See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225153868

Determining high value areas for steppe birds in Spain: Hot spots,
complementarity and the efficiency of protected areas

Article  in  Biodiversity and Conservation · November 2007

DOI: 10.1007/s10531-006-9138-2

CITATIONS READS

97 1,085

4 authors, including:

Juan Traba Eladio Luis García de la Morena

Universidad Autónoma de Madrid Biodiversity Node
161 PUBLICATIONS   2,548 CITATIONS    49 PUBLICATIONS   1,110 CITATIONS   

SEE PROFILE SEE PROFILE

Manuel B. Morales
Universidad Autónoma de Madrid
175 PUBLICATIONS   5,016 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

SFRH/BD/74840/2010 from Portuguese Science and Technology Foundation (FCT- Fundação para a Ciência e a Tecnologia) View project

I Censo Nacional de Sisón Común View project

All content following this page was uploaded by Manuel B. Morales on 28 May 2014.

The user has requested enhancement of the downloaded file.

 Running head: Hot spots for steppe birds in Spain

DETERMINING HIGH VALUE AREAS FOR STEPPE BIRDS IN SPAIN:

HOT SPOTS, COMPLEMENTARITY AND THE EFFICIENCY OF

5 PROTECTED AREAS

Juan Traba1, *, Eladio L. García de la Morena1, Manuel B. Morales1 and Francisco

Suárez1

10

1
Departamento de Ecología. Universidad Autónoma de Madrid. 28049 Madrid

15

20

*Correspondence author: Juan Traba. Dpto. de Ecología (Edificio de Biología),

Universidad Autónoma de Madrid, 28049 Madrid, Spain. Phone: 0034 91 3978005,

Fax: 0034 91 3978001. E-mail: [email protected]

1
 Summary

We have examined the distribution of 26 species of steppe birds in the Iberian

Peninsula and the Balearic Islands within a grid of 5,070 10x10 km grid cells. The

5 most valuable areas for steppe birds have been identified by selecting the upper 5% of

the 10 x 10 km grid cells after a simple ranking based on the following criteria:

species-richness, richness of rare species, rarity index, species vulnerability at

Spanish, European and Global levels, and using an index combining the previous

criteria. We have also used a heuristic algorithm to select those areas which offered

10 most complementarity. The results have been analysed on a national scale and have

been compared with those obtained by species assessments in previous status

summaries. Finally, we have performed an analysis of the coverage afforded to

locations which are valuable for steppe species by Natural Protected Areas (NPAs)

and Special Protected Areas (SPAs), and an evaluation of the potential effects of

15 temporal changes in species’ threat-status.

The combined index was the most reliable criterion for defining hotspots,

encompassing 15% of the species’ distributions within 5% of the total area

considered. This index showed a high level of geographical concordance with the

other criteria (nearly 70% of the selected grid cells coincided). Analysis of

20 complementarity delivered poorer results than simple rank-scoring. The analysis of

the efficiency of NPA network showed a very low coverage (less than 2%) of the

hotspots selected according to the Combined Index. Coverage of the SPA network

was higher (nearly 45%), although it diminished (to <35%) when only steppe-defined

SPAs were included. The geographical concordance between high value areas

25 selected using current status summaries and those chosen using earlier ones was low

2
 for European threat status (53.4%), intermediate (63.2%) for Spanish threat status and

high for SPEC status (78.1%). The Combined Index showed a high level of

geographical concordance between the old and new data (76.6%). We conclude that

automatic scoring methods (identifying hotspots) are useful for selecting valuable

5 areas and for analysing the efficiency of the network of protected natural spaces, as

well as for examining the potential effects of status changes on hotspot definition.

Furthermore, the Spanish SPA network does not cover the most important areas for

steppe birds adequately.

10 Key words: Complementarity, Hotspots, Spain, Steppe birds, Threat-status changes,

Efficiency of protected areas

3
 Introduction

Recent years have seen the development of computational techniques for identifying

high-value areas and establishing reserves, with the aims of optimising available

5 resources, identifying the minimum extent of the areas needed to ensure protection

and analysing the efficiency of the networks of protected spaces (Williams et al. 1996,

Prendergast et al. 1993a y 1993b, Lawton et al. 1994, Margules and Pressey 2000,

Kati et al. 2004, Jiguet et al. 2005, Burgess et al. 2005). These procedures, based on

simple rank-scoring on predefined criteria (sometines termed hotspot estimation) or

10 on applying algorithms (complementarity analysis), use basic, territory-based

biological information, such as presence/absence of species (Williams et al. 1996, Rey

Benayas and de la Montaña 2003) or the distribution of habitats or representative

vegetational types (Kati et al. 2004), to identify zones having the highest scores under

the criteria used (species-richness, number of rare or threatened species, landscape

15 diversity, etc.). Irrespective of the technique used, automatic methods have been

shown to be more efficient for designing networks of protected areas, in terms of the

number of suitable sites and of total area selected, than those based on ‘expert’

decisions (Pressey and Tully 1994, Pressey et al. 1997) or on the presence of key

species as indicators of high diversity (Andelman and Fagan 2000), although they

20 depend on the existence of basic territory data.

At present, one of the methods most employed to identify valuable areas is the

identification of ‘hotspots’, not as was originally proposed by Myers (1988; Myers et

al. 2000), but as it has subsequently been applied more generally to areas with high

species-diversity (Prendergast et al. 1993a 1993b, Lawton et al. 1994, Kerr 1997, Rey

25 Benayas and de la Montaña 2003, Abellán et al. 2005, Sólymos and Fehér 2005),

4
 zones containing numerous rare species of restricted distribution (Williams et al.

1996, Kerr 1997, Rey Benayas and de la Montaña 2003, Abellán et al. 2005, Sólymos

and Fehér 2005) and locations containing an abundance of threatened or vulnerable

species (Faith and Walker 1996, Rey Benayas and de la Montaña 2003, Burgess et al.

5 2005). However, it has been noted on occasion that area-selection based on hotspots

does not guarantee the inclusion of all the target species, especially when the number

of species requiring protection is very large or when they can only be included by

increasing the expanse of territory protected (Williams et al. 1996, Myers et al. 2000).

A heuristic algorithm has often been used to select the optimum number and

10 extent of areas which will guarantee the inclusion of all the target species at the same

time as minimising the amount of land needed to protect them (see Araújo and

Williams 2001, Kelley et al. 2002, Kati et al. 2004, and Sólymos and Fehér 2005 for

recent examples). This procedure, termed complementarity analysis (Vane-Wright et

al. 1991), has been found useful in the gap analysis of protected spaces (Scott et al.

15 1993, Arponen et al. 2005) and in selecting areas where the territorial expanse

required is very large and conflicts with more productive land-use interests must be

minimised (Margules and Pressey 2000).

Steppe birds, those tied to open, treeless, agrarian landscapes (Suárez et al.

1997), are one of the most threatened avian communities in Europe, as shown in

20 recent years by the marked population declines and range contractions which they

have suffered (BirdLife International 2004, Santos and Suárez 2005, Burfield 2005).

In Spain, where the steppe community is the most diverse and abundant in all Europe,

such negative tendencies are associated with two processes (Madroño et al. 2004,

Santos and Suárez 2005). On the one hand there is agricultural intensification, with

25 consequent increases in fertiliser use, pesticide application and mechanisation. On the

5
 other, there is the abandonment of large areas of marginal land, which were formerly

widely cultivated but are now experiencing an accelerated process of scrub

encroachment (Suárez-Seoane et al. 2002). The particular characteristics of these

species and their habitats demand that the expanses adequate for their conservation be

5 very extensive, so having the potential for conflict with other land-users. There is thus

a need to guarantee the selection of the best locations for incorporation into the

protected spaces network, optimising the number and/or total extent of such areas.

The Birds Directive (Directive 79/409/CEE) obliges wild bird protection and

their habitats in Europe and the establishment of a network of protected spaces to

10 conserve them. It has led to the development of a series of Special Protection Areas

(SPAs – ZEPAs in Spain) in all the member states of the European Union. SPAs in

Spain were selected on specialist advice, given the scarcity of data on species

distribution and abundance in Spain, required to apply computational techniques. The

recent atlas of Spanish breeding birds (Martí and del Moral 2003) gives information

15 on distributions on an intermediate scale (10x10 km quadrats) which allows the

systematic application of criteria to identify important areas for bird conservation.

In this study we analyse the existing data on the distribution of breeding steppe

birds in peninsular Spain and the Balearic Islands with three objectives: i) to identify

the most important areas (hotspots) for steppe birds by simple rank-scoring of

20 richness, rarity and vulnerability criteria and of an index combining these hotspots,

adressing questions such as how many steppe bird species are included in hotspots,

what average percentage of their distribution is covered by them, or what is the

geographical consistency among the criteria for hotspot definition, ii) to compare

these results with those after applying a heuristic algorithm to establish which are the

25 minimum areas (complementarity areas) required to guarantee the conservation of all

6
the breeding steppe species of the study area, and iii) to compare the selected areas

with the coverage of the existing networks of protected spaces, to assess their

effectiveness. In addition, we attempt to analyse the effect which changes in the

vulnerability-status of certain species may have on the selection of important areas.

7
 Methods

Study area

5 The study area was the whole of peninsular Spain and the Balearic Islands (Figure 1),

a total of over 500,000 km2. We used data on breeding bird presence within the

Universal Transverse Mercator (UTM) grid of 10x10 km grid cells. Quadrats whose

area was more than 50% within the sea or in extralimital countries were excluded, so

that the final analysis was limited to 5,070 grid cells. Grid cell size was always 100

10 km2 except for cells joining different UTM coordinate zones, and border cells.

Species

The 26 species included in the analysis (Appendix 1) were selected in accordance

15 with Suárez et al. (1997) on the basis of four associated criteria : i) Species typical of,

or very frequent in, the Mediterranean region, ii) ground-nesting species, iii) species

exclusive to treeless, principally flat zones, and iv) species whose principal European

population is in Spain. In addition, the list includes certain species which are not

ground-nesters, such as the Lesser Kestrel (Falco naumanni), but which we consider

20 to be clearly tied to steppe habitats by their preferential use of them (Bustamante

1997, Negro 1997). We have also included species such as the Skylark (Alauda

arvensis), which are not strictly regarded as steppe species elsewhere but which may

be unequivocally assigned to steppe ecosystems within the Iberian Peninsula (Suárez

et al. 2003a). However, we have excluded those such as the Trumpeter Finch

25 (Bucanetes githagineus) and Black Wheatear (Oenanthe leucura), which prefer

8
 semiarid habitats but are not steppe birds in the sense employed here in that they nest

exclusively in rock cavities (see, for example, Snow and Perrins 1998).

The data used here are derived from the Spanish breeding bird atlas, the Atlas

of Spanish Breeding Birds (Martí and del Moral 2003), which surveyed the whole of

5 the Spanish state using 10 x 10 km grid cells. In total we have utilised 49,294

breeding records (species-in-grid-cell records) obtained for the 26 species in the 5,070

grid cells studied.

Criteria used

10

The following criteria were used to identify ‘Hotspot’ grid cells and were applied

successively to each grid cell: species-richness, richness of rare species, rarity index

and species vulnerability at Spanish, European and Global levels. Species-richness per

grid cell was the number of species present in each, from zero up to a maximum of 26.

15 Richness of rare species was based on the total of those species which were present in

fewer than 50% of the grid cells in the study area, from zero up to a maximum of 19.

The ‘rarity’ criterion has been used previously with reference to species of

scarce distribution (Arponen et al. 2005, Sólymos and Fehér 2005, Rey Benayas and

de la Montaña 2003; Abellán et al. 2005), arranged according to their frequency of

20 occurrence in the original data. The high level of dispersion of the quadrats which

were so assessed prompted the use of an ad hoc grouping of the species into three

abundance classes, ranging from 0 (common species) to 10 (very rare species)

(Appendix 1). The rarity index per grid cell was then calculated as the sum of the

abundance classes of the species present. Only rare or very rare species add values to

25 the final score.

9
 The vulnerability or threat criterion was based on the conservation status of each

species, in terms of its abundance and distribution and the state of its populations.

This criterion has been applied on three spatial levels, assessing the Spanish threat

status (Madroño et al. 2004) whose categories are entirely based on those defined by

5 the last IUCN Red List (IUCN 2001), European threat status (BirdLife International

2004) and European vulnerability with global concern, according to SPEC category

(BirdLife International 2004). Each species was assigned a status value on a 0–10

scale (Appendix 1) for each of the three spatial levels. The ‘vulnerability’ value for

each grid cell was calculated at each of the spatial levels by adding the scores for all

10 the species present. Finally, all the information for each grid cell was expressed as a

combined index by adding together all the values for each criterion, after

standardisation of means.

Hotspot estimation

15

We have taken the highest scoring 5% (n=254, including all cells which tied at the

lowest score in this range) of the 5,070 10 x 10 km grid cells as a cut-off value for

selecting the hotspots for each criterion. This percentage has been used in previous

works (Prendergast et al. 1993a, Williams et al. 1996, Sólymos and Fehér 2005,

20 Ramírez 2004) to represent the minimum area necessary to encompass the most

important zones in a territory.

In addition, in order to evaluate variation in hotspot estimation resulting from

changes in species’ status, an analysis of geographical concordance was carried out

between the vulnerability criteria calculated from the recent status analyses (Madroño

25 et al. 2004, BirdLife International 2004) and those based on earlier ones (Blanco and

10
 González 1992, Tucker and Heath 1994). The Combined Index was also recalculated

on the basis of the latter criteria.

Complementary Areas

We have applied a rarity/complementarity-based algorithm using ResNet software

(Garson et al. 2002) to select the minimum number of grid cells which would

guarantee the presence of the complete steppe bird community. This procedure

emphasises the selection of areas which contain rare species and adds to these those

10 additional areas which most complement them (the complementarity principle, Vane-

Wright 1991). The algorithm first selects the grid cell with the highest rarity value and

repeatedly adds those grid cells which most complement the selection until the

objective of including the complete species-range is achieved (Kelley et al. 2002,

Sarkar et al. 2002). The efficiency of this software has been tested against other

15 similar algorithms (Kelley et al. 2002, after Margules et al. 1988) and it is

complemented by algorithms which ensure that the selected areas are adjacent and

contain a minimum of redundant information. See Garson et al. (2002) for more

information on ResNet.

The algorithms used for calculating complementary areas are primed with a

20 target, defined as a predetermined level of inclusion for each species. We have used

three different targets (Appendix 2): i) the minimum number of grid cells needed to

include 5% of the Spanish distribution of each species, ii) the minimum number of

grid cells needed to achieve as high a percentage of inclusion as possible for the less

widely-distributed species, and iii) the minimum number of grid cells needed to obtain

25 for each species the same percentage achieved in the Combined ‘Hot Spot’ Index.

11
 Protection Efficiency Assessment

The coverage afforded by the Natural Protected Areas (National, Natural and

5 Regional Parks: NPAs) and the Special Protected Areas (SPAs) to the hotspots

selected by applying the Combined Index was analysed, by simple intersection of

maps using a GIS application (ArcGis 9.0 software, ESRI 2000). Provided that a sub-

set of SPAs have been defined specifically by the presence of steppe birds, which has

implications on how their habitats should be managed (the conservation of SPA-target

10 species implies the conservation of their habitats, that is, steppe habitats in the present

case), the protection efficiency of that particular SPA sub-set (steppe SPAs) was also

examined. Those UTM grid cells more than 20% of whose area fell within NPAs

(n=379 grid cells) or SPAs (n=1261 grid cells) were previously selected.

15 Data analysis

The number of species present in each group of selected 10 x 10 km grid cells, and the

occurrence of each species as a percentage of all the grid cells which it occupies in

Spain, were calculated in all cases. The geographical concordance between grid cells

20 selected according to the different criteria was analysed using a Spearman rank

correlation test.

12
 Results

Maxima and geographical pattern

5 None of the 5,070 grid cells analysed held the entire steppe bird community: the

highest species-richness found per grid cell was 22 (mean ± SD: 9.43 ± 4.25). The

maximum was 15 species per grid cell if only rare species are considered (mean ± SD:

4.39 ± 3.23).

Steppe bird species-richness was greatest in the Ebro valley, the northern and

10 southern sub-mesetas (plateaus), the Extremaduran plains and the Guadalquivir valley

(Figure 2a). The principal mountain ranges and the northernmost parts of the

Peninsula held least species. Only 35 grid cells (0.7% of the total) had no steppe

species at all. A similar geographical pattern, with certain modifications, was apparent

when the richness of rare species was analysed (Figure 2b). Only 360 grid cells

15 (7.1%) had no rare species. The distribution of rare species followed the pattern

described above although it emphasised the importance of the Guadalquivir valley:

including the Doñana area, and other parts of Andalucía, as well as parts of southern

Extremadura (Figure 2c). The map (Figure 2d) obtained by applying the Spanish

threat status criterion also highlights the same general areas, although it slightly

20 minimises the importance of the Ebro valley, the uplands of the Sistema Ibérico and

the northern submeseta, and the large expanses of cereal crops of the northern

submeseta, but emphasises the southern sub-meseta, the Extremaduran plains and the

Guadalquivir valley. Application of the European threat status and SPEC criteria

again highlighted the importance of the Guadalquivir valley (Figures 2e and 2f),

25 within the same general distribution pattern described previously. The geographical

13
 consistency between the criteria used (p<0,0001 for all the Spearman rank correlation

tests between criteria) was repeated with the Combined Index, which included the

Guadalquivir valley within the zones having high-value grid cells (Figure 2g).

5 High-Value Areas for Steppe Birds in Spain

The hotspots we selected (the highest-scoring 5% of grid cells under each criterion)

took in 25 out of the 26 species in all cases. Only the Cream-coloured Courser

(Cursorius cursor) was always excluded from the selections, since it only breeds in

10 one, not very species-rich, southern grid cell. The hotspots maps obtained via the

different criteria assigned a mean of 174.8 (± 17.6) grid cells per species. Each species

was always assigned more than 25 grid cells, with the exception of the Collared

Pratincole (Glareola pratincola) which was assigned only 23 grid cells in the species-

diversity hotspots map. In addition, the hotspots took in over 14% on average of the

15 Spanish distributions of all the species, despite representing only 5% of the 5,070

analysed cells (Table 1). The Combined Index hotspots map (Figure 2h) produced the

best result, recording 25 species in 253 grid cells, representing 14.84% of the total

Spanish distribution of all the species. The rare species richness hotspots map, based

on only 19 species, achieved similar results to the other criteria (Table 1).

20 The species which were best represented by all the criteria were the most

widely-distributed ones, such as the Crested Lark (Galerida cristata) (255.86 ± 16.80

grid cells), Corn Bunting (Miliaria calandra) (255.29 ± 17.16 grid cells) or the Red-

legged Partridge (Alectoris rufa) (255.14 ± 17.58 grid cells). The most poorly

represented species were the Short-eared Owl (Asio flammeus) (30.14 ± 2.97 grid

14
 cells), Collared Pratincole (33.00 ± 8.76 grid cells) and Dupont’s Lark (Chersophilus

duponti) (55.29 ± 10.34 grid cells).

All the hotspot maps, and of course the Combined Index hotspot map (Figure

2h), showed several typical geographical clusters (the other maps are available from

5 the authors to anyone interested). These are: a) the Ebro Valley, which has some of

the most valuable grid cells, especially those located in its central range, b) the

Páramos (uplands) of the Sistema Ibérico, which hold hotspots in terms of species

diversity, diversity of rare species and the Combined Index, c) the Northern Meseta,

with two different hotspot areas: the cereal plains in the north and the limestone

10 uplands in the south, d) the Southern Meseta, with three hotspot nuclei, one straddling

its central range, another in the south, and the third in its east-southern extreme, and e)

Extremadura, with two discrete areas, La Serena in the south, and the Cáceres plains,

although neither of these last figured within the species-diversity and European threat

status hotspot maps. Elsewhere, several isolated and dispersed cells were located

15 outside these high-value nuclei, as in the Guadalquivir Valley or in the Madrid region.

The geographical concordance between hotspot maps estimated by the

Spearman rank correlation was always significant (Table 2), although the average

percentage of geographically concordant cells between the different indices was only

62.0% (Table 2). The highest mean concordance (around 69%) was between the

20 Combined Index hotspots map and the others. Some concordance between criteria

was low, as between Spanish threat status hotspots and species richness hotspots

(41.6%), and between diversity of rare species hotspots (47.8%) and European threat

status hotspots (48.7%). On the other hand, the highest geographical concordance was

between the rare species diversity and the combined index hotspots maps (83.0%).

15
 The 92 grid cells that were selected under all the criteria (high concordance cells)

showed complete geographical concordance.

Complementarity

The areas selected by applying the complementarity algorithm always encompassed

the 26 species studied but the resulting maps were less efficient than those obtained

by the simple rank-scoring methodology. As its first target the algorithm selected 232

grid cells (4.6% of the total), although these were widely dispersed across the entire

10 study area except within the Ebro valley (Figure 3a). In addition, it under-represented

some other areas which were highlighted in the hotspots analysis, such as the plains of

Extremaduran and the southern submeseta. The second target selected 781 grid cells

(15.4% of the total), significantly more than the intended 5% coverage of the Iberian

peninsula, although the areas chosen corresponded well to the best possible

15 distribution of high value areas for steppe birds (Figure 3b). The third target needed

296 grid cells in order to include 100% of the species, 29 grid cells more than those

selected using the Combined Index hotspots map. The application of this third target

gave rise to differences from the regions selected by simple rank-scoring since it did

not include grid cells in the uplands of the Sistema Ibérico, although it did incorporate

20 zones in the Guadalquivir valley (Figure 3c).

Changes in species-status

Changes in species-status according to successive national status statements in

25 Spanish Red Data Books (Blanco and González 1992, Madroño et al., 2004) and

16
 European ones (Tucker and Heath 1994, BirdLife International 2004) brought about

significant changes in the high-value areas selected by simple rank-scoring. Hence,

the overall geographical concordance between statements was low for European threat

status (53.4%), intermediate (63.2%) for Spanish threat status and high for SPEC

5 status (78.1%), showing which status appraisals had revealed the most important

changes in species-status. The Combined Index, using as it does a larger number of

criteria, showed a high level of geographical concordance between the old and new

data (76.6%).

10 Protection Efficiency Assessment

The NPA network encompassed 25 of the 26 steppe bird species, although seven of

them, apart from the Cream-coloured Courser which was not included, appeared in

fewer than 25 grid cells (0.5 % of the 5,070 10 x 10 km grid cells) within NPAs

15 (Table 3): Short-eared Owl, Lesser Short-toed Lark (Calandrella rufescens), Dupont’s

Lark, Collared Pratincole, Great Bustard (Otis tarda), Pin-tailed Sandgrouse

(Pterocles alchata) and Black-bellied Sandgrouse (Pterocles orientalis). Only 5

(2.0%) of the 253 grid cells selected in the Combined Index hotspots map fell within

the NPA network.

20 The coverage of steppe birds by the SPAs was better. Again excluding the

Cream-coloured Courser, the remaining species were all included in more than 30 grid

cells within the SPA network (Table 3), which also covered nearly 24% of the

Spanish distribution of these species. A total of 113 grid cells (44.7%) of the 253

selected by Combined Index hotspots map were included within the SPA network.

25

17
 The results altered when only those SPAs defined specifically by the presence of

steppe birds (steppe-SPAs) were included. Three species: Short-eared Owl, Lesser

Short-toed Lark and Collared Pratincole appeared in fewer than 25 grid cells in

steppe-SPAs. The steppe SPAs included 8.2% on average of the Spanish distribution

5 of the steppe bird species and only 84 (33.0%) of the grid cells included in the

Combined Index map were included within the steppe-SPA network. Similarly, only

27 (29.35%) of the 92 high concordance grid cells were in the network. When high

concordance grid cells were added to the SPAs and steppe-SPAs networks the

Spanish distribution of the species covered by the networks improved, especially for

10 the case of steppe-SPAs (26.4% and 12.1%, respectively).

18
 Discussion

The high-value areas for steppe birds in peninsular Spain identified by simple rank-

scoring techniques (hotspot location) are geographically cohesive, supported by

5 expert opinion and correspond with the appropriate biogeographical conditions. They

also show a good match with those regions which have been traditionally defined as

‘steppe’ or ‘pseudosteppe’ in Spain (Suárez et al. 1992, Suárez et al. 1997, Santos and

Suárez 2005). Algorithm-based methods have been found to be less efficient in

defining these zones, given that they have selected marginal zones or have indicated

10 the same geographical pattern as the scoring methods but at the cost of including

larger areas. Finally, the effectiveness of the networks of protected spaces for

protecting high-value steppe bird areas has been shown to be either very low (NPAs)

or moderate (SPAs), failing to cover numerous very important sites.

15 High-Value Areas for Steppe Birds in Spain

The results show that peninsular Spain is very species-rich, at least for the group of

steppe species considered by this study. Although none of the grid cells included the

entire steppe community (the maximum was 22 out of 26 species), only 35 grid cells

20 held none of them. The high average number of steppe species present per grid cell

(9.4 ± 4.3) is explicable, in part, by the excellent conditions for steppe birds offered

by a large part of Spanish territory, with great expanses devoted to cereal crops

together with extensive zones of natural shrub-steppe (including esparto Stipa

tenacissima formations) and semiarid pastures (Santos and Suárez 2005). In addition,

25 the high mean diversity values per grid cell suggest that, at least on the scale

19
 considered, the various species have a marked tendency to occur together, despite

apparent differences in habitat selection (Santos and Suárez 2005, Morales et al.

2006), which is probably due to the mosaic nature of the steppe landscapes, where

patches of crops, fallows, uncultivated land and pasture alternate.

5 The frequent presence in northern Spain and indeed in mountain areas of more

typically central-European species such as the Skylark and Hen Harrier (Circus

cyaneus) increases the species-richness of zones which have not traditionally been

regarded as steppes, as does the wide distribution of such species as the Crested Lark,

Thekla Lark (Galerida theklae), Corn Bunting and Red-legged Partridge. The status

10 of Spain as a crucial country for steppe birds is consolidated by the high frequency of

rare species (4.4 ± 3.2 species/grid cell), which were absent from only 360 grid cells

(7.1%).

The selected hotspots drew in with precision both the steppe bird community

of the study area (25 out of 26 species), and its distribution, given that just 5% of the

15 area considered encompassed over 14% of the grid cells where steppe birds occurred.

The sole species excluded, the Cream-coloured Courser, was only known to breed for

certain in one grid cell.

The high geographical concordance between criteria, something which has

also been observed in other analyses on a similar scale (Prendergast 1993a, Kerr

20 1997), allowed the consistent detection of the most valuable regions for steppe birds

in the Iberian peninsula and showed that these coincide with zones traditionally

defined as ‘steppe’ or ‘pseudosteppe’ in Spain (for a biogeographic revision, see

Santos and Suárez 2005). These, in ascending order of importance are: the plains of

Extremadura, the cereal croplands of the Northern and Southern Mesetas, the open

25 plateaus of the Sistema Ibérico, and, especially, the Ebro valley. The clumping of key

20
 sites around a few nuclei was especially notable in view of the scarcity of peripheral

high-value grid cells, which may indicate a loss of habitat quality, particularly in the

south of the peninsula, where agricultural intensification has been proportionately

greater. A possible consequence of this is the gradual disappearance of marginal

5 areas, which could generate an element of negative feedback as conservation efforts

may become concentrated on current high-value areas.

Despite the high spatial concordance between criteria, a more detailed

examination reveals differences between them, which may determine the selection of

high-value areas and the allocation of conservation funds, as a function of the

10 criterion used. Some authors have previously indicated that it may be best to apply

multiple criteria together, in order to avoid such geographical discrepancies (Hoctor et

al. 2000, Kati et al. 2004, see however Sólymos and Fehér 2005). Indeed the

Combined Index showed the highest concordance with the other criteria: around 69%

of geographic concordance. Thus, it was deemed a good synthetic indicator of high-

15 value areas. The rare species diversity hotspots criterion, considering only 19 species,

achieved similar results to the Combined Index and so could also serve as the sole

criterion. A more conservative approach could make use of the high concordance grid

cells (very hotspots), those which were selected under all the criteria.

20 Complementarity

The results of the complementary areas analysis proved to be suboptimal or poorer

than those obtained by the more parsimonius simple-rank scoring method (hotspot

identification). Applying a heuristic algorithm does not guarantee the best possible

25 solution but rather the achievement of the proposed targets (completeness in surrogate

21
 coverage), the efficient selection of the smallest area possible (Kelley et al. 2002).

Neither does it guarantee that rejected sites are less deserving of conservation than the

selected ones, given that the selection process is ‘blind’ and determined solely by the

rules of the algorithm (Pressey et al. 1997).

5 In the present case, the high dispersion of the grid cells obtained under target 1

did not correspond with what expert opinion (and hotspot maps) indicated to be the

most valuable areas. This result corresponds with those indicated by other authors,

who have indicated the propensity of heuristic algorithm methods to select areas

located at the extremes of distributions (Williams et al. 1996, Araújo and Williams

10 2001, Kati et al. 2004, Sólymos and Fehér 2005), although the high level of dispersion

of the selected grid cells reduces bias arising from the concentration of high-value

zones in just a few areas. Target 2 produced equally inefficient results, given that it

selected a significantly larger area than that obtained by identifying hotspots. With

target 3, the need to incorporate 100% of the species led to including 29 grid cells

15 more than those defined using the Combined Index (see Kati et al. 2004, Kelley et al.

2002), and resulted in the under-representation of significant steppe bird zones, such

as the uplands of the Sistema Ibérico, where there are some of the most important

concentrations of threatened species such as Dupont’s Lark.

20 Changes resulting from variation in species status

Analyses based on old and new data on species vulnerability gave rise to notable

changes in the mapping of important areas, leading to a low level of geographical

concordance between superseded and current status summaries. The low geographical

25 concordance between the European threat status map obtained after Tucker and Heath

22
 (1994) and the one obtained after BirdLife International (2004) was due to the

significant status changes shown by numerous species during the intervening period,

such as the Great Bustard, Northern Wheater (Oenanthe oenanthe) and the Corn

Bunting (which show increasing threat status) and the Collared Pratincole, Pin-tailed

5 Sandgrouse, or the Short-eared Owl (which show decreasing threat status). The

analysis of Global Concern threat status (SPEC) gave less variation between the two

data sources considered, given that only the Little Bustard (Tetrax tetrax), Northern

Wheater and Corn Bunting increased in threat status notably, so that the two sets of

SPEC maps show greater geographical concordance. Regarding Spanish threat status,

10 the differences are explicable in that the first data source indicated few species, with

none recorded in over 1,600 grid cells, compared with only 180 grid cells in the latter

summary. The intervening period has seen the inclusion in the threatened-in-Spain list

or an improvement in status of many steppe species, such as the Little Bustard, Stone

Curlew, Dupont’s Lark, the Black-eared Wheatear, the Cream-coloured Courser, the

15 Short-toed Lark (Calandrella brachydactyla) and the Lesser Short-toed Lark.

Selection of high-value sites, be it by identifying hotspots or by analyses of

complementarity, is useful for designated protected areas and, consequently, may be

relevant to fund allocation and to change conservation policies. This being so, the

choice of criteria for selecting areas requires the greatest caution, given that status

20 changes of a few species may result in significant differences in determining high-

value areas, as occurred in the case of the region of Catalonia, where the geographical

concordance between the two data sets was nil. This is especially relevant in countries

like Spain, where the conservation policies are full transferred to the regional

governments. The use of Combined Indices may moderate such undesirable outcomes

25 (Williams et al. 1996). In addition to its use in the identification of protected areas, the

23
 analysis of changes in vulnerability status can be understood as a dynamic tool to

redirect conservation strategies and funds.

Protection Efficiency Assessment

The efficiency of the NPA network has been shown to be clearly inadequate: although

the protected spaces sheltered 25 out of the 26 steppe bird species, the representation

of high-value areas was marginal, at fewer than 2% of the grid cells selected by the

Combined Index. The analysis of the SPA network gave better results, particularly

10 since most species were well represented (in over 30 grid cells) and because nearly

45% of the hotspots indicated by the Combined Index were included in SPAs.

Nevertheless, most SPAs have not been established on account of the presence of

steppe birds. This could partly explain the under-representation of several species

(Short-eared Owl, Lesser Short-toed and Dupont’s Larks, Collared Pratincole, Great

15 Bustard, and the two sandgrouses) within the SPA network. The steppe-SPA

subgroup, despite having significantly higher values under the Combined Index than

other grid cells, was still poorly effective since it included barely 33% of the most

valuable 10 x 10 km grid cells in Spain according to that index, and only 29% of the

high concordance grid cells. The protection efficiency would significantly improve if

20 these high concordance grid cells were protected, especially in the case of the steppe-

SPAs networks (from 8.2% to 12.1% on average of the Spanish distribution of the

steppe bird species).

The effectiveness of a network of protected spaces depends on the degree of

coverage offered to its target species, those which motivated the sites’ establishment.

25 So far, no NPAs have been declared on account of steppe birds, with perhaps the

24
 exception of the Villafáfila Regional Park, in the cereal croplands of the northern

submeseta. In this respect, it is important to realize that, given the high level of

overlap between SPAs and NPAs, the protection of steppe species and habitats

heavily relies on the SPA network, which has two important conservation

5 consequences. First, in Spain, the level of legal protection is much lower in SPAs than

in NPAs (for example, the only restriction to actions such as infrastructure building

and urbanization SPAs may come through the environmental impact assessment

procedure -see Suárez et al. 2003b- while in NPAs are usually limited by the

corresponding Management Plans). Second, only the small proportion of each

10 species’ distribution included in steppe-SPAs will be managed in a ‘steppe-specific’

manner. This is despite steppe species being the bird community which has shown the

severest decline both within Europe as a whole (BirdLife International 2004) and in

Spain (Santos and Suárez 2005), and despite the great extent of steppe and pseudo

steppe habitats in Spain (Suárez et al. 1992).

15 The low to moderate efficiency of the network of steppe-SPAs can only be

explained by the absence of defined methods and explicit criteria for the designation

of areas for bird protection. Such areas have frequently been designated on account of

flagship species, such as the Great Bustard, with the result that many vulnerable or

species-rich areas have remained outside the network. The establishment of protected

20 space networks on the basis of the presence of flagship or umbrella species has been

found to be poorly effective before (Kerr 1997, Andelman and Fagan 2000). It is also

possible that within Spain priority in SPA designation has been given to other bird

groups, such as forest communities, for which the effectiveness of protection has been

shown to be appreciably superior (Ramírez 2004). In any event, it must be borne in

25
 mind that just because a grid cell does not form part of the top 5% does not mean that

it lacks conservation value.

Finally, the current analysis is extremely conservative, given that the presence

of a species in a grid cell overlapped by an SPA does not necessarily mean that the

5 species was present within the SPA itself. The coincidence of species and protected

areas also does not guarantee the existence of specific protection measures, something

which is especially evident in Spain where the steppe zones find themselves

threatened by the processes of agricultural intensification, land abandonment,

urbanisation and fragmentation by large-scale infrastructure, among others. In this

10 context, the efficiency results regarding steppe-SPAs are particularly meaningful, as

they show how some of the more widely extended, but more rapidly transformed

habitats in Spain, are poorly covered by the SPA network, which has clear

implications for the conservation of steppe-bird communities.

15 Conclusions on the usefulness of the methodology

Since the Natura 2000 network of protected spaces was definitively established, the

EU member states have been obliged to evaluate its success periodically. Steppe birds

are one of the most threatened groups on a European scale, which makes them a

20 primary objective for conservation measures. The use of automatic techniques allows

the relatively easy and rapid evaluation of the success of the SPA network and other

conservation policies or actions, in addition to detecting new zones for incorporation

or where conservation measures should be maximised. To this end, scoring methods

used to define hotspots could be more effective than using heuristic algorithms, whose

25 character is more indicative than prescriptive (Pressey et al. 1997). In any event, the

26
efficacy of the methods resides in the quality of the data used (Kati et al. 2004),

whose collection and processing must be given priority.

27
 Bibliography

Abellán, P., Sánchez-Fernández, D., Velasco, J. and Millán, A. 2005. Conservation of

freshwater biodiversity: a comparison of different area selection methods.

5 Biodiv. Conserv. 14: 3457-3474.

Andelman, S.J. and Fagan, W.F. 2000. Umbrellas and flagships: Efficient

conservation surrogates or expensive mistakes? Proc. Nat. Acad. Sci. 97:

5954-5959.

10

Araújo, M.B. and Williams, P.H. 2001. The bias of complementarity hotspots towards

marginal populations. Cons. Biol. 15: 1710-1720.

Arponen, A., Heikkinen, R.K., Thomas, C.D. and Moilanen, A. 2005. The Value of

15 biodiversity in reserve selection: Representation, species weighting, and benefit

functions. Cons. Biol. 19: 2009-2014.

BirdLife International. 2004. Birds in Europe. Population estimates, trends and

conservation status. Bird Conservation Series No. 12. BirdLife International,

20 Cambridge.

Blanco, J.C. and González, J.L. (eds). 1992. Libro Rojo de los vertebrados de España.

ICONA-MAPA, Madrid.

28
 Burfield, I. 2005. The conservation status of steppic birds in Europe, in Bota, G.,

Morales, M.B., Mañosa, S. and Camprodon, J. (eds.). Ecology and Conservation

of Steppe-land Birds, pp. 119-140. Lynx Editions, Barcelona.

5 Burgess, N., Küper, W., Mutke, J., Brown, J., Westaway, S., Turpie, S., Meshack, C.,

Taplin, J. and Lovett, J. C. 2005. Major gaps in the distribution of protected

areas for threatened and narrow range Afrotropical plants. Biodivers. Cons. 14:

1877-1894.

10 Bustamante, J. 1997. Predictive models for lesser kestrel Falco naumanni distribution,

abundance and extinction in southern Spain. Biol. Cons. 80: 153-160.

ESRI. 2000. ArcGis version 9.1. Redlands, CA.

15 Faith, D.P. and Walker, P.A. 1996. Integrating conservation development:

incorporating vulnerability into biodiversity assessment of areas. Biodivers.

Cons. 5: 417-429.

Garson, J., Aggarwal, A. and Sarkar, S. 2002. ResNet Manual, ver.1.2. University of

20 Texas at Austin. Austin.

Hoctor, T.S., Carr, M.H. and Zwick, P.D. 2000. Identifying a linked reserve system

using a regional landscape approach: the Florida ecological network. Cons. Biol.

14: 984-1000.

25

29
 IUCN. 2001. IUCN Red List Categories and Criteria: Version 3.1. IUCN Species

Survival Commission. IUCN, Gland and Cambridge.

Jiguet, F., Julliard, R., Couvet, D. and Petiau, A. 2005. Modelling spatial trends bird

5 survey data: a valuable tool in biodiversity assessment. Biodivers. Cons. 14:

3305-3324.

Kati, V., Devillers, P., Dufrêne, M., Legakis, A., Vokou, D. and Lebrun, P. 2004.

Hotspots, complementarity or representativeness ? Designing optimal small-

10 scale reserves for biodiversity conservation. Biol. Cons., 120: 471-480.

Kelley, C., Garson, J., Aggarwal, A. and Sarkar, S. 2002. Place prioritization for

biodiversity reserve network design: a comparison of the SITES and ResNet

software packages for coverage and efficiency. Divers. Distrib. 8: 297-306.

15

Kerr, J.T. 1997. Species richness, endemism, and the choice of areas for conservation.

Cons. Biol. 11: 1094-1100.

Lawton, J.H., Prendergast, J.R. and Eversham, B.C. 1994. The numbers and spatial

20 distributions of species: analyses of British data, in Forey, P.L., Humphries, C.J.

and Vane-Wright, R.I. (eds). Systematics and conservation evaluation, pp. 177-

195. Clarendon Press, Oxford.

Madroño, A., González, C. and Atienza, J.C. (eds). 2004. Libro rojo de las aves de

25 España. Dirección General de la Naturaleza-SEO/BirdLife, Madrid.

30
 Margules, C.R., Nicholls, A.O. and Pressey, R.L. 1988. Selecting networks of

reserves to maximize biological diversity. Biol. Cons. 43: 63-76.

5 Margules, C.R. and Pressey, R.L. 2000. Systematic conservation planning. Nature

405: 243-253.

Martí, R. and del Moral, J. C. (eds). 2003. Atlas de las Aves Nidificantes de España.

Dirección General de la Naturaleza- SEO/BirdLife, Madrid.

10

Myers, N. 1988. Threatened biotas: “Hot Spots” in tropical forests. The

Environmentalist 8: 187-208.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. and Kent, J.

15 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853-858.

Morales, M.B., Suárez, F. and García de la Morena, E.L. 2006. Reponses des oiseaux

de steppe aux differents niveaux de mise en culture et d’intensification du

paysage agricole: une analyse comparative de leurs effets sur la densite de

20 population et la selection de l’habitat chez l’outarde canepetiere Tetrax tetrax et

l’outarde barbue Otis tarda. Rev. d’Ecol. (Terre et Vie), 61: 00-00. In press.

Negro, J. J. 1997. Lesser kestrel. BWP Update 1: 40-56.

31
 Prendergast, J., Quinn, R.M., Lawton, J.H., Eversham, B.C. and Gibbons, D.W.

1993a. Rare species, the coincidence of diversity hotspots and conservation

strategies. Nature 365: 335-337.

5 Prendergast, J., Wood, S.N., Lawton, J.H. and Eversham, B.C. 1993b. Correcting for

variation in recording effort in analyses of diversity hotspots. Biodiv. Letters 1:

39-53.

Pressey, R.L., Possingham, H.P. and Day, J.R. 1997. Effectiveness of alternative

10 heuristic algorithms for identifying indicative minimum requirements for

conservation reserves. Biol. Cons. 80: 207-219.

Pressey, R.L. and Tully, S.L. 1994. The cost of ad hoc reservation: a case study in the

Western Division of South Wales. Aust. J. Ecol. 19: 375-384.

15

Ramírez, A. 2004. Efectos geográficos, ambientales y biológicos sobre la distribución

de las aves forestales ibéricas. Ph. D. Thesis. Universidad Complutense de

Madrid. Madrid.

20 Rey Benayas, J.M. and de la Montaña, E. 2003. Identifying areas of high-value

vertebrate diversity for strengthening conservation. Biol. Cons. 114: 357-370.

Rothley, K.D. 1999. Designing bioreserve networks to satisfy multiple, conflicting

demands. Ecol. Appl. 9: 741-750.

25

32
 Santos, T. and Suárez, F. (2005) Biogeography and population trends of iberian

steppe bird. In: Bota, G.; Morales, M.B.; Mañosa, S. and Camprodon, J. (eds.).

Ecology and conservation of steppe-land birds, pp: 69-102. Lynx Edicions,

Barcelona.

Sarkar, S.; Arrarwal, A.; Garson, J.; Margules, C.R. and Zeidler, J. 2002. Place

prioritization for biodiversity content. J. Bioscience 27: 339-346.

Scott, J.M.; Davis, F.; Csuti, B.; Noss, R.; Butterfield, B.; Groves, C.; Anderson, H.;

10 Caicco, S.; D’Erchia, F.; Edwards, T.C.; Ulliman, J. and Wright, G. 1993. Gap

analysis: a geographic approach to protection of biological diversity. Wildl.

Monogr. 123.

Sólymos, P. and Fehér, Z. 2005. Conservation prioritization based on distribution of

15 land snails in Hungary. Cons. Biol. 19: 1084-1094.

Snow, D.W. and Perrins, C.M. 1998. The Birds of the Western Paleartic. Oxford

University Press, Oxford.

20 Suárez, F; Sainz, H.; Santos, T.; González-Bernáldez, F. 1992. Las estepas Ibéricas.

Ministerio de Obras Públicas y Transportes, Madrid.

Suárez, F., Naveso, M.A. and De Juana, E. 1997. Farming in the drylands of Spain:

birds of pseudosteppes. In Pain, D.J. and Pienkowski, N.W. (eds.). Farming

33
 and Birds in Europe. The Common Agricultural Policy and its Implications for

Bird Conservation, pp. 79–116. Academic Press, San Diego.

Suárez, F., Garza, V. and Morales, M.B. 2003a. Skylark Alauda arvensis densities and

5 habitat selection in Mediterranean Iberia: the role of shrub-steppes and

extensive dry-pastures. AGEE 95: 551-557.

Suárez, F., Traba, J., Morales, M.B., Arrieta, S., Herranz, J. and Oñate, J.J. 2003b.

Aves y evaluación de impacto ambiental: ¿estamos identificando el problema

10 correctamente? Ardeola 50: 295-300.

Suarez-Seoane, S., Osborne, P.E., Baudry, J. 2002. Responses of birds of different

biogeographic origins and habitat requirements to agricultural land

abandonment in northern Spain. Biol. Cons. 105: 333-344.

15

Tucker, G.M. and Heath, M.F. 1994. Birds in Europe: Their Conservation Status. Bird

Conservation Series No. 3. BirdLife International, Cambridge.

Vane-Wright, R.I.; Humphries, C.J. and Williams, P.H. 1991. What to protect?

20 Systematics and the agony of choice. Biol. Cons. 55: 235-254.

Williams, P.; Gibbons, D.; Margules, C.; Rebelo, A.; Humphries, C. and Pressey, R.

1996. A comparison of richness hotspots, rarity hotspots, and complementary

areas for conserving diversity of British birds. Cons. Biol. 10: 155-174.

25

34
 Table 1. Results of the different criteria used for selecting hotspots for steppe birds in

Peninsular Spain and Balearic Islands, across 5,070 10 x 10 km UTM grid cells. SR:

Species Richness, RSR: Rare Species Richness, RDB 2004: status sensu Spanish Red

5 Data Book (Madroño et al. 2004), EU-Threat 2004 and SPEC: status sensu BirdLife

International 2004.

EU Complementary areas

Rarity RDB Threat SPEC Combined

SR RSR Score 2004 2004 2004 Index Target1 Target2 Target3

Nº of grid cells
291 242 244 250 260 257 253 232 781 282
chosen
(5.7%) (4.77%) (4.8%) (4.9%) (5.1%) (5.1%) (5.0%) (4.58%) (15.40%) (5.56%)
(n=5,070)

Nº of species
25 25 25 25 25 25 25 26 26 26
represented
(96.2%) (96.2%) (96.2%) (96.2%) (96.2%) (96.2%) (96.2%) (100%) (100%) (100%)
(n=26)

Average

percentage of
15.6% 14.0% 14.5% 14.3% 14.6% 14.3% 14.8% 8.7% 27.9% 14.8%
distribution per

species

10

35
 Table 2. Spearman rank correlations and percentage of common cells among scores of

criteria used for estimating Hot Spots for Spanish steppe birds, across 5,070, 10x10

km grid cells in Spain. All correlations were significant (p<0.05). SR: Species

Richness, RSR: Rare Species Richness, RDB 2004: status sensu Spanish Red Data

5 Book (Madroño et al. 2004), EU-Threat 2004 and SPEC: status sensu BirdLife

International 2004).

Rarity RDB EU-Threat SPEC Combined

SR RSR Score 2004 2004 2004 Index 2004

SR 1

100

RSR 0.792 1

66.56 100

Rarity Score 0.693 0.844 1

54.62 74.19 100

RDB 2004 0.567 0.628 0.770 1

41.62 47.75 64.12 100

EU-Threat 2004 0.712 0.762 0.775 0.637 1

56.98 62.99 64.71 48.69 100

SPEC-2004 0.728 0.728 0.797 0.682 0.782 1

58.84 58.92 67.56 53.64 65.71 100

Combined Index 0.732 0.826 0.903 0.790 0.770 0.820 1

2004 59.24 71.53 83.03 66.78 64.10 70.81 100

36
 Table 3. Representation (nº of grid cells and percentage) of each species inside NPAs,

SPAs, and within steppe SPAs only, and re-evaluation after including high

concordance cells (HCC).

Steppe SPAs +

NPAs SPAs Steppe SPAs SPAs + HCC HCC

Nº Nº Nº Nº Nº

Species cells % cells % cells % cells % cells %

Circus cyaneus
33 4,71 200 28,57 53 7,57 219 31,29 72 10,29
Hen Harrier

Circus pygargus
50 2,31 403 18,59 139 6,41 456 21,03 204 9,41
Montagu’s Harrier

Falco naumanni
36 3,51 238 23,17 101 9,83 287 27,95 162 15,77
Lesser Kestrel

Alectoris rufa
274 5,90 1104 23,77 170 3,66 1159 24,96 237 5,10
Red-legged Partridge

Coturnix coturnix
169 4,58 808 21,92 155 4,21 860 23,33 218 5,91
Quail

Tetrax tetrax
20,00 1,55 264 20,50 132,00 10,25 318 24,69 197 15,30
Little Bustard

Otis tarda
5 0,95 166 31,50 111 21,06 203 38,52 159 30,17
Great Bustard

Burhinus oedicnemus
53 2,43 419 19,22 145 6,65 473 21,70 211 9,68
Stone Curlew

Cursorius cursor
0 0,00 0 0,00 0 0,00 0 0,00 0 0,00
Cream-coloured Courser*

Glareola pratincola
13 6,81 38 19,90 5 2,62 48 25,13 23 12,04
Collared Pratincole

Pterocles orientalis
13 1,55 140 16,73 117 13,98 239 28,55 176 21,03
Black-bellied Sandgrouse

37
Pterocles alchata
12 2,75 108 24,71 64 14,65 151 34,55 118 27,00
Pin-tailed Sandgrouse

Asio flammeus
1 1,37 29 39,73 17 23,29 37 50,68 27 36,99
Short-eared Owl

Alauda arvensis
127 4,51 684 24,30 130 4,62 733 26,04 188 6,68
Skylark

Chersophilus duponti
5 2,16 60 25,97 34 14,72 71 30,74 46 19,91
Dupont’s Lark

Galerida cristata
167 4,31 817 21,08 163 4,21 872 22,50 230 5,93
Crested Lark

Galerida theklae
197 6,74 726 24,82 112 3,83 772 26,39 167 5,71
Thekla Lark

Melanocorypha calandra
56 2,69 398 19,11 146 7,01 452 21,70 212 10,18
Calandra Lark

Calandrella brachydactyla
68 3,20 443 20,83 138 6,49 496 23,32 203 9,54
Short-toed Lark

Calandrella rufescens
19 7,82 58 23,87 22 9,05 77 31,69 42 17,28
Lesser Short-toed Lark

Anthus campestris
97 4,85 539 26,94 104 5,20 586 29,29 157 7,85
Tawny Pipit

Oenanthe oenanthe
127 5,45 665 28,52 123 5,27 710 30,45 178 7,63
Northern Wheatear

Oenanthe hispanica
172 5,87 647 22,08 103 3,52 697 23,79 164 5,60
Black-eared Wheatear

Cisticola juncidis
104 4,39 422 17,81 109 4,60 471 19,88 166 7,01
Fan-tailed Warbler

Sylvia conspicillata
70 5,50 287 22,55 65 5,11 319 25,06 100 7,86
Spectacled Warbler

Miliaria calandra
235 5,23 1001 22,29 169 3,76 1056 23,51 236 5,25
Corn Bunting

38
* Only breeds in 1 Peninsular grid cell.

39
 Appendix 1. Species list, number of 10x10 km grid cells occupied in Spain (Canaries

Island not included), and scores assigned for each species and criteria.

Nº Rarity RDB 04 2004 EU 2004 SPEC

Cells Score1 Score2 Threat Score3 Score4
Circus cyaneus Hen Harrier 700 10 0 2 5
Circus pigargus Montagu’s Harrier 2,168 5 7 0 1
Falco naumanni Lesser Kestrel 1,027 10 7 2 10
Alectoris rufa Red-legged Partridge 4,644 0 1 5 7
Coturnis coturnix Quail 3,686 0 1 2 5
Tetrax tetrax Little Bustard 1,288 10 7 7 10
Otis tarda Great Bustard 527 10 7 7 10
Burhinus oedicnemus Stone Curlew 2,180 5 5 7 5
Cursorius cursor Cream-coloured Courser 1 10 10 10 5
Glareola pratincola Collared Pratincole 191 10 7 5 5
Pterocles orientalis Black-bellied Sandgrouse 837 10 7 5 5
Pterocles alchata Pin-Tailed Sandgroused 437 10 7 5 5
Assio flammeus Short-eared Owl 73 10 5 2 5
Alauda arvensis Skylark 2,815 0 0 2 5
Chersophilus duponti Dupont’s Lark 231 10 10 2 5
Galerida cristata Crested Lark 3,876 0 0 2 5
Galerida theklae Thekla Lark 2,925 0 0 2 5
Melanocorypha calandra Calandra Lark 2,083 5 0 5 5
Calandrella brachydactyla Short-toed Lark 2,127 5 7 5 5
Calandrella rufescens Lesser Short–toed Lark 243 10 5 5 5
Anthus campestris Tawny Pipit 2,001 5 0 5 5
Oenanthe oenanthe Northern Wheater 2,332 5 0 5 5
Oenanthe hispanica Black-eared Wheater 2,930 0 5 2 7
Cisticola juncidis Zitting Cisticola 2,369 5 0 0 0
Sylvia conspicillata Spectacled Warbler 1,273 5 3 0 0
Miliaria calandra Corn Bunting 4,491 0 0 5 7
1
Rarity Score: Assigned thus: 0, species whose frequency of occurrence in the totality of grid cells
considered was >50%, 5, species whose frequency of occurrence was 25-50%, and 10, species whose
5 frequency of occurrence was <25%.
2
Vulnerability in Spain: Scores assigned according to the classifications in the Red Data Book 2004
(IUCN 2001, Madroño et al. 2004): 0 = Not Evaluated, 1 = Data Deficient, 3 = Least Concern, 5 =
Near Threatened, 7 = Vulnerable, and 10 = Endangered and Critically Endangered species.
3
European Vulnerability, sensu European Threat Status 2004 (BirdLife International 2004): 0 = Secure,
10 1 = Localised, 2 = Depleted, 3 = Rare, 5 = Declining, 7 = Vulnerable, and 10 = Endangered and
Critically Endangered species.

40
4
European Vulnerability with Global Concern, assigned according to SPEC category in 2004 (BirdLife
International 2004): As SPEC 1994 (Tucker and Heath 1994), but SPEC 4 is now Non SPEC Europe,
with value = 1.

41
 Appendix 2. Criteria (percentage of the distribution of each species in peninsular

Spain) used in the definition of targets for calculating complementary areas by

ResNet. Criterion 1: 5% of each species’ distribution. Criterion 2: 5% for species in

more than 2000 grid cells, 10% for species in 1000-2000 grid cells, 25% for species in

5 500-1000 grid cells, 50% for species in 250-500 grid cells, and 100% for species in

less than 250 grid cells. Criterion 3: Same as distribution achieved by the Combined

Index hotspot map.

Species Criterion 1 Criterion 2 Criterion 3

Circus cyaneus Hen Harrier 5 25 10

Circus pygargus Montagu’s Harrier 5 5 11

Falco naumanni Lesser Kestrel 5 10 19

Alectoris rufa Red-legged Partridge 5 5 5

Coturnix coturnix Quail 5 5 7

Tetrax tetrax Little Bustard 5 10 19

Otis tarda Great Bustard 5 25 33

Burhinus oedicnemus Stone Curlew 5 5 11

Cursorius cursor Cream-coloured Courser* 100 100 0

Glareola pratincola Collared Pratincole 5 100 19

Pterocles orientalis Black-bellied Sandgrouse 5 25 27

Pterocles alchata Pin-tailed Sandgrouse 5 50 38

Asio flammeus Short-eared Owl 5 100 45

Alauda arvensis Skylark 5 5 7

Chersophilus duponti Dupont’s Lark 5 100 26

Galerida cristata Crested Lark 5 5 7

Galerida theklae Thekla Lark 5 5 6

Melanocorypha calandra Calandra Lark 5 5 12

Calandrella brachydactyla Short-toed Lark 5 5 11

Calandrella rufescens Lesser Short-toed Lark 5 100 26

42
Anthus campestris Tawny Pipit 5 5 9

Oenanthe oenanthe Northern Wheatear 5 5 8

Oenanthe hispanica Black-eared Wheatear 5 5 7

Cisticola juncidis Fan-tailed Warbler 5 5 8

Sylvia conspicillata Spectacled Warbler 5 10 9

Miliaria calandra Corn Bunting 5 5 6

* Only breeds in 1 grid cell.

43
 Figure legends.

Figure 1. Map of Spain showing the main geographical and administrative regions

mentioned in the present article.

Figure 2. Patterns of richness, rarity, threat, and combined index among 10 x 10 km

5 grid cells in Spain: (a), total species richness; (b) rare species richness; (c), rarity; (d),

Spanish Red Data Book status sensu Madroño et al. (2004); (e), European threat status

sensu BirdLife International (2004); (f), SPEC category sensu BirdLife International

(2004); (g), combined index; and (h), Combined Index hot spots selecting the 5% of

all cells of highest ranking score. Maximum grid-cell scores are shown in brown. The

10 remaining non-zero score classes are shown correspondingly pink (high scoring

values), blue (median scoring values) and white (low and very low scoring values), so

that maps can be compared irrespective of units of measure.

Figure 3. Results of the complementarity analysis used to select the minimum number

of grid cells guaranteeing the presence of the complete steppe bird community based

15 on three different targets: (a) target 1, minimum number of grid cells needed to

include 5% of the Spanish distribution of each species; (b) target 2, minimum number

of grid cells needed to achieve the highest possible percentage of inclusion for the less

widely-distributed species; and (c) target 3, minimum number of grid cells needed to

obtain for each species the same percentage achieved in the Combined Index hotspot

20 map.

44
Figure 1

a)

45
 Figure 2

a)

b)

46
 Figure 2

c)

d)

47
 Figure 2

e)

f)

48
 Figure 2

g)

h)

49
 Figure 3

a)

b)

50
 c)

51

View publication stats