years to recover completely in terms of both woody and herbaceous composition,
diversity and foliar cover (Ferreira 1992).
A particularly high species richness may also be related to the sampling period, because the
abundance of biota in Mediterranean-type rivers is expected to be at its highest between
winter flooding and summer drying – both of which are characterised by low habitat
availability and the deterioration of water quality (Gasith and Resh 1999). This
intermediate period, when habitat and resource availability are at their highest, permits
the co-existence of early season colonisers and other more demanding species, thereby
eventually allowing the establishment of particularly rich and diverse river plant
assemblages (Ferreira 1994). Furthermore, the channelization of Mediterranean rivers is
generally coupled with upstream reservoir construction for flood control and irrigation
purposes – an element that leads to a reduction in high-peak floods and an increase in
summer flow and thus results in the maintenance of this relatively undisturbed and
biologically diverse period.
The floral differences between the channelized and the ‘natural’ plant assemblages along
the River Mondego are due to the dominance of cover by certain species and not to a
reduction of species richness. Significant differences in exotic and native species richness
and cover were detected between the channelized and the ‘natural’ corridor. The pattern
that was observed could result from different disturbance levels and might thus imply a
higher vulnerability on the part of the channelized stretches to colonisation by exotic
species. An ecosystem’s susceptibility to invasion and its relationship with man-made
disturbance are complex (Lepart and Debussche 1991). In addition to the type, frequency
and level of disturbance, the underlying causes of exotic richness are also related to biotic
factors such as the competitive ability of native species and the invasion potential of exotic
ones (Lonsdale 1999).
Two major features in this study area probably account for the channelized river’s higher
susceptibility to invasion by exotics. Firstly, the disturbance of the physical environment
caused by the channelization, regulation and expansion of the irrigation area during the
1980s led to new and favourable environmental conditions, including a reduction in the
intensity and frequency of scouring floods and an increase in the fine sediment and nutrient
load carried by the river. Secondly, the synergistic effect of higher propagule pressure and
the successful dispersal capability of the exotic species that were already present – namely
Azolla filiculoides, Myriophyllum aquaticum, Paspalum paspalodes and Eryngium
pandanifolium – something that has been well documented for the first two of these at
least (Ashton and Mitchell 1990). Exotic taxa frequently invade disturbed semi-natural
communities, whereas the invasion of mature, undisturbed communities may involve
overcoming the strong colonisation potential of riparian vegetation, which is inherently
reliant on vegetative propagation and propagule dispersal capabilities and fast seedling
establishment following flood events (Nilsson et al. 1993; Richardson et al. 2000).
Indeed, the very factors that support high plant species richness in riparian habitats –
namely intrinsic high disturbance features – may well also increase susceptibility to invasion
by exotics (Pyšek and Prach 1994). In rivers located in Western South Africa, the US Pacific
Northwest and South-western France, the percentage of exotic species consistently
exceeded 30% (de Ferrari and Naiman 1994; Planty-Tabacchi et al. 1996; Hood and Naiman
2000). In Arizona Stromberg et al. (1997) found that exotic cover was significantly higher on
the wettest and most frequently inundated and flood-scoured/patch types than it was on
drier, less frequently flooded terraces. Stohlgren et al. (1998) also found riparian zones in
US grassland ecosystems to be more invaded by exotic species than was the case with
upland areas.
In this study, however, exotics numbered only 10% of all species, while even lower
percentages of exotic species have been found in other Portuguese river systems (Ferreira
and Moreira 1995). A low level of exotic richness was also encountered along semi-arid
rivers of South-Eastern Spain (Tabacchi et al. 1996) in a study in which the authors
suggested that stronger native competitors are favoured by the particular hydrological
regime of Mediterranean climate areas. Although the proportion of exotic taxa is small,
exotic cover along Iberian rivers can be high on a local per-site basis, thereby suggesting
that when invadability is assessed, it is not only richness, but also estimates of abundance
such as foliar cover and biomass, that ought to be considered.
Riparian invadability by exotic taxa is probably both site-specific and dependent on three
groups of interacting factors: (i) the type of river segment, and its position in the river
system, including channel width and the lateral development of the banks, habitat
heterogeneity, the substrate in the riverbed and banks, and the flow regime; (ii) the
biological traits of the species present both locally and regionally; and (iii) the man-made
disturbances both within the river corridor and in the surrounding landscape. For example,
percentage values of exotic richness gradually increased from the heads downstream on the
Sorraia (Portugal), the Adour (France) and the McKenzie (USA), but did not do so on the
Andarax (Spain) (Ferreira 1992; Ferreira and Moreira 1995; Planty-Tabacchi et al. 1996;
Tabacchi et al. 1996). It is likely that further studies on different river systems would
contribute to a working model for riparian invadability.
In the present study the instream plots displayed a low level of floristic richness at both the
channelized and the ‘natural’ sites. In other situations a decrease in aquatic species
richness and diversity following channelization has been related to decreases in habitat
heterogeneity (Brookes 1995). However, most Iberian Mediterranean-type rivers display a
naturally low richness of euhydrophytes (Ferreira 1994; Ferreira et al. 1998b; Ferreira and
Moreira 1999), compared to that in mesic fluvial systems (Haslam 1987; Naiman et al.
1993). Truly aquatic species are few, frequently ubiquitous and highly tolerant of adverse
abiotic conditions, and tend to develop large mono-specific stands whenever their specific
requirements are met by locally available habitat conditions. No doubt part of this pattern
may result from long-term human-related disturbance.
A braided river segment such as that in the R. Mondego is predominantly characterised by

lateral and vertical interactive pathways (Ward and Stanford 1995b; Petts 2000). The first
terrace of the channelized segment is located about 2 m above water and it is likely that a
vertical connection persists between the soil surface and the water level. The second
terrace, which is positioned about 4 m above water, has lost this connection and has
become a truly terrestrial environment. Figures 9.2 and 9.4 support the evidence of a more
similar floristic composition of the plots on the second terrace and outer bank, both when
one looks solely at channelized sites and, albeit less obviously, at all sites. This relative
homogeneity of the plant composition on the second terrace can be related to variables
other than the within-site features of the river corridor, such as the current spread of
seedlings of non-riparian exotic woody species – namely boxelder, Platanus hybrida, Acacia
dealbata and Ailanthus altissima – throughout the Mondego floodplain. Seen from this
perspective, the first terrace of the channelized segment corresponded to the truly
marginal ecotone, possessing as it did a set of characteristics that dynamically interacted
across temporal and spatial scales with the adjacent ecological systems – the instream and
the outer bank (Décamps and Naiman 1990). The considerable above-ground height (4-6 m)
of the Mondego embankment results from the lowering of the original riverbed level, so the
outer terrace is directly connected to the terrestrial environment. However it also means
that the river channel itself has become isolated and it almost precludes the original lateral
interactions. Channelization changed the relative contribution of the vertical, lateral and
longitudinal interactive pathways (Ward and Stanford, 1995a).
The twelve-year period following channelization witnessed a natural recovery of the woody
vegetation, along with high native species richness on the channelized segment. The woody
assemblages of the first terrace presented a general structure and composition that were
similar to those encountered in near-by ‘natural’ conditions and also in near-by river
systems (Aguiar et al. 2000). This could be a reflex of the resilience of native
Mediterranean riparian species and their potential performance as pioneers (Gasith and
Resh 1999). Following the relative substrate and slope stabilisation of the newly created
embankment, a large set of shade-tolerant, woody climbers and hygrophyllous species
established themselves under the relatively open canopy.
After channelization, woody recovery in highly disturbed channels begins on the low and
mid-bank surfaces and tree species richness may increase with site stability (Hupp 1992). In
Arizona Stromberg (1993) found that tree species richness varied in a bell curve with the
increase of flood magnitude, with the greatest values occurring at intermediate flow levels.
In the study area channelization also favoured the establishment of dense stands of willows
near the water – something that is probably related to the decrease in scouring floods that
would otherwise wash out a large part of the seedlings and plant propagules. Willows were
also indicator species on the intermediate clustering level of the first terrace transect and
the instream plots of site 3 – a site that displayed enrichment of the substrate in sand and
silt (see also Fig. 9.4), thereby also favouring the encroachment by the willows.
Eryngium pandanifolium dominated a large part of the instream and first terrace transects.
This species, a perennial South-American member of the Umbelliferae, was first reported at
the beginning of the 20th century (Garcia 1947) in drainage channels and small rivulets of
the Mondego river lowlands, while the first herbarium specimen was collected from rice
fields in 1942. This is the only occurrence of the species anywhere in Portugal (Franco 1971)
or indeed in Europe in general (Tutin et al. 1968), and it has maintained a very low
population level over the whole of this period. Following channelization and the
improvement of the irrigation scheme, the species initiated a clear expansion along the
main river, tributaries and drainage channels, irradiating from the original foci near the
river mouth. Though no previous studies exist, there is evidence that seedling
establishment of Eryngium might be related to the increasing volume of fine sediment
deposits on the main river, which usually occur in some types or phases of regulation (Petts
2000). Because the tributary network has been isolated from the main river channel by the
large transversal embankment, thereby reducing the river system’s sediment sorting
processes, large deposits of fine materials transported by the tributaries also accumulate
along the Mondego riverine wetlands and in former meanders and ox-bow lakes.
However, the upper part of the channelized segment – especially the instream area – was
dominated by another exotic, Paspalum paspalodes (Figures 9.3 and 9.6). P. paspalodes is
one of the most ubiquitous and important invaders in Portuguese rivers and channels
(Ferreira and Moreira 1995; 1999), and penetrated the channelized segment from upstream.
This species is a C-strategist sensu Grime (1979). It is a strong competitor, possesses
extensive rhizome systems and high growth rates, produces large amounts of biomass and is
able to rapidly explore aquatic and waterlogged habitats (Wade 1990). It will soon directly
compete with E. pandanipolium for space and nutrient resources.
Small differences in substrate composition and the relative position of the transects along
the river and across the bank were the key factors that explained species composition and
distribution. The major variation in species composition is consistently transversal at both
‘natural’ and channelized sites and is related to the distance and height of the transect
from the water (as shown in the analysis of Figures 9.2, 9.4, 9.5 and 9.6). In addition,
despite the fact that the same engineering procedures and materials were used throughout
the channelization process, there were considerable differences in species composition
both within and along the channelized segment (albeit the reduction of habitat diversity
was generalised). This suggests the development of microtopographic variations and habitat
patches that have influenced vegetation establishment (Everson and Baucher 1998;
Changxing et al. 1999).
This study was financed by the Portuguese Foundation for Science and Technology’s PRAXIS
Project (3/3.1/CEG/2543/95. Francisca Aguiar was supported by a PhD grant (PRAXIS
XXI/BD/11173/97) from the same institution. We thank António Albuquerque for his
constant and valuable support during the fieldwork. We are also greatly indebted to
Dr. Kevin Murphy from Glasgow University for both his thorough revision of the manuscript
and his useful comments.
9.8 Appendices
Appendix 9.1 Indicator value of the species in the sample groups of the last level of the typology and the
abundance/presence pairs for each species on the overall groups.
Species IndVal Group

(%) A B C D E F
Instream (sites 1,2)
Azolla filiculoides 80.00 12/8
Eryngium pandanifolium 52.29 310/10 110/9 390/9 27/6
Instream (sites 4,5,6)
Paspalum paspalodes 85.49 630/15 25/3 61/5 35/3
Ranunculus peltatus 61.10 44/8 245/15 60/4
Polygonum hydropiper 55.99 12/8 245/15 90/8 66/10
Potamogeton crispus 45.33 6/2 51/8
Apium nodiflorum 44.48 15/7 81/15 46/7 16/8
Nasturtium officinalle 41.13 5/1 53/12 26/4 6/2
Plantago major 38.10 26/10 9/5 10/6
Cyperus eragrostis 34.88 96/14 66/7 98/15 2/2 1/1
Veronica anagalloides 33.25 4/4 22/9 7/3 2/2
Callitriche stagnalis 29.17 1/1 21/5 1/1
Poa trivialis 22.68 36/8 12/3 37/5 7/3 1/1
Myriophyllum aquaticum 21.70 9/5 17/9 11/3
Bidens frondosa 17.82 9/5 3/3 4/4 1/1
Cardamine hirsuta 10.67 6/2 1/1
Anagallis arvensis 9.25 8/4 1/1 7/3 7/3 2/2
Polycarpon tetraphyllum 5.28 3/3 5/5 5/1
1st terrace+instream (site 3)
Rorippa amphibia 67.74 36/7 3/3
Phragmites australis 62.27 10/1 125/8 16/4 5/1 25/3
Salix alba 61.87 25/4 420/9 420/18 10/1
Aster squamatus 50.06 9/5 41/13 66/10 47/14 15/3
Iris pseudacorus 46.84 15/3 55/7 43/8
Lythrum salicaria 40.03 72/15 90/8 64/14 26/3
Pseudognaphalium luteo-album 39.74 13/9 23/7 19/11 2/2
Calystegia sepium 35.47 7/7 34/7 32/16 25/2
Rumex conglomeratus 34.32 16/7 43/11 70/10 136/25 13/5 41/6
Ranunculus repens 30.58 11/3 23/6 37/8
Appendix 9.1 (cont’d) Indicator value of the species in the sample groups of the last level of the typology
and the abundance/presence pairs for each species on the overall groups.
(%) A B C D E F
1st terrace+instream (site 3)
Mentha aquatica 29.69 17/5 23/5 11/3
Equisetum telmateia 29.45 34/7 16/8 12/4
Lycopus europaeus 28.11 4/4 6/6 7/3 1/1 1/1
Galium palustre 24.59 17/5 20/4 3/3
Juncus articulatus 18.42 6/2 7/3 1/1
Juncus effusus 14.91 6/2 2/2 2/2
Dactylis glomerata 8.46 12/4 21/3 41/7 25/4 18/5
Galinsoga ciliata 4.05 1/1 1/1 2/2
1st terrace (sites 1,2,4,5,6)
Fraxinus angustifolia 83.99 201/22 1/1 5/1
Alnus glutinosa 62.64 360/16 5/1
Populus nigra 52.00 131/13
Brachypodium phoenicoides 47.94 26/5 156/16 5/1
Oenanthe crocata 46.96 6/2 31/11 25/5 195/25 2/2 56/9
Acer negundo 46.21 6/2 21/5 112/18
Urtica dioica 42.48 4/4 22/5 114/21 6/2 25/4
Platanus hybrida 42.25 20/2 245/12
Tradescantia fluminensis 41.64 1/1 72/9 240/8 790/22 31/3 65/7
Salix atrocinerea 40.34 15/2 240/8 595/22 50/3
Rubus ulmifolius 39.51 20/1 660/23 170/9 335/11
Parietaria punctata 29.42 1/1 5/1 80/15 15/2 27/6
Solanum sublobatum 29.25 2/2 12/3 59/18 20/8 11/2
Conyza bonariensis 27.60 20/8 13/5 86/18 31/5 11/3
Sonchus oleraceus 25.49 35/15 22/9 107/19 43/10 14/6
Hypericum undulatum 24.63 1/1 16/8 2/2
Rumex actosa 24.00 10/6
Raphanus raphanistrum 21.14 2/2 1/1 29/7 2/2
Mentha suaveolens 20.87 22/6 18/6 62/15 15/2 5/1
Carex pendula 20.85 10/2 76/11 7/3 30/5
Agrostis stolonifera 15.21 6/2 112/10 31/6 75/6
Prunella vulgaris 15.04 5/1 14/6
Stellaria media 14.82 2/2 14/5 1/1
Appendix 9.1 (cont’d) Indicator value of the species in the sample groups of the last level of the typology
and the abundance/presence pairs for each species on the overall groups
(%)
A B C D E F
2nd terrace (sites 1,2,3)
Geranium purpureum 62.39 15/2 127/20 31/6 280/15
Foeniculum vulgare 48.13 1/1 37/8 100/12 250/12
Galium aparine 45.81 7/3 5/1 104/14 2/2 109/13
Verbascum virgatum 31.25 46/5
Geranium dissectum 21.30 7/3 35/5 33/8
Mercurialis annua 21.30 2/2 11/2 18/5 66/5
Silene alba 20.28 68/11 35/9 57/8
Hedera helix spp. canariensis 19.77 82/7 10/2 110/5
Briza minor 14.25 5/1 11/3 50/3
Scrophularia canina 11.81 1/1 11/2
Rumex scutatus ssp. 9.71 16/3 11/3

induratus
Carex paniculata 7.91 15/2 7/3
Sisimbrella aspera 5.30 1/1 1/1 10/1

9.9 References
Aguiar F.C., Ferreira M.T., Moreira I.S. and Albuquerque A. 2000. Riparian types in a Mediterranean
river. Aspects of Applied Biology 58: 221-231.
Ashton P.J. and Mitchell D.S. 1990. Aquatic plants: patterns and modes of invasion, attributes of
invading species and assessment of control programmes. In: Drake J.A., Mooney H.A., di
Castri F., Groves R.H., Kruger F.J., Rejmánek M. and Williamson M. (eds), Biological
Invasions: a Global Perspective. John Wiley & Sons Ltd, Chichester, England. pp. 111-147.
Boon P.J. 1988. The impact of river regulation on invertebrate communities in the UK. Regul. Rivers:
Res. Mgmt. 2: 389-409.
Bravard J.P. and Petts G.E. 1996. Human impacts on fluvial hydrosystems. In: Petts GE and Amoros C
(eds), Fluvial Hydrosystems. Chapman and Hall, London. pp. 280-298.
Bravard J.P., Amoros C. and Patou G. 1986. Impacts of civil engineering works on the succession of
communities in a fluvial system: a methodological predictive approach applied to a section
of the Upper Rhône River. Oikos 47: 92-111.
Brookes A. 1986. Response of aquatic vegetation to sedimentation downstream from channelization

works in England and Wales. Biological Conservation 38: 351-367.
Brookes A. 1987. Recovery and adjustment of aquatic vegetation within channelization works in
England and Wales. Journal of Environmental Management 24: 365-382.
Brookes A. 1995. The importance of high flows for riverine environments. In: Harper DM and Ferguson
AJD (eds), The Ecological Basis for River Management. John Wiley & Sons Ltd., Chichester,
England. pp. 33-49.
Changxing S., Petts G.E. and Gurnell A.M. 1999. Berm development along the regulated, lower Dee,
UK. Earth Surf. Processes Landf. 24: 135-149.
de Ferrari C.M. and Naiman R.J. 1994. A multiscale assessment of the occurrence of exotic plants on
the Olympic Peninsula, Washington. J. Veg. Sci. 5:247-258.
Décamps H. and Naiman R.J. 1990. Towards an ecotone perspective. In: Naiman R.J. and Décamps H.
(eds), The ecology and management of aquatic-terrestrial ecotones. Man and the Biosphere
Series, 4, UNESCO, Paris. pp. 1-5.
Décamps H. and Tabacchi E. 1992. Species richness in vegetation along river margins. In: Giller P.S.,
Hildrew A.G. and Raffaelli D.G. (eds), Aquatic Ecology. Scale, Pattern and Process.
Blackwell Scientific Publications, Oxford. pp. 1-20.
climate. In: Groves R.H. and di Castri F. (eds), Biogeography of Mediterranean Invasions.
Dufrêne M. and Legendre P. 1997. Species assemblages and indicator species definition: the need of
an asymmetrical and flexible approach. Ecological Monographs 67: 345-366.
Duvel W., Volkmar R.D., Specht W.I. and Johnson F.W. 1976. Environmental impact of river
channelization. Water Research Bulletin 12: 799-812.
Everson D.A., and Boucher D.H. 1998. Tree species-richness and topographic complexity along the
riparian edge of the Potomac river. Forest Ecology and Management 109: 305-314.
Ferreira M.T. 1992. Ecologia e Dinâmica das Comunidades de Macrófitos Lóticos da Bacia Hidrográfica
do Sorraia. Interferências dos Ecossistemas Agrários Envolventes. Unpublished PhD Thesis.
Agronomy Institute, Technical University, Lisbon.
Mediterranean agricultural systems: In: Pyšek P., Prach K., Rejmanek M. and Wade M.
(eds), Plant Invasions - General Aspects and Special Problems, SPB Academic Publishing,
Amsterdam. pp.117-127.
Ferreira M.T., Catarino L. and Moreira I. 1998a. Aquatic weed assemblages in an Iberian drainage
channel system and related environmental factors. Weed Research 38: 291-300.
Ferreira M.T., Godinho F.N. and Cortes R.M. 1998b. Macrophytes in a southern iberian river. Vehr.
Alismataceae-Iridaceae. Escolar Editora, Lisboa.
Gramineae. Escolar Editora, Lisboa.
Ed. Autor, Lisboa.
Franco J.A. (ed.) 1984. Nova Flora de Portugal (Continente e Açores). 2. Clethraceae-Compositae. Ed.
Autor, Lisboa.
Garcia J. 1947. Algumas novidades para a flora ibérica. Ann. Jard. Bot. Madrid 6(2): 421-438.
responses to predictable seasonal events. Annu. Ver. Ecol. Syst., 30:51-81.
Gregory S.V., Swanson F.J., McKee W.A. and Cummins K.W. 1991. An ecosystem perspective of
riparian zones. BioScience 41:540-551.
Grime J.P. 1979. Plant strategies and vegetation processes. John Wiley & Sons Ltd. Chichester, UK.
Haslam S.M. 1987. River Plants of Western Europe. The macrophytic vegetation of water courses of
the European Economic Community. Cambridge University Press. Cambridge.
Higler L.W.G. 1993. The riparian community of north-west European lowland streams. Freshwater
Biology 29: 229-241.
Hood G.W. and Naiman R.J. 2000. Vulnerability of riparian zones to invasion by exotic vascular plants.
Hupp C.R. 1992. Riparian vegetation recovery patterns following stream channelization: a geomorphic
perspective. Ecology 73(4): 1209-1226.
Inoue M. and Nakagoshi N. 2001. The effects of human impact on spatial structure of the riparian
vegetation along the Ashida River, Japan. Landscape and Urban Planning 53: 111.121.
Iversen T.M., Kronvang B., Madsen B.L., Markmann P. and Nielsen M.B. 1993. Re-establishment of
Danish streams: restoration and maintenance measures. Aquatic Conservation: Marine and
Freshwater Ecosystems 3: 73-92.
Regul. Rivers: Res. Mgmt. 7:279-289.
Lepart J. and Debussche M. 1991. Invasion processes as related to succession and disturbance. In:
Groves R.H. and di Castri F. (eds), Biogeography of Mediterranean Invasions. Cambridge
University Press, Cambridge. pp. 159-177.
Lonsdale W.M. 1999. Global patterns of plant invasions and the concept of invasibility. Ecology 80(5):
1522-1536.
MacCarthy D.T. 1985. The adverse effects of channelization and their amelioration. In: Alabaster J.S.
(ed), Habitat Modification and Freshwater Fisheries. Butterworth, London. pp. 83-97.
Murphy K.J., Fox A.M. and Hanbury R.G. 1987. A multivariate assessment of plant management
impacts on macrophyte communities in a Scottish canal. Journal Applied Ecology 24: 1063-
1079.
Naiman R.J. and Décamps H. 1997. The ecology of interfaces: riparian zones. Annu. Ver. Ecol. Syst.
28: 621-65.
biodiversity. Ecol. Appl. 3: 209-212.
Nilsson C., Nilsson E., Johansson M.E., Dynesius M., Grelsson G., Xiong S., Jansson R. and Danvind M.
419-431.
Petersen R.C., Petersen L.B.-M. and Lacoursiere J. 1992. A building block model for stream
restoration. In: Boon P.J., Calow P. and Petts G.E. (eds), River Conservation and
Management. John Wiley & Sons Ltd., Chichester. pp. 293-309.
Petts G.E. 2000. A perspective on the abiotic processes sustaining the ecological integrity of running
waters. Hydrobiologia 422/423: 15-27.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., de Ferrari C. and Décamps H. 1996. Invasibility of
Pollock M.M., Naiman R.J. and Hanley T.A. 1998. Plant species richness in riparian wetlands-a test of
biodiversity theory. Ecology 79(1): 94-105.
Pyšek P. and Prach K. 1994. How important are rivers for supporting plant invasions? In: de Waal L.,
Child L., Wade P.M. and Brock J. (eds), Ecology and Management of Invasive Riverside
Plants. John Wiley & Sons Ltd., Chichester. pp. 19-26.
Rao K.S. and Pant R. 2001. Land-use dynamics and landscape change pattern in a typical
Ecosystems and Environment: 86:113-123.
Richardson D.M., Pyšek P., Rejmanek M., Barbour M.G., Panetta F.D. and West C.J. 2000.
Distributions 6:93-107.
Rohlf F.J. 1993. NTSYS-Pc, Numerical Taxonomy and Multivariate Analysis System, Version 1.80.
Exeter Software, Setauket, NY.
Shields Jr. F.D.S. and Hoover J.J. 1991. Effects of channel restabilization on habitat diversity,
Twentymile Creek, Mississippi. Regul. Rivers: Res. Mgmt. 6:163-181.
Stohlgren T.J., Bull K.A., Otsuki Y., Villa C.A. and Lee M. 1998. Riparian zones as havens for exotic
plant species in central grasslands. Plant Ecology 138: 113-125.
Stromberg J.C. 1993. Instream flow models for mixed deciduous riparian vegetation within a semiarid
region. Regul. Rivers: Res. Mgmt. 8: 225-235.
Swales, S. 1988. Fish populations of a small lowland channelized river in England subject to long-term
river maintenance and management work. Regul. Rivers: Res. Mgmt. 2: 403-506.
Tabacchi E, Planty-Tabacchi A-M, Salinas, MJ and Décamps H. 1996. Landscape structure and diversity
in riparian plant communities: a longitudinal comparative study. Regul. Rivers: Res. Mgmt.
12: 367-390.
ter Braak C.J.F. 1985. Correspondence analysis of incidence and abundant data: properties in terms
of a unimodal response model. Biometrics 41:859-873.
ter Braak C.J.F. 1990. CANOCO version 3.1. Update notes. Agricultural Mathematics group,
Wageningen.
ter Braak C.J.F and Smilauer P. 1998. CANOCO. Reference manual and user´s guide to Canoco for
Ithaca, NY, USA.
Tutin T.G, Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. and Webb D.A.
(eds). 1968. Flora Europaea. 2. Rosaceae-Umbelliferae (Dicotyledoneae). Cambridge
University Press, Cambridge.
Tutin T.G., Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. and Webb D.A.
(eds.) 1980. Flora Europaea. 5. Alismataceae-Orchidaceae (Monocotyledoneae). Cambridge
Wade P.M. 1990. General biology and ecology of aquatic weeds. In: Pieterse A.H. and Murphy K.J.
(ed), Aquatic Weeds. The Ecology and Management of Nuisance Aquatic Vegetation, Oxford
University Press, Oxford. pp. 17-30.
Wade P.M. and Edwards R.W. 1980. The effect of channel maintenance on the aquatic macrophytes of
the drainage channels in the Monmouthshire Levels, South Wales, 1870-1976. Aquatic
Botany 8: 307-322.
Ward J.V., Tockner K. and Schiemer F. 1999. Biodiversity of floodplain river systems: ecotones and
connectivity. Regul. Rivers: Res. Mgmt. 15: 125-139.
Ward J.V. and Stanford J.A. 1995a. Ecological connectivity in alluvial river ecosystems and its
disruption by flow regulation. Regul. Rivers: Res. Mgmt. 11: 105-119.
Ward J.V. and Stanford J.A. 1995b. The serial discontinuity concept: extending the model to
floodplain rivers. Regul. Rivers: Res. Mgmt. 10: 159-168.
Wissmar R.C. and Swanson F.J. 1990. Landscape disturbances and lotic ecotones. In: Naiman R.J. and
Décamps H. (eds), The Ecology and Management of Aquatic-terrestrial Ecotones. Man and
Capítulo 10
Padrões de distribuição de riqueza e
cobertura de plantas exóticas e nativas
ao longo de um rio Ibérico semi-árido
e do seu leito de cheia
Chapter 10 Patterns of exotic and native plant

richness and cover along a semi-arid Iberian river and
across its floodplain*
*
SUBMITTED TO: Plant Ecology, with authors: F.C. Aguiar, M.T. Ferreira and A. Albuquerque
Capítulo 10 - Patterns of Plant Invasions on Iberian Floodplains | 233
10.1 Abstract
Patterns of native and exotic species richness and cover were examined together with the
composition of species assemblages along a typical Mediterranean river and across its
floodplain. Major goals were to: (1) relate the differences between native and exotic species
patterns to landscape features and local environment, and (2) determine ecosystem
properties, including species diversity, that influence the invasibility of riparian habitats, in
order to predict invasions.
Vegetation data were gathered from three landforms in six valleys: frequently flooded banks
(FF), infrequently flooded valley floors (IF), and rarely flooded higher floodplains (RF). A
total of 287 vascular plant species were identified, but only 26 taxa were exotic. A
hierarchical classification showed that species composition varied across the floodplain,
rather than along the river. Exotic species richness was frequently low at the plot level, but
increased in the valley and on the higher floodplains, in opposition to patterns of exotic
cover. Exotic species cover was substantial on FF landforms; however, nitrophilous native
species predominated on IF with a dynamic water-land ecotone, strong grazing pressure and
a high disturbance level, although well adapted Mediterranean shrubs occupied sites with
coarse substrates. On RF a group of native and exotic ruderals varied in cover according to
site features and manifold land use. We found positive correlations between native and
exotic species richness across landforms, thus supporting the hypothesis that in riparian
habitats richer communities are more prone to invasions. Micro-site features, species
invasiveness and biotic interactions seemed to be determinant in explaining invasibility
differences in Iberian river systems.
Key words: exotic species, species richness, river systems, biological traits, Portugal.
10.2 Introduction
Riparian areas are known to be highly vulnerable to invasion by exotic species, especially
when subjected to human-induced disturbances (Aguiar et al. 2001; Hood and Naiman 2000).
Disturbance and landscape structure probably both affect different stages of the invasion
process and lead to different responses in terms of invasive spread (With 2002; Prieur-
Richard and Lavorel 2000). However, both a given ecosystem’s susceptibility to invasion
(alias its invasibility) and also its opposing “ecosystem resistance” are determined by
numerous biotic and abiotic factors that interact across a wide range of temporal and spatial
scales (Londsale 1999; Heikkinen 1996). A large body of literature is devoted to
understanding the factors that contribute to invasibility on both a small scale and the
landscape level, as well as to invasiveness – i.e. invaders’ biological attributes (Prieur-
Richard and Lavorel 2000). Despite this, to date few studies have looked at a combination of
species composition, richness patterns and dominance interactions between native and
exotic species.
Although invasion ecologists have frequently seen riparian corridors as attractive case-study
ecosystems, few have attempted to relate landscape connectivity, invasibility, and diversity
across the floodplain (Hood and Naiman 2000; Mouw and Alaback 2003). Some studies suggest
that semi-arid landscapes from riparian zones support high diversity and are more susceptible
to invasion than adjacent upland areas (Planty-Tabacchi et al. 1996; Stohlgren et al. 1998;
Hood and Naiman 2000). However, in a comparative study of channelized and semi-natural
river lowlands in Central Portugal, Aguiar et al. (2001) found that the major variation in
species composition and exotic richness was consistently transversal at both ‘natural’ and
channelized sites. Moreover, exotic and native species richness was higher at upland sites in
the near-natural riparian ecosystem than at in-stream ones, as it was in an intermediate area
in the highly disturbed segment. We sampled vegetation from a typical Mediterranean
riparian landscape in southern Iberia, which displayed a relatively homogeneous floral
distribution (Ferreira et al. 2001), albeit influenced by distinct levels of disturbance and high
habitat heterogeneity. This study attempts to contribute to our knowledge about both the
spatial patterns of Mediterranean plant assemblages from managed floodplains and the
predictability of richness-environment relationships. On the other hand, we also aimed to
relate exotic cover and the potential predictor variables, and to estimate the importance of
the diverse biological species traits on the success of invasions.
10.3 Materials and methods
10.3.1 Study area
The Guadiana River is one of the longest watercourses in the Iberian Peninsula and flows
westward through South-Central Spain and South-Eastern Portugal to the Atlantic Ocean
(Fig. 10.1, upper right). The flow regime is subject to erratic interannual fluctuations that
are typical of a Mediterranean river. High floods usually occur in fall or early winter, with a
gradually declining flow and a subsequent drying during late spring and summer that lead to
harsher habitat conditions.
0 10 20 30 km
Figure 10.1 Iberian Peninsula (upper right) with indication of the Guadiana basin. Portuguese part of the
Guadiana basin (bottom left) with the location of the sampling sites.
Relief and climate gradients throughout the basin are relatively low, but the geological
scenery is complex (Table 10.1). Most of the basin is made up of metasedimentary and
metavolcanic rocks that belong to the Hercinic Iberian basement, but calcareous areas are
also present, as are some Cenozoic and Quaternary deposits. Land occupation is
predominantly rural and devoted to cereal cultivation, semi-natural sylvopastoral grasslands,
olive orchards and holm oak and cork oak woodlands. Irrigation crops such as orchards, rice,
tomatoes, maize and sunflowers occupy the rich alluvial soils on the irrigation perimeters.
Marginal areas with steep slopes and stony soils are occupied by sclerophyllous
Mediterranean shrubs. Human settlements are small (few have more than 2,000 inhabitants)
and scattered. Industrial facilities are small and dotted throughout the basin and most are
related to livestock (cattle, sheep, pigs), cork processing and the olive oil industry. Livestock
grazing, mining, dam building and water abstraction are the main impacts on the stream
channel and adjacent riparian areas.
Site locations were chosen along the main course of Guadiana River, ranging from 140 m at
the Spanish border to 20 m at the “Pulo do Lobo Anticlinory Formation” (Fig. 10.1, bottom
left). Selection criteria included differences in land-use, valley geomorphology, dominant
substrate, available plant habitat and human disturbance level.
Table 10.1 summarises the main environmental features of the surveyed areas. Average
annual temperatures were around 16.0-16.8 ºC and average rainfall was always below 750
mm per year (P-Clima 1999). Channel-width ranged from 150 m on the uppermost stream
reach to 25 m at the lowest sampling site (the latter due to a geomorphological constraint).
Juromenha is dominated by fine substrate elements [silt, lime and fine particulate organic
matter (FPOM)], whereas Estrela and Mourão are mainly composed of thin layers of silty sand
over gravel and Moura and Pedrogão of a mixture of sand, silt and FPOM. The Pulo do Lobo
site, which is a steep stream reach, is formed of schist and resistant quartzites and phyllites.
Agricultural practices and livestock grazing were common at all the locations, although those
most impacted by cultivation and grazing were Juromenha, Moura and Pedrogão. The
transversal and longitudinal river profile has been also modified with time – e.g. by
watermills at Moura, a weir at Pulo do Lobo and gravel and sand extractions at Juromenha
and Mourão.
10.3.2 Sampling procedures
A nested-survey sampling was undertaken in late spring (May and early June) of 1998. First of
all, three valley locations or “floodplain landforms” [as defined for semi-arid riparian areas
by Harris (1999)] were identified at each of the six selected sites: inundated and frequently
flooded banks (FF), infrequently flooded valley floors (IF), and unflooded or rarely flooded
higher floodplains (RF). The limits of the valley landforms were determined for each site
using data from the National Hydrological Institute (unpublished), oral statements from local
professional fishermen and also expert on-site observations. At each valley location three
longitudinal transects with ten 2x2-m contiguous plots were laid out. The transepts were
placed across the valley locations in order to cover the whole floral and habitat variability. A
dismantleable square quadrat made of a wooden frame divided into 1-m2 quarters by cross-
wires of string was used to better estimate foliar cover. Percentage cover of each species
was recorded by two experts and then combined, as was the cover of three vegetation
strata: trees, shrubs, and herbaceous plants. Total foliar covers of native and exotic species
were also estimated in loco at each plot and at each transept, with a slight error due to a
few exotic unidentified species. Species that could not be identified in the field were
collected and identified at the Higher Institute of Agronomy’s herbarium. It was not possible
to identify about 2% of all the plant species surveyed, mainly due to missing flower parts.
Species identification followed Tutin et al. (1980; 1993), Franco (1971; 1984), Franco and
Rocha-Afonso (1994; 1998; 2003), and taxonomic nomenclature was updated with Castroviejo
et al. (1993; 1997) and Talavera et al. (1999) when necessary. Exotic plants (synonyms: alien
plants, non-native plants, non-indigenous plants) were considered to be those species that
had recently or remotely been introduced as a result of intentional or accidental events due
to human activity [as per Richardson et al. (2000)].
Distance to source, altitude, valley width and geological background were obtained from a
GIS (Geographic Information System), using 1:25 000 digital maps. A geostatistical
programme [P-Clima software, version 1.2 (1999)] was used to obtain interpolated
meteorological parameter values for the sampling sites. Land-use and wetted channel width
were recorded on site, as was distance to water for each transept (Table 10.1). The
proportions of substrate types were visually estimated for each plot using the sampling
square quadrat and considering four particle-size classes: large substrate: >64 mm (bedrock,
cobbles, and boulders); medium substrate: 2 - 64 mm (gravel and pebble); small substrate:
0.2 - 2 mm (sand); and fine substrate (silt, lime, and FPOM).
Variables related to human disturbance and the use of the floodplain were also considered:
(i) physical alterations of the transversal or longitudinal river profile; occupation of the
floodplain by (ii) irrigation crops, orchards or rice fields; by (iii) forestry; and by (iv) human
settlements, roads, and industrial areas; (v) point pollution sources; (vi) trophic level; (vii)
gravel and sand extraction; (viii) livestock grazing; and (ix) physical evidence of human-
induced disturbance of riverbanks and valley floor for recreation. All these variables were
located on a scale with three degrees of magnitude, from 0 (absent) to 2 - highly disturbed,
polluted, perturbed or influenced by human activities or landscape vegetation. For each site
the first six anthropogenic variables were additively combined and weighted with those
related to direct interferences on vegetation (grazing, mining, and physical alterations
related to local use of fluvial and riparian zones) at each valley location, so as to obtain a
unique variable termed “human disturbance” (range 0-18).
10.3.3 Data analysis
A species-by-plot matrix was constructed in such a way as to list the 540 plots from upstream
to downstream sites and from the instream to the upland floodplain landforms. Cumulative
plant species cover was noted for the 40-m2 transepts and for the entire valley location, and
was corrected for the total sampling area. This data was classified using NTSYS-pc 2.0 (Rohlf
1993), with the Bray-Curtis coefficient of dissimilarity, and the complete clustering method,
in order to determine whether floral variations were larger between or within sites. Prior to
analysis a log transformation [log10 (x+1)] of species cover was used and the available
habitat was corrected. Only species with absolute frequency ≥ 4 across all sampling plots
were retained for analysis (176 out of 287 species). The IndVal 2.0 programme (Dufrêne and
Legendre 1997) was then applied to determine the indicator species and species assemblages
that characterised the groups of site-valley locations which had been obtained using the
previous typology. Indicator species are defined as those that are the most characteristic of a
given typological group and are most frequently and abundantly found at the majority of
sites belonging thereto. For each species i in each site group j, the ‘Indicator Value’
(IndValij) is the product of Aij, which is the mean abundance of a species i at the sites in
group j compared to all the groups in the study, and Bij, which is the relative frequency of
occurrence of species i at sites in group j (Dufrêne and Legendre 1997). The index (multiplied
by 100) is at its maximum (100%) when a given species is observed at all the sites of a certain
group and at no others. A random procedure with 499 permutations was used to reallocate
sites within each site group and served to test the significance of the maximum indicator
value of a species i, which is the largest value for IndVali observed in all groups j belonging to
the typology in question.
To better understand the spatial distribution patterns of exotic species within each site, the
average frequency of each exotic species was determined per valley location. Average
frequencies of a species at the valley location of a given site were calculated by dividing the
number of plots on which a species was noted and the total number of plots sampled per
valley location. Lifespan, Raunkjaer life form, usual habitat and origin of exotic species were
also taken into account to describe exotic species assemblages. Species richness values were
calculated for each plot, both for the entire flora and for native and exotic species groups,
and were averaged for each valley location. Analysis of variance (ANOVA) was used to
compare species richness and species cover between sites at each valley location (n=30) and
between valley locations (n=180). When the within-site variation was significant, multiple
comparisons of means were tested using the Tukey test (Zar 1984).
Pearson product-moment correlations between species richness and species cover and the
environmental variables (channel width, valley width, altitude, distance to source, average
annual rainfall, average annual temperature) were determined. We also explored the
relationships between the number and cover of the native plant species and the number and
cover of the exotic plant species using linear regression models, with separate analyses for
the valley locations. Kriging – a geostatistical gridding method – was used to visualize the
spatial distribution trends of exotic species cover and substrate characteristics at the most
invaded sites. Stepwise forward multiple regressions were used to assess the ability of native
species richness and cover, human disturbance and substrate particle-size to predict exotic
species richness and cover. In order to avoid multi-colinearity, variables that were
significantly correlated (p<0.01) within each valley location were not added to the model.
Durbin-Watson statistics were also taken into account in order to test the independence of
residuals. All statistical analyses were conducted with STATISTICA/w 6.0 (StatSoft Inc. 2001),
except for the Kriging contour maps, which were generated using SURFER version 7.02
(Golden Software Inc. 2000).
10.4 Results
10.4.1 Overview
A total of 287 vascular plant species from sixty-one families were identified. Infrequently
flooded (IF) and unflooded or rarely flooded (RF) areas were the most species-rich, hosting
196 and 211 vascular plant species respectively, in contrast to the inundated and frequently
flooded areas (FF), where only 35% of the total surveyed species were noted.
Twenty-six exotic species were recorded. They followed a spatial trend that was similar to
the total species one, with 23 and 19 species recorded at IF and RF respectively, and 11
species encountered at the FF landform. Of the exotic species, only five were considered
riparian – i.e., related to aquatic, waterlogged or wet habitats (Table 10.2). Of the total
taxa, around 27% were riverine species, including eighteen helophytes and one hydrophyte,
sensu Raunkjaer. Approximately 70% of the exotic species recorded were considered
infrequent – i.e. present at less than 5% of the total sampling plots. 73% of all the recorded
species were considered ‘low frequency’. The most frequent exotic species were the
perennials Paspalum paspalodes L. and Aster squamatus (Sprengel) Hieron., which occurred
at more than a quarter of all plots, and the annuals Conyza albida Sprengel, C. bonariensis
(L.) Cronq., Bidens frondosa L., Datura stramonium L. and Amaranthus retroflexus L. All of
these species are considered ruderals – that is, species which are typically present on
wasteland, cultivated ground, meadowland and roadsides.
10.4.2 Vegetation patterns along the river basin and across the floodplain
The dendrogram obtained via the NTSys is presented in Fig. 10.2. A 90% dissimilarity level
was taken as the cut-off point. Relevant indicator species (i.e. maximum IndVal >50%,
p≤0.01) for each clustering level are also shown. Maximum indicator values (between
brackets) were ranked by decreasing order within each group of indicator species. The
results showed the presence of three groups – which were determined by aggregating the
different valley locations from the different sites – in an orderly transverse (upland-water)
manner. Dissimilarity was high within-site and relatively low between sites. The highest
floristic difference – 99% of dissimilarity – occurred between the ‘RF group’ and the more
wetted areas (‘IF’ plus ‘FF’ groups). Juromenha-FF and Pedrogão-IF were exceptions to these
clusters and demonstrated higher floral similarities to the subsequent upper areas. Ruderals
and species that are typical of the Mediterranean landscape were characteristic of the ‘RF
group’, whereas hygrophilous and nitrophilous species (e.g. Polygonum hydropiper L.,
Xanthium strumarium L., P. distichum) were associated with FF and IF locations.
An uncharacteristic species assemblage of both terrestrial and hygrophilous taxa represents

the ‘IF group’, while a single native species, Cyperus longus L., characterizes the other
clustering node.
Hordeum murinum (99.9) Polycarpon tetraphyllum (91.8) Aster squamatus * (83.9)

Tolpis barbata (97.4) Plantago lanceolata (89.8) Mentha pulegium (76.2)
Poa trivialis (95.5) Sinapis arvensis (88.0) Dittrichia viscosa (75.4)
Phalaris minor (95.2) Carthamus lanatus (85.2) Rumex bucephaloporus (74.4)
Paronychia argentea (93.8) Pulicaria paludosa (84.4) Hedypnois cretica (71.0) RF GROUP
RF: ALL SITES
+
Apium nodiflorum (82.1) Atriplex hastata (51.7)
Bidens frondosa* (78.8) Cyperus eragrostis* (51.4) IF: PEDROGÃO
Melilotus segetalis (59.9) Plantago major (50.5)
Solanum nigrum* (57.6) Xanthium spinosum (50.3) IF GROUP
Securinega tinctoria (52.5) IF: ALL SITES
EXCEPT
PEDROGÃO
+
FF: JUROMENHA
Polygonum hydropiper (98.2)
Polygonum persicaria (89.2)
Xanthium strumarium (83.0) FF GROUP
Paspalum paspalodes* (78.8)
Lythrum salicaria (72.7)
Cyperus longus (93.1) FF: ALL SITES
Tamarix africana (72.7) EXCEPT
JUROMENHA
0.99 0.88 0.77 0.66 0.54 0.43
Bray-Curtis dissimilarity coefficient
Figure 10.2 Hierarchical grouping of sites per valley location from the complete clustering method using Bray-
Curtis similarity coefficient and selected indicator species (maximum IndVal ≥ 50%, P≤ 0.01) for each level of
the typology. Asterisks denote exotic species. Valley location: FF – inundated and frequently flooded; IF –
infrequently flooded; RF – rarely flooded or unflooded.
10.4.3 Patterns of species richness and species cover
Apart from the relative floristic homogeneity encountered along the river, we found signi-
ficant evidence of variation in plant richness and cover between sites within each valley
location (Table 10.3). At the plot level, total species richness attained a high of forty-two
species per 4 m2 quadrat, with a maximum recorded value of seven exotic species and thirty-
nine native species, although average values for exotic species ranged from 0.1 to 3.1 per
valley location (n=30), and for native species from 3.2 to 20.5. IF and RF landforms were
found to be richer in native and exotic species than FF (Table 10.3, Fig. 10.3), along with the
total taxa results. However, exotic cover was consistently greater at FF landforms, in
contrast to the low percentages of cover in the valley floor and upper areas.
Table 10.3 Results of ANOVA of native and exotic species richness and cover (means ± SD) at the valley
locations of the sampling sites. Different superscript letters within each valley location (n=30) indicate
significant differences between sites (p < 0.05, Tukey’s test).
Valley Site Species richness (nº) Species cover (%)

location Native Exotic Native Exotic
Juromenha 14.5 ± 3.3a 0.7 ± 0.7 a 75.2 ± 27.4a 2.1 ± 3.5a
Mourão 4.5 ± 3.8b 1.3 ± 0.7 b 25.2 ± 24.8b 54.1 ± 38.1b
Estrela 4.2 ± 3.9 b 0.1 ± 0.3 c 26.2 ± 28.1b 0.2 ± 0.6a

Frequently flooded
Moura 3.2 ± 3.4 b 1.0 ± 0.2 a 12.9 ± 17.7 b 64.0 ± 36.7 b
Pedrogão 3.5 ± 3.1 b 1.3 ± 0.5 b 29.9 ± 27.0b 24.8 ± 19.3c
Pulo do Lobo 4.8 ± 6.1 b 0.6 ± 1.0a,c 10.4 ± 11.5 b 2.3 ± 4.5 a
Juromenha 18.0 ± 4.8a,b 3.1 ± 1.4a,b 85.1 ± 19.1a 21.3 ± 20.3a
Mourão 12.7 ± 10.7a 1.7 ± 1.9 b 71.6 ± 26.1a 4.9 ± 9.5b
15.1 ± 9.1a,b 1.2 ± 1.2 b 45.1 ± 25.0 b 1.8 ± 2.6 b

Infrequently flooded
Estrela
Moura 11.8 ± 5.5a 1.1 ± 0.9 b 95.0 ± 11.7 c 2.2 ± 2.5b
Pedrogão 20.5 ± 3.6 b 1.8 ± 0.9 b 76.5 ± 15.2a 2.0 ± 1.1b
Pulo do Lobo 15.9 ± 9.8a,b 2.0 ± 1.5a,b 38.1 ± 26.7 b 3.0 ± 3.4b
Juromenha 19.5 ± 3.2 1.7 ± 1.1a,d 81.5 ± 19.2a 11.1 ± 15.9a,c

Unflooded or rarely flooded
Mourão 18.1 ± 3.9 2.7 ± 1.3c 68.2 ± 27.5a,b 13.1 ± 20.3c
Estrela 15.8 ± 3.0 0.5 ± 0.6b,d 84.2 ± 11.0 a 0.7 ± 0.8b
Moura 19.4 ± 7.8 1.9 ± 1.0a,c 65.8 ± 26.5a,b 2.6 ± 1.6a,b
Pedrogão 18.3 ± 3.1 1.4 ± 0.8a,d 72.3 ± 19.1a,b 1.7 ± 1.1b
Pulo do Lobo 17.1 ± 5.7 1.0 ± 0.8a,d 59.3 ± 23.8 b 1.2 ± 1.3b
20
18 Exotic species
Native species
16 c'
Species richness (nº)

14
b'
12
10
8
6
4 a'
2
a b b
0
80
70
c'
60
Species cover (%)
b'
50
40
30
20 a'
a
10
0 b b
FF IF RF
Valley location
Figure 10.3 Comparison of species richness and species cover for exotic and native species at the three valley
locations. Means ± 0.95 confidence intervals are represented. Columns with different letters indicate
significant differences at P>0.01, using Tukey post-hoc comparison tests between valley locations within each
group of species (native or exotic).
ANOVA also revealed significant differences in native species cover between the three valley
locations, with cover values increasing towards the dryer upland areas. Exotic cover was
substantial in the inundated and frequently flooded Mourão, Moura and Pedrogão areas and
the infrequently flooded sampled areas at Juromenha, but decreased considerably in the
remaining dryer areas (Fig. 10.4).
FF
30
100
25
80
20
60
15
40
10
5 20
Species cover (superficial percentage cover)

Species richness (number of species)
0
0
IF
40 100
35
80
30
25 60
20
40
15
10
20
5
0 0
RF
35 100
30
80
25
20 60
15 40
10
20
5
0 0
Pedrogão
Mourão
Pedrogão
Juromenha
Moura
Mourão
Juromenha
Moura
Pulo do Lobo
Pulo do Lobo
Estrela
Estrela
Site
Figure 10.4 Medians, 25/75 th quartiles, minimum and maximum values of native and exotic species richness
and species cover at the three valley locations from the six studied sites.
Table 10.4 describes the vegetation by listing the most prominent species at the six sites
across the three sampled landforms. Dominant species were determined by considering: (i)
frequency of occurrence greater than 50% per site and valley location; and (ii) foliar cover
greater than 30% on at least 25% of the plots surveyed at each valley location. On the
landform scale, only two exotic species and eleven native species met the “dominant-species
criteria”, however at some plots other species presented high cover (>70%) – e.g. Cuscuta
monogyna Vahl, Xanthium spinosum L., D. stramonium, Atriplex hastata L. and Hordeum
murinum L. The most noteworthy exotic species was the successful knotgrass, P. distichum –
a mat-forming perennial with creeping rhizomes and stollons, which was dominant at FF
landforms and gradually declined with distance from the water. Two perennial natives with
Eurasian origins – the grass Cynodon dactylon (L.) Pers, and Cyperus longus – shared
supremacy of cover at places where Paspalum was poorly represented: namely Juromenha,
Pedrogão and Estrela. However, both P. distichum and C. longus were restricted to moist
soils, while Cynodon was common across the floodplain. Assemblages of diverse herbaceous
annual and perennial native species occupied the IF and RF sites, but Tamarix africana, and
the Ibero-African shrubby spurge Flueggea tinctoria (L.) G.L. Webster were nonetheless also
abundant in the infrequently inundated areas. Of the native dominant species, the most
widespread were Xanthium strumarium and Spergularia purpurea (Pers.) G. Don.fil.,
although they only head the cover ratings at a few sites.
10.4.4 Predicting invasibility
Cover of exotic species was only noteworthy at some of the sampled sites; we have therefore
used those sites to examine the relationship between species richness and cover of native
and exotic species and to evaluate the predictive value of environmental factors on the
invasibility of ecosystems.
On the floodplain scale, the species richness of native and exotic plants was positively and
consistently correlated (r = 0.60, p < 0,000001). We also observed that losses of native cover
were significantly related to exotic cover (r = -0,53, p < 0,000001). These general patterns
that we observed for the floodplain were maintained across all the landforms, but significant
relationships were only observed in the frequently flooded areas and on the valley floor for
the number of species, and at the RF transepts for cover (Fig. 10.5).
There was no significant bivariate correlation >|0.3| between the landscape descriptors we
examined and the exotic species richness and cover. However, multiple forward regression
analyses using small-scale variables (i.e. habitat factors) were highly significant and
explained 43%, 52% and 24% of the variance (adjusted R2) for exotic species richness, and
23%, 51% and 48% for exotic cover, at FF, IF and RF landforms respectively (Table 10.5). Of
the eight potentially predictive factors examined at the plot level, only the proportion of
fine and small particle-size class substrate and the cover and richness of native plants were
significantly correlated with the number and cover of exotic plants (BETA values). In the
flooded and moist riparian areas the best predictor for exotic plant enrichment was native
plant richness, whereas the proportion of fine substrate elements was the major factor for
exotic species cover and the subsequent decline in native species cover. The forward
procedure discarded disturbance level, shade and distance from water within each valley
location from all the regression analyses (p<0.01).
FF Mourão/Moura/Pedrogão
5
100
r2 = 0,35; r = 0,59, p = 0,000000001; r2 = 0,02; r = -0,14, p = 0,2045
4
y = 0,87 + 0,09*x 80
3 60
2 40
1 20
0
0
-2 0 2 4 6 8 10 12 14
IF Juromenha/Mourão
7 80
r2 = 0,43; r = 0,66, p = 0,00000001; r2 = 0,051; r = -0,22, p = 0,0841;
6 70
y = 0,34 + 0,14*x
5 60
Exotic species richness (number)
4 50
Exotic species cover (%)
40
3
30
2
20
1
10
0
0
-1
0 5 10 15 20 25 30 35 40 45
RFJuromenha/Mourão
8
100
r2 = 0,40; r = -0,63, p = 0,00000007;
7
r2 = 0,03; r = 0,17, p = 0,1888
6 80 y = 51,9 - 0,50*x
5
60
4
3
40
2
1 20
0
0
-1
8 10 12 14 16 18 20 22 24 26 28 0 20 40 60 80 100
Native species richness (number) Native species cover (%)
Figure 10.5 Relationship of native species richness to exotic species richness, and of native species cover to
exotic species cover for each valley location; data from sites with average exotic cover lower than 4% (n=30
sampling plots) were removed from the analyses. Correlation coefficients, r square and significance levels are
shown for each scatterplot. Linear regression equations are shown for significant correlations. Some black dots
correspond to more than one sampling plot with the same value.
10.5 Discussion
10.5.1 Spatial patterns of plant species richness and cover
Overall, exotics represented only 9% of recorded species. Analogous results have been
reported by studies using equivalent sampling methodologies at other Portuguese water-
courses: namely around 10% for the Mondego lowlands (Aguiar et al. 2001), 14% for the
woody flora of the Tagus River basin, (Aguiar et al. 2000) and 8 % for the Tagus flora as a
whole. Even lower rates of richness have been recorded for the entire Guadiana fluvial
system (4%) (Ferreira et al. 2002) and other southern Iberian basins: 4% for the Sado River
basin (Aguiar et al. unpublished data) and around 6% in some Algarve river basins (Ferreira et
al. unpublished data). Exotic richness was also low in fluvial corridors of South-Eastern Spain,
where it ranged from around 6% to 8% (Tabacchi et al. 1996). In contrast, along semi-arid
rivers in South Africa, Western North America and South-Western France the proportion of
exotic taxa ranges from 20% to 30% (Planty-Tabacchi et al. 1996; Hood and Naiman 2000).
Some authors have suggested that the low number of exotics in Iberia is related to a lower
level of recent human disturbance, together with the fact that native competitors are well
adapted to the area’s torrential hydrological regimes (Tabacchi et al. 1996). However, a high
level of recent perturbations by humans is common throughout this basin (Ferreira et al.
2002), so type, frequency and gradient could be important features in a comprehensive study
on exotic richness trends. Besides the disturbance regime and the resistance of the
ecosystem, the underlying causes of exotic richness may be related to the attributes of both
the native and the exotic species and to propagule pressure (Lonsdale 1999).
Patterns of invasibility in Iberian rivers should not only include exotic richness trends, but
also estimates of abundance such as foliar cover (Aguiar et al. 2001). In fact, heavy invasions
in Mediterranean fluvial corridors are frequently due to just one or a small number of strong
exotic competitors, and it is common for there to be no observations of consistent positive
relationships between exotic richness and exotic cover. In opposition to richness and native
cover trends, we detected a decline in exotic cover between the frequently flooded area and
the other landforms, with a notably higher rate of cover in inundated areas. Contrasting
responses by native and exotic species were also encountered in grassy vegetation and were
attributed to the different evolutionary histories of natives and exotics, which were due to
pre-adaptations to agricultural land-use (McIntyre and Lavorel 1994). Besides finding the
highest rate of exotic richness on the infrequently flooded valley floor, we also noted an
increase in native (and total) species richness with distance from water. This contrasts with
the reports by Stohlgren et al. (1998) of a higher exotic richness in riparian zones than in
upland areas in the central grasslands of the USA, as well with the observations of Planty-
Tabacchi et al. (1996) in Southwest France and the North-Western USA. In our study this
spatial pattern was probably related to the existence of a greater connectivity between the
flooded and the infrequently flooded landforms than between the latter and upland areas
(Fig. 10.2). In fact, we observed a higher similarity in species composition between wetted
areas, with a dispersion of hygrophilous and nitrophilous species throughout those water-
enriched habitats. In addition, we only found large numbers of terrestrial ruderals and
Mediterranean herbaceous or sclerophyllous shrubby species on the rarely flooded landforms.
Remarkably, about 85% of these terrestrial species were infrequent or even site-specific.
Nevertheless, floodplain-terrestrial connectivity provides riparian areas with a higher number
of generalist species than specialist ones, thus helping to explain the overlapping of species
ranges between riparian corridors and uplands (Mouw and Alaback 2003). Other reasons for
the fact that upland areas contributed more to the total flora include the heterogeneous
land-use of adjacent ecosystems (pastures, olive groves, woodlands, and irrigation crops),
and the intricate geomorphological background along the river, which permits a large and
diverse species pool.
10.5.2 Predicting invasibility in Iberian riparian floodplains
Our results support the “rich get richer” theory, which predicts an increase in the exotic
species pool in areas with high levels of native species richness, in opposition to Elton’s
proposition (1958), which suggests that negative relationships such as niche displacement and
competitive exclusion are particularly significant. This ecological paradigm and experimental
evidence were reviewed by Prieur-Richard and Lavorel (2000) and have recently been
discussed by Stohlgren et al. (2003). In the present study the strongest correlation was found
in the infrequently flooded areas, which correspond to a lotic ecotone – i.e., an intermediate
functional entity that combines the aquatic ecosystem and the adjacent terrestrial areas
(Pinay et al. 1990). Transitional areas that are driven by edge effects are known to be
extremely dynamic in space and time and this is crucial for species fluxes. Terrestrial-aquatic
ecotones are usually more biologically diverse than the adjacent ecosystems and play a
central role in the regulation of interpatch dynamics and in providing favourable sites for
species dispersion (Petts 1990).
The majority of the exotic species found in this study and in the riparian corridors of Portugal
were terrestrial, but the most successful invaders were aquatic or riparian – e.g. Eichhornia
crassipes (Mart.) Solms-Laub., Myriophyllum aquaticum (L.) Moench., Azolla filiculoides
Lam., Eryngium pandanifolium Cham. and Schlecht., Arundo donax L., Cyperus sp.,
Paspalum distichum (Ferreira and Moreira 1995, Aguiar et al. 2001). In our study area human-
induced disturbance has persisted throughout historical times and has involved agricultural
land-use of the river valleys, mining and grazing. It is likely that with time, “weedy native”
nitrophilous species such as Cynodon dactylon and Xanthium strumarium become successful
dominants in riparian habitats. These species may be reduced or even replaced by
hygrophilous exotic plants, especially when water availability is not a limiting factor. In fact,
species traits for dominant native and exotic species are quite similar and this is particularly
obvious for the dominant species in frequently flooded areas (Table 10.4, Appendix 10.1).
Perennial lifespan, vegetative propagation by stollons and rhizomes, and having various
dispersal strategies seem to be determinant factors in success in moist habitats. There is also
is evidence of a high rate of propagule influx by the foremost exotic species, P. distichum,
which were detected at all sites, and the differences in native/exotic dominance in a per-
site basis may be a consequence of habitat features, rather than diverse disturbances or
dispersal barriers. This provides supporting evidence to suggest that along with water
availability and micro-scale features, this is one explanation for the diverse susceptibility of
sites and landforms to invasion. Despite the substantial differences in landscape features,
such as land-use and geomorphology, these variables were not significant factors in exotic
species richness and cover. Micro-site properties and the biotic interactions discussed earlier
were the most effective variables in predicting the richness and cover of exotic flora. The
enrichment of fine elements on the substrate – e.g. silt, lime and fine particulate organic
matter (FPOM) – favour the establishment of exotic plants, which are generally adapted to
high nutrient levels (McIntyre and Lavorel 1994). On a channelized Iberian Mediterranean
river, small differences in substrate composition also explained much of the variation in
species composition and distribution (Aguiar et al. 2001). The organic and enriched sediment
is probably artificial and related to agricultural practices. For example, it is clear that at the
Juromenha sampling site exotic cover increased as the proportion of sediment rose (Fig.
10.6). In opposition, in FF and IF areas at the Estrela and Pulo do Lobo sampling sites, for
instance, there is a high proportion of exposed bedrock and coarse substrates respectively,
and preferential colonization was performed by native herbaceous species and shrubs that
are adapted to harsh conditions.
Exotic
FF
Longitudinal distribution of sampling plots (m)
IF RF cover (%)
18
100
16
90
14 80
12 70
60
10
50
8 45
6 40
30
4
25
2
20
0 15
-2 10
-2 2 6 10 14 18 22 26 30 0
Distance from water (m)
Figure 10.6 Kriging map of exotic species cover and fine substrate at the Juromenha sampling site. Proportions
of fine substrate are labelled and overlaid on the shaded contour map of exotic cover. The midpoint of sampling
plots is indicated ( +).
An unexpected result of the predictive analysis was the non-responsiveness of exotic species
richness and cover to the disturbance gradient. Reasons for this may in part be
methodological; the disturbance level was a qualitative and additive variable and thus
subject to inaccuracy when assessing partial components. On the other hand, in the
Guadiana Basin anthropogenic disturbance is long-lasting (Ferreira et al. 2001) and could
understate local and smaller-scale disturbances. Complementary studies on habitat
properties – e.g. moisture gradient and nutrient soil content – are needed and would
probably be decisive to an understanding of invasibility and ecosystem resistance in
Mediterranean riparian corridors.
10.6 Acknowledgments
(POCTI/MGS/42584/2001). F.C. Aguiar was supported by a grant (PRAXIS XXI/BD/11173/97)
from the same institution. We are grateful to Patricia Gonzalez and Dalila Espírito-Santo for
their assistance with plant identification. We also thank M. Cristina Duarte for her revision of
species traits.
10.7 Appendices
Appendix 10.1 Raunkjaer life form, lifespan, propagation mode and dispersal strategies, invasive status, and
ecology of the dominant species (sources of information: Castroviejo et al., 1993, 1997, Ditomaso and Healy,
2003, Fernandes, 1957; Franco 1971; 1984, Franco and Rocha-Afonso, 1994, 1998, 2003; Holm et al., 1991; Pinto
da Silva, 1971; Ridley, 1990; Talavera et al., 1999). The preferential landform where a certain species was
found in the present study is also noted. *p – perennial or biannual; a – annual.
Seed or Preferential
Raunkjaer Life Propagation Invasive landform
Dominant species propagule Ecology
life form span* mode status (present
dispersal study)
Exotic
geophyte/ Hydrochory
Seeds, stollons,
Paspalum distichum L. hemicrypto- p Zoochory x Hygrophilous FF
rhizomes
phyte Anthropochory
Phyla canescens (Kunth) Hydrochory
Chamae- Stem fragments, Hygrophilous,
Greene p Zoochory - IF
phyte (seeds?) nitrophilous
Anthropochory
Native
Cyperus longus L. Hydrochory
geophyte p Seeds, rhizomes x Hygrophilous FF
Anthropochory
Polygonum hydropiper
Seeds Hydrochory Hygrophilous,
L. therophyte a - FF
Zoochory nitrophilous
Xanthium strumarium L. Hydrochory

Hygrophilous,
therophyte a Seeds Zoochory x FF
nitrophilous
Anthropochory
Cynodon dactylon (L.)
Anemochory
Pers. Hemicrypto- Seeds, stollons, Zoochory
p x Nitrophilous IF
phyte rhizomes Anthropochory
Flueggea tinctoria (L.) nano-
G.L. Webster Seeds,
phanero- p Hydrochory - Hygrophilous IF
phyte stem fragments
Scirpus holoschoenus L. Hemicrypto- Hydrochory
p Seeds, rhizomes - Hygrophilous IF
phyte Anemochory
Tamarix africana Poiret (Hydrochory?) Hygrophilous,
micro- Seeds, salt tolerant,
phanerophy p Anemochory x IF
stem fragments facultative
te Anthropochory phreatophyte
Imperata cylindrica (L.)
Anemochory
Raeuschel Hemicrypto- Seeds,
p Zoochory x - RF
phyte rhizomes
Anthropochory
Hydrochory
Juncus bufonius L.
therophyte Anemochory
p Seeds - Hygrophilous RF
Zoochory
Anthropochory
Retama monosperma nano-
(L.) Boiss. phanero- p Seeds Anemochory - - RF
phyte
Spergularia purpurea
Psamophilous
(Pers.) G. Don. fil. therophyte a Seeds Anemochory x RF
10.8 References
Aguiar F.C., Ferreira M.T., Moreira I.S. and Albuquerque A. 2000. Riparian types in a Mediterranean
river. Aspects of Applied Biology 58: 221-231.
Aguiar F.C., Ferreira M.T. and Moreira I. 2001. Exotic and native vegetation establishment following
channelization of a western Iberian river. Regul. Rivers: Res. Mgmt. 17: 509-526.
Castroviejo S., Aedo C., Benedí C., Laínz M., Muñoz-Garmendia F., Nieto-Feliner G. and Paiva J. 1997.
Flora Iberica: Plantas Vasculares de la Península Ibérica e Islas Baleares. 8. Haloragaceae-
Euphorbiaceae, Real Jardín Botánico, CSIC, Madrid, Spain.
Castroviejo S., Aedo C., Cirujano S., Laínz M., Montserrat P., Morales R., Muñoz- Garmendia F., Navarro
C., Paiva J. and Soriano C. 1993. Flora Iberica: Plantas Vasculares de la Península Ibérica e
Islas Baleares. 3. Platanaceae-Plumbaginaceae (partim)-Capparaceae, Real Jardín Botánico,
CSIC, Madrid, Spain.
Deutschewitz K., Lausch A., Kühn I. and Klotz S. 2003. Native and alien plant species richness in
relation to spatial heterogeneity on a regional scale in Germany. Global Ecology &
Ditomaso J.M. and Healy E.A. 2003. Aquatic and Riparian Weeds of the West. University of California.
Agriculture and Natural Resources, Publication 3021, California, USA.
Dufrêne M. and Legendre P. 1997. Species assemblages and indicator species definition: the need of an
asymmetrical and flexible approach. Ecological Monographs 67: 345-366.
Elton C.S. 1958. The ecology of invasions by animals and plants. Methuen, London, England.
Fernandes R. 1957. Notas sobre a Flora de Portugal. Bol. Soc. Broteriana 31: 183-217.
Mediterranean agricultural systems. In: Pysek P., Prach K., Rejmanek M. and Wade M. (eds),
Plant Invasions - General Aspects and Special Problems, SPB Academic Publishing,
macrophytes in a Southern Iberian River. Verh. Internat. Verein. Limnol. 27 : 3833-3837.
Ferreira M.T., Godinho F.N. and Cortes R.M. 1998. Macrophytes in a southern Iberian river. Vehr.
approach. Arch. Hydrobiol. 155(1), 121-145.
Franco J.A. and Rocha-Afonso M.L. 1994; 1998; 2003. Nova Flora de Portugal (Continente e Açores).
3(1) Alismataceae-Iridaceae; 3(2) Gramineae; 3(3) Juncaceae-Orchidaceae. Escolar Editora,
Lisbon.
Franco J.A. (ed). 1971; 1984. Nova Flora de Portugal (Continente e Açores). 1. Lycopodiaceae-
Umbelliferae.; 2. Clethraceae-Compositae. Author Edition, Lisbon.
Golden Software Inc., 2000. SURFER (Surface Mapping System), version.7.02.
Harris R.R. 1999. Defining reference conditions for restoration of riparian plant communities: examples
from California, USA. Environmental Management 24(1): 55-63.
Heikkinen R.K. 1996. Predicting patterns of vascular plant species richness with composite variables: a
meso-scale study in Finnish Lappland. Vegetatio 126: 151-165.
Holm L.G., Plucknett D.L., Pancho J.V. and Herberger J.P. 1991. The world's worst weeds. Distribution
and biology. Krieger Publishing Company, Malabar, Florida.
Hood W.G. and Naiman R.J. 2000. Vulnerability of riparian zones to invasion by exotic vascular plants.
1522-1536.
McIntyre S. and Lavorel S. 1994. Predicting richness of native, rare and exotic plants in response to
521-531.
Nilsson C., Ekblad A., Dynesius M., Backe S., Gardfjell M., Carlberg B., Hellqvist S. and Jansson R.1994.
A comparison of species richness and traits of riparian plants between a main river channel
and its tributaries. Journal of Ecology 82: 281-295.
P-Clima. 1999. Software Provedor de Dados Climatológicos para Portugal. Versão 1.2. Projecto ALTENER
XVII/4.1013/Z/96-092. (ed.) R. Aguiar, INETI/DER, Lisbon.
Petts G.E. 1990. The role of ecotones in aquatic landscape management. In: Naiman R.J. and Décamps
H. (eds), The Ecology and Management of Aquatic-terrestrial Ecotones, Eds. Man and the
Biosphere Series, Paris, pp. 227-261.
Pinto da Silva A.R. 1971. Les plantes synanthropiques au Portugal et aux Açores. Boissiera 19: 297-303.
Pinay G., Décamps H., Chauvet E. and Fustec E. 1990. Functions of ecotones in fluvial systems. In:
Naiman R.J. and Décamps H. (eds), The Ecology and Management of Aquatic-terrestrial
Ecotones, Eds. Man and the Biosphere Series, Paris pp. 141-169.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., De Ferrari C. and Décamps H. 1996. Invasibility of
Prieur-Richard A.-H. and Lavorel S. 2000. Invasions: the perspective of diverse plant communities.
Richardson D.M., Pysek P., Rejmanek M., Barbour M.G., Panetta F.D. and West C.J. 2000.
Rohlf F.J. 1993. NTSYS-Pc, Numerical Taxonomy and Multivariate Analysis System, Version 1.80. Exeter
Ridley H.N. 1990. The dispersal of plants throughout the world (reprint). Bishen Singh Mahendra Pal
Singh, Dehra Dun, India.
Sax D.F. Native and naturalized plant diversity are positively correlated in scrub communities of
California and Chile. Diversity and Distributions 8: 193-210.
StatSoft, Inc. 2001. STATISTICA (data analysis software system), version 6.
Stohlgren T.J., Barnett D.T. and Kartesz J.T. 2003. The rich get richer: patterns of plant invasions in
the United States. Front. Ecol. Environ. 1(1) : 11-14.
Stohlgren T.J., Binkley D., Chong G.W., Kalkhan M.A., Schell L.D., Bull K.A., Otsuki Y., Newman G.,
Bashkin M. and Son Y. 1999. Exotic plants invade hot spots of native plant diversity.
Stohlgren T.J., Bull K.A., Otsuki Y., Villa C.A. and Lee M. 1998. Riparian zones as havens for exotic
plant species in the central grasslands. Plant Ecology 138: 113-125.
diversity in riparian plant communities: a longitudinal comparative study. Regul. Rivers: Res.
Mgmt. 12: 367-390.
Talavera S., Aedo C., Castroviejo S., Romero-Zarco C., Sáez L., Salgueiro F.J. and Velayos M. 1999.
Flora Iberica: Plantas Vasculares de la Península Ibérica e Islas Baleares. 7(1). Leguminosae
(partim), Real Jardín Botánico, CSIC, Madrid, Spain.
Tutin T.C., Burges N.A., Chater A.O., Edmonson J.R., Heywood V.H., Moore D.M., Webb D.A.,
Valentine D.H. and Walters S.M. 1993. Flora Europaea. 1.Psilotaceae- Platanaceae.
Cambridge University Press. Cambridge.
Tutin T.C., Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. and Webb D.A. 1980.
Flora Europaea. 5. Alismataceae-Orchidaceae (Monocotyledonae). Cambridge University
Press. Cambridge.
With K.A. 2002. The landscape ecology of invasive spread. Conservation Biology 16(5): 1192-1203.
Zar J.H. 1984. Biostatistical Analysis, 2nd. ed. Prentice Hall Inc., New Jersey, USA.
Capítulo 11
Padrões de invasão do graminhão (Paspalum
distichum L.) em habitats ripícolas portugueses
Chapter 11 Invasibility patterns of knotgrass

(Paspalum distichum) in Portuguese riparian habitats*
*
ACCEPTED BY: Weed Technology, with authors: F.C. Aguiar, M.T. Ferreira, A. Albuquerque and I.
Bernez
Capítulo 11 - Invasibility Patterns of Knotgrass in Riparian Habitats | 263
11.1 Abstract
Spatial patterns of the exotic riverine species knotgrass (Paspalum distichum L.) were
examined in Mediterranean river basins in Southwestern Iberia. The major goals were to
assess the degree of invasibility of riparian habitats by this species and to determine the
influence of environmental factors and human-induced disturbances on both the landscape
and the habitat scales. The present study demonstrates knotgrass’s ability to invade
riparian habitats in Portuguese freshwater ecosystems. However, most of the spatial
variation of the knotgrass cover seemed to be driven by local factors, such as fine sediment
enrichment and the fragmentation of riparian woods, and by other anthropogenic
interferences in relation to both the fluvial system and the surrounding landscape.
Nomenclature: knotgrass, Paspalum distichum L., PASDS.
Key words: exotic species, human disturbances, environmental variables, Mediterranean

Basin.
11.2 Introduction
Knotgrass, an exotic perennial grass originally from tropical areas of America, has been
widely introduced into moist grasslands of Portugal. On its native range, it is usually
considered to be a desirable component of the ecosystem, in that it controls the erosion of
banks, ditches, and beach dunes, is used as a forage grass (DiTomaso and Healy 2002) and
helps to rehabilitate overgrazed land. However, in the Mediterranean Basin, once
established on stream banks and in riparian corridors, it frequently forms dense, monotypic
stands at the expense of native grass species. Knotgrass soon becomes a common weed in
orchards, vineyards and fields of maize, rice and other irrigation crops (Guillerm et al.
1990). Most research to date has been targeted at its biology, chemical control and
integrated pest management.
In the Mediterranean basin, knotgrass was initially mentioned in France, where it was
deliberatly introducted as a crop in Bordeaux region in 1802 (Le Floc’h 1991). In Portugal it
was first reported in 1887 at Tagus bank margins (Silva 1940), and described in 1909 in one
of the first Portuguese floras (“Flora Portuguesa” of G. Sampaio). Until the nineteen
seventies its distribution was confined to areas of coastal and alluvial floodplains in the
main basins of Portugal and at rice fields and irrigation crops (Franco and Rocha-Afonso
1980). Recently, knotgrass infestations in rice fields seemed to decrease (Vasconcelos et al.
1999), however successful invasions of non-crop habitats, such as silt-rich margins of
lagoons, and wetlands (Costa et al. 1999), channelized fluvial corridors (Aguiar et al. 2001),
highly disturbed bank margins (Ferreira and Moreira 1995), and semi-natural communities in
riparian habitats (Franco and Rocha Afonso 1998; Ferreira et al. 2001, 2002) have pointed
to the increasing importance of knotgrass as an environmental weed. Even so, none of these
studies have evaluated the present range of this species or focused on the potential factors
that enhance its spread in river habitats. To better understand patterns of invasibility, we
analysed the frequency and cover of knotgrass in relation to environmental features, and
the man-induced disturbances that are presumed to account for exotic plant species
invasibility in riparian ecosystems.
11.3 Materials and Methods
Fieldwork was conducted at 336 sites during the late spring / early summer of 1994-1995
(Sado basin), 1997-1998 (Tagus basin), 1999 (Guadiana basin), and 2000 (Algarve, Mira and
Oeste basins). The studied basins possess a Mediterranean climate with an important intra
and inter-annual hydrological variability and long-lasting anthropogenic interferences,
especially those related to agricultural land-use. Each sampling site comprised a 100-m-long
stretch, and lateral limits for the sampling area were the approximate maximum annual
wetted width of the river corridor, which generally includes the riparian galleries. Sites
were spread throughout the basins on 2nd- to 5th-order streams sensu Strahler (1957), and
selected such that they generally spanned the gradient of environmental conditions and of
antropogenic disturbances. Stream ordering based on confluent points provides a measure
of position within the drainage network, and permits comparison between similar
hydromorphological systems. Particular habitat characteristics are displayed in the
Mediterranean basin by first order streams, which usually dry early in spring and present
low flow in winter (Gasith and Resh 1999), and by outsized fluvial corridors or streams near
to the estuarine areas (usually 6th- order streams or higher), therefore they were not
included on the present study.
The location of the Iberian Peninsula, of Portugal and the studied basins are represented on
the right side of the Fig. 11.1, and the number of sampling sites surveyed for each basin (n)
is shown in the columns of the left side. All species found in water and on both margins and
inner banks were identified and allocated a relative cover class (6 abundance classes: 0:
absent; 1: <5%; 2: >5%-30%; 3:>30-50%; 4:>50%-75%; 5: >75%), similar to the Braun-Blanquet

scale (1932). In order to identify the main factors that contributed to knotgrass presence
and cover, regional and environmental data were gathered from a Portuguese GIS database,
from lithological and topographical maps, or recorded in the field. Climate data were
generated from a geostatistical climate programme (P-Clima 1999).
5 Median 25%-75% Min-Max
4
Knotgrass cover
3
n=48
n= 9
2
n=48
n=126
n=28
1
n=77
0
Tagus Guadiana Mira
Oeste Algarve Sado
River basin
Figure 11.1 Median, 25/75th quartiles, minimum and maximum values for knotgrass cover in the studied river
basins. The number of sampling sites (n) is indicated in quartile bars for each basin. The location of the
Iberian Peninsula, Portugal and the studied basins are represented on the right side of the figure.
Variables related to human disturbance and the use of the drainage basin were also
considered. At the river segment level (a reach of one to ten kilometres, according to the
size of the river), variables included an evaluation of the impact of human settlements,
industries and roads, and of the artificial barriers (dams, weirs, culverts), the dominance of
irrigated crops, orchards and pastures, and the condition of the riparian woody assem-
blages. At the site level, the disturbance of the banks and riverbed by cattle and people
was visually assessed during site surveys, and variables such as the sediment load, nutrient
and organic inputs, and the toxic condition of the site were moderated for each site with
the Portuguese Water Resources Information System, a database from the Portuguese Water
Institute that stores and processes hydrological, geographical and geological information.
All these variables were evaluated on a scale with five degrees of magnitude, from 1- no
obvious alterations to 5 - highly disturbed, polluted, perturbed or influenced by human
activities. Further information on the potential predictive variables is included in Table
11.1.
Table 11.1 Characterization of regional, environmental and anthropogenic variables, and respective scale
level.
Variables Description and evaluation Scale level Source of data

Regional
Latitude and Measured from the downstream beginning River basin Digital topo-
longitude (degrees) of the site. graphical maps1)
Distance from Distance from source to the sampling site, River basin SNIRH2)
source (km) measured along the river.
Strahler’s order River hierarchical classification, based on River basin SNIRH
number (2-5) confluent points of the streams. (head-
water =1; two headwaters joined together
=2; two streams of 2nd- order =3,…)
Altitude (m) The altitude of the site above average sea River basin Digital topo-
level. graphical maps1)
Environmental
Permanent flow Dummy variable: 1= streams that normally River basin Field work,
regime (0-1) flow yearlong and have well defined SNIRH
channels.
Yearly average Interpolated rainfall and temperature River P-clima
rainfall (mm), yearly values at the sampling site. segment software3)
average
temperature (ºC)
Calcareous Dominance of calcareous and carbonate River Lithological
geological rocks. Opposed to siliceous geological segment maps4)
background (0-1) background.
Gradient slope (‰) Slope of the riverbed along the stream River Digital
reach. segment topographical
maps1)
Wetted width (m) Width of the stream channel averaged from Sampling site Field work
ten measured transects.
Stream depth (m) Depth of the stream channel, averaged Sampling site Field work
from ten points, using a graduated dip-net
pole.
Conductivity Water conductivity, measured with field Sampling site Field work
(µS/cm) probes.
Table 11.1 (cont.’d) Characterization of regional, environmental and anthropogenic variables, and
respective scale level.
Variables Description and evaluation Scale level Source of data

Superficial substrate Visual evaluation of proportions of Sampling site- Field work
(%) particle-size classes: plot level
1. bedrock; 2. large substrate; 3.
medium substrate; 4. sand; 5. lime,
silt, and fine particulate organic matter.
Tree shade (%) Projection of the canopy cover on the Sampling site Field work
sampling reach, estimated by visual
evaluation.
Land use
Intensive agriculture 1= dominance of irrigated crops, River segment Field work
(0-1) orchards and pastures on the valley
floor.
Urbanisation (1-5) Impact of urban land occupation, from River basin and Field work,
inexistent to severe impact. River segment SNIRH
Human-induced disturbances
Riparian State of riparian formations (from near- River segment Field work
fragmentation (1-5) natural state to major deviation from
reference conditions)
Bank modification Physical alterations of the river Sampling site Field work
(1-5) transversal profile (from near-natural
morphological conditions to major
alterations)
Connectivity (1-5) Impact of artificial barriers (from no River basin and SNIRH and field
barriers to severe broken fluvial river segment work
linkages).
Upstream dam (0-1) 1= existence of artificial lenthic water River segment SNIRH
bodies upstream from the sampling site
Sediment load (1-5) Deviation from reference conditions River segment Field work,
(from “natural” to major artificial SNIRH
increase)
Nutrients and organic Deviation from normal undisturbed River segment Field work,
inputs (1-5) conditions (from reference conditions SNIRH
to major deviations)
Toxic conditions Deviation from normal undisturbed Sampling site Field work,
(1-5) conditions (from reference conditions SNIRH
to major deviations)
1) digital topographical maps (1:25 000) from Portuguese Army Geographic Institute (IGeoE)
2) SNIRH - Portuguese Water Resources Information System from the Portuguese Water Institute (INAG).
3) P-Clima. 1999. Software provider of climatological data for Portugal. Ver. 1.2. Project ALTENER
XVII/4.1013/Z/96-092. (Ed.) R. Aguiar, National Institute of Engineering and Industrial Technology (INETI) /
Renewable Energy Department, Lisbon, Portugal.
4) lithological maps (1:50 000) from Portuguese Geographic Institute (IGeo)
As a first approach to understand the distribution patterns of knotgrass in relation to the

substrate features, data was collected from three longitudinal transects with ten 2 m x 2 m
contiguous plots at three sampling sites in the Guadiana basin, where knotgrass presented
considerable differences in abundance. Percentage cover of knotgrass and proportions of
substrate types were visually estimated for each plot using a sampling square quadrat and
considering five particle-size classes: bedrock; large substrate (> 64 mm, including cobbles
and boulders); medium substrate (2-64 mm, gravel and pebble); sand (0.2-2 mm), and lime,
silt, and fine particulate organic matter (< 0.2 mm).
The invasiveness of knotgrass and its frequency of occurrence were analysed and
comparisons of its cover among basins were performed using the Kruskal-Wallis One-Way
ANOVA (p<0.001). Data were log transformed [log (x+1)]. For each exotic species i in each
basin j, the ‘Frequency of Occurrence’ (FOij) is the quotient of the number of sites where
the given species was observed in the basin j (nij) and the total number of sites surveyed at
the basin (nj), expressed in percentage. The average frequency of occurrence of a species i
(FOi ) is the quotient of the total number of sites where a given species was found (ni), and
of the total number of surveyed locations (N=336), expressed in percentage.
The FO of knotgrass and of the other exotic species was also combined with its abundance
and dominance to evaluate its invasiveness. We considered a site to be invaded when a
single exotic species developed persistent, naturalised populations and presented a
percentage cover of 30% to 50% or higher in the sampling area. This threshold level was
defined intuitively by a group of Portuguese experts in limnology and ecology. Accordingly,
the FOi>30%, is the number of sites where a species i was found in the surveyed basin, or in
the total sampling area, and gives a measure of the invasion potential of a species.
Dunn’s test made it possible to identify the pairs of basins with significant differences
(p<0.05) of knotgrass cover. Data from the Mira river basin were discarded from the ANOVA
due to the low number of sampling sites. Possible relationships between knotgrass presence
and cover and the underlying factors were estimated with the Pearson Product-Moment
Correlation, and with Spearman Rank Correlation for interval data and dummy variables. All
statistical analyses were conducted with STATISTICA/w 6.0 (StatSoft Inc. 2001).
11.4 Results and Discussion
Forty-four exotic species were recorded in the studied basins. Table 11.2 presents origin,
Raunkiaer life-form classification system (1934) and life span of the exotic species found.
Table 11.2 Origin, Raunkiaer life form and life span of the exotic species recorded on the sampling sites.
Exotic species mentioned on the Portuguese legislation (Ministério do Ambiente 1999) as invasive species are
note with a *.
Specie Origin Raunkjaer life form Life

span
Abutilon theophrasti Medik. South Asia therophyte a
• Acacia dealbata Link SE Australia; Tasmania mesophanerophyte p
• Acacia longifolia (Andrews) Willd. Australia microphanerophyte p
• Acacia melanoxylon R. Br. in Aiton SE Australia; Tasmania megaphanerophyte p
Acer negundo L. North America mesophanerophyte p
• Ailanthus altissima (Miller) Swingle China mesophanerophyte p
Amaranthus blitoides S. Watson North America therophyte a
Amaranthus hybridus L. Tropical / subtropicalAmerica therophyte a
Amaranthus retroflexus L. North America therophyte a
Arundo donax L. Central (?) and South Asia geophyte p
Aster squamatus (Sprengel) Hieron Central and South America hemicriptophyte p
• Azolla filiculoides Lam. North and South America hydrophyte p
Bidens aurea (Aiton) Sherff Central America therophyte a
Bidens frondosa L. America therophyte a
Callitriche cribrosa Schotsman North Africa hydrophyte p
Castanea sativa Miller Oriental Mediterranean region mesophanerophyte p
Chenopodium ambrosioides L. Tropical America therophyte a
Chrysanthemum segetum L. SW Asia therophyte a
• Conyza bonariensis (L.) Cronq. South America therophyte a
Cyperus eragrostis Lam. Tropical America geophyte p
• Datura stramonium L. America therophyte a
• Eichhornia crassipes (Mart.) Solms-Laub Tropical America hydrophyte p
Eucalyptus camaldulensis Dehnh. Australia mesophanerophyte p
Eucalyptus globulus Labill. ssp. globulus SE Australia; Tasmania megaphanerophyte p
Ficus carica L. SW Asia microphanerophyte p
Galinsoga ciliata (Raf.) S.F. Blake from Mexico to Chile therophyte a
• Galinsoga parviflora Cav. South America therophyte a
Gomphocarpus fruticosus (L.) W.T. Aiton South Africa nanophanerophye p
Helianthus annuus L. North America therophyte a
Lindernia procumbens (Krock.) Philcox Europe; Asia therophyte a
Myriophyllum aquaticum (Vell.) Verdc. South America hydrophyte p
• Oxalis pes-caprae L. South Africa geophyte p
Paspalum distichum L. Tropical geophyte p
Phytolacca americana L. North America hemicryptophyte p
Prunus cerasus L. SW Asia microphanerophyte p
Prunus domestica L. Caucasus microphanerophyte p
Ricinus communis L. Tropical regions nanophanerophyte p
• Robinia pseudoacacia L. North America mesophanerophyte p
Salix viminalis L. Central and East Europe mesophanerophyte p
Salix babylonica L. China macrophanerophyte p
Senecio petasitis (Sims) DC. México camaephyte p
• Tradescanthia fluminensis Velloso from SE Brasil to Argentina camaephyte p
Vitex agnus-castus L. South Europe nanophanerophyte p
Zantedeschia aethiopica (L.) Spreng. South Africa geophyte p
The invasive status considered on the Portuguese legislation (Ministério do Ambiente 1999)
was also noted. Exotic species were present in approximately 81% of the sampling sites,
however around 70% of them were present in less than 5% of the total sites.
Identical knotgrass cover patterns were found in the studied basins (Fig. 11.1). The Sado
basin presented consistently lower median covers (Dunn’s test, p<0.05), though notable
knotgrass cover and a high frequency of occurrence were found along the main course of
the river. A study of the nearby Guadiana basin showed that species assemblages were
relatively homogeneous from one hydrological year to another, especially in the case of
resilient grasses and graminoids, which include knotgrass (Ferreira et al. 2001).
Nevertheless, the difference in invasibility could be partially related to the year of
sampling. In the Sado basin sampling followed a 4-year period of extremely low rainfall,
whereas knotgrass is usually abundant in locations with water availability, such as the
alluvial floodplains located mainly along the main river course.
Table 11.3 locates the exotic invasions on the studied basins by listing the most prominent
species, and the respective frequency of occurrence.
Table 11.3 Total frequency of occurrence (FO) and frequency of occurrence of the exotic species with cover
above 30% (FO>30% ) at least in one sampling site.
Species FO (%) FO>30% (%)
T O G A M S T O G A M S
Acacia dealbata y - - y y - y - - y -
Acacia melanoxylon y y - y y - y - -
Arundo donax y y y y y
Azolla filiculoides y - y - - y y - y - - y
Bidens frondosa y y y - y -
Cyperus eragrostis y y y y y y
Datura stramonium y y y y - y y -
Eichhornia crassipes y - - - - y y - - - y
Eucalyptus camaldulensis y - y y - y - y -
Eucalyptus globulus - - - - - y - - - - - y
Lindernia dubia y - - - - - y - - - - -
Myriophyllum aquaticum y - - - - - y - - - - -
Oxalis pes-caprae y y y y y
Paspalum distichum y
Ricinus communis - y - y - - - - y - -
Salix babylonica y - - - - - y - - - - -
T-Tagus basin; O-Oeste basin; G-Guadiana basin; A-Algarve basin; M-Mira basin; S-Sado basin.
y FO ≤ 25% FO >25-50% FO ≥50-75% FO ≥75%

Sixteen species exhibit an invasive ability on riparian ecosystems, although the majority
only heads the cover ratings at a few sites. Results showed that knotgrass and giant reed
(Arundo donax L.) were consistently widespread in the studied basins and displayed higher
invasive ability on the sampling area when compared with other exotic species that were
also present.
Fig. 11.2 compares the FO of knotgrass and giant reed in the studied basins. Only in Oeste
basin, giant reed surpasses the knotgrass invasiveness. Floodplains in Oeste region are cha-
racterized by numerous human settlements and small farmlands occupied by irrigated
crops, orchards and vineyards, leading to a patch mosaic pattern. Giant reed was often
planted to erosion control, to confine properties, as windbreaks, and as tutor of creeper-
cultivated plants, such as the common beans (Phaseolus vulgaris L.). Nevertheless, in ripa-
rian corridors of highly disturbed landscapes, it typically develops dense stands, displacing
native vegetation and increasing flooding and siltation (Aguiar et al. 1996).
90
80
Frequency of occurrence (%)
70
60
50
40
30
20
10
0
Tagus Guadiana Mira Average
Oeste Algarve Sado
River basin
Total FO of knotgrass (%) FO of knotgrass with cover > 30%
Total FO of giant reed (%) FO of giant reed with cover > 30%
Figure 11.2 Comparison of the frequency of occurrence (FO) of knotgrass and of giant reed per basin and at
all sites. FO with species cover above 30% for the both species was also represented.
No bivariate relation was found between the giant reed cover and the knotgrass cover
(Spearman correlation =0.04; P<0.05), or between their presences (Spearman correlation
=0.01; P<0.05). However, other studies in Portuguese riparian habitats revealed some
competitive interactions between knotgrass and other exotic species, such as the Eryngium
pandanifolium Cham. and Schlecht., (Aguiar et al. 2001) or with native species with similar
species traits, such as the bermuda-grass [Cynodon dactylon (L.) Pers.] (F. Aguiar
unpublished data).
Knotgrass had already been found to be amongst the most successful invaders of aquatic or
riparian habitats in Portugal (Aguiar et al. 2001; Ferreira and Moreira 1995; Ferreira et al.
2002), though some other exotic species such as water hyacinth, Eichhornia crassipes
(Mart.) Solms-Laub., parrotfeather, Myriophyllum aquaticum (L.) Moench., pacific
mosquitofern, Azolla filiculoides Lam., Eryngium pandanifolium Cham. and Schlecht., and
the giant reed, Arundo donax L. can be locally more detrimental (Ferreira and Moreira
1995; Aguiar et al. 1996; Aguiar et al. 2001).
Knotgrass cover and occurrence were consistently associated with areas of irrigated crops
and urban and industrial areas, as well as with human-related disturbances. While broad
geographic variables, such as altitude, climate, and geological background were not
significant (Table 11.4), revealing a considerable ecological plasticity of this specie.
However, distance to source and the river hierarchical typology displayed significant
positive relations with knotgrass invasibility patterns. Headwater streams with intermittent
flow and notable gradient river slopes were less invaded, in opposition to higher order
streams, with permanent flow regime. In general, in the sampling area, upper reaches are
less impacted by agriculture, and are frequently surrounded by Mediterranean shrubland
and thus displaying high “naturalness” rates of its flora (Planty-Tabacchi et al. 1996). In
opposition, lowland floodplains are frequently devoted to high-demanding irrigated crops,
such as maize, rice and tomato, which increase the nutrient soil content via fertilizers
(particularly nitrates), and promote the expansion of knotgrass, a typical nithrophilous
species. The enrichment of fine elements on the substrate – e.g. silt, lime and fine
particulate organic matter and sediment load – also contributes to a favourable
environment for the exotic species, which are generally adapted to high nutrient levels
(McIntyre and Lavorel 1994). In the present study, the increase of fine substrate elements
was also related with the knotgrass spread (Pearson Product Moment Correlation=0.46;
P<0.01). Moreover, agricultural practices such as milling can lead to colonization of other
areas and increase the invasion potential by spreading the rhizomes and stolon fragments of
knotgrass. The presence of human settlements, and other artificial features, including
roads and bridges, increase the number of visitors on the riparian areas, thus promoting the
dispersal of propagules and seeds by anthropocory. Canopy openings that naturally result
from the torrential flow regime or the intentional fragmentation of streamside woody
vegetation for human purposes (e.g. water abstraction, increase of cropland area) are usual
in Portuguese fluvial corridors (F. Aguiar and Ferreira unpublished data), and play an
important role to maintain herbaceous and hydrophyte species assemblages.
Table 11.4 Summary statistics for the correlations relating the presence and cover of knotgrass to regional,
environmental and anthropogenic variables. * = p < 0.05; **= p < 0.01; ns= non-significant, p=0.05 ; nv=no
variance.
Scale level Knotgrass
Variables presence cover
River Basin
Latitude (in degrees) -0,26** ns
Distance from source (in km) 0,20* ns
Strahler’s order number 0,20* 0,19*
Urbanization (1-5) 0,23* 0,26**
Connectivity (1-5) -0,22* ns
1)
River Segment
Permanent flow regime (0-1) -0,25** -0,27**
Gradient slope (‰) -0,27** -0,22*
Sediment load (1-5) ns 0,20*
Intensive agriculture (0-1) 0,33** 0,35**
Riparian fragmentation (1-5) 0,21* 0,19*
Sampling site/habitat level

Tree shade (%) ns -0,19*
”)
Lime, silt, and FPOM (%) nv 0,46**
Toxic conditions (1-5) 0,23* ns
Conductivity (µS/cm) 0,25** ns
Nutrients and organic inputs (1-5) 0,26** 0,27**
1) river segment: 1 km for small rivers (catchment <100 km2); 5 km for medium-sized rivers (100-1000 km2) and 10
km for large rivers (>1000 km2).
2) from the nested-sampling survey in the Guadiana basin (4m2 quadrat plots, n=90)
However, the balance between woody and herbaceous vegetation by shading, depends on
the fragmental patterns of riparian galleries and particularly on the extension of treeless
reaches. Knotgrass cover is favoured by exposure to light, which stimulates germination of
seeds (DiTomaso and Healy 2002) and promotes the sprouting, rooting and the growth of the
rhizomes, stem and shoot segments (Huang et al. 1987; Liu et al. 1991). Dawson and Kern-
Hansen (1978) studied the effect of natural and artificial shade on bank and aquatic
vegetation of lowland streams and concluded the possibility of using shade as a
management technique to control weed growth in riparian areas.
The lack of connectivity, i.e. the alteration of the natural flow seemed to be less
favourable to knotgrass establishment and spread. Species migrate from upstream regions
and colonize the floodplain, so large daily variations of flow in these river systems also
affect the colonization and spread of knotgrass, as well as the water dispersal of seeds and
the propagule influx of the species. Floodplain lateral connectivity also allows species
exchange between regional and local species pools (Mouw and Alaback 2003) and is
probably as fundamental as the longitudinal connectivity of the river.
Although these findings provide some information about the present status of the spread of
knotgrass and related environmental determinants, site-based studies are needed to
understand population dynamics and achieve some degree of predictive capability. Micro-
site features and biotic interactions seemed to be determinant in explaining invasibility
differences in Portuguese riparian habitats, thus research needs point to quantitative
analysis of habitat features, e.g. soil nutrient and soil organic matter contents, together
with studies on competitive relations between knotgrass and the native and exotic species
assemblages.
This study was financed by the Portuguese Foundation for Science and Technology
(POCTI/MGS/42584/2001). Francisca Aguiar was supported by a PhD grant (PRAXIS
XXI/BD/11173/97) and Ivan Bernez by a post-doctoral grant (SFRH/BPD/1513/2000) from
the same institution. The comments and revision performed by two anonymous reviewers
greatly improved the final version of the manuscript.
11.6 References
Aguiar F., Moreira I. and Ferreira M.T. 1996. A percepção da vegetação aquática infestante pelas
entidades gestoras dos recursos hídricos (English abstract title: Perception of aquatic weed
problems in Portugal by water resources managers). Revista de Ciências Agrárias 19: 35-56.
Aguiar F.C., Moreira I. and Ferreira M.T. 2001. Exotic and native vegetation establishment following
channelization of a western Iberian river. Regulated Rivers: Research and Management 17:
509-526.
Braun-Blanquet J. 1932. Plant Sociology: The study of plant communities. (English translation), Mc-
Graw-Hill, New York.
Costa J.C., Capelo J., Espírito-Santo M.D., Lousã M., Monteiro A., Mesquita S., Vasconcelos M.T. and
Moreira I. 1999. Plant communities of the lagoons of the Portuguese Coastal Superdistrict –
a multivariate approach. Hydrobiologia 415: 67-75.
Dawson F.H. and Kern-Hansen U. 1978. Aquatic weed management in natural streams: the effect of
shade by the marginal vegetation. Verh. Internat. Verein. Limnol. 20: 1429-1434.
DiTomaso J.M. and Healy E.A. 2002. Aquatic and Riparian Weeds of the West. University of California
Agriculture and Natural Resources. Publication 3421. Oakland, US.
Ferreira M.T. and Moreira I.S. 1995. The invasive component of a river flora under the influence of
Mediterranean agricultural systems. In: Pysek P., Prach K., Rejmánek M. and Wade M.
(eds.), Plant Invasions - General Aspects and Special Problems, SPB Academic Publishing.
approach. Archiv für Hydrobiologie 155:121-145.
macrophytes in a Southern Iberian River. Verh. Internat. Verein. Limnol. 27: 3833-3837.
Franco J. A. and Rocha-Afonso M.L. 1998. Nova Flora de Portugal (Continente e Açores). 3.
Gramineae. Escolar Editora, Lisboa.
Franco J.A. and Rocha-Afonso M.L. 1980. Distribuição em Portugal das principais infestantes. In:
Proceedings of the I Congresso de Fitiatria e Fitofarmacologia, Lisboa. pp.1-22.
responses to predictable seasonal events. Annu. Rev. Ecol. Syst. 30: 51-81.
Guillerm J. L., Le Floc’h E., Maillet J. and Boulet C. 1990. The invading weeds within the Western
Mediterranean Basin. In: di Castri F., Hansen A.J. and Debussche M. (eds), Biological
invasions in Europe and in the Mediterranean Basin. Kluwer Academic Publ., Dordrecht. pp.
61-84.
Huang W.Z., Hsiao A.I. and Jordan L. 1987. Effects of temperature, light and certain growth
regulation substances on sprouting, rooting and growth of single-node rhizome and shoot
segments of Paspalum distichum L. Weed Research 27: 57-67.
Le Floc’h, E. 1991. Invasive plants of the Mediterranean Basin. In: Groves R.H. and di Castri F. (eds.),
Biogeography of Mediterranean invasions. Cambridge University Press, Cambridge. pp. 67-
80.
Liu S.H., Hsiao A.I. and Quick W.A. 1999. The influence of leaf blade, nutrients, water and light on
the promotion of axillary bud growth of isolated single-node stem segments of Paspalum
distichum L. Weed Research 31: 385-394.
McIntyre S. and Lavorel S. 1994. Predicting richness of native, rare and exotic plants in response to
habitat and disturbance variables across a variegated landscape. Conservation Biology 8:
521-531.
Ministério do Ambiente. 1999. Decreto-Lei nº 565/99 de 21-12-1999, Espécies não indígenas da flora e
da fauna. Diário da Républica I Série A, 295:9100-9114.
P-Clima. 1999. Software Provedor de Dados Climatológicos para Portugal. Ver. 1.2. Projecto ALTENER
XVII/4.1013/Z/96-092. (Ed.) R. Aguiar, INETI / DER, Lisbon, Portugal.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., de Ferrari C. and Décamps H. 1996. Invasibility of
Raunkiaer C. 1934. The Life Forms of Plants and Statistical Geography. Oxford Clarendon Press,
Oxford.
Sampaio G. 1909. Flora Portuguesa. Imprensa Moderna, Porto.
Silva A.R.P. 1940. O género Paspalum em Portugal. Agronomia Lusitana 2: 5-23.
StatSoft, Inc. 2001. STATISTICA (data analysis software system), version 6.
Strahler A.N. 1957. Quantitative analysis of watershed geomorphology. Trans. American Geophysical
Union 38: 913-920.
Vasconcelos T., Tavares M. and Gaspar N. 1999. Aquatic plants in the rice fields of the Tagus Valley,
Portugal. Hydrobiologia 415: 59-65.
Capítulo 12
Flora exótica e endémica em locais de
referência e de não-referência de corredores
fluviais do Centro e Sul de Portugal
Chapter 12 Exotic and endemic flora on reference

and non-reference sites from Iberian fluvial systems*
*
IN PREPARATION, with authors: F.C. Aguiar, M.T. Ferreira and A. Albuquerque.
Capítulo 12 – Exotic and Endemic Flora on Reference and Non-reference Sites | 279
12.1 Aims
Exotic plant species from aquatic and wet habitats of Portuguese freshwater systems were
compared between reference conditions, i.e. near-natural river corridors, and non-
reference conditions in order to detect differences in the susceptibility of the river
ecosystem to invasion. The occurrence and abundance of endemic species was also studied.
Key words: Mediterranean Basin, Portugal, freshwater ecosystems, rivers, invasive species,
species richness, human disturbance.
12.2 Methods
We recorded vegetation at 272 sampling sites from two river types of centre and south of
Portugal, having differences in the geological background. Species identification followed
Franco (1971; 1984), Franco and Rocha-Afonso (1994; 1998; 2003), Tutin et al. (1993; 2001),
and taxonomic nomenclature was updated with Castroviejo et al. (1993; 1997) and Talavera
et al. (1999) when necessary. Abiotic variables and human disturbance metrics were
collected from a Geographic Information System (GIS) database or recorded on site.
Selection criteria for reference site-groups included the physical alterations of riverbanks
and valley floor, state of the woody formations, artificial sediment load, and water quality
parameters. Possible relationships among the observed values of richness and cover of both
exotic and endemic species and the underlying factors were estimated with Pearson
Product-Moment Correlations, and with Spearman´s Correlation for interval data and for
dummy variables. Comparisons of exotic and endemic species richness and cover among
groups of river types from reference/non-reference conditions were performed using t-
tests, and Mann-Whitney U tests for non-normal data.
12.3 Results and discussion
A total of 568 taxa were identified, of which 28 were endemic and 44 were exotic (see
Appendices 12.1 and 12.2, respectively). Exotic species were present in 91% of sites,
whereas the endemic species occurred in about 45% of the sampling locations. In the fluvial
corridors of the south-western Iberia (Corbacho et al. 2003) and throughout the
Mediterranean Basin (Décamps et al. 1988), the riparian vegetation has been fragmented
and species assemblages were modified, therefore reference conditions also reflect past
and present human disturbances (Aguiar et al. 2000, Ferreira et al. submitted). The most
abundant exotic species were perennials, such as the rhizomatous grasses, knotgrass
(Paspalum distichum L.), and giant reed (Arundo donax L.), and the hydrophytes water-
hyacinth (Eichhornia crassipes (Mart.) Solms-Laub) and pacific mosquitofern (Azolla
filiculoides Lam.). Though always present on reference sites, exotic species are consistently
more spread in non-reference sites on both river types, with calcareous background
presenting a higher invasibility (Fig. 12.1). In siliceous river types, the richness and cover of
endemic species were significantly lower in the non-reference sites than they were in the
reference ones, whereas no differences were detected between the corresponding
calcareous groups. In addition, there were no significant differences in exotic and endemic
species richness and cover between the two river types, whether from the reference and
the non-reference groups. Reference conditions on calcareous river types were scarcely
recorded. The high quality of the soils of the floodplains supports intensive agricultural
systems, such as irrigations crops and rice fields, and have a favourable location (mainly
littoral) for urban settlements, thus representing a highly disturbed landscape. Exotic
richness and cover (Table 12.1) were positively related with the direct man-induced
disturbances within the fluvial system, for instance the physical alteration of bank margins,
and with the use of the floodplain, such as the ‘urban occupation’, the ‘intensive
agriculture’, and the ‘nutrient inputs’. Endemic species also respond to anthropogenic
variables, rather than to geographical variables, with its richness and cover increasing
towards lower human impacts on fluvial systems and related floodplains.
Exotic species
100 10
Siliceous Calcareous
Outliers
80 8
a a
60 6
a a
40 4
a a
Species richness (nº.)

a a 2
Species cover (%)
20 a'
a' a' b'
b' b'
a' a'
0 0
Endemic species
10
4
8
3
6
2
4 a a
a a
a a
2 1
a a
a' a' b' a' a' a' b' a'
0 0
Reference Non-reference Reference Non-reference
River group River group
Figure 12.1 Comparison of endemic and exotic species richness and cover for reference and non-reference
site-groups. Median, non-outlier ranges, and outliers are represented. Columns with different letters indicate
significant differences at P>0.01, a and b between river types (calcareous and siliceous rivers), and a’ and b’
between river groups (reference and non-reference sites).
Table 12.1 Summary statistics for the correlations relating the richness and cover of exotic and endemic
species to regional, environmental and anthropogenic variables. * = p < 0.05; **= p < 0.01; ns= non-
significative, p=0,05. For details on variables description, see Aguiar et al. (accepted).
Variables and landscape metrics Exotic Endemic

richness cover richness cover
Longitude (degrees) 0.43* 0.40* -0.30* -0.30*
Distance from source (km) 0.17* 0.12* ns ns
Urban occupation (1-5) 0.34** 0.49** -0.32* -0.24*
Intensive agriculture (1-5) 0.29** 0.40** ns -0.23*
Regulation (1-5) -0.42** -0.27** 0.19* 0.22*
Tree shade (%) ns -0.16* 0.19* 0.23*
Riparian fragmentation (1-5) 0.36** 0.34** -0.25 ns
Sediment load (1-5) 0.29** 0.36** ns ns
Conductivity (µS/cm) ns 0.40* -0.27* -0.28*
Lime, silt, and FPOM (%) ns 0.23* -0.17* ns
Toxic conditions (1-5) 0.35** 0.42** -0.21* -0.20*
Physical bank alteration (1-5) 0.33** 0.35** -0.26* -0.19*
Nutrients and organic inputs (1-5) 0.30** 0.43** ns ns
12.4 Final considerations
There is a large evidence of the capability of some exotic species in the displacement of
native and endemic flora and being responsible for diversity loss through monospecific
spread on river margins and aquatic ecosystems. Further efforts must be done in Portuguese
riparian areas concerning the protection of the few (relatively) preserved riparian habitats
by lowering the direct and indirect land-use pressures on riparian corridors and preventing
future invasions. Mitigation measures and restoration projects on impoverished riparian
corridors and fluvial systems must also consider the geomorphological and abiotic setting,
as well as the current floristic composition. In addition, significant differences between
geological river types were detected in the present study, and therefore requiring different
management options, and diverse land use decision-making.
This study was financed by the Portuguese Foundation for Science and Technology
(POCTI/MGS/42584/2001). Francisca Aguiar was supported by a PhD grant (PRAXIS
XXI/BD/11173/97) from the same institution.
12.6 Appendices1
Appendix 12.1 Type of endemism, usual habitat and life span of the endemic species recorded in the sampling
sites.
Specie Type of Usual Life

endemism habitat* span**
Calamintha baetica Boiss. & Reuter European t p
Callitriche brutia Petagna Lusitanian a p
Cistus populifolius L. ssp. populifolius European t p
Cistus psilosepalus Sweet European t p
Clematis campaniflora Brot. Iberian r a
Cytisus striatus (Hill) Rothm. European t p
Galium broteroanum Boiss. & Reuter European r p
Gladiolus illyricus Koch ssp. reuteri (Boiss.) Coutinho Iberian r p
Gratiolia linifolia Vahl Lusitanian r p
Herniaria lusitanica Chaudhri European t a
Hypericum elodes L. European r p
Juncus acutiflorus Ehrh. Ex Hoffm. ssp. rugosus (Steudel) Coutinho Iberian r p
Lavandula luisieri (Rozeira) Rivas-Martínez European t p
Lavandula pedunculata (Miller) Cav. ssp. lusitanica (Rozeira) Franco European t p
Linaria spartea (L.) Willd. European t a
Lotus subbiflorus Lag. ssp. castellanus (Boiss. & Reuter) P.W. Ball. European t a
Marsilea batardae Launert Iberian r p
Myosotis lusitanica Schuster Lusitanian r p
Narcissus fernandesii G. Pedro Iberian r p
Narcissus jonquilla L. Iberian r p
Prunus spinosa L. ssp. insititoides (Fic. & Coutinho) Franco Lusitanian t p
Pterocephalus intermedius (Lag.) Coutinho European t p
Pulicaria paludosa Link in Schrader European r a
Rhynchosinapis hispida (Cav.) Heywood European t a
Salix salviifolia Brot. ssp. australis Franco Lusitanian r p
Teucrium scorodonia L. ssp. scorodonia European t p
Vincetoxicum nigrum (L.) Moench. European r p
Wahlenbergia hederacea (L.) Rchb. European r p
*t – terrestrial; r – riparian; a – aquatic

**p – perennial or biannual; a – annual
1
Sources of information of Appendices 12.1 and 12.2: Castroviejo et al. (1993; 1997), Franco (1971; 1984),
Franco and Rocha-Afonso (1994; 1998; 2003), Talavera et al. (1999) and Tutin et al. (1993; 2001).
Appendix 12.2 Origin, Raunkjaer life form and life span of the exotic species recorded in the sampling sites.
Exotic species mentioned on the Portuguese legislation (Ministério do Ambiente 1999) as invasive species are note
with an •. p – perennial or biannual; a – annual.
Specie Origin Raukiaer life form Life

span
Abutilon theophrasti Medik. South Asia therophyte a
• Acacia dealbata Link SE Australia; Tasmania mesophanerophyte p
• Acacia longifolia (Andrews) Willd. Australia microphanerophyte p
• Acacia melanoxylon R. Br. in Aiton SE Australia; Tasmania megaphanerophyte p
Acer negundo L. North America mesophanerophyte p
• Ailanthus altissima (Miller) Swingle China mesophanerophyte p
Amaranthus blitoides S. Watson North America therophyte a
Amaranthus hybridus L. Tropical and subtropical therophyte a
Amaranthus retroflexus L. North America therophyte a
Arundo donax L. Central (?) and South Asia geophyte p
Aster squamatus (Sprengel) Hieron Central and South America hemicriptophyte p
• Azolla filiculoides Lam. North and South America hydrophyte p
Bidens aurea (Aiton) Sherff Central America therophyte a
Bidens frondosa L. America therophyte a
Callitriche cribrosa Schotsman North Africa hydrophyte p
Castanea sativa Miller Oriental Mediterranean mesophanerophyte p
Chenopodium ambrosioides L. Tropical America therophyte a
Chrysanthemum segetum L. SW Asia therophyte a
• Conyza bonariensis (L.) Cronq. South America therophyte a
Cyperus eragrostis Lam. Tropical America geophyte p
• Datura stramonium L. America therophyte a
• Eichhornia crassipes (Mart.) Solms- Tropical America hydrophyte p
Eucalyptus camaldulensis Dehnh. Australia mesophanerophyte p
Eucalyptus globulus Labill. ssp. SE Australia; Tasmania megaphanerophyte p
Ficus carica L. SW Asia microphanerophyte p
Galinsoga ciliata (Raf.) S.F. Blake from Mexico to Chile therophyte a
• Galinsoga parviflora Cav. South America therophyte a
Gomphocarpus fruticosus (L.) W.T. South Africa nanophanerophye p
Helianthus annuus L. North America therophyte a
Appendix 12.2 (cont.’d) Origin, Raunkjaer life form and life span of the exotic species recorded in the
sampling sites. Exotic species mentioned on the Portuguese legislation (Ministério do Ambiente 1999) as invasive
species are note with an •. p – perennial or biannual; a – annual.
Specie Origin Raukiaer life form Life

span
Lindernia procumbens (Krock.) Philcox Europe; Asia therophyte a
Myriophyllum aquaticum (Vell.) Verdc. South America hydrophyte p
• Oxalis pes-caprae L. South Africa geophyte p
Paspalum distichum L. Tropical geophyte p
Phytolacca americana L. North America hemicryptophyte p
Prunus cerasus L. SW Asia microphanerophyte p
Prunus domestica L. Caucasus microphanerophyte p
Ricinus communis L. Tropical regions nanophanerophyte p
• Robinia pseudoacacia L. North America mesophanerophyte p
Salix viminalis L. Central and East Europe mesophanerophyte p
Salix babylonica L. China macrophanerophyte p
Senecio petasitis (Sims) DC. México camaephyte p
• Tradescanthia fluminensis Velloso from SE Brasil to Argentina camaephyte p
Vitex agnus-castus L. South Europe nanophanerophyte p
Zantedeschia aethiopica (L.) Spreng. South Africa geophyte p
12.7 References
Aguiar F.C., Ferreira M.T., Albuquerque A. and Bernez I. accepted. Invasibility Patterns of Knotgrass
(Paspalum distichum L.) in Portuguese Riparian Habitats. Weed Technology.
Aguiar F.C., Ferreira M.T. Moreira I.S. and Albuquerque A. 2000. Riparian types on a Mediterranean
Castroviejo S., Aedo C., Benedí C., Laínz M., Muñoz-Garmendia F., Nieto-Feliner G. and Paiva J. 1997.
Flora Iberica: Plantas Vasculares de la Península Ibérica e Islas Baleares. 8. Haloragaceae-
Euphorbiaceae, Real Jardín Botánico, CSIC, Madrid, Spain.
Castroviejo S., Aedo C., Cirujano S., Laínz M., Montserrat P., Morales R., Muñoz- Garmendia F.,
Navarro C., Paiva J. and Soriano C. 1993. Flora Iberica: Plantas Vasculares de la Península
Ibérica e Islas Baleares. 3. Platanaceae-Plumbaginaceae (partim)-Capparaceae, Real Jardín
Botánico, CSIC, Madrid, Spain.
Corbacho C., Sánchez J.M. and Costillo E. 2003. Patterns of structural complexity and human
Agriculture, Ecosystems & Environment 95: 495-507.
Ferreira M.T., Aguiar F. and Nogueira C. submitted. Changes in riparian integrity over space and time:
influence of environment and land use.
Franco J.A. and Rocha-Afonso M.L. 1994; 1998; 2003. Nova Flora de Portugal (Continente e Açores).
3(1) Alismataceae-Iridaceae; 3(2) Gramineae; 3(3) Juncaceae-Orchidaceae. Escolar
Editora, Lisbon.
Franco J.A. (ed). 1971; 1984. Nova Flora de Portugal (Continente e Açores). 1. Lycopodiaceae-
Umbelliferae.; 2. Clethraceae-Compositae. Author Edition, Lisbon.
Ministério do Ambiente. 1999. Decreto-Lei nº 565/99 de 21-12-1999, Espécies não indígenas da flora e
da fauna. Diário da Républica I Série A, 295: 9100-9114.
Talavera S., Aedo C., Castroviejo S., Romero-Zarco C., Sáez L., Salgueiro F.J. and Velayos M. 1999.
Flora Iberica: Plantas Vasculares de la Península Ibérica e Islas Baleares. 7(1). Leguminosae
(partim), Real Jardín Botánico, CSIC, Madrid, Spain.
Tutin T.C., Burges N.A., Chater A.O., Edmonson J.R., Heywood V.H., Moore D.M., Webb D.A.,
Valentine D.H. and Walters S.M. 1993. Flora Europaea. 1.Psilotaceae- Platanaceae.
Cambridge University Press. Cambridge.
Tutin T.C., Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. and Webb D.A.
2001. Flora Europaea. 5. Alismataceae-Orchidaceae (Monocotyledonae). Cambridge
University Press. Cambridge.
Parte V
Conclusões e considerações finais
Part V Conclusions
and final remarks
O melhor de uma verdade é o que dela nunca se chega a saber.
Virgílio António Ferreira (1916-1996)

Capítulo 13
Discussão de resultados e conclusões
Chapter 13 Discussion and conclusions

292 | Parte V - Conclusões e Considerações Finais
Capítulo 13 – Discussão de Resultados e Conclusões | 293
13.1 Introdução
Os trabalhos realizados no âmbito desta dissertação permitiram reconhecer a importância

das galerias ribeirinhas e da vegetação ripícola na integridade ecológica dos sistemas
fluviais, bem como explorar o potencial preditivo da vegetação em face a perturbações
antrópicas e ambientais. Paralelamente, o estudo dos padrões de distribuição de grupos
florísticos, em bacias com diferenças a nível ambiental, geográfico e de ocupação do solo,
permitiram discutir as diferenças da vegetação ripícola face a pressões antrópicas, e a
variações geográficas, habitacionais e morfogénicas. Em particular, os estudos da
invasibilidade dos ecossistemas aquáticos e ripícolas e a avaliação do potencial de invasão
da flora exótica em sistemas fluviais portugueses contribuíram para uma discussão
fundamentada em relação aos perfis florísticos1 encontrados noutros sistemas fluviais
mediterrânicos. Identificaram-se algumas prioridades de investigação na sequência dos
resultados obtidos e respectiva discussão e reconheceram-se necessidades de amostragem,
nas quais se inclui a consideração de variáveis potencialmente interessantes de modo a
aumentar o grau de conhecimento do sistema fluvial nas nossas condições.
13.2 Composição e integridade das galerias ribeirinhas em sistemas fluviais

mediterrânicos
13.2.1 Padrões de riqueza e composição
Os ecossistemas ripícolas dos sistemas fluviais portugueses suportam comunidades vegetais

lenhosas com baixa riqueza, aliada a uma considerável homogeneidade florística. Tomando
o caso de estudo da bacia do Tejo (Aguiar et al. 2000 -capítulo 3), registou-se uma riqueza
total de 54 espécies (árvores, arbustos e lianas) e valores da ordem das seis a oito espécies
lenhosas por local amostrado. Trabalhos realizados na Península Ibérica apresentaram
1
Os nomes vulgares das espécies referidas neste capítulo constam dos trabalhos de Moreira et al. (2002) e
Catarino et al. (2001), complementados por Rocha (1996) para espécies de fácies terrestre. Seguiram-se as floras
de Tutin et al. (2001), Franco e Rocha-Afonso (2003) e a revisão taxonómica de Duarte et al. (2004), para
actualização dos nomes científicos das espécies, que entretanto foram alterados no decurso deste trabalho. Na
primeira vez que a espécie é mencionada, indicam-se os sinónimos que foram utilizados em capítulos anteriores e
o respectivo classificador entre parênteses.
valores de riqueza em espécies lenhosas ripícolas igualmente baixos, e.g. no troço médio da
Bacia do Guadiana (Corbacho et al. 2003), e sudeste de Espanha (Tabacchi et al. 1996;
Salinas et al. 2000) enquanto que outras regiões de clima mediterrânico apresentaram
valores de riqueza consideravelmente mais elevados, com registos de espécies lenhosas
superiores a 160 espécies no Sudoeste da França (Tabacchi e Planty-Tabacchi 1990;
Tabacchi et al. 1990) e de cerca de duas centenas na América do Norte (Tabacchi et al.
1996). Acresce que, se nos reportarmos apenas às espécies associadas ao meio lótico, a
riqueza em espécies lenhosas diminui para 26 espécies recenseadas e quatro a cinco
espécies por local na parte portuguesa da bacia do Tejo. A frequência de ocorrência da
maioria das espécies de fácies terrestre é muito baixa, pelo que não deverão ser
consideradas como verdadeiros componentes dos habitats ripícolas. A presença destas
espécies pode ser devida a plantações como são exemplo a oliveira (Olea europaea L. var.
europaea), o castanheiro (Castanea sativa Miller) e espécies dos géneros Eucalyptus e
Pinus. Por outro lado, nas margens de cursos de água temporários em zonas de forte
estiagem surgem espécies características dos matos mediterrânicos como Cistus sp., Myrtus
communis L., Pistacia lentiscus L. que ocorrem pontualmente ou cobrem grande parte das
margens. Por último, espécies exóticas invasoras, como o ailanto [Ailanthus altissima (Mill.)
Swingle] e espécies do género Acacia, colonizam margens dos cursos de água e tendem à
formação de comunidades monoespecíficas. Estas espécies estão associadas à perturbação
humana no sistema fluvial e áreas envolventes e são uma das razões do empobrecimento em
espécies nas galerias. De facto, uma das principais causas apontadas para os baixos valores
encontrados é a perturbação humana recente e histórica no sistema fluvial, ecótonos
ripícolas e zonas confinantes, sobretudo devido à regularização e à ocupação agrícola e
urbana do solo (Aguiar et al. 2000, Corbacho et al. 2003).
Numa primeira tentativa de eleição de locais de referência, avaliou-se a perturbação física

no perfil transversal e longitudinal dos troços amostrados e seleccionaram-se locais com
menor alteração. Neste universo, confirmou-se uma subida do número médio de espécies
lenhosas. Por outro lado, é frequente a ocorrência de bosques depauperados, com
dominância e elevada densidade de uma espécie, como é o caso dos amiais ribeirinhos, que
são comuns em solos aluviais e periodicamente inundados, sem estiagem acentuada e dos
salgueirais arbustivos, frequentes em margens aluvionares de cursos de água torrenciais. No
entanto, um padrão similar pode ser encontrado devido a interferências antrópicas, como é
o caso dos salgueirais na bacia do Divor (Ferreira 1994) e Sorraia (Ferreira 1992). Este
exercício explicativo dos padrões de riqueza em sistemas ibéricos conduzem à necessidade
de uma abordagem conjunta dos padrões espaciais de integridade e de composição
florística das galerias ripícolas.
Apenas grandes assimetrias geográficas e climáticas permitiram a separação de galerias com

claras diferenças florísticas. As variáveis ambientais e antrópicas explicaram uma pequena
percentagem da variabilidade intrínseca da composição florística, e revelaram um gradiente
hidrotérmico e hierárquico importante, separando locais de maior altitude, com substratos
rochosos, menor temperatura e maiores escoamentos e precipitações, dos locais com maior
estiagem, maior largura do canal e maiores números de ordem. Contudo, a afiliação
geográfica dos tipos florísticos encontrados não é evidente. O efeito de escala poderá
influenciar o estudo dos determinantes ambientais e a definição de grupos florísticos, pelo
que a consideração de áreas mais vastas poderá conduzir a gradientes ambientais mais
diferenciados e a uma maior coesão de grupos, como os encontrados por Mykestad e Birks
(1993) em estudos de distribuição de espécies de Salix na Europa. Por outro lado, nesta
dispersão está implícita a influência da perturbação humana, confirmada pela aplicação da
mesma metodologia de tratamento de dados aos locais menos perturbados, que permitiu
uma melhor diferenciação regional dos grupos florísticos e uma maior distinção em termos
composição.
Este primeiro trabalho da dissertação (capítulo 3) permitiu reconhecer a necessidade de

estudo de locais ou condições de referência, e depreender que esta selecção é uma etapa
crítica no processo, uma vez que há fortes indícios da ausência de bosques ribeirinhos
pristinos na Península ibérica (Gallego-Fernández et al. 1999), e poucos podem ser
considerados semi-naturais ou de referência. Por sua vez, na Europa, estima-se que mais de
90% dos leitos de cheia aluvionares foram ou são cultivados, e apenas uma pequena parte
destes sistemas permanece com as funções e dinamismo semelhantes ao que poderia ser
considerado natural (Tockner e Stanford 2002). Acresce ainda, que o referencial histórico,
ou melhor, paleoecológico em sistemas fluviais europeus é quase inexistente, dificultando o
desenho de modelos conceptuais de sistemas de referência (Brown 2002).
13.2.2 Padrões de continuidade e integridade
O padrões de continuidade e integridade das galerias ribeirinhas foram estudados em sub-

bacias do Rio Tejo com uma elevada variabilidade ambiental e de uso do solo envolvente
(capítulo 4).
De um modo geral, este estudo revelou uma baixa integridade e uma fragmentação
generalizada das galerias ribeirinhas, conjuntamente com uma reduzida largura média da
zona ripária. Para além da variabilidade interbacia, os padrões de fragmentação variaram
bastante ao longo dos sistemas fluviais. Os sistemas fluviais mediterrânicos apresentam
padrões de fragmentação e conectividade singulares, devidas à escassez de água e à

perturbação humana, mas também resultantes da complexa geomorfologia (Malanson e
Cramer 1999). De facto, a cursos intermitentes de cabeceiras e substrato rochoso
correspondem zonas ripícolas naturalmente desprovidas de árvores, suportando apenas
vegetação arbustiva ou herbácea. Por outro lado, zonas de vegetação arbórea densa e
contínua encontradas a jusante de albufeiras (e.g. albufeira da Bouçã no Rio Zêzere),
resultante de invasão da zona ribeirinha por espécies exóticas como a mimosa (Acacia
dealbata Link), podem conduzir a ilações erróneas sobre a integridade das galerias
ribeirinhas, se não se atender a estudos complementares de composição florística.
Em sistemas fluviais mediterrânicos, a vegetação lenhosa ripícola estrutura-se em faixas de

reduzida largura, apresentando uma variabilidade importante ao longo do sistema fluvial,
em resposta do gradiente ambiental e das perturbações no sistema fluvial e no meio
envolvente. De facto, e à parte dos constrangimentos topográficos, geológicos e climáticos,
os valores encontrados foram baixos, mesmo quando comparados com outros sistemas de
características mediterrânicas, deixando antever uma maior degradação. Os valores médios
encontrados por bacia variam desde 2,9 metros a 9,8 metros, com um máximo absoluto de
24 metros. Por outro lado, em zonas temperadas frias, com uma envolvente fortemente
agrícola e urbanizada, foram encontrados valores da ordem dos 130 metros de
desenvolvimento lateral da floresta ripícola (Roth et al. 1996).
A avaliação dos padrões espaciais de desenvolvimento lateral das galerias ribeirinhas na

Bacia do Tejo (capítulo 4), permitiu reconhecer a heterogeneidade montante-jusante a
nível da das bacias amostradas. Evidenciaram-se três padrões espaciais de variação
longitudinal: um grupo de bacias caracterizado por uma maior densidade e largura ripícola
nos troços fluviais médios, nas quais se incluem as dos rios Pônsul, Canha e Tera; um
segundo caracterizado por um aumento contínuo ao longo do gradiente longitudinal do
sistema (Divor, Meimoa e Nabão); e um último cujos padrões de elevada perturbação
artificial, nomeadamente regularização e canalização, parecem contribuir para a
manifestação de padrões irregulares de continuidade e largura da galeria ripícola,
sobretudo nos troços médios e inferiores (bacias dos rios Zêzere e Trancão). Parte destes
padrões são explicados pela diferenciação florística das galerias, aliada aos
constrangimentos ambientais. Nas cabeceiras Noroeste da Bacia do Tejo, estendem-se
galerias naturalmente esparsas e estreitas, compostas por arbustos como a urze-branca
(Erica arborea L.), o sabugueiro (Sambucus nigra L.), e o amieiro-negro (Frangula alnus
Mill.), enquanto que nos troços de cabeceira das bacias da margem esquerda a sudeste,
galerias igualmente esparsas são compostas por espécies resilientes e adaptadas a
condições de maior aridez como a tamargueira (Tamarix africana Poir.) e o loendro (Nerium
oleander L.). Em oposição, nos leitos de cheia dos troços médios da Ribeira de Canha e nas
cabeceiras do Rio Divor desenvolvem-se bandas contínuas e densas de salgueiros (sobretudo
Salix alba L. e S. salviifolia Brot.) relacionadas, entre outros factores, com a importação e
retenção de sedimentos ricos em nutrientes, provenientes da agricultura intensiva de
regadio ou dos arrozais das zonas envolventes.
As galerias ripícolas estudadas formam uma intrincada malha sequencial e aparentemente

aleatória de fragmentos de galeria bem constituída em ambas as margens, de troços sem
galeria, de zonas com árvores isoladas ou galeria esparsa em ambas as margens e de troços
com galeria contínua em apenas uma margem. As interrupções são frequentes, com médias
de 3,5 a 9 vezes de mudança entre classes de continuidade de galeria por unidade de
amostragem de 3 km, atingindo máximos de 11 a 19 mudanças de classe. A galeria
encontra-se ausente em mais de um terço do corredores fluviais estudados à excepção da
bacia da Ribeira de Canha, ultrapassando os 50% nos rios Pônsul, Zêzere e Trancão. As
galerias esparsas ou compostas por árvores em grupos isolados são frequentes quer nas
zonas de cabeceira, quer em troços médios e inferiores do sistema fluvial. Parte deste
padrão é explicado por limitações naturais sobretudo nas cabeceiras, no entanto é usual o
corte de vegetação arbórea, a dessecação da vegetação na zona ripícola, entre outras
perturbações que induzem à alteração dos padrões naturais de continuidade da galeria. As
galerias contínuas ocupam menos de um quarto do comprimento total dos corredores
fluviais, e encontram-se geralmente dispersas em pequenos fragmentos ao longo dos troços
médio e inferior dos sistemas fluviais, o que revela uma potencial capacidade de suporte de
galerias de maior integridade.
As análises de componentes principais revelaram importantes diferenças entre bacias, tanto

a nível climático e hidrogeomórfico, separando claramente as bacias nordeste das as do Sul
e Centro-oeste, como em relação ao uso do solo, com o destaque de bacias cuja paisagem é
homogénea e dominada por montados, (Canha, Tera, Divor e Pônsul). Na área envolvente
dos corredores fluviais das bacias de Meimoa, Zêzere, Trancão e Nabão (transeptos de 1000
metros), é frequente a ocorrência de pelo menos três tipos diferentes de uso do solo dos
dez considerados. Além disso, são já o resultado do somatório de várias parcelas, revelando
uma malha complexa de uso do solo, decorrente, em parte, de estruturas fundiárias de
dimensão reduzida. A presença de caminhos e corredores de acesso ao longo da zona
ripícola é também comum.
Numerosos trabalhos, realizados em várias regiões do globo, apontam para a importância do

uso da terra nos padrões de fragmentação da vegetação ripícola, como por exemplo
Décamps et al. (1988), Jongman (1992), Lepart e Debussche (1992), Allen e O’Connor
(2000), Haslam (2000), Rao e Pant (2001), Inoue e Nakagoshi (2001), Corbacho et al. (2003),
Snyder et al. (2003). No estudo desenvolvido (capítulo 4) o uso do solo e os factores
ambientais explicaram, na mesma ordem de grandeza, uma pequena parte da variabilidade
dos padrões de fragmentação e continuidade das galerias, enquanto que a variabilidade
florística é melhor explicada por factores ambientais. Acresce que a contribuição do uso do
solo na explicação da composição das galerias é diminuta e não significativa. No entanto, a
variabilidade explicada em conjunto pelas variáveis ambientais e pelo uso do solo é
significativa em ambas as análises, apesar de apresentar valores reduzidos. Quando
analisamos o tipo de variáveis ambientais que determinam a composição e a fragmentação
das galerias, sobressai o maior peso das variáveis geográficas na determinação da
composição, como a altitude, precipitação, temperatura, humidade relativa seguidas de
variáveis locais como a modificação artificial do perfil transversal do sistema fluvial e
presença de afloramentos rochosos. Em oposição, variáveis de posicionamento longitudinal
do troço (distância à nascente, número de ordem) e variáveis locais (largura do canal,
origem geológica do substrato) explicaram melhor a variabilidade dos padrões de
integridade das galerias. Curiosamente os tipos de utilização do solo (agrícola, florestal e
urbana) que contribuíram para a explicação da variabilidade florística e estrutural das
galerias são similares, nomeadamente os troços e unidades de amostragem envolvidos por
“montados de sobro e azinho”, destacam-se de unidades dominadas por “matos e pastagens
semi-naturais” e de zonas com predominância de “hortas, pomares, vinhas e olivais”. Além
destas variáveis, na fragmentação das galerias há uma contribuição da presença de áreas
industriais e urbanas nas zonas envolventes, enquanto que na definição florística das
galerias não se demonstrou uma influência significativa.
Estes resultados suportam a ideia da existência de constrangimentos hidrogeomórficos e

climáticos nos ecótonos ripícolas mediterrânicos e salientam o potencial de previsão das
características do habitat na determinação da estrutura e composição da vegetação
ribeirinha. No entanto, reconhece-se, com Allen e O’ Connor (2000), que as complexas
relações entre factores ambientais e sistemas adjacentes tornam difícil a avaliação efectiva
do papel da perturbação humana na zona ripícola e sistema fluvial. Os ecótonos ripícolas
mediterrânicos dependem consideravelmente das especificidades dos ecossistemas
adjacentes, e a estrutura e padrões de fragmentação são uma expressão da vegetação
arbórea constrangida pelo uso milenar do solo e do próprio sistema fluvial. As
conectividades lateral, longitudinal e vertical do sistema fluvial são afectadas
indirectamente por numerosas e diversas intervenções humanas na zona ripícola, com
consequências na integridade ecológica do corredor fluvial e ripícola, como demonstrado
por Brierley et al. (1999).
Em conclusão, as galerias ripícolas estruturam-se em cordões de largura reduzida e

frequentemente entrecortados, e com uma grande variabilidade longitudinal, potenciada
por acções locais de pequena escala, como o corte de vegetação, plantações, pastoreio,
fogo, construção de represamentos, de difícil quantificação e análise em estudos de grande
escala espacial, como é o estudo de sub-bacias e de bacias hidrográficas. Estas
perturbações locais podem ter uma influência de grandeza semelhante ou superior às
perturbações causadas a um nível hierárquico mais elevado, como a regularização do
sistema fluvial (McIntyre e Hobbs 1999).
13.2.3 Padrões de alteração espaço-temporais
No estudo do capítulo 5, sobre alterações espaço-temporais das galerias ripícolas e dos

ecossistemas adjacentes nas sub-bacias do Rio Tejo (Trancão, Nabão e Canha), foram
identificadas modificações significativas no uso do solo no período considerado (1951-1995),
acompanhadas por diferenças na integridade das galerias ripícolas nos troços amostrados.
No entanto, parece ser necessária uma extensa alteração no uso do solo, para imprimir uma
modificação na cobertura, largura ou fragmentação das galerias, como acontece no Trancão
(1965-1989), onde um aumento substancial da área urbanizada correspondeu a uma redução
da extensão ocupada por vegetação arbórea ripícola, e na ribeira de Canha (1951-1995),
onde o aumento da área de culturas de regadio e da zona urbana se traduziu numa maior
cobertura e largura da galeria, mas também numa maior fragmentação. Contudo, as
consequências da alteração do uso do solo são menos evidentes no Rio Nabão, onde o
aumento da área de pinhal, conseguida à custa da diminuição da ocupação por montados e
matos e o aumento da área urbana não produziram efeitos significativos na largura,
continuidade e cobertura da vegetação lenhosa ribeirinha. Nesta bacia, a dispersão dos
agregados habitacionais e industriais e a presença de hortas e pomares de pequena
dimensão parece ter um efeito atenuador do impacte da pressão humana, ou seja, com
menores efeitos globais directos quando comparados com Canha e Trancão, cuja
distribuição da área urbana é mais concentrada. De um modo geral, as galerias ripícolas
demonstraram ter uma apreciável resistência a perturbações indirectas resultantes de
comutações no uso do solo, no período considerado. Os resultados sugerem que a
proximidade e a magnitude de parcelas com usos distintos do solo são factores
preponderantes na alteração dos padrões de integridade das galerias ripícolas.
Uma grande parte da variabilidade biológica não é explicada pelos factores ambientais e do
uso do solo, à semelhança dos resultados de outros trabalhos realizados em sistemas fluviais
ibéricos (Ferreira 1994; Ferreira e Moreira 1999; Ferreira et al. 1998, 2002). Uma
importante componente pode resultar de efeitos cumulativos de pressão antrópica

continuada na generalidade dos sistemas fluviais mediterrânicos (di Castri 1990; Lepart e
Debussche 1992). No presente trabalho, a floresta ripícola restringe-se a faixas de largura
reduzida, frequentemente descontínuas e fragmentadas – um padrão mantido ao longo do
tempo, não obstante as modificações no uso do solo e as alterações de continuidade da
galeria.
Contudo, este estudo é apenas uma imagem instantânea de um período de 35-45 anos,
enquanto que a perturbação humana nesta região é exercida há mais de 6000 anos a.C.,
desde o período Neolítico, com alterações extremas no tecido florestal. No entanto, ainda
não está demonstrado se os padrões florísticos e estruturais observados resultaram de uma
estrutura existente antes da ocupação humana, ou se são resultado de uma co-evolução
devido à prolongada pressão antrópica.
Estudos cronológicos sobre alterações recentes no uso do solo e nas galerias ripícolas, assim
como análises paleo-ecológicas deveriam ser prioridades de investigação para a
caracterização da vegetação ripícola e dos respectivos padrões de distribuição espacial. No
entanto, dificuldades na obtenção de material de base e a morosidade implícita na
aquisição de dados têm contribuído para uma escassez generalizada destes trabalhos.
Alguns estudos sugerem a existência de florestas ripícolas com maior desenvolvimento na
Bacia do Mediterrâneo, cuja redução nos últimos cem anos tem sido particularmente
intensa (Gallego-Fernández et al. 1999). As comunidades vegetais existentes teriam sofrido
fragmentação e redução em área ao longo de milhares de anos, causada por um maior
stresse hídrico, com perda da conectividade vertical entre a superfície e o sistema
subsuperficial e com consequências na diminuição de água na zona radicular, a qual teria
induzido à morte da vegetação ripícola por dessecação. As comunidades contemporâneas
seriam vestígios de formações ripícolas pré-históricas. Esta hipótese é concordante com o
baixo número de espécies lenhosas encontrada quando comparada com regiões climáticas
semelhantes (e.g. Szaro 1990), o reduzido número de tipologias florísticas e a baixa
discriminação geográfica (Aguiar et al. 2000 – capítulo 3; Salinas et al. 2000), e com a
aparente resiliência em face a alterações significativas no uso do solo adjacente ou a
intervenções directas no sistema fluvial, como a canalização e linearização (Aguiar et al.
2001- capítulo 9).
13.3 Qualidade ecológica e potencial preditivo da vegetação ripícola em

sistemas fluviais mediterrânicos
13.3.1 Contribuição das galerias para a sustentabilidade de comunidades biológicas em

sistemas mediterrânicos
O coberto lenhoso ripícola tem um papel determinante no suporte de várias comunidades,

como mamíferos e aves (Spackman e Hughes 1995), macroinvertebrados (Tait et al. 1994),
peixes (Roth et al. 1996; Jones et al. 1999), répteis e anfíbios (Brito et al. 2002), não só
como fonte alimento, mas também para repouso, refúgio, local de reprodução ou como
corredor migratório (Naiman e Décamps 1997). Deste modo, a composição e os padrões de
cobertura e fragmentação dos corredores ripícolas são de extrema importância na
integridade ecológica e sustentabilidade de ecossistemas fluviais e ribeirinhos (Malanson e
Cramer 1999; Allen e O’Connor 2000) e são condicionantes da diversidade ecológica.
Este estudo (capítulo 6), incidiu em particular, no contributo da composição e estrutura da

vegetação lenhosa ripícola e dos factores ambientais para a distribuição de grupos tróficos
funcionais de macroinvertebrados bentónicos. Os detritos orgânicos provenientes da
vegetação lenhosa ripícola como folhas e pequenos ramos - geralmente designados na
Língua inglesa por ‘coarse particulate organic matter’ (CPOM = material orgânico
reconhecível de diâmetro superior a 1 mm) - são uma fonte de alimento importante para
organismos aquáticos fluviais (Conners e Naiman 1984; Benfield 1997), e têm uma grande
influência nas cadeias tróficas de sistemas fluviais mediterrânicos (Cortes et al. 1995). Por
outro lado, a vegetação ripícola promove a heterogeneidade de habitats aquáticos como
resultado do ensombramento e da deposição de material lenhoso no canal e influencia a
distribuição de comunidades de macroinvertebrados (Hall et al. 1994).
Este grupo de organismos é particularmente sensível a factores ambientais (Hawkins et al.

1982; Townsend et al. 1983; Ormerod e Edwards 1987; Collier 1995; Wright 1995) e ao
coberto ripícola (Downes et al. 1993) e vários trabalhos têm relacionado a composição das
florestas ribeirinhas e a perturbação humana com os padrões comunitários destes
organismos (e.g. Hawkins et al. 1982; Rundle et al. 1992; Tait et al. 1994). A estrutura
ripícola parece ser o factor mais determinante para a entrada de detritos na generalidade
dos sistemas fluviais, quer seja directamente por queda no canal, quer transportada
lateralmente a partir das margens (Naiman e Décamps 1997). No entanto, em sistemas
fluviais mediterrânicos a par da estrutura, a cobertura da vegetação ripícola é um factor
determinante (Maamri et al. 1994). Nos rios intermitentes do Sul da Peninsula Ibérica, com
acentuada sazonalidade de caudais, a composição e estrutura da vegetação tem

características peculiares, incluindo uma cobertura considerável de arbustos de folha
persistente e esclerofílica, além de espécies caducifólias, como freixos, salgueiros e
amieiros.
A par da composição florística, a composição em macroinvertebrados em sistemas fluviais

mediterrânicos também tem características distintas dos cursos de zonas temperadas
húmidas, com proporções geralmente elevadas de quironomídeos (Coimbra et al. 1996;
Pinto e Morais 1998). O presente estudo (capítulo 6) corroboraram-se estes resultados,
revelando comunidades de macroinvertebrados aquáticos empobrecidas (4-29 taxa, média
17.6) e dominadas por espécies ou grupos que toleram a poluição orgânica e a perturbação
humana (sobretudo quironomídeos). Quanto a grupos tróficos funcionais, encontraram-se
essencialmente colectores de depósito (5.4 - 100%, média 58.1), e em menor proporção
colectores-filtradores (0 - 81.3%, média 14.1), trituradores (0 - 58.3%, média 12.1),
predadores (0-51.1%, média 10.6) e raspadores (0-25.5%, média 5.0). A elevada proporção
de colectores de depósito está provavelmente relacionada com a abundância de partículas
orgânicas de pequena dimensão (FPOM) resultantes da agricultura praticada nas áreas
confinantes com a zona ripícola, e que geralmente têm um considerável valor nutritivo
(Shepard e Minshall 1981). A cobertura por vegetação herbácea é frequentemente elevada
em corredores fluviais do Sudoeste da Peninsula Ibérica (Ferreira e Moreira 1999), e pode
constituir uma fonte de alimento suplementar facilmente acessível por via detrítica ou por
herbivoria. Por outro lado, o regime intermitente de caudais com a ocorrência sazonal de
sistemas de características lênticas induz a longas permanências do material orgânico
detrítico na zona pelágica, i.e. na massa de água propriamente dita (Gasith e Resh 1999).
Estas características parecem ser desfavoráveis à população de organismos trituradores,
que estão preferencialmente adaptados a sistemas reófilos, mais oxigenados e de
características lóticas, onde estruturas de retenção, como raízes e troncos conduzem à
acumulação de depósitos lenhosos (Prochaza et al. 1991; Friberg 1997; Flory e Milner 1999).
Pouca atenção tem sido devotada aos efeitos da estiagem nos grupos tróficos funcionais
(Vidal-Abarca et al. 1992). Contudo Gasith e Resh (1999) sugerem que as características
habitacionais desenvolvidas durante a estiagem em sistemas fluviais mediterrrânicos
diminuem o papel dos trituradores na cadeia alimentar, em relação a regiões mais húmidas.
Por outro lado, nos sistemas lóticos, prevalecem frequentemente hábitos alimentares
generalistas (Mihuc 1997). Geralmente é aceite que espécies taxonomicamente próximas
pertencem à mesma guilda, mas é usual a ocorrência de classificações incorrectas (Hawkins
e MacMahon 1989). De facto são escassos os dados de autoecologia das espécies de
macroinvertebrados em sistemas mediterrânicos, e pouco se conhece sobre o uso múltiplo
de recursos alimentares por estes organismos.
As variáveis que mais contribuíram para a explicação na distribuição e composição das

comunidades de macroinvertebrados bentónicos foram as associadas à estrutura e
composição das galerias ribeirinhas, nomeadamente a cobertura arbórea total, o
ensombramento do canal pelas canópias e a abundância de freixos (Fraxinus angustifolia
Vahl ssp. angustifolia) e de amieiros (Alnus glutinosa Mill.). Os amieiros formam
usualmente galerias contínuas e densas, enquanto que os freixos ocorrem geralmente
isolados ou em pequenos grupos com outras espécies ripícolas. Este padrão está
provavelmente associado à participação da folhada e de outros fragmentos da vegetação
ripícola na cadeia alimentar do meio aquático. Porém, e à semelhança de estudos
realizados noutros sistemas ibéricos (González et al. 2003), não foi possível associar a
distribuição de trituradores (e de um modo geral, dos macroinvertebrados) com a
distribuição espacial de CPOM. Este facto, está provavelmente associado à deficiente
qualidade nutritiva da maioria das galerias ripícolas estudadas, e à escassa disponibilidade
de recursos alimentares para o consumo por invertebrados aquáticos. Algumas espécies
lenhosas que fazem parte das galerias têm folhas escamosas e imbricadas (e.g., Erica
arborea e Tamarix africana), esclerofílicas ou coreáceas (e.g. Crataegus monogyna Jacq.)
ou com elevada pubescência (e.g. Salix atrocinerea Brot. e S. salviifolia), materiais que
não são, de um modo geral, apreciados por invertebrados bentónicos (Anderson e Sedell
1979). Apenas a abundância de freixos e amieiros foi relacionada significativamente com a
distribuição das populações de macroinvertebrados, e em particular com a abundância de
trituradores. Estas duas espécies ripícolas têm folhas que se decompõem facilmente e
oferecem uma dieta alimentar de boa qualidade para macroinvertebrados (Bärlocher 1985).
Das vinte e uma variáveis abióticas consideradas, apenas a condutividade, distância à

nascente e número de meses estivais sem caudal foram determinantes para a explicação da
variabilidade biológica. Os trabalhos de Cortes et al. (1998) revelaram resultados similares
na Bacia do Guadiana, enquanto que em águas superficiais das regiões nortenhas de
Portugal, com características mediterrânicas menos acentuadas, os agrupamentos de
macroinvertebrados são mais coesos e mostram uma maior relação com as variáveis
ambientais (Graça et al. 1989; Cortes 1992). Os elevados níveis de perturbação natural são
característicos dos sistemas fluviais mediterrânicos (Gasith e Resh 1999) e podem compelir
à existência generalizada de comunidades resilientes, oportunísticas e relativamente
persistentes, que apresentam uma fraca relação com as variáveis ambientais que
geralmente influenciam os ecossistemas fluviais (Puig et al. 1991).
Em suma, mais de 50% da variabilidade das comunidades de macroinvertebrados não é

explicada pelas variáveis (bióticas e abióticas) estudadas. A imprevisibilidade do regime
hidrológico dos sistemas mediterrânicos combinada com as exigentes necessidades hídricas
(sobretudo para irrigação) provocam uma elevada variabilidade de caudal e uma aridez
típica de corredores fluviais semi-áridos (Puig et al. 1991; Poff et al. 1997; Gasith e Resh
1999). A implementação de caudais ecológicos, e a manutenção de uma cobertura ripícola
aceitável, com a preservação de bosques ribeirinhos formados por espécies ripícolas como o
freixo e o amieiro, são elementos-chave na conservação e gestão de cursos ibéricos de
características semi-áridas.
13.3.2 Potencial preditivo da vegetação macrofítica mediterrânica
Numerosos trabalhos comprovam que os macróficos respondem à perturbação humana e

podem ser utilizados como indicadores da qualidade da água ou da qualidade ecológica do
sistema fluvial (Seddon 1972; Whitton 1979; Wiegleb 1981; Newbold e Holmes 1987),
reflectindo os efeitos da eutrofização (Demars e Harper 1998), da poluição orgânica
(Caffrey 1987), da regularização (Nilsson e Jansson 1995) e do desvio do curso de água
(Stromberg e Patten 1990). Embora as comunidades florísticas em sistemas mediterrânicos
sejam relativamente homogéneas e apresentem uma grande resiliência em face a
perturbações naturais e artificiais, há evidência de um potencial de alteração dos padrões
comunitários em situações de degradação. Os hidrófitos, i.e. plantas aquáticas estritas, tem
sido o grupo mais vulgarmente utilizado em países europeus, uma vez que a resposta a
perturbações efectivas no sistemas fluvial é francamente superior a de outros grupos como
as espécies emergentes ou higrofíticas. No entanto, a flora ibérica é pobre em hidrófitos e
a vegetação ripícola herbácea é dominada por emergentes e espécies tolerantes a
oscilações sazonais de caudal. Deste modo, as variações florísticas em termos de
composição ou de descriptores vegetacionais, como a riqueza total, cobertura em exóticas,
número de espécies ripícolas, são importantes na determinação da eficácia da utilização de
plantas ripícolas como bioindicadoras da qualidade ecológica de sistemas fluviais
mediterrânicos.
Uma das características mais notórias do clima mediterrânico é a grande redução da

precipitação nos meses de Verão, com repercussões na composição florística e na cobertura
das espécies durante a estiagem. Em sistemas fluviais e zonas ripícolas desconhece-se a
magnitude das cambiantes florísticas associadas à variabilidade climática sazonal e
interanual. Foi realizada a comparação florística e análise de métricas biológicas como a
riqueza total, número de espécies raras, riqueza em espécies terrestres, em anos com
valores de precipitação significativamente diferentes (1994 e 1996), e em duas estações
(Primavera 1996 e Verão 1997) no curso principal do Rio Guadiana e nos tributários (parte
portuguesa da bacia) (capítulo 7).
Nesta área, foi encontrado um menor elenco florístico na Primavera, aliado a um menor
número de espécies inventariadas por local. Por outro lado, o número de espécies terrestres
é significativamente superior nos inventários de Verão, correspondendo a uma utilização da
zona ripícola por espécies oportunistas e adaptadas a condições de secura, com origem
provável nas zonas envolventes. Apesar das diferenças de regime pluviométrico, não foram
detectadas diferenças significativas entre os inventários realizados no Verão em anos
distintos. Parâmetros como a riqueza total e percentagem de espécies infrequentes
(espécies que ocorrem em menos de dois inventários) também não foi alterada nos
diferentes anos e períodos de amostragem. Contudo, o hidrófito Marsilea batardae Launert,
espécie endémica rara foi apenas encontrado na Primavera. As amostragens tardias (pleno
Verão), embora com maior elenco florístico são pobres em espécies anuais higrófitas e em
hidrófitos. Por outro lado, na fase inicial de Primavera, parte das espécies anuais e vivazes
higrófitas ainda não se desenvolveram. Assim, pode concluir-se que o período de final de
Primavera e início de Verão (Maio – início de Junho) será o período óptimo de crescimento e
o mais apropriado para amostragens florísticas em corredores fluviais nesta região,
sobretudo se for o caso de amostragens únicas.
As comunidades vegetais ribeirinhas da bacia do Rio Guadiana são pobres em espécies e

apresentam uma elevada percentagem em elementos terrestres e em espécies pouco
frequentes e de pequena cobertura. São dominadas por grupos de espécies ubíquas, não
obstante as condições distintas de disponibilidade de água em anos hidrológicos distintos.
Exemplos destas espécies são as emergentes Cyperus longus L., Eleocharis palustris (L.)
Roem & Schult., Scirpoides holoschoenus (L.) Soják (=Scirpus holoschoenus L.),
Schoenoplectus lacustris (L.) Palla (=Scirpus lacustris L.), e Typha latifolia L., a exótica
2
rizomatosa Paspalum distichum L. [= P. paspalodes (Michx.) Scribn.], e as espécies
lenhosas Fraxinus angustifolia Vahl ssp. angustifolia, Nerium oleander e Flueggea tinctoria
(L.) G.L. Webster [= Securinega tinctoria (L.) Rothm.]. Resultados semelhantes foram
relatados no Verão de 1994 nesta região (Ferreira et al. 1998) numa amostragem mais
extensa, podendo este padrão ecológico ser considerado próprio destas comunidades
biológicas, e não necessariamente devido a alterações humanas contemporâneas, como a
regularização de caudais. Deste modo, tanto a variabilidade hidrológica interanual como a
perturbação hidrológica exógena, poderá não ser detectada de imediato em amostragens
florísticas, a não ser que esta perturbação se mantenha por vários anos.
2
O nome actual consta da segunda edição do volume 5 da Flora Europaea (Tutin et al. 2001).
13.3.3 Os macrófitos como bioindicadores da qualidade ecológica de sistemas fluviais
O uso de macrófitos para a avaliação do estado ecológico de sistemas fluviais, bem como na
monitorização biológica apoia-se na análise da presença e abundância de certas espécies ou
de grupos de espécies, ou na sua ausência, que respondendo de forma significativa e
detectável às alterações antrópicas, denunciam condições ambientais particulares. No
trabalho do capítulo 8, a aproximação multivariada para a avaliação do estado ecológico de
sistemas fluviais foi aplicada com sucesso a comunidades vegetais do sudeste da Peninsula
Ibérica, permitindo a definição de ecótipos e dos respectivos locais de referência, com
apoio na análise de correspondências canónicas e na classificação aglomerativa hierárquica.
Os procedimentos metodológicos e a escolha da época de amostragem de macrófitos fluviais

têm sido apontados como aspectos críticos para o sucesso da avaliação da qualidade
ecológica em meio aquático utilizando métodos de análise multivariada (Bunn e Davies
2000; Wright 2000). Idealmente, os atributos ecológicos de um dado sistema devem ser
relativamente estáveis ao longo do tempo e as comunidades ou agrupamentos biológicos
deverão ser persistentes e apresentar uma estrutura determinística (Wright 2000). Os rios
intermitentes são caracterizados por fraca persistência das condições ambientais derivada
das alterações do regime hidrológico, comunicando um grau de previsibilidade limitada às
comunidades bióticas aquáticas (Gasith e Resh 1999; Bunn e Davies 2000). Em alguns casos,
como relatado em Humphrey e Storey (2000) houve a necessidade de desenvolver modelos
multivariados para diferentes épocas do ano. Contudo, neste estudo, e com base nos
trabalhos realizados anteriormente em sistemas semi-áridos (ver 13.3.2), as amostragens
foram realizadas antes da estiagem, na altura de crescimento óptimo para a maioria dos
macrófitos, permitindo o estudo das comunidades macrofíticas. Para outros sistemas
fluviais como no Norte do país será necessário aferir a metodologia e determinar o melhor
período de amostragem, ou os períodos de amostragem preferíveis.
O delineamento da amostragem inclui outras decisões metodológicas fundamentais, como a

definição do comprimento do troço de amostragem e dos respectivos limites laterais. Um
troço de amostragem que englobe o canal e a área geralmente inundada pelas cheias
anuais, com uma dimensão longitudinal de 250 m foi adoptado na maioria dos trabalhos da
dissertação, e foi demonstrada ser suficiente por Ferreira e Moreira (1999) para garantir a
amostragem da diversidade florística e habitacional presente nos sistemas fluviais do
sudeste da Península Ibérica. No entanto, há que tomar em consideração a existência de
espécies de feição terrestre, uma componente importante nas comunidades de macrófitos
em sistemas fluviais intermitentes. A eliminação desta componente biológica resulta
necessariamente numa considerável redução da riqueza específica total e local; todavia, a
sua inclusão conduz a uma menor previsibilidade dos padrões de distribuição das plantas
ripícolas, uma vez que são espécies de ocorrência geralmente fortuita e localizada (Ferreira
1994; Ferreira e Moreira 1999).
Algumas espécies exóticas ripícolas, como a cana (Arundo donax L.), o graminhão
(Paspalum distichum), e o junção (Cyperus eragrostis L.) são comuns nesta região e
desenvolvem com frequência populações persistentes e naturalizadas nos corredores
fluviais, pelo que a sua exclusão dos modelos multivariados poderá induzir a resultados
deturpados. Por outro lado, a consideração da componente exótica prejudica
potencialmente a eleição das condições de referência, dada a sua ocorrência generalizada
em toda a bacia. Porém, os locais de referência seleccionados através da análise
multivariada e utilizando os factores de perturbação antrópica, apresentam uma baixa
riqueza e cobertura em espécies exóticas. Em oposição, a dominância de um local por
espécies exóticas associa-se a uma forte perturbação antrópica no vale de cheia (Ferreira e
Moreira 1995) ou no canal (ver 13.4.1).
No estudo do capítulo 8, utilizaram-se 19 variáveis abióticas e nove antrópicas para a

análise fundamentada dos três ecótipos constituídos. Em particular, a disponibilidade
hídrica e o substrato geológico relacionaram-se com o uso do solo no vale de cheia e
determinaram em conjunto os padrões de distribuição espacial da vegetação. As
comunidades macrofíticas expressam-se em escalas espaço-temporais distintas e
apresentam diferentes respostas às alterações ambientais (Habersack 2000). Na
monitorização ecológica ou na bioavaliação de ecossistemas, o número de variáveis
consideradas deve ser o mais elevado possível, a não ser que estudos de grande escala
tenham previamente aferido os factores mais influentes para as comunidades de cada
ecótipo (e.g. Holmes et al. 1998). No decorrer do tratamento de dados, permitiu-se um
certo grau de subjectividade, como é o caso da definição do nível de partição dos
dendogramas resultantes da classificação, que determinaram os ecótipos e os respectivos
locais de referência. Tradicionalmente, as aproximações multivariadas envolvem a selecção
prévia de locais de referência, ou seja os potencialmente menos alterados, com base nas
condições geomorfológicas e fisico-químicas. Para tal, será necessária a existência de um
volume de informação ambiental actualizada e fiável, a qual nem sempre está disponível
para rios ibéricos (Alba-Torcedor e Pujante 2000). Para ultrapassar esta limitação, os locais
de referência foram identificados por análise multivariada, após a definição dos ecótipos e
dos factores ambientais influentes. A utilização de análise multivariada para a selecção dos
locais de referência foi sugerida por alguns autores como Gasith e Resh (1999), e poderá ser
uma alternativa adequada em regiões com ecótipos pouco definidos e com escassez de
dados ambientais.
A Bacia Mediterrânica foi uma das primeiras regiões do globo a sofrer os efeitos adversos
das actividades humanas, sobretudo as relacionadas com a desflorestação, o pastoreio e a
agricultura (di Castri 1991). Para além das alterações históricas, o grau de perturbação
exógena recente é comum a toda a bacia e interfere na definição inicial de ecótipos, além
de contribuir para a propagação de ruído ao longo do procedimento metodológico proposto,
e dificultar a definição das condições de referência. Por exemplo, no ecótipo “calcário”,
todos os locais apresentaram um grau considerável de perturbação, devido à elevada
capacidade de uso agrícola destes solos, e a divisão entre locais de referência e de teste foi
pouco clara. Ecótipos com baixa variabilidade, quer seja por serem constituídos por um
universo de locais preservados ou perturbados, dificulta a aplicação deste método, uma vez
que o grau de dissimilaridade é baixo. Outra questão, é a existência de número mínimo de
locais a considerar em cada ecótipo, abaixo do qual os vectores de perturbação não são
detectáveis. Ambos os casos estiveram patentes neste estudo: o ecótipo “silicioso-sul”
apresentou um gradiente de perturbação diminuto e, em consequência, a definição das
condições de referência foi dificultada, enquanto que o ecótipo “silicioso-norte”, apesar de
ter um baixo número de locais, incluía sítios com elevada regularização, permitindo uma
melhor distinção entre os locais de referência e os de teste. A aplicação deste método à
escala da suprabacia poderá ser interessante, na medida em que irá aumentar o número de
locais por ecótipo e provavelmente alargar o gradiente de perturbação antrópica, com
melhor definição de condições de referência.
13.4 Invasibilidade, causalidade abiótica e perturbação antrópica em sistemas

fluviais mediterrânicos
13.4.1 Padrões de estabelecimento da vegetação exótica e nativa num rio mediterrânico

canalizado
As obras de engenharia hidráulica tradicional, nas quais se inclui a canalização são

geralmente responsáveis pelos maiores impactes antropogénicos nos ecossistemas fluviais
(Brookes 1988). As rectificações do perfil longitudinal e transversal afectam a maioria das
características e dos processos hidrológicos e geomórficos, quer no troço canalizado, quer a
jusante deste (Hupp 1992), e geralmente ocasionam uma redução da heterogeneidade física
do leito e das margens, aceleram os processos de erosão e alteram os padrões de
sedimentação e de escoamento. Em consequência, ao nível do ecossistema, o sistema
fluvial experimenta uma redução da heterogeneidade habitacional e da diversidade

ecológica (MacCarthy 1985; Swales 1988; Shields Jr. e Hoover 1991). Enquanto que os
efeitos da canalização nas componentes hidráulica e física do sistema são amplamente
estudados, poucos trabalhos se reportaram ao estudo da alteração da composição florística
e da recuperação da vegetação após a canalização (Bravard et al. 1986; Brookes 1986;
1995; Hupp 1992). Entre as consequências ecológicas sobre a flora dos corredores fluviais,
inclui-se a perca de diversidade específica e de comunidades pioneiras (Bravard et al. 1986)
e alterações na composição florística (Wade e Edwards 1980; Murphy et al. 1987). Todavia,
em algumas situações, a canalização teve apenas pequenas repercussões a longo prazo
(Duvel et al. 1976), ou induziu a acréscimos na biomassa vegetal, mas sem alteração
significativa da composição florística (Brookes 1986).
Em Portugal, tais alterações nunca foram demonstradas ou quantificadas, sendo o Baixo

Mondego uma zona excelente para o seu estudo. O segmento fluvial abaixo do açude de
Coimbra foi sujeito a linearização e aprofundamento, colocação de travessões, reforço dos
taludes com a construção de dois terraços e à homogeneização do substrato do leito e
margens. Após um período de recuperação de doze anos, a distribuição, composição e
cobertura da vegetação aquática e ripícola foi estudada ao longo do segmento canalizado
do Rio Mondego e comparada com a vegetação de um segmento a montante, como
referencial das condições seminaturais anteriores à canalização, como se apresenta no
capítulo 9.
Os resultados do estudo do capítulo 9, indicaram uma riqueza total de espécies lóticas no

rio canalizado semelhante à encontrada no referencial ‘seminatural’, e relativamente
elevada quando comparada com outros rios ibéricos de características mediterrânicas mais
acentuadas (e.g. Ferreira 1994; Ferreira et al. 1998; Ferreira e Moreira 1999). As planícies
aluvionares tendem a apresentar uma elevada resiliência à perturbação e curtos períodos
de recuperação (Wissmar e Swanson 1990). Ademais, em rios mediterrânicos é esperada
uma persistência elevada da riqueza de espécies como reflexo da adaptação evolucionária
dos organismos à elevada heterogeneidade espaço-temporal (Gasith e Resh 1999). Por
exemplo, a vegetação ripícola do Rio Sorraia, principal afluente da margem esquerda do Rio
Tejo, também sujeito a canalização e aprofundamento do leito e revolução das margens,
levou apenas quatro anos a recuperar em termos de composição, cobertura e diversidade
em espécies herbáceas e lenhosas (Ferreira 1992). Acresce que, a canalização neste caso e
de um modo geral, em rios mediterrânicos está associada à regularização a montante para
controlo de cheias ou para regadio, um factor que conduz à redução dos picos de cheia e
aumenta o caudal estival, resultando na manutenção de um período mais alargado de
menor perturbação hidrológica e biológica.
As diferenças florísticas entre o rio canalizado e o “natural” são devidas à dominância em

termos de cobertura de determinadas espécies e não à redução da riqueza específica. O
número de espécies comuns aos dois elencos é de 98 espécies num total de 160 e 150
espécies no rio canalizado e no “natural”, respectivamente; um número relativamente
elevado em face a condições ambientais distintas. Com efeito, detectaram-se diferenças
significativas de riqueza e cobertura em espécies exóticas e nativas por local entre ambos
os segmentos estudados. No troço canalizado a cobertura superficial média em exóticas foi
de 41.5% para o meio aquático e de 62.5% e 38.0% para o 1º e 2º terraços respectivamente.,
enquanto que no troço “natural” de referência, os valores médios de abundância em
exóticas foram de 3.1% no canal, 29.7% no talude interior e 15.2% no exterior. Em oposição,
a cobertura em nativas é significativamente superior nos taludes dos locais de referência e
nos terraços dos segmentos do rio canalizado. Este padrão resultou provavelmente da maior
vulnerabilidade do troço canalizado à invasão por espécies exóticas, devido ao tipo,
frequência e grau de pressão antrópica a que esteve sujeito, quando comparada com a
perturbação do troço ‘seminatural’. Para além do factor antrópico, as causas subjacentes à
riqueza em espécies exóticas estão relacionadas com o potencial de invasão e capacidade
competitiva das espécies exóticas em causa (Lonsdale 1999).
A elevada invasibilidade dos ecossistemas fluviais do segmento canalizado deve-se

sobretudo a dois factores causais interligados. Em primeiro lugar, as perturbações físicas
endógenas causadas pela canalização, e os efeitos indirectos da expansão da rede de valas
e canais de rega (efectuada durante a década de oitenta), contribuíram para a criação de
condições ambientais novas (e favoráveis), incluindo a redução da intensidade e frequência
das cheias e o enriquecimento em nutrientes e sedimentos. Em segundo lugar, o
estabelecimento e colonização por taxa exóticos está associado ao efeito sinérgico de uma
elevada pressão de propágulos aliada a uma maior possibilidade de dispersão das espécies
exóticas já presentes, nomeadamente as espécies Azolla filiculoides Lam., Myriophyllum
aquaticum (Vell.) Verdc., P. distichum e Eryngium pandanifolium Cham. e Schlecht. e
Tradescantia fluminensis Velloso, processo este bem documentado noutros sistemas pelo
menos para as duas primeiras espécies (Ashton e Mitchell 1990).
As espécies exóticas lóticas invadem frequentemente comunidades perturbadas, enquanto

que a invasão de comunidades seminaturais maduras necessitam de um forte potencial de
colonização, o qual está associado à possibilidade de propagação vegetativa, à eficiência e
variedade dos mecanismos de dispersão de propágulos e sementes e ao rápido
estabelecimento das plântulas após as cheias (Nilsson et al. 1993; Richardson et al. 2000).
Com efeito, os factores que suportam uma elevada riqueza total nos habitats ripários –
nomeadamente as características associadas à perturbação (natural ou exógena) – podem

aumentar a susceptibilidade de um ecossistema à invasão (Pyšek e Prach 1994).
Neste estudo, as espécies exóticas apenas atingem os 10% do total de espécies

inventariadas, enquanto que percentagens ainda mais baixas foram referenciadas noutros
sistemas fluviais portugueses (ver subcapítulo 13.4.2). Embora a proporção de taxa exóticos
seja diminuta, as espécies exóticas podem dominar as comunidades vegetais lóticas,
impondo a apreciação conjunta da riqueza e de estimativas de abundância, (e.g. cobertura
foliar ou biomassa) na avaliação da invasibilidade local de ecossistemas mediterrânicos. A
invasibilidade de ecossistemas ripícolas está na dependência de factores locais específicos e
de um grupo de factores interactuantes a considerar: (i) o tipo de troço fluvial e a sua
posição no sistema lótico, incluindo a largura do canal e o desenvolvimento lateral dos
taludes, a heterogeneidade habitacional, o substrato do leito e taludes, e o regime
hidrológico; (ii) os atributos biológicos das espécies existentes no local e na região; e (iii) e
a perturbação humana no rio e nos ecossistemas confinantes.
No estudo apresentado no capítulo 9, os transeptos localizados na massa de água revelaram

uma baixa riqueza, tanto no sistema canalizado como no “natural”. Noutros sistemas, os
decréscimos na riqueza florística e na diversidade ecológica após a canalização, foram
relacionados com a homogeneização dos habitats (Brookes 1995). No entanto, é usual
valores baixos de riqueza em espécies aquáticas estritas na maior parte dos sistemas
mediterrânicos ibéricos (Ferreira 1994; Ferreira et al. 1998; Ferreira e Moreira 1999),
quando comparada com sistemas fluviais europeus de clima húmido (Haslam 1987; Naiman
et al. 1993). O número de hidrófitos é baixo, sendo eles com frequência ubíquos e muito
tolerantes a condições abióticas adversas, e tendendo a desenvolver comunidades
monoespecíficas extensas, quando os requisitos mínimos específicos são alcançados. Sem
dúvida, que este padrão pode ser uma consequência das constantes perturbações humanas
nos ecossistemas lóticos.
O primeiro terraço do segmento canalizado situa-se a cerca de 2 metros acima do nível da

água e é provável que a conectividade vertical persista entre a superfície e a zona hiporreica.
Contudo, o segundo terraço, posicionado a 4 metros acima da água, perdeu forçosamente a
ligação ao meio aquático e converteu-se num habitat terrestre efectivo, após a alteração da
tipologia transversal do sistema. Os resultados obtidos suportam a evidência de uma
composição similar ao longo deste transepto. Esta relativa homogeneidade da flora local pode
ser relacionada com outros factores que não os intrínsecos do corredor fluvial, tais como a
dispersão de propágulos de espécies exóticas lenhosas de fácies terrestre, como acontece
com a mimosa e o ailanto. Nesta perspectiva, o primeiro terraço do segmento canalizado
corresponde ao verdadeiro écotono ripícola, possuindo um conjunto de características que

interagem de forma dinâmica ao longo de escalas espaço-temporais com os ecossistemas
adjacentes – o meio aquático e o talude exterior (Décamps e Naiman 1990).
A notável altura acima da água (4-6 metros) dos taludes marginais resultaram do
aprofundamento do leito original, e deste modo o talude exterior contacta directamente com
o ambiente terrestre envolvente. Todavia, o canal fluvial votado ao isolamento, perdeu
grande parte das conexões laterais originais. A canalização alterou a contribuição relativa das
comunicações verticais, longitudinais e laterais do sistema lótico (Ward e Stanford, 1995),
com consequente perda da interpenetração florística transversal, geralmente presente com
maior ou menor expressão nas comunidades seminaturais.
O período de doze anos após a canalização, permitiu uma recuperação natural da vegetação
lenhosa ripícola, acompanhada por uma elevada riqueza em espécies nativas no segmento
canalizado. As comunidades ripícolas lenhosas apresentaram uma estrutura e composição
semelhante às galerias ribeirinhas do referencial de comparação no Rio Mondego, que poderá
ser um bom indicador da resiliência das espécies nativas ripícolas e do seu potencial como
pioneiras em sistemas mediterrânicos (Gasith e Resh 1999). Após a estabilização do substrato
e do declive dos novos taludes, prosperaram sob o coberto lenhoso, um grande número de
espécies umbrófilas, de fanerófitos escandentes e higrófitos. Igualmente a canalização
favoreceu o estabelecimento de salgueiros junto à água, o que poderá estar relacionado com
a diminuição dos caudais extremos, que usualmente arrastavam os propágulos e as plântulas
recém-formadas.
A espécie Eryngium pandanifolium, designada localmente por piteirão, dominava uma grande
parte dos transeptos do canal e do 1º terraço. Esta espécie, originária da América do Sul, foi
inventariada pela primeira vez no início do século XX em canais de drenagem do Baixo
Mondego, perto da Figueira da Foz (Garcia 1947), enquanto que o primeiro exemplar
herborizado data de 1942, e foi colhido nos arrozais da planície aluvionar. Esta espécie só
ocorre em Portugal no Baixo Mondego (Franco 1971), e não se conhecem ocorrências noutras
regiões da Europa (Tutin et al. 1968). Antes do projecto de regularização e canalização do Rio
Mondego esta espécie manteve níveis populacionais relativamente baixos, expandindo-se
rapidamente no curso principal, tributários, valas de terra e canais de irrigação, após as obras
de canalização e de melhoramento da rede de irrigação, a partir do foco inicial perto da foz
do rio. Embora não existam estudos prévios, há evidência de que o estabelecimento das
plântulas de Eryngium esteja relacionado com o aumento de sedimentos ricos em nutrientes,
que ocorre com frequência em determinadas fases do processo de regularização (Petts 2000).
No entanto, o coberto vegetal do troço superior do segmento canalizado – sobretudo na área

frequentemente inundada – era dominado por outra espécie exótica, P. distichum, cuja
introdução terá sido realizada a montante ou a partir dos campos agrícolas adjacentes. Esta
espécie inclui-se no grupo das mais ubíquas e invasoras dos canais e corredores fluviais
portugueses (Ferreira e Moreira 1995; 1999); é um C-estrategista sensu Grime (1979), ou seja
uma espécie competidora eficaz, com um extenso sistema rizomatoso e elevadas taxas de
crescimento, produz volumes elevados de biomassa e é capaz de colonizar rapidamente quer
os habitats ripícolas húmidos, quer o meio aquático (Wade 1990). Poderá vir a competir
directamente com o Eryngium por espaço e nutrientes.
Na explicação dos padrões de distribuição e de composição florística concorreram as

pequenas diferenças na composição do substrato do leito e margens, e o posicionamento
longitudinal e transversal dos transeptos amostrados. A maior variação na composição das
comunidades vegetais foi encontrada ao longo do gradiente de humectação do sistema, tanto
no troço canalizado, como no “natural”, com uma zonação transversal clara (fácies aquática-
fácies terrestre).
Apesar do ressecionamento do perfil do rio, do aprofundamento e linearização, da

homogeneização do material do leito e margens, com a inevitável perda de heterogeneidade
habitacional, evidenciaram-se consideráveis diferenças na flora amostrada ao longo do
segmento canalizado. Este facto sugere que as pequenas variações microtopográficas e no
mosaico estrutural de habitats (como a interposição de travessões nos segmentos superiores)
influenciaram o estabelecimento da vegetação (Everson e Baucher 1998; Changxing et al.
1999), conjuntamente com a capacidade de recuperação do ecossistema e com a capacidade
colonizadora e invasora de taxa nativos e exóticos.
13.4.2 Padrões de riqueza e cobertura de espécies exóticas em sistemas fluviais de

Os corredores fluviais, pelo seu dinamismo espacial e temporal, constituem casos de estudo
atractivos no âmbito das invasões ecológicas. Apesar de existir um considerável volume de
informação, poucos trabalhos tentaram relacionar a conectividade lateral, a invasibilidade
dos ecossistemas e a diversidade ao longo do gradiente longitudinal e transversal dos
sistemas lóticos (Hood e Naiman 2000; Mouw e Alaback 2003). Alguns estudos sugerem que
as zonas ripícolas em sistemas de características semi-áridas suportam uma maior
diversidade florística e são mais susceptíveis à invasão que as áreas adjacentes (Planty-
Tabacchi et al. 1996; Stohlgren et al. 1998; Hood e Naiman 2000).
Na sequência dos resultados do estudo de invasibilidade dos sistemas lóticos do Baixo

Mondego (capítulo 9), o trabalho apresentado no capítulo 10, desenvolvido num sistema
fluvial de características semi-áridas (Rio Guadiana) pretende esclarecer os padrões
espaciais das comunidades vegetais e a capacidade de previsão implícita nas relações
riqueza específica-habitat. Por outro lado, relaciona-se a cobertura exótica com as
potenciais variáveis ambientais que determinam a invasibilidade dos ecossistemas, e
exploram-se os atributos biológicos determinantes no sucesso das invasões.
No elenco florístico global, as espécies exóticas representaram apenas 9% das espécies

inventariadas. Resultados análogos foram relatados noutros corredores fluviais portugueses
em estudos que utilizaram metodologias de amostragem semelhantes: nomeadamente cerca
de 10% no vale de cheia do Rio Mondego (Aguiar et al. 2001- capítulo 9), 14% para a flora
lenhosa ripícola (e 8% para a flora total) da parte portuguesa da bacia do Tejo (Aguiar et al.
2000 - capítulo 3). Taxas de riqueza ainda menores foram anotadas em bacias de
características mediterrânicas mais acentuadas, como na parte portuguesa do Guadiana
(4%) (Ferreira et al. 2002 - capítulo 8) e noutras bacias no Sudoeste da Península Ibérica: 4%
para a bacia do Sado (Aguiar et al. dados não publicados) e cerca de 6% para bacias do
Algarve (Ferreira et al. dados não publicados). Valores da mesma ordem de grandeza (entre
6 e 8%) foram encontrados em rios do Sudeste da Espanha (Tabacchi et al. 1996), onde os
autores sugeriram que as espécies nativas de características competitivas são favorecidas
pela singularidade do regime hidrológico mediterrânico e por um baixo nível de perturbação
humana recente, quando comparados com outros sistemas de características climáticas,
geomórficas e históricas distintas, como nos rios semi-áridos da África do Sul, das regiões
Oeste da América do Norte, e Nordeste dos Estados Unidos e no Sudoeste da França, onde a
proporção de taxa exóticos variou entre 20% e 30% (de Ferrari e Naiman 1994; Planty-
Tabacchi et al. 1996; Hood e Naiman 2000). Porém, é comum na parte portuguesa da Bacia
do Guadiana, um grau elevado de perturbação antrópica (Ferreira et al. 2002 - capítulo 8),
assim o tipo frequência e magnitude da perturbação podem ser factores-chave no estudo de
padrões de riqueza em espécies exóticas, aliados ao potencial de resistência do
ecossistema, ao potencial de invasão das espécies em causa e às consequências ecológicas
decorrentes da acentuada estiagem.
O estudo fundamentado dos padrões de invasibilidade em rios ibéricos deverá resultar da

análise conjunta da presença e da abundância de espécies exóticas, como evidenciado no
subcapítulo anterior. De facto, as infestações por espécies exóticas em corredores fluviais
mediterrânicos são causadas, de um modo geral, por uma espécie ou por um pequeno
número de espécies muito competitivas, e é comum não se detectarem relações positivas
entre a riqueza e a cobertura em espécies exóticas. No capítulo 10, e em oposição aos

padrões de riqueza e cobertura em espécies nativas, verificou-se que a cobertura em
exóticas diminui significativamente desde a massa de água até ao leito de cheia, com uma
cobertura elevada na zona pelágica e área marginal geralmente inundada. Resultados
semelhantes na vegetação de prados húmidos foram atribuídos aos diferentes historiais
evolutivos das espécies nativas e exóticas, derivados de pré-adaptações ao uso agrícola e
pascícola dos solos (McIntyre e Lavorel 1994).
Para além de uma maior riqueza exótica na zona ripícola que no canal, verificou-se um
aumento na riqueza nativa (e total) com a distância à água. Em contrapartida, os trabalhos
de Stohlgren et al. (1998) e Planty-Tabacchi et al. (1996) relatam uma maior riqueza no
corredor fluvial que no vale de cheia. As razões prováveis são a existência de uma maior
conectividade entre a área inundada e a área afectada pelas cheias sazonais, do que entre
esta e as zonas de cota superior no vale de cheia. Com efeito, foi detectada uma maior
similaridade na composição florística nas áreas mais húmidas, com a ocorrência
generalizada de comunidades de higrófitos e nitrófilos. Acresce que, a maior parte das
espécies terrestres ruderais, das plantas mediterrânicas herbáceas e das espécies arbustivas
esclerofílicas só foram inventariadas no leito de cheia, e que entre elas, cerca de 85% são
consideradas pouco frequentes ou mesmo de ocorrência única. Por outro lado, e na
sequência dos trabalhos de Mouw e Alaback (2003), a existência de conectividade entre o
écotono ripícola e o meio terrestre, permite que as zonas ripícolas possuam um maior
número de espécies generalistas do que especialistas, o que explica um certo grau de
sobreposição entre as espécies dos corredores ripícolas e as do leito de cheia. Outras razões
para a maior contribuição do vale de cheia para a flora total, incluem a variedade de uso do
solo nos ecossistemas adjacentes (pastagens, olivais, montados, culturas de regadio), e a
intrincada geomorfologia existente ao longo do sistema, que permite uma maior diversidade
potencial de espécies.
Os resultados apresentados no capítulo 10 suportaram a hipótese de que os locais com

maior riqueza florística são preferencialmente invadidas por espécies exóticas, em oposição
à proposição de Elton (1958), que sugeria que as relações negativas, como a ocupação de
nichos disponíveis e a exclusão competitiva eram factores inquestionáveis na explicação de
invasões por espécies exóticas. Este paradigma ecológico, bem como as evidências
experimentais foram revistas por Prieur-Richard e Lavorel (2000) e recentemente
discutidas em Stohlgren et al. (2003). Além de que a maior correlação foi encontrada na
zona afectada pelas cheias sazonais, correspondente ao écotono lótico, uma entidade
intermédia funcional, que pactua com o ecossistema aquático e as áreas terrestres
adjacentes. As áreas de transição que mantêm efeitos de fronteira, são extremamente

dinâmicas no espaço e no tempo e são cruciais para o intercâmbio de espécies.
A maior parte das plantas exóticas encontradas nos corredores fluviais são terrestres, mas
as invasoras são aquáticas ou ripícolas – e.g. jacinto-aquático [Eichhornia crassipes (Mart.)
Solms-Laub.], pinheirinha-de-água (Myriophyllum aquaticum), azola [Azolla filiculoides],
piteirão (Eryngium pandanifolium), cana (Arundo donax), graminhão (Paspalum distichum)
(Ferreira e Moreira 1995, Aguiar et al. 2001). Nesta área de estudo (bacia do Guadiana) a
actuação humana persistiu ao longo do tempo e envolve o uso agrícola e pascícola dos vales
de cheia e a extracção mineral. É possível, que nestas situações, plantas nativas nitrófilas
como a grama [Cynodon dactylon (L.) Pers.] se convertam em infestantes ambientais de
sucesso em habitats ribeirinhos. Estas espécies podem reduzir ou mesmo substituir as
exóticas higrófilas, sobretudo quando a disponibilidade de água é um factor limitativo.
De facto, os atributos biológicos das espécies nativas e exóticas dominantes são

semelhantes, o que é particularmente óbvio quando nos reportamos às áreas
frequentemente inundadas (ver 10.5.2). No sucesso das infestações por espécies exóticas e
nativas em habitats húmidos parecem ser determinantes: a forma de vida perene, a
propagação vegetativa por estolhos e rizomas e dispor de várias estratégias de dispersão.
Também se evidencia um elevada pressão de propágulos da espécie exótica dominante, P.
distichum, e as diferenças entre a dominância entre nativas e exóticas à escala do troço
fluvial, são provavelmente uma consequência das características do habitat ripário, em vez
de diversos tipos de perturbação ou de aspectos como barreiras à dispersão.
Um resultado inesperado da análise da capacidade de predição dos ecossistemas, foi a

correlação não significativa entre a riqueza e cobertura em espécies exóticas e o gradiente
de perturbação. As razões podem ser em parte metodológicas, uma vez que se utilizou uma
variável antrópica aditiva, baseada em estimativas qualitativas, e portanto, sujeita a
incorrecções propagadas pelos vários componentes parciais. Por outro lado, podem ter sido
subestimadas algumas alterações a nível local ou mesmo do habitat.
Apesar das diferenças substanciais no uso do solo e a geomorfologia, estas variáveis não
foram factores significativos para a determinação da riqueza e cobertura em exóticas. As
propriedades do micro-habitat e as interacções bióticas foram as variáveis mais efectivas na
previsão da riqueza e cobertura da flora exótica. O enriquecimento em elementos finos do
substrato – e.g. limo, argila e partículas finas de matéria orgânica – favorecem o
estabelecimento de plantas exóticas, que geralmente estão adaptadas a um elevado nível
de nutrientes (McIntyre e Lavorel 1994). A título de exemplo, refira-se que na Juromenha é
clara a relação entre o aumento de elementos finos no substrato e a cobertura em exóticas,

enquanto que no Pulo do Lobo e na Estrela, com uma proporção elevada de cascalho e
afloramentos rochosos, respectivamente, a colonização foi preferencialemnte realizada por
uma comunidade nativa de espécies herbáceas e arbustivas adaptadas a condições adversas.
13.4.3 Padrões de invasibilidade de ecossistemas mediterrânicos pelo graminhão (Paspalum

distichum L.)
Os padrões de distribuição da espécie exótica Paspalum distichum, apresentados no

capítulo 11, foram estudados em bacias do Centro e Sul de Portugal. Os principais
objectivos deste estudo foram a avaliação do grau de invasibilidade dos habitats ripários
por esta espécie e a determinação da influência dos factores ambientais e da perturbação
humana nos sistemas fluviais e ecossistemas adjacentes.
A informação existente sobre a área de distribuição e expansão do graminhão, e em geral

das espécies exóticas ripícolas, é escassa e fragmentária. Na Bacia Mediterrânica, a espécie
P. distichum, originária da América tropical, foi mencionada pela primeira vez em França,
onde foi deliberadamente introduzida como cultura na região de Bordeaux em 1802 (Le
Floc’h 1991). Em Portugal, foi encontrado pela primeira vez em 1887 nas margens do Tejo
(Silva 1940), e descrito numa das primeiras Floras Portuguesas (“Flora Portuguesa” de G.
Sampaio, 1909).
Uma primeira abordagem espacial ecográfica foi apresentada por Franco e Rocha-Afonso
(1980), e permitiu reconhecer que, até aos anos setenta, esta espécie estava confinada às
áreas costeiras e vales aluvionares das principais bacias de Portugal, infestando arrozais e
culturas e regadio. Recentemente, as infestações em arrozais parecem ter sofrido um
decréscimo (Vasconcelos et al. 1999), enquanto que a invasão de zonas não agricultadas,
tem sido relatada em várias habitats dulçaquícolas, como lagoas (Costa et al. 1999),
corredores fluviais canalizados (Aguiar et al. 2001- capítulo 9), margens muito alteradas de
sistemas lóticos (Ferreira e Moreira 1995), e comunidades seminaturais de habitas ripários
(Franco e Rocha Afonso 1998; Ferreira et al. 2001, 2002 - capítulos 7 e 8), onde esta
espécie forma frequentemente densos tapetes em zonas com elevado teor de humidade
edáfica, penetrando mesmo em áreas inundadas de pouca profundidade.
O graminhão e a cana foram as espécies exóticas de maior frequência nas bacias estudadas
(Tejo, Oeste, Sado, Guadiana, Mira e Algarve) e demonstraram, de um modo geral, uma
maior invasibilidade quando comparadas com outras espécies exóticas presentes na área.
Apenas na bacia das Ribeiras do Oeste, a cana evidencia um carácter invasor mais
consistente que o graminhão. De notar, que apesar desta espécie ser reconhecida como
uma das invasoras com maior sucesso nos habitats aquáticos e ribeirinhos de Portugal
(Aguiar et al. 2001; Ferreira e Moreira, 1995; Ferreira et al. 2002), algumas espécies como
o jacinto-aquático, a pinheirinha-de-água, a azola, o piteirão e também a cana podem ter
efeitos mais adversos, apesar da menor amplitude de distribuição (Ferreira e Moreira 1995;
Aguiar et al. 1996; Aguiar et al. 2001). As bacias das Ribeiras do Oeste são caracterizadas
por numerosos aglomerados populacionais e por um complexo mosaico de pequenas
parcelas ocupadas por diversas culturas hortícolas, vinhas, pomares e culturas arvenses. É
comum nesta região a plantação da cana para protecção contra a erosão hídrica, para
delimitar propriedades, como cortinas de abrigo, ou ainda como tutor de culturas
trepadoras. No entanto, em corredores ripícolas muito intervencionados, esta espécie
desenvolve tipicamente densos cordões contínuos, competindo com as comunidades nativas
e aumentando o risco de cheias e a carga de sedimentos (Aguiar et al. 1996). Não foram
encontradas correlações significativas entre a cobertura de cana e graminhão, ou entre as
suas ocorrências. Todavia, estudos anteriores em habitats ripários evidenciaram potenciais
interacções competitivas do Paspalum e de outras espécies exóticas, como o piteirão
(Aguiar et al. 2001- capítulo 9) ou nativas, como a grama (capítulo 10).
A cobertura e frequência do graminhão foi associada significativamente à presença de

culturas de regadio e áreas industriais e urbanas, assim como a outras interferências
humanas tanto a nível do sistema fluvial, como no vale de cheia. A expansão da espécie
parece estar sobretudo relacionada com factores locais em vez de geográficos como a
altitude, o substrato geológico ou variáveis climáticas, o que revela uma considerável
plasticidade ecológica desta espécie.
No entanto, a distância à nascente e outras variáveis tipológicas hierárquicas foram

correlacionadas positivamente com os padrões de invasão do Paspalum. As cabeceiras de
rios com fluxo intermitente e declives acentuados foram menos invadidos, em contraste
com os rios de maior número de ordem e com regime hidrológico permanente. Em oposição,
as planícies aluvionares são com frequência dedicadas a culturas de regadio, como o milho,
o arroz e o tomate, bastante exigentes em nutrientes. A aplicação de fertilizantes em
excesso dá origem à lixiviação dos elementos mais móveis, como o azoto (e também o
potássio), que provocam poluição difusa e promovem a expansão do graminhão, uma
espécie nitrófila típica. De modo semelhante, zonas ricas em elementos finos no substrato,
foram relacionadas com maiores extensões de graminhão. Além disso, algumas práticas
agrícolas, como a fresagem, podem levar ao aumento do potencial invasor e à ampliação da
área infestada desta espécie por dispersão de fragmentos de rizomas e estolhos. A
presença de áreas urbanas e outras estruturas artificiais, como pontes, atravessamentos e

caminhos aumentam o número de visitantes na zona ribeirinha, promovendo também a
dispersão de propágulos por antropocoria e a modificação das características físicas do
ecossistema, entre outras.
A fragmentação da galeria ripícola pode ser um processo natural, como resultado dos
efeitos do regime torrencial, ou intencional para facilitar o acesso ao canal para a rega ou
para aumento da área de cultivo, entre outras razões. A falta de continuidade na galeria
ripícola é comum nos corredores fluviais portugueses (capítulo 4), e desempenha um papel
crucial na manutenção de comunidades herbáceas e hidrofíticas. No entanto, o equilíbrio
entre a vegetação arbórea e arbustiva e a herbácea pelo ensombramento, depende dos
padrões de fragmentação das galerias ripícolas e em particular, das extensões sem galeria.
A cobertura superficial do graminhão é favorecida pela exposição à luz, que estimula a
germinação de sementes (DiTomaso e Healy 2002) e promove o abrolhamento das gemas, o
enraizamento e o crescimento de rizomas, caules e segmentos apicais (Huang et al. 1987;
Liu et al. 1991).
Embora os resultados obtidos tenham aumentado o conhecimento acerca da área de

expansão desta espécie e respectiva causalidade abiótica, serão desejáveis estudos locais
para compreender a dinâmica populacional e obter um grau consistente de capacidade de
previsão do ecossistema. As características a nível do micro-habitat parecem ser
determinantes na explicação das diferenças de invasibilidade nos habitats ripários
portugueses, pelo que será necessária uma melhor quantificação das características do
ecossistema, como o teor em nutrientes e material orgânico, conjuntamente com estudos
de interacção competitiva com outras espécies nativas e exóticas.
13.4.4 Padrões do estabelecimento de espécies exóticas e endémicas em locais de

referência e não-referência
Não se conhecem estudos biogeográficos de grande escala em Portugal envolvendo a

invasão de plantas ripícolas e a susceptibilidade dos ecossistemas dulçaquícolas. Com base
nos estudos apresentados no capítulo 12 apresentam-se algumas considerações sobre esta
primeira abordagem ecográfica transbacia da flora exótica e endémica de sistemas fluviais
portugueses.
Com o objectivo principal de reconhecer as diferenças de invasibilidade dos ecossistemas

ripícolas, a vegetação exótica e nativa foi estudada em locais locais com perturbação
antropogénica mínima e em locais de não-referência, no Centro e Sul de Portugal, em duas

tipologias geológicas distintas, que se pensam ser importantes em sistemas ibéricos, na
sequência de trabalhos anteriores (capítulos 3 e 8): rios de substrato silicioso e calcário.
Embora os locais de referência tenham sido definidos após a inventariação fitoecológica,
não se recorreram a métodos de análise multivariada, mas à análise pericial da informação
ecológica existente e recolhida no campo. Este processo permitiu identificar algumas
lacunas na informação biológica, fisiográfica, fisico-química e geomórfica, cujo
conhecimento poderá ser colmatado noutros estudos, como a bioavaliação de sistemas
fluviais.
Foram identificadas 568 espécies, das quais 44 são exóticas e 28 endémicas. As espécies
exóticas estão presentes em 91% dos locais, enquanto que as endémicas ocorrem em cerca
de 45% do total de troços amostrados. Confirmou-se o potencial invasor das gramíneas
rizomatosas, como o graminhão e a cana, e dos hidrófitos jacinto-aquático e a azola. A
cobertura em espécies exóticas é significativamente mais elevada no grupo de locais
perturbados que nos locais de referência, em ambos os tipos geológicos, com o grupo de
substrato calcário a apresentar coberturas significativamente superiores aos rios de
substrato silicioso. No grupo do tipo silicioso, a riqueza e a cobertura em espécies
endémicas é significativamente menor nos locais perturbados que nos de referência,
enquanto que no tipo calcário, não foram detectadas diferenças significativas. Também não
foram detectadas diferenças significativas entre a riqueza e cobertura de espécies
endémicas entre os dois ecótipos, considerando respectivamente os locais de referência e
não-referência.
A riqueza e cobertura em exóticas estão correlacionadas positivamente com as

perturbações antrópicas no sistema fluvial, como por exemplo o uso do solo agrícola e
urbano no vale de cheia e o aumento da concentração em nutrientes. Por sua vez, as
espécies endémicas respondem inversamente às variáveis antrópicas, com a riqueza e a
cobertura a aumentar com a diminuição das pressões humanas nos ecossistemas adjacentes
e no sistema fluvial. A capacidade de algumas espécies exóticas substituírem a flora nativa
e endémica é evidenciada neste trabalho, para além de outros efeitos na redução da
diversidade através da expansão monoespecífica em ecossistemas aquáticos e ribeirinhos.
Nos ecossistemas dulçaquícolas portugueses deverá ser realizado um maior esforço de

protecção dos (poucos) habitats preservados existentes, por redução das pressões directas e
indirectas no uso do solo e nos corredores fluviais, bem como na prevenção de invasões
futuras. As medidas de mitigação e projectos de restauração de corredores ripícolas
degradados devem considerar o cenário geomórfico e abiótico presente, assim como a
composição florística “natural” da região. Ademais, foram detectadas diferenças

significativas entre os dois ecótipos geológicos, e portanto deverão requerer opções de
gestão e de planeamento distintas.
Aguiar F., Moreira I. e Ferreira M.T. 1996. A percepção da vegetação aquática infestante pelas
entidades gestoras dos recursos hídricos. Revista de Ciências Agrárias 19: 35-56.
Aguiar F.C., Ferreira M.T., Moreira I. e Albuquerque A. 2000. Riparian types in a Mediterranean basin.
Aspects of Applied Biology 58: 221-232.
Aguiar F.C., Moreira I. e Ferreira M.T. 2001. Exotic and native vegetation establishment following
channelization of a western Iberian river. Regulated Rivers: Research and Management. 17:
509-526.
Alba-Torcedor J. e Pujante A.M. 2000. Running water biomonitoring in Spain: opportunities for a
predictive model. Em: Wright J.F., Sutcliffe D.W. e Furse M.T. (eds), Assessing Biological
Quality of Freshwaters. RIVPACS and other techniques. The Freshwater Biological
Association, Ambleside, Cumbria. pp. 207-217.
Allen A.P. e O’Connor R.J. 2000. Interactive effects of land-use and other factors on regional bird
distributions. Journal of Biogeography 21:889-900.
Anderson N.H. e Sedell J.R. 1979. Detritus processing by macroinvertebrates in stream ecosystems.
Annual Review of Entomology 24: 351-357.
Ashton P.J. e Mitchell D.S. 1990. Aquatic plants: patterns and modes of invasion, attributes of
invading species and assessment of control programmes. Em: Drake J.A., Mooney H.A., di
Castri F., Groves R.H., Kruger F.J., Rejmánek M. e Williamson M. (eds), Biological
Invasions: a Global Perspective. John Wiley & Sons Ltd, Chichester, England. pp. 111-147.
Barbour M.T., Gerritsen J., Snyder B.D. e Stribling J.B. (eds). 1999. Rapid bioassessment protocols for
use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and Fish. 2nd
Ed. EPA/841-B-98-010. - EPA Office of Water, Washington DC.
Bärlocher F. 1985. The role of fungi in the nutrition of stream invertebrates. Botanical Journal of the
Linnean Society 91: 83-94.
Benfield E.F. 1997. Comparison of litterfall input to streams. Journal of the North American
Bravard J.P., Amoros C. e Patou G. 1986. Impacts of civil engineering works on the succession of
communities in a fluvial system: a methodological predictive approach applied to a section
of the Upper Rhône River. Oikos 47: 92-111.
Brierley G.J., Cohen T., Fryirs K. e Brooks A. 1999. Post-European changes to the fluvial
geomorphology of Bega catchment, Australia: implications for river ecology. Freshwater
Biology 41: 839-848.
Brito A., Cortes R. e Ferreira M.T. 2002. Capítulo 5. Vertebrados tetrápodes dulçaquícolas. Em: I.
Moreira, M.T. Ferreira, R. Cortes, P. Pinto e P.R. Almeida (eds), Ecossistemas Aquáticos e
Ribeirinhos. Ecologia, Gestão e Conservação. Instituto da Água, MCOTA. pp. 5.1-5.15.
Brookes A. 1986. Response of aquatic vegetation to sedimentation downstream from channelization

works in England and Wales. Biological Conservation 38: 351-367.
Brookes A. 1995. The importance of high flows for riverine environments. In: Harper DM and Ferguson
AJD (eds), The Ecological Basis for River Management. John Wiley & Sons Ltd., Chichester,
England. pp. 33-49.
Brown A.G. 2002. Learning from the past : paleohydrology and paleoecology. Freshwater Biology 47:
817-829.
Bunn S.E. e Davies P.M. 2000. Biological processes in running waters and their implications for the
assessment of ecological integrity. Hydrobiology 422-423: 61-70.
Caffrey J. 1987. Macrophytes as indicators of organic pollution in Irish rivers. Em: Richardson D. (ed.),
Biological Indicators of Pollution, Royal Irish Academy, Dublin.
Catarino L., Moreira I., Ferreira T. e Duarte M.C. 2001. Plantas aquáticas infestantes de valas e
canais. ISAPress, Lisboa.
Changxing S., Petts G.E. e Gurnell A.M. 1999. Berm development along the regulated, lower Dee, UK.
Earth Surf. Processes Landf. 24: 135-149.
Coimbra C.N., Graça M.A.S. e Cortes R.M. 1996. The effects of a basic effluent on macroinvertebrate
community structure in a temporary Mediterranean river. Environmental Pollution 3: 301-
307.
Collier K.J. 1995. Environmental factors affecting the taxonomic composition of aquatic
macroinvertebrate communities in lowland waterways of Northland, New Zealand. New
Zealand Journal of Marine and Freshwater Research 4: 453-465.
Conners M.E. e Naiman R.J. 1984. Particulate allochtonous inputs: relationships with stream size in
undisturbed watersheds. Canadian Journal of Aquatic Sciences 41: 1437-1484.
Corbacho C., Sánchez J.M. e Costillo E. 2003. Patterns of structural complexity and human
Cortes R.M. 1992. Seasonal patterns of benthic communities along the longitudinal axis of river
systems and the influence of abiotic factors on the spatial structure of those communities.
Archiv für Hydrobiologie 126: 85-103.
Cortes R.M., Graça M., Vingada J.N. e Oliveira S.V. 1995. Stream typology and dynamics of leaf
processing. Annales de Limnologie 31: 119-131.
Cortes R.V., Ferreira M. T. e Godinho F.N. 1998. Benthic community organization in an intermittent
lotic system. Verhandlungen der Internationalen Vereinigung für theoretische und
Costa J.C., Capelo J., Espírito-Santo M.D., Lousã M., Monteiro A., Mesquita S., Vasconcelos M.T. e
Moreira I. 1999. Plant communities of the lagoons of the Portuguese Coastal Superdistrict –
a multivariate approach. Hydrobiologia 415: 67-75.
de Ferrari C.M. e Naiman R.J. 1994. A multiscale assessment of the occurrence of exotic plants on the
Olympic Peninsula, Washington. J. Veg. Sci. 5:247-258.
Décamps H. e Naiman R.J. 1990. Towards an ecotone perspective. Em: Naiman R.J. e Décamps H.
(eds), The ecology and management of aquatic-terrestrial ecotones. Man and the Biosphere
Series, 4, UNESCO, Paris. pp. 1-5.
Décamps H., Fortune M., Gazelle F. e Patou G. 1988. Historical influence of man on the riparian
Demars B.O.L. e Harper D.H. 1998. The aquatic macrophytes of an English lowland river system:
assessing response to nutrient enrichment. Hydrobiologia 384: 75-88.
climate. Em: Groves R.H. e di Castri F. (eds), Biogeography of Mediterranean Invasions.
Di Castri F., Hansen A.J. e Debussche M. (eds.) 1990. Biological invasions in Europe and in the
Mediterranean Basin. Kluwer Academic Publ., Dordrecht.
DiTomaso J.M. e Healy E.A. 2002. Aquatic and Riparian Weeds of the West. University of California
Agriculture and Natural Resources. Publication 3421. Oakland, US.
Duarte M.C., Moreira I. e Ferreira M.T. 2004. Flora de ecossistemas aquáticos e ribeirinhos
portugueses: delimitação taxonómica, tipológica e espacial. Recursos Hídricos 25: 67-94.
Duvel W., Volkmar R.D., Specht W.I. e Johnson F.W. 1976. Environmental impact of river
channelization. Water Research Bulletin 12: 799-812.
Elton C.S. 1958. The ecology of invasions by animals and plants. Methuen, London, England.
Everson D.A. e Boucher D.H. 1998. Tree species-richness and topographic complexity along the
riparian edge of the Potomac river. Forest Ecology and Management 109: 305-314.
Verhandlungen der Internationalen Vereinigung für theoretische und angewandte
Ferreira M.T. e Moreira I.S. 1995. The invasive component of river flora under the influence of
Mediterranean agricultural systems. Em: Pyšek P., Prach K., Rejmanek M. e Wade M.
(eds.), Plant Invasions - General Aspects and Special Problems, SPB Academic Publishing,
Ferreira M.T. e Moreira I.S. 1999. River plants from an Iberian basin and environmental factors
Ferreira M.T., Albuquerque A., Aguiar F.C. e Catarino L.F. 2001. Seasonal and yearly variations of
macrophytes in a Southern Iberian River. Verh. Internat. Verein. Limnol. 27: 3833-3837.
Ferreira M.T., Albuquerque A., Aguiar F.C. e Sidorkewicz N. 2002. Assessing reference sites and
approach. Arch. Hydrobiol. 155(1): 121-145.
Ferreira M.T., Godinho F.N. e Cortes R.M. 1998. Macrophytes in a southern iberian river. Vehr.
Ferreira M.T.M.C.C. 1992. Estrutura e dinâmica das comunidades de macrófitos lóticos da bacia
hidrográfica do Rio Sorraia. Interferências dos ecossistemas agrários envolventes.
Dissertação de Doutoramento em Engenharia Florestal. Instituto Superior de Agronomia,
Lisboa.
Flory E.A. e Milner A.M. 1999. Influence of riparian vegetation on invertebrate assemblages in a
recently formed stream in Glacier Bay National Park, Alaska. Journal of the North
Franco J. A. e Rocha-Afonso M.L.. 1998. Nova Flora de Portugal (Continente e Açores). 3. Gramineae.
Escolar Editora, Lisboa.
Ed. Autor, Lisboa.
Franco J.A. e Rocha-Afonso M.L. 1980. Distribuição em Portugal das principais infestantes. Em: Actas
do I Congresso de Fitiatria e Fitofarmacologia, Lisboa. pp.1-22.
Franco J.A. e Rocha-Afonso M.L. 2003. Nova Flora de Portugal (Continente e Açores). 3(3) Juncaceae-
Orchidaceae. Escolar Editora, Lisbon.
Friberg N. 1997. Benthic invertebrate communities in six Danish forest streams: Impact of forest type
on structure and function. Ecography 20: 19-28.
Gallego-Fernández J.B., García-Mora M.R. e García-Novo F. 1999. Small wetlands lost: a biological
199.
Garcia J. 1947. Algumas novidades para a flora ibérica. Ann. Jard. Bot. Madrid 6(2): 421-438.
Gasith A. e Resh V.H. 1999. Streams in Mediterranean climate regions: abiotic influences and biotic
51-81.
González J.M., Basaguren A. e Pozo J. 2003. Macroinvertebrate communities along a third-order

Iberian stream. Ann. Limnol.. –Int. J. Lim. 39(4). 287-296.
Graça M.A.S., Fonseca D.M. e Castro S.T. 1989. The distribution of macroinvertebrate communities in
two Portuguese rivers. Freshwater Biology 22: 297-308.
Grime J.P. 1979. Plant strategies and vegetation processes. John Wiley & Sons Ltd. Chichester, UK.
Hall S.J., Rafaelli D. e Thrush S.F. 1994. Patchiness and disturbance in shallow water benthic
assemblages. Em: Giller P.S., Hildrew A.G. e Rafaelli D.G. (eds), Aquatic ecology. Scale,
pattern and processes. Blackwell Science, London, UK. pp. 333-375.
Haslam S. 1987. River Plants of Western Europe. Cambridge University Press, Cambridge.
Hawkins C.P. e MacMahon J.A. 1989. Guilds: the multiple meanings of a concept. Annual Review of
Entomology 34: 423-451.
Hawkins C.P., Murphy M.L. e Anderson N.H. 1982. Effects of canopy, substrate composition and
gradient on the structure of macroinvertebrate communities in Cascade Range streams of
Oregon. Ecology 63: 1840-1856.
Holmes N.T.H., Boon P.J. e Rowell T.A. 1998. A revised classification system for British rivers based
on their aquatic plant communities. Aquatic Conservation: marine and freshwater
Hood G.W. e Naiman R.J. 2000. Vulnerability of riparian zones to invasion by exotic vascular plants.
Huang W.Z., Hsiao A.I. e Jordan L. 1987. Effects of temperature, light and certain growth regulation
substances on sprouting, rooting and growth of single-node rhizome and shoot segments of
Paspalum distichum L. Weed Res. 27: 57-67.
Humphrey C.L. e Storey A.W. 2000. AUSRIVAS: operator sampling processing errors and temporal
variability - implications for model sensibility. Em: Wright J.F., Sutcliffe D.W. e Furse M.T.
Hupp C.R. 1992. Riparian vegetation recovery patterns following stream channelization: a geomorphic
perspective. Ecology 73(4): 1209-1226.
Inoue M. e Nakagoshi M. 2001. The effects of human impact on spatial structure of the riparian
vegetation along the Ashida river, Japan. Landscape and Urban Planning 53: 111-121.
Jones III E.B.D., Helfman G.S., Harper J.O. e Bolstad P.V. 1999. Effects of riparian forest removal on
fish assemblages in southern Appalachian streams. Conservation Biology 13(6): 1454-1465.
Regulated Rivers: Research and Management 7: 279-289.
Le Floc’h, E. 1991. Invasive plants of the Mediterranean Basin. Em: Groves R.H. e di Castri F. (eds.),
Biogeography of Mediterranean invasions. Cambridge University Press, Cambridge. pp. 67-
80.
Lepart J. e Debussche M. 1992. Human impact on landscape patterning. Em: Hansen A.J. e di Castri F.
(eds.), Landscape Boundaries. Consequences for Biotic Diversity and Ecological Flows,
Ecological Studies 92,. Springer-Verlag, New-York, USA. pp. 76-106.
Liu S.H., Hsiao A.I. e Quick W.A. 1999. The influence of leaf blade, nutrients, water and light on the
promotion of axillary bud growth of isolated single-node stem segments of Paspalum
distichum L. Weed Res. 31: 385-394.
1522-1536.
Maamri A., Chergi H. e Pattee E. 1994. Allochthonous input of coarse particulate organic matter to a
Moroccan mountain stream. Acta Oecologica 15: 495-508.
MacCarthy D.T. 1985. The adverse effects of channelization and their amelioration. Em: Alabaster
J.S. (ed), Habitat Modification and Freshwater Fisheries. Butterworth, London. pp. 83-97.
Malanson G.P. e Cramer B.E. 1999. Landscape heterogeneity, connectivity, and critical landscapes for
conservation. Diversity and Distributions 5: 27-39.
McIntyre S. e Hobbs R.A. 1999. A framework for conceptualising human effects on landscapes and its
relevance to management and research models. Conservation Biology 13(6): 1282-1292.
McIntyre S. e Lavorel S. 1994. Predicting richness of native, rare and exotic plants in response to
521-531.
Mihuc T.B. 1997. The functional trophic role of lotic primary consumers: generalist versus specialist
strategies. Freshwater Biology 37: 455-462.
Moreira I., Ferreira M.T., Aguiar F. e Duarte M.C. 2002. Capítulo 4. Plantas infestantes e invasoras de
ecossistemas dulçaquícolas. Em: I. Moreira, M.T. Ferreira, R. Cortes, P. Pinto e P.R.
Almeida (eds), Ecossistemas Aquáticos e Ribeirinhos. Ecologia, Gestão e Conservação.
Instituto da Água, MCOTA. pp. 4.1-4.17.
Mouw J.E.B. e Alaback P.B. 2003. Putting floodplain hyperdiversity in a regional context: an
Biogeography 30: 87-103
Murphy K.J., Fox A.M. e Hanbury R.G. 1987. A multivariate assessment of plant management impacts
on macrophyte communities in a Scottish canal. Journal Applied Ecology 24: 1063-1079.
Myklestad A. e Birks H.J.B. 1993. A numerical analysis of the distribution patterns of Salix L. species
in Europe. Journal of Biogeography 20:1-32.
Naiman R.J. e Décamps H. 1997. The ecology of interfaces: riparian zones. Annual Review of Ecology
and Systematics 28: 621-658.
Naiman R.J., Décamps H. e Pollock M. 1993. The role of riparian corridors in maintaining regional
biodiversity. Ecol. Appl. 3: 209-212.
Newbold C. e Holmes N.T.H. 1987. Nature conservation: water quality criteria and plants as water
quality monitors. Water Pollution Control 86: 345-364.
Nilsson C. e Jansson R. 1995. Floristic differences between riparian corridors of regulated and free-
flowing boreal rivers. Regulated Rivers: Research & Management 11: 55-66.
Nilsson C., Nilsson E., Johansson M.E., Dynesius M., Grelsson G., Xiong S., Jansson R. e Danvind M.
419-431.
Ormerod S.J. e Edwards R.W. 1987. The ordination and classification of macroinvertebrate
assemblages in the catchment of the River Wye in relation to environmental factors.
ownes B.J., Lake P.S. e Schreiber E.S.G. 1993. Spatial variation in the distribution of stream
invertebrates: implications of patchiness for models of community organisation.
Petts G.E. 2000. A perspective on the abiotic processes sustaining the ecological integrity of running
waters. Hydrobiologia 422/423: 15-27.
Pinto P. e Morais M. 1998. Colonisation curves of macroinvertebrate fauna in a temporary river (river
Degebe, Guadiana basin, Portugal). Verhandlungen der Internationalen Vereinigung für
theoretische und angewandte Limnologie 26: 113-115.
Planty-Tabacchi A.-M., Tabacchi E., Naiman R.J., de Ferrari C. e Décamps H. 1996. Invasibility of
Poff N.L., Allan J.D., Bain M.B., Karr J.R., Prestegaard K.L., Richter B.D., Sparks R.E., e Stromberg
J.C. 1997. The natural flow regime. A paradigm for river conservation and restoration.
Prieur-Richard A.-H. eLavorel S. 2000. Invasions: the perspective of diverse plant communities.
Prochaza K., Stewart B.A. e Davies B.R. 1991. Leaf litter retention and its implications for shredder
distribution in two headwater streams. Archiv für Hydrobiologie 120: 315-325.
Puig M. A., Aboal M. e Soston A. 1991. New approaches to Mediterranean fluvial communities.
Oecologia Aquatica 21: 13-20.
Pyšek P. e Prach K. 1994. How important are rivers for supporting plant invasions? Em: de Waal L.,
Child L., Wade P.M. e Brock J. (eds), Ecology and Management of Invasive Riverside Plants.
John Wiley & Sons Ltd., Chichester. pp. 19-26.
Rao K.S. e Pant R. 2001. Land use dynamics and landscape change pattern in a typical
Ecosystems and Environment 86: 113-123.
Richardson D.M., Pyšek P., Rejmanek M., Barbour M.G., Panetta F.D. e West C.J. 2000. Naturalization
and invasion of alien plants: concepts and definitions. Diversity and Distributions 6:93-107.
Rocha F. 1996. Nomes vulgares de plantas existentes em Portugal. Protecção da Produção Agrícola,
ed. Especial. Direcção Geral de Protecção das Culturas, Lisboa.
Roth N.E., Allan J.D. e Eriksson D.L. 1996. Landscape influences on stream biotic integrity assessed at
multiple spatial scales. Landscape Ecology 11(3): 141-156.
Rundle S.D., Lloyd E.C. e Ormerod S.J. 1992. The effects of riparian management and physico-
chemistry on macroinvertebrate feeding guilds and community structure in upland British
streams. Aquatic Conservation: Marine and Freshwater Ecosystems 2: 309-324.
Salinas M.J., Blanca G. e Romero T. 2000. Evaluating riparian vegetation in semi-arid Mediterranean
35.
Sampaio G. 1909. Flora Portuguesa. Imprensa Moderna, Porto.
Seddon B. 1972. Aquatic macrophytes as limnological indicators. Freshwater Biology 2: 107-130.
Shepard R.B. e Minshall G.W. 1981. Nutritional value of lotic species compared with allochtonous
materials. Archiv für Hydrobiologie 90: 467-488.
Shields Jr. F.D.S. e Hoover J.J. 1991. Effects of channel restabilization on habitat diversity,
Twentymile Creek, Mississippi. Regulated Rivers: Research and Management. 6:163-181.
Silva A.R.P. 1940. O género Paspalum em Portugal. Agronomia Lusitana 2: 5-23.
Snyder C.D., Young J.A., Villella R. e Lemarié D.P. 2003. Influences of upland and riparian land use
patterns on stream biotic integrity. Landscape Ecology 18: 647-664.
Spackman S.C. e Hughes J.W. 1995. Assessment of minimum stream corridor width for biological
conservation: species richness and distribution along mid-order streams in Vermont, USA.
Biological Conservation 71: 325-332.
Stohlgren T.J., Barnett D.T. e Kartesz J.T. 2003. The rich get richer: patterns of plant invasions in
the United States. Front. Ecol. Environ. 1(1) : 11-14.
Stohlgren T.J., Bull K.A., Otsuki Y., Villa C.A. e Lee M. 1998. Riparian zones as havens for exotic plant
species in the central grasslands. Plant Ecology 138: 113-125.
Stromberg J.C. e Patten D.T. 1990. Riparian vegetation instream flow requirements: a case study
from a diverted stream in the eastern Sierra Nevada, California. Environmental
Swales, S. 1988. Fish populations of a small lowland channelized river in England subject to long-term
river maintenance and management work. Regulated Rivers: Research and Management. 2:
403-506.
Tabacchi E. e Planty-Tabacchi A. 1990. Evolution longitudinale de la vegetation du corridor de

l’Adour. Botanica Pirenaico-Cantabrica 5: 455-468.
Tabacchi E., Planty-Tabacchi A. e Décamps O. 1990. Continuity and discountinuity of the riparian
vegetation along a fluvial corridor. Landscape Ecology 5 (1): 9-20.
Tabacchi E., Planty-Tabacchi A.-M., Salinas M.J. e Décamps H. 1996. Landscape structure and
diversity in riparian plant communities: a longitudinal comparative study. Regulated
Rivers: Research & Management 12: 367-390.
Tait C.K., Li J.L., Lamberti G.A., Pearsons T.N. e Li H.W. 1994. Relationships between riparian cover
and the community structure of high desert streams. Journal of the North American
Tockner K. e Stanford J.A. 2002. Riverine flood plains: present state and future trends. Environmental
Conservation 29(3): 308-330.
Townsend C.R., Hildrew A.G. e Francis J. 1983. Community structure in some southern streams: the
influence of physico-chemical factors. Freshwater Biology 13: 521-544.
Tutin T.G, Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. e Webb D.A. (eds).
1968. Flora Europaea. 2. Rosaceae-Umbelliferae Cambridge University Press, Cambridge.
Tutin T.G., Heywood V.H., Burges N.A., Moore D.M., Valentine D.H., Walters S.M. e Webb D.A. (eds.)
2001. Flora Europaea. 5. Alismataceae-Orchidaceae (Monocotyledoneae). Cambridge
Vasconcelos T., Tavares M. e Gaspar N. 1999. Aquatic plants in the rice fields of the Tagus Valley,
Portugal. Hydrobiologia 415: 59-65.
Vidal-Abarca M.R., Suarez M.L. e Ramirez-Diaz L. 1992. Ecology of Spanish semi-arid streams.
Limnetica 8: 151-160.
Wade P.M. 1990. General biology and ecology of aquatic weeds. Em: Pieterse A.H. e Murphy K.J. (ed),
Aquatic Weeds. The Ecology and Management of Nuisance Aquatic Vegetation, Oxford
University Press, Oxford. pp. 17-30.
Wade P.M. e Edwards R.W. 1980. The effect of channel maintenance on the aquatic macrophytes of
the drainage channels in the Monmouthshire Levels, South Wales, 1870-1976. Aquatic
Botany 8: 307-322.
Ward J.V. e Stanford J.A. 1995. The serial discontinuity concept: extending the model to floodplain
rivers. Regulated Rivers: Research and Management. 10: 159-168.
Whitton, B.A. 1979. Plants as indicators of river water quality. Em: James A. e Evison L. (eds.)
Biological Indicators of Water Quality. John Wiley & Sons, Chichester. pp. 5.1-5.34.
Hydrobiologia 79: 91-100.
Wissmar R.C. e Swanson F.J. 1990. Landscape disturbances and lotic ecotones. Em: Naiman R.J. e
Décamps H. (eds), The Ecology and Management of Aquatic-terrestrial Ecotones. Man and
Wright J. 1995. Development and use of a system for predicting the macroinvertebrate fauna in
flowing waters. Australian Journal of Ecology 20: 181-197.
Wright J.F. 2000. An introduction to RIVPACS. Em: Wright J.F., Sutcliffe D.W. e Furse M.T. (eds),
Capítulo 14
Considerações Finais
Chapter 14 Final remarks

Capítulo 14 – Considerações Finais | 333
14.1 Nota final
Os resultados e conclusões dos estudos desenvolvidos (Partes II a IV) poderão ter aplicações
práticas, para além da esfera da investigação, designadamente na recuperação de galerias
ribeirinhas, na monitorização ecológica e na bioavaliação do estado dos corredores lóticos,
e servir de apoio à concepção de modelos sobre a invasão de ecossistemas dulçaquícolas de
carácter mediterrânico. Estes trabalhos permitiram igualmente:
i. colmatar lacunas de conhecimento das inter-relações entre os componentes

florísticos e os abióticos e antrópicos dos corredores fluviais de características
mediterrânicas;
ii. aumentar o conhecimento das respostas da vegetação a alterações naturais e

artificiais do regime hidrológico;
iii. contribuir para a caracterização das galerias ripícolas, nomeadamente nos padrões
espaço-temporais de composição, riqueza florística e de integridade ecológica;
iv. reconhecer o papel das galerias ripícolas mediterrânicas na integridade ecológica e

biótica fluvial;
v. explicar padrões de invasibilidade com base nas características habitacionais,

geomorfogénicas e hidrológicas, e explorar o potencial de resistência à perturbação
dos ecossistemas ripícolas mediterrânicos e os factores determinantes no proce4sso
de invasão por espécies exóticas;
vi. contribuir para a proposta de métricas a utilizar na bioavaliação da qualidade

ecológica em sistemas fluviais portugueses.
Foram utilizadas metodologias de amostragem, que poderão ser modelares em estudos com
objectivos similares ou noutras aplicações. Foram particularmente bem sucedidas:
i. a inventariação fito-ecológica de troços fluviais para estudos à escala da bacia e

transbacia (capítulos 3, 4, 6, 7, 8, 11 e 12);
ii. a recolha de dados por interpretação de fotografia aérea, com utilização de

unidades de amostragem sectoriais e transeptos, e complementadas com trabalho
de campo (capítulos 4 e 5);
iii. o estudo dos padrões de invasibilidade de ecossistemas aquáticos e ripícolas,

utilizando amostragens imbricadas, com o posicionamento de transeptos, blocos e
inventários contíguos ao longo do perfil tipológico transversal e longitudinal sistema
(capítulos 9 e 10).
Também no tratamento de dados, sublinham-se alguns dos aspectos inovadores em análise

ecológica de macrófitos fluviais, como:
i. a determinação das espécies indicadoras (Dufrêne e Legendre 1997) de grupos

tipológicos decorrentes da análise classificativa hierárquica;
ii. a utilização da análise parcial de correspondências canónicas (ter Braak 1990;

Borcard et al. 1992, Rodriguez e Magnan 1995) de modo a conhecer a contribuição
dos vários componentes (bióticos, abióticos, antrópicos) na explicação da
variabilidade biológica;
iii. a utilização do método multivariado na selecção de locais de referência e na

bioavaliação ecológica de sistemas fluviais, com cálculo do desvio ecológico entre
locais teste e locais de referência;
iv. a utilização de programas de geoestatítica, nomeadamente do método de Kriging

(Golden Software Inc. 2000) para a visualização e determinação de relações
cobertura e riqueza – factores abióticos e previsão de invasões.
Resumem-se alguns destes pontos, que poderão ser importantes no desenvolvimento de

trabalhos futuros:
i. o uso de escalas semi-quantitativas de abundância do coberto vegetal é vantajoso

em estudos de grande escala espacial; no entanto há toda a vantagem em
considerar classes menos abrangentes em estudos auto-ecológicos ou na bioava-
liação, de modo a assegurar a significância do tratamento de dados e um análise
mais pormenorizada da variabilidade biológica. É um bom exemplo o sucesso da
aplicação da avaliação percentual do coberto vegetal, nos estudos de invasibilidade
no rio Mondego e no curso principal do Rio Guadiana;
ii. o comprimento dos troços de amostragem nos trabalhos à escala da bacia foi de 250
metros, o que parece ser exagerado para a maior parte dos segmentos fluviais
estudados. Aponta-se o valor de 100 metros como suficiente para a reflectir a
diversidade biológica e habitacional; no entanto, carece de maior esforço
prospectivo a escolha do troço mais representativo do local;
iii. na selecção de locais de referência e em estudos ecológicos de vegetação devem

ser revistos os factores de perturbação locais e ao nível do habitat, em conjunção
com os descriptores ambientais e antrópicos a nível regional ou da bacia;
iv. parte da variabilidade biológica não explicada deve-se, entre outros factores já
mencionados, ao uso de variáveis qualitativas, nalguns casos com uma avaliação
demasiado subjectiva. A título de exemplo, o trabalho dos padrões de invasividade
deveria ter dados analíticos sobre parâmetros fundamentais, como o conteúdo em

nitratos ou o teor de matéria orgânica;
v. deverá ser utilizada a georreferenciação dos inventários, e a utilização de sistemas

de informação geográfica na obtenção de dados e na interpretação de resultados;
avanços tecnológicos estes, só disponíveis na fase final deste trabalho;
vi. por resolver fica a ausência de dados paleoecológicos sobre as florestas ripícolas, o
que dificulta a utilização de referenciais históricos.
14.2 Outros trabalhos futuros: a continuação e o desafio
No culminar deste tipo de trabalho, surge a tentação de tecer algumas conjecturas sobre a
continuidade futura destes estudos ou se é uma “página encerrada” num universo de novas
prioridades.
De facto, alguns pontos ficaram em aberto, como o estudo de perturbações locais nas
galerias ripícolas. Parte da variabilidade dos padrões de integridade e continuidade das
galerias, bem como da composição florística não foram explicados pelas variáveis
ambientais consideradas. Concretizando: os vários usos do solo nos ecossistemas adjacentes
foram estudados globalmente nos primeiros 1000 metros do vale ou até à linha de cumeada;
poderá o uso do solo envolvente na área contígua, digamos nos primeiros 100 ou 200
metros, ser determinante para a explicação dos padrões encontrados? Estudos ao nível do
segmento fluvial, com a análise de pressões humanas episódicas, como pastoreio sazonal da
zona ripícola, cortes pontuais de vegetação arbórea, historial agrícola da propriedade
confinante poderiam revelar aspectos importantes neste contexto.
A abertura permitida no estudo conjunto da vegetação ripícola e das populações de

macroinvertebrados, e os interessantes resultados obtidos, sustentam a vontade em ampliar
o âmbito de estudo, nomeadamente com a participação em equipas multidisciplinares. Um
exemplo concreto será a utilização dos resultados sobre a integridade ecológica das galerias
ripícolas: face à estrutura fragmentária e aos valores reduzidos de largura ripária, estarão
as espécies ribeirinhas (avifauna, mamíferos) adaptadas a esta estrutura linear constrangida
pela envolvente agrícola, florestal e urbana? Qual o valor mínimo de largura ripária e de
taxa de fragmentação suportável por outras comunidades ripícolas mediterrânicas? Um
corredor ripícola com vegetação formada por grupos de árvores isoladas em ambas as
margens apresenta uma maior sustentabilidade que corredores com galeria ripícola apenas
numa margem? Em que medida, a presença, distribuição e cobertura de espécies arbóreas

exóticas nos corredores ripícolas promove ou retrai a diversidade de outras comunidades.
Estes estudos contactam com aspectos da conservação de sistemas lóticos, e poderão ser
importantes a nível técnico-científico em acções de mitigação e recuperação de
ecossistemas ribeirinhos.
Um dos trabalhos mais cativantes desta dissertação foi o estudo dos padrões de
invasibilidade no leito intervencionado do Rio Mondego. Este vale esteve sujeito a cheias
catastróficas e a inundação prolongada, com a destruição de grande parte das infra-
estruturas da Obra de Aproveitamento Hidroagrícola do Baixo Mondego. Passados 3 anos
após as cheias e cinco anos após o trabalho de campo realizado, seria interessante retomar
as amostragens, no sentido de acompanhar a evolução das espécies exóticas em condições
ambientais novas, sobretudo em relação ao piteirão e ao graminhão, bem como a
recuperação da galeria recém-formada.
A reabilitação de sistemas fluviais degradados, nomeadamente pela restruturação das

galerias e restabelecimento da conectividade lateral e longitudinal será um outro ponto de
interesse, sobretudo em estudos sobre o potencial colonizador das espécies nativas
pioneiras em relação às espécies exóticas. As amostragens realizadas para os vários
trabalhos desta dissertação permitiram a observação colateral de aspectos importantes,
entre eles a elevada abundância de plântulas de espécies arbóreas ripícolas (e.g. freixos e
salgueiros) nas zonas marginais frequentemente inundadas. Estudos efectivos sobre a
capacidade de recuperação dos sistemas, com a avaliação laboratorial e no campo do
potencial de regeneração do ecossistema, do banco de sementes e dos factores influentes
poderá ser uma boa via de estudo, na sequência desta dissertação.
Finalmente, e de acordo com o cenário actual de alterações climáticas à escala global,

numerosos estudos (a nível nacional e internacional) estão em curso sobre o seu potencial
efeito em comunidades bióticas. Em Portugal, não se conhecem ainda modelos preditivos
quer da flora nativa ripícola, quer dos taxa exóticos, e em que medida as alterações
climáticas contribuem para o agravamento das invasões biológicas. Este projecto parece ser
mais tentador e desafiador do que os acima referidos, talvez por ser o mais distante
conceptualmente dos trabalhos desenvolvidos, tanto a nível metodológico como no
tratamento de dados e discussão de resultados. Ele poderia envolver acções de
monitorização de comunidades vegetais ou de determinadas espécies-chave indicadoras,
além de proposta fundamentada de medidas de mitigação dos efeitos das alterações
climáticas.
Borcard D.P. Legendre e P. Drapeau. 1992. Partialling out the spatial component of spatial variation.
Ecology 73: 1045-1055.
Dufrêne M. e Legendre P. 1997. Species assemblages and indicator species definition: the need of an
asymmetrical and flexible approach. Ecological Monographs 67: 345-366.
Golden Software Inc., 2000. SURFER (Surface Mapping System), version.7.02.
Rodriguez M.A. e Magnan P. 1995. Applications of multivariate analysis in studies of the organisation
and structure of fish and macroinvertebrate communities. Aquatic Sciences 57: 199-216.

Tese-Phd 2004 Tese PH.D

Enviado por

Direitos autorais:

Formatos disponíveis

Tese-Phd 2004 Tese PH.D

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Tese-Phd 2004 Tese PH.D

Enviado por

Direitos autorais:

Formatos disponíveis

Universidade Técnica de Lisboa

Instituto Superior de Agronomia

Francisca Constança Frutuoso de Aguiar

Vegetação ripícola em sistemas fluviais mediterrânicos.

Doutoramento em Engenharia Florestal

Presidente Reitor da Universidade Técnica de Lisboa

Dissertação apresentada neste Instituto para obtenção do grau de Doutor

To my father and to my son

… roots and branches of myself

ao contrário do que se julga e pôs a correr foi um

Capítulo 2 Ecologia da vegetação ripícola mediterrânica: uma síntese 13

2.2.2 A flora e as classificações 18

2.2.3 Perspectivas do ecossistema 22

2.2.4 Invasão ecológica 26

2.3.1 O clima mediterrânico e o regime hidrológico 28

2.3.2 Composição e estrutura das galerias ripícolas 29

2.3.3 Atributos e funções das galerias ribeirinhas e das zonas ripícolas 32

2.3.4 As galerias ribeirinhas e as actividades humanas 35

2.3.5 A vegetação ripícola como indicador de integridade ecológica 42

2.4 Proposta de estudo 44

Parte II As galerias ribeirinhas em sistemas fluviais mediterrânicos: 57

Capítulo 3 Tipologia de galerias ribeirinhas numa bacia Mediterrânica 59

Chapter 3 Riparian types in a Mediterranean basin

Capítulo 4 Influência antrópica na composição e integridade das galerias ribeirinhas na 77

5.1 Abstract 109

Parte III Factores ambientais determinantes da vegetação ripícola e 133

6.1 Abstract 137

Capítulo 7 Variações sazonais e anuais de macrófitos em sistemas fluviais semi-áridos 155

Chapter 7 Seasonal and yearly variations of macrophytes in a southern Iberian River

7.1 Introduction 157

Capítulo 8 Selecção de locais de referência e avaliação da qualidade ecológica da 167

8.1 Abstract 169

Parte IV Biodiversidade da vegetação ripícola e invasão por plantas 197

9.1 Abstract 201

9.2 Introduction 201

Capítulo 10 Padrões de distribuição de riqueza e cobertura de plantas exóticas e nativas 231

10.1 Abstract 233

Capítulo 11 Padrões de invasão do graminhão (Paspalum distichum L.) em habitats 261

11.1 Abstract 263

Capítulo 12 Flora exótica e endémica em locais de referência e de não-referência de 277

12.1 Aims 279

Parte V Conclusões e considerações finais 289

Capítulo 13 Discussão de resultados e conclusões 291

13.1 Introdução 293

13.2.2 Padrões de continuidade e integridade 295

13.2.3 Padrões de alteração espaço-temporais 299

13.3.3 Os macrófitos como bioindicadores da qualidade ecológica de sistemas fluviais 306

Capítulo 14 Considerações finais 331

14.1 Nota final 333

14.2 Outros trabalhos: a continuação e o desafio 335

14.3 Referências bibliográficas 337

2.7 Enquadramento e síntese dos principais pontos de partida e discussão e da proposta de 45

4.3 Spatial variation of riparian width for the studied corridors. 91

À Professora Teresa Ferreira, orientadora desta dissertação, pelo estreito acompanha-

Ao Professor Ilídio Moreira, co-orientador desta dissertação, por facultar experiência e

Ao Salvador e ao meu filho, pela sã convivência com os meus papéis e “problemas

Ao meu irmão, Ricardo Aguiar, pela preciosa colaboração dada no processamento e

Á minha cunhada, Rute Mendes, agradeço todo o interesse demonstrado e em especial o

Ao Engenheiro António Albuquerque, pela sua amizade e pelo acompanhamento nas

Perturbação [disturbance] remoção da vegetação competidora e PE, AEE