Mangrove restoration, a case for an evidence based approach

ResearchOnline@JCU This file is part of the following reference: Machin, James (2015) Mangrove restoration, a case for an evidence based approach. PhD thesis, James Cook University. Access to this file is available from: http://researchonline.jcu.edu.au/46580/ The author has certified to JCU that they have made a reasonable effort to gain permission and acknowledge the owner of any third party copyright material included in this document. If you believe that this is not the case, please contact [email protected] and quote http://researchonline.jcu.edu.au/46580/ “MANGROVE RESTORATION, A CASE FOR AN EVIDENCE BASED APPROACH” JAMES MACHIN BSc. PGDip. THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY WITHIN THE COLLEGE OF MARINE AND ENVIRONMENTAL SCIENCE, JAMES COOK UNIVERSITY OCTOBER 2015 Acknowledgements Firstly, I would like to thank the staff at the Mangrove Research Centre, Ranong, specifically Nan for helping me on long hard days in the field and Wijarn Meepol for keeping me sane while in Ranong and letting me stay at the centre on my monthly visits from Bangkok. Thanks also to Betsy Jackes and Joe Holtum for their initial support for the project remotely while I was still based in Thailand and support for the pilot study, project design and initial write up early on. I am grateful to Michelle Waycott for her vision, support and suggestion to upgrade from my Masters and guiding me through the upgrade process. Thanks also to Simon Robson for his advice on the statistical aspects of the study, not looking too surprised when I turned up at JCU again in 2014 after a bit of a break and support all through 2014/2015 in the last intense phase of analysis and writing, reviewing innumerable drafts and helping with seminar preparation. I would also like to thank Marcus Sheaves for his introduction to multivariate statistics, reviewing drafts and general support and to Norm Duke for his advice and sharing of many references and reviewing drafts. My first introduction to issues surrounding mangrove restoration came from hearing Robin Lewis present at a conference in Ranong that I helped organise when working at the Ranong Mangrove Research Centre as an Australian Volunteer Abroad. I have had the opportunity to keep in touch and work with Robin over the years and he remains an inspiration as a passionate advocate for more effective approaches to mangrove restoration around the world. I am very grateful towards my employer DHI Singapore for agreeing to support my move back to Australia to complete the thesis full time. Finally, last but not least, my loving wife Orapan and my boys Jake and Dylan for their understanding especially in the last year, it wouldn’t have happened without your love and support, I know it’s been difficult but worth it in the end! i Abstract Mangrove forest restoration projects undertaken globally to date have had significant issues and limitations and “most attempts to restore mangroves often fail completely or fail to achieve the stated goals” (Lewis, 2005). The root cause of many of the issues associated with mangrove restoration is that knowledge of the underlying factors controlling spatial patterns in mangroves is often lacking or not integrated into design of restoration projects, impeding the successful restoration of these critical ecosystems. To improve the success of restoration and address many of the issues associated with the current paradigm for mangrove restoration, there is a need to develop an evidence based approach to the planning and implementation of mangrove restoration projects. The current thesis demonstrates the need for an evidence based approach to mangrove restoration through implementation of case studies in lower intertidal mangrove communities of the Ngao river, located in the Kraburi river estuary, Ranong province, southern Thailand and development of a series of recommendations on how to incorporate this improved evidence base into different phases of mangrove restoration projects. The case studies target identified gaps in knowledge of the different stages of the mangrove tree life cycle through: observational studies on mature tree forest composition and spatial patterns in two forest types; observational studies and experiments examining the dispersal and development of mangrove propagules and seeds and establishment of released propagules in experimental enclosures; and experiments examining the survival and growth of transplanted seedlings in the two lower intertidal forest communities, with distance from the Ngao river and inside and outside canopy gaps. Chapter 2 of the thesis documents the findings of the first case study and showed that the two lower intertidal forest communities included in the study, the “Aegiceras forest” and the “Sonneratia-Avicennia forest”, were low in mangrove tree species diversity and differed based on species similarity in assemblages of mature trees and differences in forest structure parameters of dominant species. Within each of the two forest types, assemblages of mature trees also differed with distance from the Ngao river with the study confirming the presence of distinct spatial ii patterns or zones within each of the two forest types. Duration of tidal inundation was found to be an important determinant of forest structure in the two forest types, significantly correlated with assemblages of mature trees as well as key forest structure parameters. These correlations and the differences in natural assemblages of mature mangrove trees between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest”, suggest that assemblages of mature trees are related to hydroperiod and duration of inundation but that forest communities in both the “Aegiceras forest” and “Sonneratia-Avicennia forest” could have a broad tolerance to hydroperiod characteristics of the two forest types. Within the two forest types, the physical effects associated with more exposed locations close to the Ngao river appear to determine within forest spatial patterns, with A.corniculatum seemingly less tolerant of these more exposed locations than A.alba and S.alba but a better competitor than A.alba and S.alba at more sheltered sites inland. Chapter 2 also provided information on the demographics of mature trees in the two forest types, which suggest that secondary succession has possibly taken place within the “Aegiceras forest” in the past with A.corniculatum trees taking over from S.alba and A.alba trees, a pattern which is possibly currently being repeated in the “Sonneratia-Avicennia forest”. These patterns of secondary succession have not previously been documented before in these lower intertidal communities. The second case study, described in Chapter 3 focused on propagules and developing seedlings and had two components. The first component was a propagule dispersal study which showed that assemblages of propagules and developing seedlings differed between the two lower intertidal forest types, as did the propagule and immature seedling forest structure parameters of the three dominant species A.corniculatum, A. alba and S.alba. Assemblages of propagules and developing seedlings also differed with distance from the Ngao river with the greatest density of A.corniculatum propagules and seedlings in the “Aegiceras forest” found in the middle zone of the forest and the lowest density found in the zone immediately adjacent to the Ngao river and inland. A.alba propagules and seedlings had contrasting spatial patterns and were confined to quadrats closest to the Ngao river. In the “Sonneratia-Avicennia forest”, A.corniculatum seedling density was also related to the distance from the Ngao river, highest in the zone adjacent to the river and iii lowest inland. The propagule dispersal study also found clear seasonal variation in propagule and seedling abundance of the dominant species and seasonal differences between forest types, with seedlings of the dominant species tending to persist across the year in the “Sonneratia-Avicennia forest” but not in the “Aegiceras forest”. The second component of Chapter 3 was a propagule release experiment which showed that a greater proportion of A. corniculatum propagules released into enclosures developed into mature seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, a greater proportion of A.alba propagules developed into mature seedlings in the “Sonneratia-Avicennia forest”. The rate of propagule development was also related to distance from the Ngao river. For A. corniculatum the highest proportion was found in the middle of the forest dominated by A.corniculatum trees, while in contrast, a greater proportion of A.alba propagules developed into mature seedlings in plots further away from the Ngao river outside the area dominated by mature A.alba trees. The results suggest that differences in assemblages of propagules and seedlings between the two forest types are closely related to the mature forest composition and the Importance value of conspecific trees in the canopy. Within the two forest types, the results also support the secondary role of duration of inundation, with physical effects associated with more exposed locations close to the Ngao river appearing to determine within forest distribution patterns of A.corniculatum propagules and seedlings which were seemingly less tolerant to exposed locations than propagules and seedlings of A.alba and S.alba. The results described in Chapter 3 also suggest differences in reproductive strategy amongst species and provides additional evidence of secondary succession through observations of: (1). A wide distribution of A.corniculatum propagules and seedlings across the study area; and (2). persistence of A.corniculatum propagules and seedlings over an annual period in the “SonneratiaAvicennia forest”. These results add weight to the theory espoused in Chapter 2 of the thesis that A.corniculatum is expanding its range in the “Sonneratia-Avicennia” forest, replacing the dominant S.alba and A.alba, perhaps in the same manner as this species has done in past in the “Aegiceras forest”. iv The third case study documented in Chapter 4, was a transplant study focusing on A. corniculatum and S.alba seedlings. Results showed that survival and growth of transplanted seedlings of both species was greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. S.alba seedlings showed a particularly strong preference for environmental conditions in the “Sonneratia-Avicennia forest” type. Survival and growth of seedlings of both species also increased with distance from the Ngao river in both forest types. This trend was especially pronounced for f A.corniculatum seedlings. The proportion of transplanted seedlings of both species surviving also differed with canopy type, with a greater proportion of seedlings surviving in light gaps than under the forest canopy. Canopy type had less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings, which showed a strong preference for light gaps as opposed to locations under the forest canopy. Canopy type had less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest” and also had less of an effect on survival of transplanted seedlings in the quadrats closest to the Ngao river than it did further inland. In both forest types the survival and growth of transplanted seedlings of both species was correlated with the percentage of time that they were inundated by tides, and the survival and growth of both species increased as the proportion of the day in which the site was inundated decreased. Clear differences in survival and growth of transplanted seedlings of the two species appears to be related to significantly different levels of light availability in the two forest types resulting from the differing forest structure of these two lower intertidal communities, the denser, more closed canopy of the “Aegiceras forest” and the more open canopy “Sonneratia-Avicennia forest” as reported in Chapter 2 of this thesis. Within the two forest types, a secondary role of duration of tidal inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to influence seedling survival and growth patterns, both of which were significantly greater in plots away from the Ngao river than plots close to it. A.corniculatum seedlings were seemingly less tolerant to these more exposed locations than those of S.alba. The effects of light level and physical disturbance on seedling establishment, although not unequivocal, are consistent with results from the literature and also provide further insights into v tolerances of the two lower intertidal species to environmental conditions typically observed at restoration sites. Chapter 5 of the thesis summarises the results of the case studies and makes recommendations for how to incorporate each aspect of the improved evidence base arising from the case studies, into different phases of mangrove restoration projects. A table making explicit links between the case studies and restoration is included as was a concept for a decision support system for restoration of lower intertidal forests in the form of a flow chart based on results of studies of mature forest, propagules and seeds and mature seedlings. The flow chart demonstrates how each piece of evidence could potentially be applied in practice to enable the best decisions on which mangrove restoration approach to adopt to be made in an objective manner. For example information on hydroperiod, combined with information on site exposure and level of natural recruitment at the site would result in a specific recommendation for restoration of a site with either A.corniculatum, S.alba or A.alba seedlings where natural recruitment at the selected site was insufficient. Incorporated together with the improved evidence base on lower intertidal forests developed through the three case studies, the decision support system can potentially serve as a practical tool for the integration of scientific knowledge about mangrove ecosystems into restoration planning and ultimately improve the results of restoration and assist in the recovery of these essential ecosystems. vi Table of Contents Acknowledgments i Abstract ii Table of Contents vi List of Tables ix List of Figures xi 1. 1.1 1.2 1.2.1 1.2.2 1.3 General Introduction .............................................................................................. 1-1 Introduction .................................................................................................................... 1-1 Overall thesis objective ................................................................................................... 1-6 Case studies..................................................................................................................... 1-6 A logical framework for restoration decision making ..................................................... 1-8 References....................................................................................................................... 1-8 2. Spatial patterns of forest structure in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand ....................................................... 2-1 Abstract ........................................................................................................................... 2-1 Introduction .................................................................................................................... 2-4 Methods .......................................................................................................................... 2-7 Study sites ....................................................................................................................... 2-7 Mangrove forest composition, status and management ............................................... 2-8 Soils ................................................................................................................................. 2-8 Climate ............................................................................................................................ 2-9 Tidal regime..................................................................................................................... 2-9 Surface elevation and hydroperiod............................................................................... 2-10 Forest structure............................................................................................................. 2-12 Data analysis ................................................................................................................. 2-18 Results ........................................................................................................................... 2-19 Surface elevation and hydroperiod............................................................................... 2-20 Forest structure............................................................................................................. 2-24 Discussion ...................................................................................................................... 2-57 Spatial and seasonal variation in hydroperiod in the two forest types ........................ 2-57 Variation in species composition and spatial patterns of forest structure in the two forest types ................................................................................................................... 2-58 Change dynamics .......................................................................................................... 2-61 Environmental parameters responsible for observed patterns of community structure ........................................................................................................................ 2-63 Conclusions ................................................................................................................... 2-66 References..................................................................................................................... 2-70 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.2.7 2.2.8 2.3 2.3.1 2.3.2 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.5 2.6 3. Dispersal and early development of mangrove propagules and seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern vii 3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.3.5 3.3.6 3.4 3.4.1 3.4.2 3.4.3 3.5 3.6 3.7 Thailand ................................................................................................................. 3-1 Abstract ........................................................................................................................... 3-1 Introduction..................................................................................................................... 3-3 Methods .......................................................................................................................... 3-4 Study sites ....................................................................................................................... 3-4 Experimental method ...................................................................................................... 3-4 Measurement parameters .............................................................................................. 3-9 Data analysis .................................................................................................................. 3-12 Results ........................................................................................................................... 3-14 Environmental parameters ........................................................................................... 3-14 Variation in composition and distribution of propagules and seedling assemblages with forest type ........................................................................................ 3-15 Variation in propagule and seedling assemblages with month of monitoring ............. 3-25 Variation in propagule and seedling assemblages with distance from the Ngao river within forest types ................................................................................................ 3-31 Relationship between mangrove propagule and seedling assemblages and environmental parameters .................................................................................... 3-41 Enclosure experiments .................................................................................................. 3-50 Discussion ...................................................................................................................... 3-61 Composition and distribution of propagules and seedling assemblages ...................... 3-61 Development of experimentally released propagules .................................................. 3-68 Environmental factors responsible for observed patterns in natural propagule and seedling assemblages and experimentally released propagules .......... 3-68 Conclusions.................................................................................................................... 3-74 Appendices .................................................................................................................... 3-79 References ..................................................................................................................... 3-88 4. Survival and growth of transplanted mangrove seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand ................... 4-1 Abstract ......................................................................................................................................... 4-1 4.1 Introduction..................................................................................................................... 4-3 4.2 Methods .......................................................................................................................... 4-4 4.2.1 Study sites ....................................................................................................................... 4-4 4.2.2 Species description .......................................................................................................... 4-4 4.2.3 Experimental method ...................................................................................................... 4-5 4.2.3.1 Collection of plant material ............................................................................................. 4-5 4.2.3.2 Nursery pre-treatment .................................................................................................... 4-5 4.2.3.3 Growth of seedlings in the nursery ............................................................................... 4-10 4.2.3.4 Transplanting of mature mangrove seedlings ............................................................... 4-11 4.2.3.5 Field monitoring ............................................................................................................ 4-11 4.2.4 Measurement parameters ............................................................................................ 4-11 4.2.5 Data analysis .................................................................................................................. 4-12 4.2.5.1 Survival .......................................................................................................................... 4-12 4.2.5.2 Growth........................................................................................................................... 4-14 4.3 Results ........................................................................................................................... 4-14 viii 4.3.1 4.3.1.1 4.3.1.2 4.3.1.3 4.3.2 4.3.2.1 4.3.2.2 4.3.2.3 4.3.2.4 4.4 4.4.1 4.4.1.1 4.4.1.2 4.4.1.3 4.4.2 4.4.3 4.5 4.6 Survival .......................................................................................................................... 4-14 Forest and species ......................................................................................................... 4-14 Distance from the Ngao river ........................................................................................ 4-17 Canopy type .................................................................................................................. 4-22 Growth .......................................................................................................................... 4-24 Time............................................................................................................................... 4-24 Forest and species ......................................................................................................... 4-24 Distance from the Ngao river ........................................................................................ 4-29 Canopy type .................................................................................................................. 4-30 Discussion ...................................................................................................................... 4-33 Survival and growth of mangrove seedlings ................................................................. 4-34 Differences in survival and growth between forest types ............................................ 4-35 Differences in survival and growth with distance from the Ngao river ........................ 4-36 Differences in survival and growth with canopy type................................................... 4-37 Change dynamics .......................................................................................................... 4-38 Environmental factors responsible for observed patterns of survival and growth ...... 4-39 Conclusions ................................................................................................................... 4-41 References..................................................................................................................... 4-44 5. 5.1 5.2 5.2.1 Summary and implications for mangrove restoration .............................................. 5-1 Introduction……….............. .............................................................................................. 5-1 Summary of technical chapters....................................................................................... 5-1 Chapter 2: Spatial patterns of forest structure in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand........................................ 5-1 Chapter 3: Dispersal and early development of mangrove propagules and seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand ............................................................................................. 5-3 Chapter 4: Survival and growth of transplanted mangrove seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand....... 5-6 Implications for restoration ............................................................................................ 5-7 A concept for a decision support system for restoration of lower intertidal forests ... 5-13 References………………... ................................................................................................. 5-18 5.2.2 5.2.3 5.3 5.3.1 5.4 List of Tables Table 2-1: Definitions of hydroperiod terminology used in the current chapter. ........................ 2-11 Table 2-2: Definitions of univariate forest structure terminology used in the current chapter... 2-14 Table 2-3: Variation in tidal duration and frequency between and within the two forest types. 2-20 Table 2-4: Predicted inundation frequency and duration in each forest type showing variation across the calendar year. .............................................................................................................. 2-24 Table 2-5: The six mangrove tree species recorded in the two forest types in the study area.... 2-26 Table 2-6: ANOSIM comparisons and SIMPER analysis of mangrove community structure between the “Aegiceras forest” and the “Sonneratia-Avicennia forest” confirming differences in community structure in the two forest types. ................................................................................................. 2-27 Table 2-7: Summary of forest structure parameters in the two forest types for mangrove trees (stem diameter≥2.5cm). ............................................................................................................... 2-31 ix Table 2-8: Stem size of saplings and trees of the six species in the two forest types................... 2-36 Table 2-9: ANOSIM comparisons and results of SIMPER analysis of mangrove tree community structure amongst groups within the “Aegiceras forest”. ............................................................ 2-38 Table 2-10: Forest structure parameters of the five species found in nMDS ordination groups in the “Aegiceras forest” ......................................................................................................................... 2-41 Table 2-11 Spearman’s rank correlation coefficient of hydroperiod parameters along two nMDS ordination axes for mangrove vegetation. .................................................................................... 2-44 Table 2-12: ANOSIM comparisons of mangrove tree community structure amongst groups within the “Sonneratia-Avicennia forest” ................................................................................................ 2-49 Table 2-13: Forest structure parameters of the five species found in nMDS ordination groups in the “Sonneratia-Avicennia forest” ....................................................................................................... 2-52 Table 2-14 Spearman’s rank correlation coefficient of hydroperiod parameters along two nMDS ordination axes for mangrove vegetation. .................................................................................... 2-56 Table 3-1: Summary table highlighting links between overall study objectives and survey design 3-8 Table 3-2: Definitions of environmental parameters as applied to the current chapter.............. 3-10 Table 3-3: Definitions of univariate propagule and seedling parameters in the current chapter. 3-11 Table 3-4: Definitions of parameters used to describe propagule development in experimental enclosures in the current chapter. ................................................................................................ 3-12 Table 3-5: Eigenvalues for six environmental variables in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” showing that a large proportion of variation in the environmental variables was accounted for by PC1 and PC2................................................................................ 3-15 Table 3-6: Eigenvectors for six environmental variables in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” highlighting the direction of the correlation between environmental variables and the axes of the PCA. ................................................................................................ 3-15 Table 3-7: Composition of mangrove propagules of the six species in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” ....................................................................................................... 3-17 Table 3-8: Composition of mangrove seedlings of the six species in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” ....................................................................................................... 3-17 Table 3-9: Summary of ANOSIM and SIMPER results showing significant differences in community structure in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” based on propagule and seedling density data..................................................................................................................... 3-20 Table 3-10: ANOSIM and SIMPER comparisons of mangrove propagule and seedling assemblages with distance from the Ngao river in the “Aegiceras forest”. ....................................................... 3-33 Table 3-11: ANOSIM and SIMPER comparisons of mangrove propagule and seedling assemblages in the “Sonneratia-Avicennia forest”. ............................................................................................... 3-35 Table 3-12: Summary of BIOENV statistical analysis showing correlation between mangrove propagule and seedling assemblages in the two forest types and environmental parameters ... 3-42 Table 3-13: The proportion of propagules of A.corniculatum, A.alba and S.alba propagules released into enclosures reaching each development stage two months after release. ............................ 3-51 Table 3-14: Results of logistic regression analysis of the proportion of propagules released into enclosures at different development stages two months after release in the two forest types. 3-51 Table 3-15: Proportion of propagules released into enclosures at each development stage two months after release ..................................................................................................................... 3-54 Table 3-16: Results of logistic regression analysis of the proportion of propagules released into enclosures at different development stages two months after release based on distance from the x Ngao river in the two forest types. ............................................................................................... 3-55 Table 4-1: Survival and growth parameters used during field monitoring ................................... 4-11 Table 4-2: Variation in the proportion of transplanted (a) A. corniculatum and (b) S. alba seedlings surviving three, six and twelve months after transplanting. ........................................................ 4-16 Table 4-3: Results of logistic regression analysis predicting the probability of survival of transplanted seedlings at month three, six and twelve based on species and forest type.......... 4-16 Table 4-4: Variation in the survival of (a) A. corniculatum and (b) S. alba seedlings with distance in the two forest types three, six and twelve months after transplanting. ..................................... 4-19 Table 4-5: Results of logistic regression analysis of survival of transplanted seedlings at month three, six and twelve based on distance from the Ngao river ...................................................... 4-19 Table 4-6: Variation in survival of transplanted A. corniculatum and S. alba seedlings in light gaps and under the forest canopy three, six and twelve months after transplanting.......................... 4-23 Table 4-7: Results of logistic regression predicting the probability of survival of transplanted seedlings at three, six and twelve months after transplanting based on canopy type. ............... 4-23 Table 4-8: Results of repeated measures ANOVA analysis comparing mean height increment of transplanted seedlings at month three and six based on time since planting, species, forest type, distance from the Ngao river and canopy type. ........................................................................... 4-26 Table 4-9: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings at three, six and twelve months after transplanting in the two forest types.. ................................. 4-27 Table 4-10: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings three, six and twelve months after transplanting with distance from the Ngao river in the two forest types. Data are pooled across canopy treatments. ........................................................... 4-29 Table 4-11: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings at three, six and twelve months after transplanting in different canopy types ........................... 4-32 Table 5-1: Summary of applications of knowledge developed through the thesis to restoration of degraded lower intertidal forests in the immediate region ........................................................... 5-8 List of Figures Figure 1-1: Schematic diagram of the relationship between the three experimental chapters in the thesis ......................................................................................................................................... 1-7 Figure 2-1: The location of the Ngao river in Ranong province, Thailand: (a). Thailand and surrounding countries;( b). Ranong province; (c). the Kraburi estuary………………………………………2-7 Figure 2-2: Monthly rainfall patterns in Ranong province (1991-2001)………………………….………….. 2-9 Figure 2-3: Predicted tidal range in the Kraburi river estuary, Ranong province, Thailand in 2002 showing Mean High Water Springs (MHWS), Mean High Water Neaps (MHWN), Mean Low Water Neaps (MLWN) and Mean Low Water Springs (MLWS) ............................................................... 2-10 Figure 2-4: The technique used to measure the level of tidal inundation across experimental transects. Measurement of tidal inundation at a site: (a) dyed tape is attached to stakes; (b) water washes dye out of the tape at high tide; (c) height reached by the high tide is measured. ........ 2-11 Figure 2-5: Aerial photo of the estuary showing paired transects in the two forest types . ........ 2-12 Figure 2-6: View of the “Aegiceras forest” (a) looking into the forest from the Ngao river, (b) looking towards the Ngao river from within the forest and the “Sonneratia-Avicennia forest”, (c) from looking inland and (d) within the forest looking towards the Ngao river at low tide .......... 2-13 Figure 2-7: Experimental quadrats used to measure forest structure within each forest type and transect. ....................................................................................................................................... 2-16 xi Figure 2-8: Layout of 50 m experimental transects in the two forest types showing 10 m x 10 m quadrats ....................................................................................................................................... 2-17 Figure 2-9: (a) Higher surface elevation, (b) Higher mean inundation duration and (c) Constant Inundation frequency with distance from the Ngao river across transects in the “Aegiceras forest” relative to the “Sonneratia-Avicennia forest” ............................................................................... 2-23 Figure 2-10: (a) Predicted tidal level over the 2002 calendar year, (b) mean inundation duration and (c) mean inundation frequency of experimental quadrats in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. ...................................................................................................... 2-25 Figure 2-11: Association of experimental quadrats in the two forest types based on mangrove tree stem density (no. ha-1). (a) nMDS ordination from Bray Curtis similarities on square root transformed stem density data .................................................................................................... 2-28 Figure 2-12: Difference in selected mangrove forest structure parameters of mature mangrove trees of different mangrove species in the “Aegiceras forest” (orange) and the “SonneratiaAvicennia forest” (green)............................................................................................................... 2-32 Figure 2-13: Significantly higher mean stem height, stem diameter, total basal area, total stem density and Importance value of A.corniculatum trees in the “Aegiceras forest” compared to the “Sonneratia-Avicennia forest”. ...................................................................................................... 2-34 Figure 2-14: Significantly higher mean stem diameter of (a) mature trees (when all species are pooled)(b) A.alba trees and (c) relative density and Importance value of S.alba trees in the “Sonneratia-Avicennia forest” compared to the “Aegiceras forest”. ........................................... 2-35 Figure 2-15: Difference in stem size class distributions of mangrove species in (a) the “Aegiceras forest” and (b) the “Sonneratia-Avicennia forest”. ....................................................................... 2-37 Figure 2-16: Association of experimental quadrats in the “Aegiceras forest” based on total tree stem density (no. ha-1). ................................................................................................................ 2-39 Figure 2-17: (a) Vegetation profiles of experimental transects in the “Aegiceras forest. (b) Comparison of forest structure parameters of the five species found in nMDS ordination groups in the “Aegiceras forest” ................................................................................................................... 2-42 Figure 2-18: Variation in stem diameter and total stem density of A.corniculatum trees between nMDS ordination groups in the “Aegiceras forest”. ...................................................................... 2-44 Figure 2-19: Significant correlation between nMDS ordination axis 2 and Inundation duration and frequency in the “Aegiceras forest” .............................................................................................. 2-45 Figure 2-20: Bubble plots and biplots overlaid on the nMDS plots of mangrove tree composition based on stem density in the “Aegiceras forest”. ......................................................................... 2-46 Figure 2 21: Significant correlations between forest structure parameters and Inundation duration in the “Aegiceras forest”. .............................................................................................................. 2-47 Figure 2-22: Association of experimental quadrats in the “Sonneratia-Avicennia forest” based on stem density ................................................................................................................................. 2-50 Figure 2-23: (a) Vegetation profiles of experimental transects in the “Sonneratia-Avicennia forest” (b) Comparison of forest structure parameters of the five species found in nMDS ordination groups in the “Sonneratia-Avicennia forest”............................................................................................. 2-53 Figure 2-24: Comparison of significantly different forest structure characteristics of S.alba, A.alba and R.mucronata trees in nMDS ordination groups in the “Sonneratia-Avicennia forest”. ......... 2-55 Figure 2-25: Observed and predicted patterns of secondary succession in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”.(a) Aegiceras forest showing possible past forest profile and existing forest, (b) Sonneratia-Avicennia forest showing existing and predicted forest profile. .. 2-62 xii Figure 2-26: Visual summary of differences in univariate forest structure parameters (a) between the two forest types and (b) within each of forest types ............................................................. 2-68 Figure 3-1: Distribution of 2 m x 2 m quadrats across transects in the two forest types. .............. 3-6 Figure 3-2: Layout of 50m transects in the two forest types showing 2 m x 2 m quadrats ........... 3-6 Figure 3-3: Example of 2 m x 2 m quadrats located (a) Adjacent to the Ngao river in the “Aegiceras forest”, and (b) In the middle zone of the “Sonneratia-Avicennia forest” towards the Ngao river.3-7 Figure 3-4: Soil surface of one quadrat containing germinating propagules of A. corniculatum. .. 3-8 Figure 3-5: Experimental enclosure used in the study showing (a). dimensions of experimental enclosure, (b). actual enclosure in the field in the “Sonneratia-Avicennia forest”. ....................... 3-9 Figure 3-6: Principal Components Analysis (PCA) plot showing variation in environmental variables (a) across the two forest types, and (b) with distance from the Ngao river. Principal component axes 1 and 2 cumulatively account for 76.7% of the total variation present. ............................. 3-16 Figure 3-7: nMDS ordination showing clear differences in community structure between the two forest types based on: (a) propagule and (b) seedling density data. ........................................... 3-19 Figure 3-8: Variation in mean density of propagules and seedlings of the lower intertidal species found in the “Aegiceras forest” (orange colour) and the “Sonneratia-Avicennia forest” (green colour) over the twelve month monitoring period. ...................................................................... 3-22 Figure 3-9: Significant differences in mean density, relative density and relative frequency of (a) A.corniculatum, (b) S.alba and (c) A.alba propagules and seedlings between the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. ........................................................................................ 3-23 Figure 3-10: Comparison of the mean density of: (a) propagules and (b) seedlings of the lower intertidal mangrove species in the “Aegiceras forest” in each month of monitoring. . ............... 3-26 Figure 3-11: Comparison of mean density of: (a) propagules and (b) seedlings of the lower intertidal specialist species in the “Sonneratia-Avicennia forest” with month of monitoring.. ... 3-29 Figure 3-12: nMDS ordinations of:(a) propagule and(b) seedling density data showing association of experimental quadrats in the “Aegiceras forest” with distance from the Ngao river.............. 3-32 Figure 3-13: (a) Vegetation profiles of experimental transects in the “Aegiceras forest”, (b) bar graphs showing variation in mean propagule and seedling density with distance from the Ngao river and (c) box plots showing significant differences in the density and frequency of propagules and seedlings of lower intertidal species with distance from the Ngao river .............................. 3-37 Figure 3-14: (a) Vegetation profiles of experimental transects in the “Sonneratia-Avicennia forest”, (b) bar graphs showing variation in mean propagule and seedling density with distance from the Ngao river and (c) box plots showing significant differences in the density of A.corniculatum propagules and seedlings with distance from the Ngao river ...................................................... 3-41 Figure 3-15: Biplots of (a) propagules and (b) seedlings in the two forest types ......................... 3-42 Figure 3-16: Correlations between duration of tidal inundation and the density and relative density of propagule and seedlings recorded in quadrats over the twelve month monitoring period. ... 3-45 Figure 3-17: Correlations between the total stem diameter of mature trees and the density and relative density of propagule and seedlings recorded in experimental quadrats in the two forest types over the twelve month monitoring period.. ....................................................................... 3-48 Figure 3-18: Correlations between the Importance value of mature trees and the density and relative density of propagule and seedlings recorded in experimental quadrats in the two forest types over the twelve-month monitoring period.. ....................................................................... 3-49 Figure 3-19: Proportion of propagules of A.corniculatum, A.alba and S.alba recorded at each development stage two months after release into experimental enclosures in the “Aegiceras xiii forest” (orange) and “Sonneratia-Avicennia forest” (green). ....................................................... 3-52 Figure 3-20: Proportion of released propagules developing as mature seedlings two months after release into experimental cages with distance away from the river in the “Aegiceras forest”(Orange); and “Sonneratia-Avicennia forest” (Green). ...................................................... 3-56 Figure 3-21: Correlation between the duration of inundation (%) and the proportion of propagules of the same species developing into different development stages two months after release into experimental cages in the: (a) “Aegiceras forest” (Orange); and (b) “Sonneratia-Avicennia forest” (Green). ....................................................................................................................................... 3-58 Figure 3-22: Correlation between Importance values of mature trees and the proportion of propagules developing into different development stages two months after release into experimental cages ....................................................................................................................... 3-60 Figure 3-23: Visual summary of differences in univariate forest structure parameters (a) between the two forest types and (b) within each of forest types .............................................................. 3-75 Figure 4-1: Illustrations of (i) S. alba and (ii) A. corniculatum (a) branchlet, (b)fruit, (c) flower (d) longitudinal section of flower (e) stamen and (f) stems, and (iii) A. alba (a) habit, (b) branch with flowers and fruit, (c) flower seen from top and side (d) fruit and (e) buds ................................... 4-6 Figure 4-2: Ripe propagules/ fruit of (a) A. corniculatum, (b) A. alba and (c) S. alba prior to collection and transport to the experimental nursery. ................................................................... 4-7 Figure 4-3: Seedlings of (a) A. alba, (b) S. alba and (c) A. corniculatum seedlings in the experimental nursery prior to transplanting. ...................................................................................................... 4-11 Figure 4-4: Tagged A. corniculatum seedlings ready to be transported to the field. ................... 4-11 Figure 4-5: Distribution of blocks of transplanted (a) A. corniculatum, (b) A. alba, and (c) S. alba seedlings in the two forest types. ................................................................................................. 4-12 Figure 4-6: Block of transplanted seedlings at field site. .............................................................. 4-13 Figure 4-7: The proportion of transplanted A.corniculatum and S.alba seedlings surviving in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” at three, six and twelve months.. ..... 4-17 Figure 4-8: The proportion of transplanted (a) S.alba and (b) A.corniculatum seedlings surviving distance from the Ngao river in the “Aegiceras forest ” and the “Sonneratia-Avicennia forest”....................................................................................... 4-20 Figure 4-9: The proportion of transplanted a) S.alba and b) A.corniculatum seedlings surviving in light gaps and under the forest canopy in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. ....................................................................................................................................... 4-24 Figure 4-10: Height increment of (a) S. alba and (b) A. corniculatum seedlings three, six and twelve months from transplanting to the “Aegiceras forest” and “Sonneratia-Avicennia forest”. ......... 4-28 Figure 4-11: Variation in the mean height increment of transplanted (a) S.alba and (b) A, corniculatum seedlings three, six and twelve months after transplanting at different distances from the Ngao river in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”…. ............ 4-30 Figure 4-12: Variation in the mean height increment of (a) S. alba and (b) A. corniculatum seedlings at three, six and twelve months after transplanting to light gaps and under the forest canopy in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. .......................................................... 4-33 Figure 4-13: Visual summary of differences in univariate forest structure parameters (a) between two forest types, (b) with distance from the Ngao river (c) with canopy type. ............................ 4-43 Figure 5-1: Flow chart summarising the recommended steps incorporating the evidence base..5-17 developed under the project to a hypothetical mangrove restoration project xiv 1. General Introduction 1.1 Introduction Mangrove forests are well recognised for their ecological and economic values and the role that they play in providing a broad range of direct and indirect goods and ecosystem services (Barbier et al., 2008, Tomlinson, 1987). These values include support for coastal fisheries (Aburto-Oropeza et al., 2008, Rönnbäck, 1999, Mumby et al., 2004); a source of subsistence fuel, charcoal, fibers and tannins for local communities (Bandaranayake, 1998, Warren-Rhodes et al., 2011), maintenance of coastal water quality and the values of neighbouring ecosystems (Schaffelke et al., 2005, Wolanski et al., 1997) and provision of important habitat for a wide range of fauna species (Nagelkerken et al., 2008, Buelow and Sheaves, 2015). Mangrove forests are also well recognised for their role in coastal protection and specifically wave and storm surge attenuation and erosion control (Mazda et al., 2006, McIvor et al., 2012, Winterwerp et al., 2005). Linked to many of the above values, mangroves play an essential role in protecting the livelihoods of coastal populations, especially in developing countries (Dahdouh-Guebas et al., 2006). More recently, the carbon sequestration values of mangrove forests and soils have been better understood in view of climate change and increasing CO2 levels in the atmosphere, with mangroves now recognised as among the most carbon-rich forests in the tropics, with above and below ground carbon stocks estimated to exceed those held within terrestrial ecosystems including rainforests (Donato et al., 2011, McLeod et al., 2011). When these direct and indirect ecosystem services are combined, the dollar value of the services provided by the world’s mangroves has been estimated at more than US$1,648 billion annually (Costanza et al., 1997). Notwithstanding their importance, mangrove forests continue to experience a variety of pressures and are declining as a result of unsustainable utilization and habitat conversion worldwide (Valiela et al., 2001, Duke et al., 1997). While statistics vary, it is generally acknowledged that between 20% and 35% of the world’s mangroves have been lost since 1980 (FAO, 2007) and that mangroves continued to be lost at a rates of approximately 1% per year, rates 3.5 times greater than overall global forest losses (Polidoro et al., 2010, Giri et al., 2011, Van Lavieren et al., 2012). Losses of 1-1 mangroves have been particularly prevalent in south-east Asia where it is estimated that over half of the forests in the region have been lost in the last seventy years (Aksornkoae, 1993). The global decline of mangrove forests has been attributed generally to conversion to agriculture, aquaculture, tourism, urban development and overexploitation (Alongi, 2002, Giri et al., 2011). In south-east Asia, the main causes of loss of mangrove forests are described as conversion to pond aquaculture (shrimp and fish), clear felling for woodchip and pulp production, land clearing for urban and port development and human settlements and harvest of timber for domestic use (FAO, 2007). Future threats to mangroves are expected as a result of climate change which is anticipated to exacerbate existing pressures as a result of a rise in relative sea levels and water temperatures and an increase in the frequency and severity of severe storms (Gilman et al., 2008, Alongi, 2008, IPCC, 2007). Increased awareness of the importance of mangrove forests (particularly for commercial and recreational fisheries in the USA and Australia) and coastal protection in many developing countries and recognition of the unsustainable nature of these significant losses, has led to renewed efforts to protect and restore them (Bosire et al., 2008, Field, 1998). Additional impetus for mangrove restoration globally has come from the large area of abandoned shrimp ponds in former mangrove areas available for restoration (Stevenson, 1997, Matsui et al., 2010, Barbier and Cox, 2004) as well as the emergency recovery efforts after oil spills (Duke et al., 1997) and natural disasters such as the Asian Tsunami (Dahdouh-Guebas et al., 2005, Danielsen et al., 2005), and destructive typhoons in the Philippines (Primavera et al., 2011). Additional mangrove restoration initiatives, such as those in Florida, USA, have had as their motivation increased legal requirements for restoration of mangrove forests impacted by dumping of dredge spoil and impoundment for mosquito control (Brockmeyer et al., 1996) and government requirements that any projects impacting on mangrove areas (e.g through filling or land reclamation) are required to restore an equivalent area of mangrove elsewhere (Lewis, 2000). More recently, mangrove restoration programs have arisen from the realisation that mangroves can potentially play an important role in shoreline stabilisation as an alternative to traditional hard engineering solutions such as sea walls, through so called “hybrid engineering” initiatives or in combination with these traditional engineering focused techniques (Winterwerp et al., 2005, Schmitt et al., 2013). 1-2 Major mangrove restoration programs have been undertaken in more than in twenty countries (Field, 1998), led by national governments often with the support of international donors such as the European Union, Asian Development Bank, and the World Bank and United Nations organisations (e.g. FAO, UNEP, UNDP and UNESCO). A number of Non-government organisations have also undertaken significant restoration programs (e.g IUCN, WWF, Mangrove Action Project and Wetlands International in Thailand and Indonesia) alongside national governments, local community groups or private companies (as part of Corporate Social Responsibility Programs). Some examples of major programs are the World Bank’s Coastal Wetlands Protection and Development Project which planted 28,000ha of mangroves across the coastal belt of the Mekong delta (Streever, 1999), a Thai government restoration initiative program over five years planting 40,000ha of forest (Suwannodom et al. (1998) cited in Erftemeijer and Lewis (2000)) and government programs in the Philippines such as the World Bank’s Central Visaya’s project which planted more than more than 44,000 ha of forest (Primavera et al., 2011). Although there have been some successful examples, such as the case study of large scale mangrove restoration in Bangladesh restoring 120,000ha over a 25 year period (Saenger and Siddiqi, 1993), the majority of mangrove forest restoration projects undertaken to date worldwide have had significant issues and limitations which have been the subject of a number of comprehensive reviews at the global (Lewis, 2000, Lewis, 2005, Field, 1998) and national scale e.g. for the Philippines (Primavera and Esteban, 2008), for Thailand (Suwannodom et al., 1998) and for the US state of Florida (Crewz and Lewis, 1991). In summary, at the global scale, Lewis (2005) states that “most attempts to restore mangroves often fail completely or fail to achieve the stated goals” while National programs mentioned above have recorded similar results, with the large scale Thai mangrove restoration program regarded as largely unsuccessful (Suwannodom et al., 1998) as were the large scale government restoration projects in the Philippines which were reported to have poor growth rates and low long term survival rates (Primavera and Esteban, 2008). The reasons for the lack of success of the majority of mangrove restoration projects undertaken to date are diverse but include: 1-3 Inadequate site assessment and selection processes. Issues raised regarding site selection processes revolve around the fact that: (a) Often the cause of the original loss of mangroves or stressor affecting the health of the mangrove is not known before restoration begins (Rovai et al., 2012) and (b) Site conditions are often not well understood or studied which in many cases has resulted in selection of inappropriate restoration sites. For example, mangroves in the Philippines and Thailand were reported to have been planted in sandy substrates of exposed coastlines and in sandflats and seagrass beds which did not previously support mangroves (Erftemeijer and Lewis, 2000, Primavera and Esteban, 2008). Similarly, issues with site selection in Florida were noted by Crewz and Lewis (1991) who reported the major causes of failure of mangrove restoration projects as: improper planting elevations, improper slopes and site drainage, inferior substrates and site location problems. Other case studies of restoration of mangroves in response to oil spills note that planting was carried out at sites where natural recovery was likely to occur and therefore planting was deemed unnecessary, and in fact detrimental to forest recovery due to the invasive nature of planting activities (e.g site clearance and trampling of soils) (Duke et al., 1997). Planting of mangrove seedlings as the default restoration option. In many mangrove restoration projects there has been a tendency to only use nursery raised planted seedlings in mangrove restoration, adopting traditional forestry based silviculture techniques which focused on nursery raising seedlings and planting them out at regular intervals across the restoration site (Lewis, 2000, Thampanya et al., 2002, Samson and Rollon, 2008). This approach has alternatively been called “tree focused restoration” or “gardening” by previous authors (Kaly and Jones, 1998, Lewis, 2005). The exclusive focus on planted seedlings disregards the benefits of use of the range of potential restoration approaches which amongst others include: (1) Hydrologic restoration involving removal of impediments to tidal flow to restore hydrologic reconnections to the surrounding water system (e.g dikes) (Brockmeyer et al., 1996, Turner and Lewis, 1996); (2) Planting of nurse species such as salt marsh plants to facilitate sediment accretion and natural mangrove recruitment (Lewis, 2005), (3) Distribution of seeds/ propagules to address propagule limitations (Lewis et al., 2005, Qureshi, 1996) and (4) Hybrid engineering models where low cost structures such as bamboo groynes are used to facilitate sediment accretion and natural recruitment of mangroves (Winterwerp, 2005). 1-4 Planting of inappropriate species. A characteristic of many unsuccessful mangrove restoration projects is that mangrove species planted were often not suited for the environmental conditions at the planting sites (Crewz and Lewis, 1991, Primavera et al., 2011). Typically, mangrove species selected are those of commercial value for timber such as those from the family Rhizophoraceae in preference to other species typically playing a colonising role such as Avicennia and Sonneratia which may in many cases more tolerant of environmental conditions typically experienced at restoration sites such as abandoned shrimp farms (Matsui et al., 2010, Kitaya et al., 2002). Lack of adequate monitoring and documentation of the success or otherwise of restoration efforts. The paucity of information about the success or otherwise of mangrove restoration projects around the world has been acknowledged as a major impediment in improvement of the success of mangrove restoration globally (Field et al., 1998). Where monitoring of mangrove restoration projects does occur, a common criticism is that the time available for monitoring is often shorter than the time required for the function of the ecosystem to be restored (McKee and Faulkner, 2000, Bosire et al., 2003). For example, many projects undertaken in the Philippines have initially been considered “successful” based on tree growth rates and forest cover but longer term development of project sites has actually been poor, with low structural complexity and recolonisation by natural recruits, factors which makes these forests less resilient and vulnerable to disturbance from natural disasters (e.g typhoons) and possibly future sea level rise (Salmo et al., 2013, Alongi, 2008, Walters, 2000). In these cases, the short term nature of post planting monitoring was insufficient to identify the above short comings of the current approach to restoration. Related to post restoration planting are the lack of quantifiable restoration objectives and success criteria of many restoration projects which can be used to objectively evaluate project success (Saenger, 2002, Lewis, 2009). The root cause of many of the above technical issues and the current paradigm for mangrove restoration arises from a combination of: (1). A lack of knowledge of the basic drivers of mangrove forest structure, dispersal and early development of propagules and growth and survival of seedlings (Field, 1998, Lewis, 2009); and (2). A lack of a coherent framework for incorporating scientific knowledge about mangrove ecosystems into the design of mangrove restoration projects. 1-5 Without the above, mangrove restoration practitioners tasked to restore degraded mangrove ecosystems are to some extent “working in the dark” when planning and implementing mangrove restoration projects. 1.2 Overall thesis objective To improve success of restoration and address many of the issues associated with the current paradigm for mangrove restoration, there is a need to develop an evidence based approach to the planning and implementation of mangrove restoration projects. This thesis demonstrates the need for an evidence based approach to mangrove restoration through development of: (1) a series of case studies aimed at improving understanding of the natural functioning of lower intertidal mangrove ecosystems including their natural spatial distribution patterns and the critical environmental factors underlying these (Ellison, 2000, McKee et al., 2007, Duke et al., 1998, Friess et al., 2012) and (2). A logical framework for incorporation this knowledge into restoration decision making. 1.2.1 Case studies Case studies were implemented in the lower intertidal mangrove community of the Kraburi river estuary, Ranong province, southern Thailand, dominated by the three species: Aegiceras corniculatum (L.) Blanco, Avicennia alba Blume and Sonneratia alba J. Smith (the nomenclature used follows Tomlinson (1987)). These species are regarded as lower intertidal specialist species, and are often found growing together in the outer mangrove fringe (Santisuk, 1983). Unfortunately, the spatial patterns within lower intertidal mangrove communities in southern Thailand and the species that characterise them, are generally not well understood, having traditionally receiving little research attention. The paucity of knowledge about the functioning of these communities creates impediments to the successful restoration of these communities which have been degraded by economic activities such as shrimp farming and tin mining (Aksornkoae, 1993). The evidence developed through case studies will therefore contribute to the scientific knowledge pertaining to lower intertidal communities in southern Thailand and the immediate region. 1-6 Case studies were designed to be holistic in nature, moving from observational studies at the level of the mature mangrove tree to observational and experimental studies of mangrove propagules and seedlings (Figure 1-1). Chapter 2 of the thesis describes the result of observational studies and surveys aimed at describing the forest composition and spatial patterns in two distinct mature lower intertidal mangrove communities and background environmental conditions including climatic conditions and tidal patterns characteristic of the study area. Figure 1-1: Schematic diagram of the relationship between the three experimental chapters in the thesis In doing so, the chapter seeks to answer the questions: “What is the hydroperiod at the two sites and how does this vary spatially and seasonally across the sites?”; “What are the natural spatial patterns of forest structure within lower intertidal communities in the Ngao river?”; and “How do spatial patterns of forest structure in the two forest communities relate to the hydroperiod in these communities?”. Information presented in this chapter will also serve as a useful framework for the subsequent experimental components of this thesis focusing on dispersal and establishment of mangrove propagules and seedlings within these communities. 1-7 Chapter 3 of the thesis describes the results of experiments examining the dispersal and development of mangrove propagules and seeds across the tidal gradient, seeking to answer the questions: “What are the spatial and seasonal patterns of propagule and seedling assemblages in the two lower intertidal forest types?”; “Does protection from predation and physical disturbance influence the development of transplanted propagules in the two forest types?”; and “How are these spatial patterns related to variation in environmental parameters between forest types and with distance from the Ngao river within each forest type?” The subsequent chapter of the thesis (Chapter 4) describes experiments undertaken to develop an understanding of the factors affecting the post establishment stage of the mangrove seedling through analysis of the relative ability of transplanted mature mangrove seedlings to survive and grow under different conditions of hydroperiod, physical environmental conditions and light. Specifically, the chapter seeks to answer the questions: “Does growth and survival of transplanted mangrove seedlings vary across the two different lower intertidal communities?”; and “Does variation in environmental parameters between forest types, with distance from the Ngao river within each forest type and with canopy type influence the rate of survival and growth of transplanted seedlings?” 1.2.2 A logical framework for restoration decision making The final chapter of the thesis (Chapter 5) summarises the results of the three case studies and makes practical recommendations for mangrove restoration practitioners (through a summary table and concept decision support system), on how the new evidence obtained from case studies can be used to inform and ultimately improve the results of restoration to assist in the recovery of these essential ecosystems. 1.3 References ABURTO-OROPEZA, O., EZCURRA, E., DANEMANN, G., VALDEZ, V., MURRAY, J. & SALA, E. 2008. Mangroves in the Gulf of California increase fishery yields. Proc Natl Acad Sci U S A, 105, 10456-9. AKSORNKOAE, S. 1993. Ecology and management of mangroves. IUCN. 1-8 ALONGI, D. M. 2002. Present state and future of the world's mangrove forests. Environmental Conservation, 29. ALONGI, D. M. 2008. Mangrove forests: Resilience, protection from tsunamis, and responses to global climate change. Estuarine, Coastal and Shelf Science, 76, 1-13. BANDARANAYAKE, W. M. 1998. Traditional and medicinal uses of mangroves. Mangroves and Salt Marshes, 2, 133–148. BARBIER, E. B. & COX, M. 2004. An Economic Analysis of Shrimp Farm Expansion and Mangrove Conversion in Thailand. Land Economics 80, 389-407. BARBIER, E. B., KOCH, E. W., SILLIMAN, B. R., HACKER, S. D., WOLANSKI, E., PRIMAVERA, J., GRANEK, E. F., POLASKY, S., ASWANI, S., CRAMER, L. A., STOMS, D. M., KENNEDY, C. J., BAEL, D., KAPPEL, C. V., PERILLO, G. M. & REED, D. J. 2008. Coastal ecosystem-based management with nonlinear ecological functions and values. Science, 319, 321-3. BOSIRE, J. O., DAHDOUH-GUEBAS, F., KAIRO, J. G. & KOEDAM, N. 2003. Colonization of nonplanted mangrove species into restored mangrove stands in Gazi Bay, Kenya. Aquatic Botany, 76, 267-279. BOSIRE, J. O., DAHDOUH-GUEBAS, F., WALTON, M., CRONA, B. I., LEWIS, R. R., FIELD, C., KAIRO, J. G. & KOEDAM, N. 2008. Functionality of restored mangroves: A review. Aquatic Botany, 89, 251-259. BROCKMEYER, R. E., JR., REY, J. R., VIRNSTEIN, R. W., GILMORE, R. G. & EARNEST, L. 1996. Rehabilitation of impounded estuarine wetlands by hydrologic reconnection to the Indian River Lagoon, Florida (USA). Wetlands Ecology and Management, 4, 93-109. BUELOW, C. & SHEAVES, M. 2015. A birds-eye view of biological connectivity in mangrove systems. Estuarine, Coastal and Shelf Science, 152, 33-43. COSTANZA, R., D'ARGE, R., DE GROOT, R., FARBER, S., GRASSO, M., HANNON, B., LIMBURG, K., NAEEM, S., O'NEILL, R., PARUELO, J., RASKIN, R., SUTTON, P. & VAN DEN BELT, M. 1997. Nature, 387, 253. CREWZ, D. W. & LEWIS, R. R. 1991. An Evaluation of Historical Attempts to Establish Emergent Vegetation in Marine Wetlands in Florida. Florida Sea Grant Technical Publication Gainesville: Florida. DAHDOUH-GUEBAS, F., COLLIN, S., LO SEEN, D., RONNBACK, P., DEPOMMIER, D., RAVISHANKAR, T. & KOEDAM, N. 2006. Analysing ethnobotanical and fishery-related importance of mangroves of the East-Godavari Delta (Andhra Pradesh, India) for conservation and management purposes. J Ethnobiol Ethnomed, 2, 24. DAHDOUH-GUEBAS, F., JAYATISSA, L. P., DI NITTO, D., BOSIRE, J. O., LO SEEN, D. & KOEDAM, N. 2005. How effective were mangroves as a defence against the recent tsunami? Curr Biol, 15, R443-7. DANIELSEN, F., SØRENSEN, M., OLWIG, M., VAITHILINGAM, S., PARISH, F., BURGESS, N., HIRAISHI, T., KARUNAGARAN, V., RASMUSSEN, M., HANSEN, L., QUARTO, A. & SURYADIPUTRA, N. 2005. The Asian Tsunami A Protective Role for Coastal Vegetation. SCIENCE, 310, 643. 1-9 DONATO, D. C., KAUFFMAN, J. B., MURDIYARSO, D., KURNIANTO, S., STIDHAM, M. & KANNINEN, M. 2011. Mangroves among the most carbon-rich forests in the tropics. Nature Geoscience, 4, 293-297. DUKE, N., BALL, M. & ELLISON, J. 1998. Factors influencing biodiversity and distributional gradients in mangroves. Global Ecology & Biogeography Letters, 7, 27-47. DUKE, N. C., PINZÓN M, Z. S. & PRADA T, M. C. 1997. Large-Scale Damage to Mangrove Forests Following Two Large Oil Spills in Panama1. Biotropica, 29, 2-14. ELLISON, A. M. 2000. Mangrove Restoration: Do We Know Enough? Restoration Ecology, 8, 219229. ERFTEMEIJER, P. & LEWIS, R. I. 2000. Planting mangroves on intertidal mudflats: habitat restoration or habitat conversion? In: SUMANTAKUL, V., HAVANOND, S., CHAROENRAK, S. & AMORNSANGUANSIN, J. (eds.) Enhancing Coastal Ecosystem Restoration for the 21st century. Proceedings of a Regional Seminar East for Southeast Asian Countries: Ecotone VIII. Ranong, Thailand: Royal Forest Department of Thailand, Bangkok. FAO 2007. The world's mangroves 1980-2005. FAO Forestry Papers. Rome. FIELD, C., OSBORN, J., HOFFMAN, L., POLSENBERG, J., ACKERLY, D., BERRY, J., BJÖRKMAN, O., HELD, A., MATSON, P. & MOONEY, H. 1998. Mangrove biodiversity and ecosystem function. Global Ecology & Biogeography Letters, 7, 3-14. FIELD, C. D. 1998. Rehabilitation of Mangrove Ecosystems-An Overview. Marine Pollution Bulletin 37, 383-392. FRIESS, D. A., KRAUSS, K. W., HORSTMAN, E. M., BALKE, T., BOUMA, T. J., GALLI, D. & WEBB, E. L. 2012. Are all intertidal wetlands naturally created equal? Bottlenecks, thresholds and knowledge gaps to mangrove and saltmarsh ecosystems. Biological Reviews, 87, 346-366. GILMAN, E. L., ELLISON, J., DUKE, N. C. & FIELD, C. 2008. Threats to mangroves from climate change and adaptation options: A review. Aquatic Botany, 89, 237-250. GIRI, C., OCHIENG, E., TIESZEN, L. L., ZHU, Z., SINGH, A., LOVELAND, T., MASEK, J. & DUKE, N. 2011. Status and distribution of mangrove forests of the world using earth observation satellite data. Global Ecology and Biogeography, 20, 154-159. IPCC 2007. Climate change 2007: synthesis report. Contribution of Working Groups I, II and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change In: PACHAURI, R. & REISINGER, A. (eds.). Geneva, Switzerland. KALY, U. L. & JONES, G. P. 1998. Mangrove Restoration: A Potential Tool for Coastal Management in Tropical Developing Countries. Ambio, 27, 656-661. KITAYA, Y., JINTANA, V., PIRIYAYOTHA, S., JAIJING, D., YABUKI, K., IZUTANI, S., NISHIMIYA, A. & IWASAKI, M. 2002. Early growth of seven mangrove species planted at different elevations in a Thai estuary. Trees, 16, 150-154. LEWIS, R. I. 2000. Ecologically based goal setting in mangrove forest and tidal marsh restoration. Ecological Engineering, 15, 191-198. 1-10 LEWIS, R. R. 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecological Engineering, 24, 403-418. LEWIS, R. R. 2009. Methods and Criteria for Successful Mangrove Forest Restoration. In: GERADO, M., PERILLO, E., WOLANSKI, E., CAHOON, D. & BRINSON, M. M. (eds.) Coastal Ecosystems: An integrated ecosystem approach. Elsevier. LEWIS, R. R., HODGSON, A. B. & MAUSETH, G. S. 2005. Project facilitates the natural reseeding of mangrove forests (Florida). Ecological Restoration, 23, 276-277. MATSUI, N., SUEKUNI, J., NOGAMI, M., HAVANOND, S. & SALIKUL, P. 2010. Mangrove rehabilitation dynamics and soil organic carbon changes as a result of full hydraulic restoration and re-grading of a previously intensively managed shrimp pond. Wetlands Ecology and Management, 18, 233-242. MAZDA, Y., MAGI, M., IKEDA, Y., KUROKAWA, T. & ASANO, T. 2006. Wave reduction in a mangrove forest dominated by Sonneratia sp. Wetlands Ecology and Management, 14, 365-378. MCIVOR, A. L., MÖLLER, I., SPENCER, T. & SPALDING, M. 2012. Reduction of wind and swell waves by mangroves. The Nature Conservancy and Wetlands International. MCKEE, K. L. & FAULKNER, P. L. 2000. Restoration of Biogeochemical Function in Mangrove Forests. Restoration Ecology, 8, 247-259. MCKEE, K. L., ROOTH, J. E. & FELLER, I. C. 2007. MANGROVE RECRUITMENT AFTER FOREST DISTURBANCE IS FACILITATED BY HERBACEOUS SPECIES IN THE CARIBBEAN. Ecological Applications, 17, 1678-1693. MCLEOD, E., CHMURA, G. L., BOUILLON, S., SALM, R., BJÖRK, M., DUARTE, C. M., LOVELOCK, C. E., SCHLESINGER, W. H. & SILLIMAN, B. R. 2011. A blueprint for blue carbon: toward an improved understanding of the role of vegetated coastal habitats in sequestering CO2. Frontiers in Ecology and the Environment, 9, 552-560. MUMBY, P. J., EDWARDS, A. J., ARIAS-GONZALEZ, J. E., LINDEMAN, K. W., BLACKWELL, P. G., GALL, A., GORCZNSKA, M. I., HARBORNE, A. R., PESCOD, C. L., RENKEN, H., WABNITZ, C. C. & LLEWELLYN, G. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature, 427, 533. NAGELKERKEN, I., BLABER, S. J. M., BOUILLON, S., GREEN, P., HAYWOOD, M., KIRTON, L. G., MEYNECKE, J. O., PAWLIK, J., PENROSE, H. M., SASEKUMAR, A. & SOMERFIELD, P. J. 2008. The habitat function of mangroves for terrestrial and marine fauna: A review. Aquatic Botany, 89, 155-185. POLIDORO, B. A., CARPENTER, K. E., COLLINS, L., DUKE, N. C., ELLISON, A. M., ELLISON, J. C., FARNSWORTH, E. J., FERNANDO, E. S., KATHIRESAN, K., KOEDAM, N. E., LIVINGSTONE, S. R., MIYAGI, T., MOORE, G. E., NGOC NAM, V., ONG, J. E., PRIMAVERA, J. H., SALMO, S. G., SANCIANGCO, J. C., SUKARDJO, S., WANG, Y. & YONG, J. W. 2010. The loss of species: mangrove extinction risk and geographic areas of global concern. PLoS One, 5, e10095. PRIMAVERA, J. H. & ESTEBAN, J. M. A. 2008. A review of mangrove rehabilitation in the Philippines: successes, failures and future prospects. Wetlands Ecology and Management, 16, 345-358. 1-11 PRIMAVERA, J. H., ROLLON, R. N. & SAMSON, M. S. 2011. The Pressing Challenges of Mangrove Rehabilitation. 217-244. QURESHI, M. T. 1996. Restoration of mangroves in Pakistan. Restoration of Mangrove Ecosystems. Okinawa, Japan: International Society of Mangrove Ecosystems (ISME). RÖNNBÄCK, P. 1999. The ecological basis for economic value of seafood production supported by mangrove ecosystems. Ecol. Econ., 29, 235. ROVAI, A. S., SORIANO-SIERRA, E. J., PAGLIOSA, P. R., CINTRÓN, G., SCHAEFFER-NOVELLI, Y., MENGHINI, R. P., COELHO-JR, C., HORTA, P. A., LEWIS, R. R., SIMONASSI, J. C., ALVES, J. A. A., BOSCATTO, F. & DUTRA, S. J. 2012. Secondary succession impairment in restored mangroves. Wetlands Ecology and Management, 20, 447-459. SAENGER, P. 2002. Mangrove ecology, silviculture and conservation. Kluwer Academic Publishers. SAENGER, P. & SIDDIQI, N. A. 1993. Land from the sea: The mangrove afforestation program of Bangladesh. Ocean & Coastal Management, 20, 23-39. SALMO, S. G., LOVELOCK, C. E. & DUKE, N. C. 2013. Assessment of vegetation and soil conditions in restored mangroves interrupted by severe tropical typhoon ‘Chan-hom’ in the Philippines. Hydrobiologia, 733, 85-102. SAMSON, M. S. & ROLLON, R. N. 2008. Growth Performance of Planted Mangroves in the Philippines: Revisiting Forest Management Strategies. AMBIO: A Journal of the Human Environment, 37, 234-240. SANTISUK, T. 1983. Taxonomy and Distribution of Terrestrial Trees and Shrubs in the Mangrove Formations in Thailand. . Natural History Bulletin, Siam Society. , 31 63-91. SCHAFFELKE, B., MELLORS, J. & DUKE, N. C. 2005. Water quality in the Great Barrier Reef region: responses of mangrove, seagrass and macroalgal communities. Mar Pollut Bull, 51, 27996. SCHMITT, K., ALBERS, T., PHAM, T. T. & DINH, S. C. 2013. Site-specific and integrated adaptation to climate change in the coastal mangrove zone of Soc Trang Province, Viet Nam. Journal of Coastal Conservation, 17, 545-558. STEVENSON, N. J. 1997. Disused shrimp ponds: Options for redevelopment of mangroves. Coastal Management, 25, 425-435. STREEVER, W. J. 1999. An international perspective on wetland rehabilitation, Springer Science & Business Media. SUWANNODOM, S., SIRIBOON, S. & PAPHAVASIT, N. 1998. An approach to rehabilitate mangrove forests in Thailand. Fifth International Conference on Remote Sensing for Marine and Coastal Environments. San Diego, California. THAMPANYA, U., VERMAAT, J. E. & DUARTE, C. M. 2002. Colonization success of common Thai mangrove species as a function of shelter from water movement. Marine Ecology Progress Series, 237, 111-120. TOMLINSON, P. B. 1987. The Botany of Mangroves. 1-12 TURNER, R. E. & LEWIS, R. R., III 1996. Hydrologic restoration of coastal wetlands. Wetlands Ecology and Management, 4, 65-72. VALIELA, I., BOWEN, J. L. & YORK, J. K. 2001. Mangrove Forests: One of the World's Threatened Major Tropical Environments. BioScience, 51, 807-815. VAN LAVIEREN, H., SPALDING, M., ALONGI, D., KAINUMA, M., CLÜSENER-GODT, M. & ADEEL, Z. 2012. Securing the Future of Mangroves. A Policy Brief., UNU-INWEH. WALTERS, B. B. 2000. Local Mangrove Planting in the Philippines: Are Fisherfolk and Fishpond Owners Effective Restorationists? Restoration Ecology, 8, 237-246. WARREN-RHODES, K., SCHWARZ, A.-M., BOYLE, L. N., ALBERT, J., AGALO, S. S., WARREN, R., BANA, A., PAUL, C., KODOSIKU, R., BOSMA, W., YEE, D., RÖNNBÄCK, P., CRONA, B. & DUKE, N. 2011. Mangrove ecosystem services and the potential for carbon revenue programmes in Solomon Islands. Environmental Conservation, 38, 485-496. WINTERWERP, J. C., BORST, W. G. & DE VRIES, M. B. 2005. Pilot Study on the Erosion and Rehabilitation of a Mangrove Mud Coast. Journal of Coastal Research, 212, 223-230. WOLANSKI, E., SPAGNOL, S. & LIM, E. B. 1997. The importance of mangrove flocs in sheltering seagrass in turbid coastal waters. Mangroves and Salt Marshes, 1, 187-191. 1-13 2. Spatial patterns of forest structure in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand Abstract Knowledge of the underlying factors controlling spatial patterns in mangrove ecosystems is fundamental for restoration of mangrove communities damaged or degraded by natural or anthropogenic influences. Unfortunately, the spatial patterns within lower intertidal mangrove communities are not well understood which creates impediments to their successful restoration when degraded by economic activities. Results of an observational study carried out in two lower intertidal mangrove communities of the Ngao river, in the Kraburi river estuary, Ranong province, southern Thailand show that the two forest types were low in mangrove tree species diversity with only six species found in total with S.alba, A. alba and A.corniculatum the dominant species. Community assemblages of mature trees differed between the two forest types based on species similarity, largely due to the greater densities of A.corniculatum trees in the “Aegiceras forest”. When all species were pooled, trees (based on tree stem diameter) were significantly larger in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” while total stand basal area and stand stem density were not significantly different. For individual species, the stem diameter, height, total stand basal area and Importance value of the dominant species, A.corniculatum were greater in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, S.alba and A. alba trees were larger (greater stem diameters) and more important (higher Importance values) in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. Stem density, when broken down by stem diameter size classes suggest differences in the demographics of mature trees in the two forest types, with significant differences in the proportion of stems falling into different size classes in the two forest types. The “SonneratiaAvicennia forest” had a greater proportion of larger A.alba trees (stem diameter ≥ 10cm) than the “Aegiceras forest” while in contrast, the “Aegiceras forest” had a greater proportion of smaller A.corniculatum trees (stem diameter<10cm) as compared to the “Sonneratia-Avicennia forest”. The differences in demography of the two forest types is supported by observations of S. alba trees in the “Aegiceras forest” which had heavily right skewed stem diameter size classes dominated by trees with large stem diameters with few saplings, indicative of a mature population 2-1 of trees which is not regenerating. The distribution of stem diameter size classes of A. corniculatum in the same forest type, in contrast, had a left skewed population indicating that this species is regenerating and maintaining itself. In the “Sonneratia-Avicennia forest”, the distribution of stem diameter size classes of A.corniculatum was also left skewed while S.alba and A.alba had a broader range of stem diameter size classes. In addition to between forest variation, Chapter 2 of the thesis also highlighted within forest differences in assemblages of mature trees with distance from the Ngao river, confirming the presence of distinct spatial patterns or zones. Further analysis of univariate forest structure parameters of mangrove vegetation in these “zones” in the “Aegiceras forest” showed that in the “Aegiceras forest”, tree densities and stem diameter were significantly different between “zones”. In the “Sonneratia-Avicennia forest”, however, distance from the Ngao river appeared to play a less important role in defining spatial patterns. Measurement of hydroperiod parameters confirmed that surface elevation was significantly lower and duration of inundation significantly higher in the “Aegiceras forest” than in the “SonneratiaAvicennia forest”. Surface elevation and hydroperiod within the two forest types showed similar spatial trends, with both inundation duration and frequency decreasing with distance inland from the Ngao river and corresponding increases in elevation. In both forest types, there was only slight seasonal variation in mean inundation frequency and mean inundation duration across the calendar year. The duration of inundation was significantly correlated with assemblages of mature trees and positively correlated with key forest structure parameters of S.alba, A.corniculatum and negatively correlated with key forest structure parameters of A.alba. These correlations and the differences in natural assemblages of mature mangrove trees between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest”, suggest that assemblages of mature trees are related to hydroperiod and duration of inundation but that forest communities in both the “Aegiceras forest and “SonneratiaAvicennia forest” could have a broad tolerance to hydroperiod characteristics of the two forest types. Within the two forest types, the physical effects associated with more exposed locations close to the Ngao river appears to determine within forest spatial patterns with A.corniculatum 2-2 trees seemingly less tolerant of these more exposed locations than A.alba and S.alba trees but a better competitor than A.alba and S.alba trees in inland more sheltered sites. Information on forest structure obtained in Chapter 2 suggests that secondary succession has possibly taken place within the “Aegiceras forest” where A.corniculatum trees have invaded the forest at some stage, outcompeting S.alba trees in all quadrats except for the outer and most inner quadrats. In contrast, in the “Sonneratia-Avicennia forest”, it appears that A.corniculatum trees are expanding their range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest”. The above inferences about the environmental factors responsible for observed spatial patterns form a useful base on which to develop hypotheses for further controlled experiments in subsequent chapters of the thesis. (Platt, 1964) 2-3 2.1 Introduction Mangrove forests are commonly characterised by one or more species of salt and flood tolerant woody plants distributed in discrete mixed species zones between mean sea level and the spring high tide and generally confined to protected tropical and subtropical coastlines (Chapman, 1976, Macnae, 1969, Watson, 1928, Snedaker, 1982, Duke et al., 1998, Alongi, 2002). Across the IndoPacific region spatial patterns in mangroves typically show “Aegiceras, Avicennia and Sonneratia occupying the lowest intertidal zones; various species of Bruguiera and Rhizophora in the midintertidal zones; and Heritiera, Xylocarpus and numerous other species in the highest intertidal” (Smith, 1992, Watson, 1928). Spatial patterns in mangroves are well described in the mangrove literature including those in Australia (Semeniuk, 1983, Macnae, 1966), Papua New Guinea (Robertson et al., 1991), Indonesia (Hinrichs et al., 2008), and Thailand (Aksornkoae, 1993). In Western Australia, for example, Semeniuk (1983) found that a seaward assemblage of Avicennia and Sonneratia exists which is followed by a zone of Rhizophora. Similarly, Macnae (1966) reported zonation patterns in northern Australia where: “a landward fringe which may be either (a) forested or (b) colonized by (1) Avicennia, and by halophytes; (2) Ceriops thickets; (3) Bruguiera forests; (4) Rhizophora forests; and (5) seaward fringe of Avicennia and Sonneratia. Further afield, in southern Thailand, Miyawaki et al. (1985), Mochida et al. (1999) and Santisuk (1983) have described spatial patterns as changing from forests dominated by the presence of the three lower intertidal specialist species: Sonneratia alba J. Smith and Avicennia alba Blume and Aegiceras corniculatum (L.) Blanco at the seaward fringe to R. apiculata dominated forest, then R. apiculata and B. gymnorrhiza mixed forest, and finally X. moluccensis dominated forest at the point furthest inland. Research in northern Australia (Bunt, 1996) and in the Bangladesh Sundarbans (Ellison et al., 2000) has questioned the simplicity of these described spatial patterns noting that clear cut zonation patterns do not seem to occur in all locations and the phenomena cannot be generalized. A major factor attributed to increased complexity in spatial patterns of mangrove forests especially in northern Australia is the influence of salinity which, in combination with the physiological tolerance of each species, influences species range across estuaries (Duke et al., 1998). The underlying reasons for spatial patterns in mangrove forests such as those described above, have been the topic of vigorous discussion in the scientific literature for many years, with 2-4 numerous hypotheses developed and experiments carried out to elucidate observed spatial patterns. Critical factors proposed in the literature include the interspecific differences in response of the different life stages of the mangrove tree life cycle (propagules and seeds, seedlings and mature trees) to physical, chemical and biological conditions that change with tidal inundation (Smith, 1992, Snedaker, 1982). These include wave and tidal action (Wolanski et al., 2013, Hutchings and Saenger, 1987); salinity (Ball and Pidsley, 1995, Clarke and Hannon, 1970); soil redox potential (Nickerson and Thibodeau, 1985); waterlogging (Giglioli and Thornton, 1965, Clarke and Hannon, 1970); light levels and interspecific competition (Smith, 1987); factors affecting dispersal of propagules (Rabinowitz, 1978b) and the influence of predators (Smith, 1987, Smith et al., 1989, McKee, 1995). Other authors have proposed that the structure of mangrove forests is in part explained by within forest gap dynamics where lightning strikes or hurricanes create forest gaps that are then filled by different species depending on the community composition at the surrounding forest (Putz and Chan, 1986, Duke, 2001). Interactions between the above factors have also been noted to influence spatial patterns within mangrove forests (Smith, 1992). Knowledge of the underlying factors controlling spatial patterns in mangrove ecosystems is fundamental for restoration of mangrove communities damaged or degraded by natural or anthropogenic influences and can ultimately be applied to mangrove restoration through for example, matching of mangrove species with known environmental conditions at restoration sites or modifying the hydrology at the site to facilitate natural recovery. Unfortunately, the spatial patterns within some mangrove communities are not well understood. This creates impediments to their successful restoration when degraded by economic activities such as charcoal harvesting, shrimp farming and tin mining (Aksornkoae, 1993), factors which led to the loss of approximately 50% of the mangroves in southern Thailand since 1960 (Plathong, 1998), cited in Field (1998). A particular gap in knowledge relates to the spatial patterns within lower intertidal mangrove communities along the Ngao river, in the Kraburi river estuary, Ranong province, which have traditionally received little research attention in the past. In addition, where research has been conducted, quantification of spatial patterns through statistical testing has rarely occurred with most studies limited to simple descriptions of observed patterns (Ellison et al., 2000). Noted gaps in knowledge regarding these lower intertidal communities in southern Thailand have led to the planting of species not suited to the environmental conditions at 2-5 restoration sites, a disregard for hydrologic conditions at sites (which is important given the disturbed hydrologic conditions often associated with shrimp farms), and the tendency to only use seedlings instead of propagules or seeds in mangrove restoration work (Matsui et al., 2010). In this chapter, I describe an observational study of lower intertidal mangrove forests of the Ngao river, Ranong province, southern Thailand, as a first step in developing an improved understanding of the functioning of these important communities. The aim of the chapter is to describe the background environmental conditions present within the Ngao river estuary including climatic conditions, soil characteristics, tidal patterns and information on the known composition of mangrove forests in the estuary. The species composition and spatial patterns of forest structure across the tidal gradient are then described in two distinctly different lower intertidal mangrove communities and explanations given as to environmental factors responsible for these observations. Specifically, in this chapter I propose to answer the following questions:  “What is the hydroperiod at the two sites and how does this vary spatially and temporally across the sites?”;  “What are the natural spatial patterns of forest structure within lower intertidal communities in the Ngao river?”; and  “How do spatial patterns of forest structure in the two forest communities relate to hydroperiod in these forests?” Information presented in this chapter provides useful background data and helps frame hypotheses for subsequent experimental components of this thesis that focus on dispersal and establishment of natural mangroves propagules and seedlings (Chapter 3) and growth and survival of transplanted seedlings within these lower intertidal communities (Chapter 4). Importantly the chapter provides practical information (e.g. that relating to species composition, density and basal area of mature trees and their distribution across the two forest types) for restoration practitioners designing and implementing restoration programs within or in the immediate vicinity of lower intertidal communities and making decisions on the selection of mangrove species to be planted, or those relating to planting density and planting locations. In addition, the chapter also demonstrates a relatively simple method for an observational study of a reference forest which can be replicated by mangrove restoration practitioners. 2-6 2.2 Methods Study sites The study was conducted in the mangrove forests of the Ngao river (9o50’N, 98o35’E), a shallow, mangrove fringed tidal creek located in Ranong province, southern Thailand (Figure 2-1). The Ngao river is one of many interconnecting waterways that make up the Kraburi estuary, a large mangrove estuary of approximately 11.5 km2 in area formed in the delta of the Kraburi river which defines the border between Thailand and neighbouring Myanmar (Chunkao et al., 1985), (Macintosh et al., 2002). The mangrove forests of the Ngao river form part of continuous coastal belt of mangroves along the shoreline of Ranong province and cover a total area of 19,308 ha (1993 data) (Khemnark, 1995). Figure 2-1: The location of the Ngao river in Ranong province, Thailand: (a). Thailand and surrounding countries;( b). Ranong province; (c). the Kraburi estuary. (c) (b) (a) 2-7 Mangrove forest composition, status and management Previous studies of the mangrove forests within the Ngao river estuary recorded the presence of a total of 24 species of mangrove trees, shrubs and vines with the most common species being A. alba, B. cylindrica, B. parviflora, Ceriops tagal, R. apiculata, R. mucronata, S. alba, and Xylocarpus granatum (Macintosh et al., 1991). A later study by Mochida et al. (1999) identified four mangrove communities as follows: (1) Sonneratia alba – Avicennia alba community, (2) Rhizophora apiculata community, (3) R. apiculata – Bruguiera gymnorrhiza community, (4) Ceriops tagal – Xylocarpus spp. community along an intertidal gradient from the seaward fringe to inland. Focusing on lower intertidal communities, some limited studies in the Ngao river estuary have found that the forests are typically dominated by three species: A. corniculatum (L.) Blanco, A. alba Blume and S. alba J. Smith (Santisuk, 1983). Mangrove forests along the Ngao river are predominantly secondary regrowth forests previously harvested for production of charcoal and/or disturbed by shrimp farming or tin-mining activity (Komiyama et al., 1992) although the sites selected for the current study are not believed to have been impacted by the above activities and are located far from known disturbed areas. The forests within the Ngao river estuary are now managed by the Thai Department of Marine and Coastal Resources and specifically the Ranong Mangrove Forest Research Center (RMFRC) for their conservation and research value. The forests have also been recognised for their ecological, economic and social importance by UNESCO through the establishment of the Ranong Man and the Biosphere reserve in 1997 covering an area of more than 30,000 ha (Thulstrup, 1998, Macintosh et al., 2002). Soils Soils along the Ngao river are reported to show variable properties as a result of previous disturbances by tin mining and shrimp farming, with the most common having a moderately high clay content (20-50%), moderate organic content (5–20 %) and moderate cation exchange capacity (10-25 meq. per 100 g). Soils impacted by tin mining activities have a moderate to high sand content (10-80 %) underneath a layer of silty clay of low organic content (1-5 %) and low cation exchange capacity (1-10 meq. per 100 g) (Macintosh et al., 1991). 2-8 Climate Ranong is the wettest province in Thailand receiving an average of 4200 mm rainfall per year (from 45 years record), with rainfall recorded on an average of 197 days per year (Macintosh et al., 1991). Local climatic patterns are influenced by two monsoon seasons; the southwest monsoon from April to September and northeast monsoon from October to March. October is the wettest month (Average of 785 mm) and March the driest (Average of 3.1 mm)(Macintosh et al., 1991)(Figure 2-2). The average annual temperature in Ranong is 26.7o C, with a minimum in January (20.6o C) and the maximum experienced in March (34.5o C) (Macintosh et al., 1991). Figure 2-2: Monthly rainfall patterns in Ranong province (averaged over the period 1991-2001). Tidal regime The tidal regime in the Ngao river is semi-diurnal. Mean Seal Level is 2.34 m, and the predicted annual minimum and maximum is 0.34cm and 4.34 m respectively. The tide chart for 2002 is displayed in Figure 2-3. 2-9 Figure 2-3: Predicted tidal range in the Kraburi river estuary, Ranong province, Thailand in 2002 showing Mean High Water Springs (MHWS), Mean High Water Neaps (MHWN), Mean Low Water Neaps (MLWN) and Mean Low Water Springs (MLWS) (Source: Royal Thai Survey Department) Month Surface elevation and hydroperiod An approximation of the surface elevation of the ground level and hydroperiod parameters at 5m points along each of the experimental transects was obtained following the method described by English et al. (1997). This method is also described in other recent mangrove studies (Macintosh et al., 2002, Ashton et al., 2003). Strips of cotton tape dyed with red water soluble food colouring were attached to wooden stakes installed at five metre intervals along each transect prior to the spring tide in October 2002. The dye on the cotton strips was washed away by the rising high tide, leaving a clear mark at the point reached by the tide (Figure 2-4). As recommended by English et al. (1997), the height measured at the watermark was then corrected for the “creep” or capillary movement of water up the cloth (Figure 2.4) as well as for any difference between the base of the stakes used for measurement and the general ground level. The method of hydroperiod calculation did not use actual tide measurements. Comparison of this data with the known maximum water height above chart datum for that day as indicated in the published tide tables as a benchmark, allowed calculation of elevation of each point above chart datum. Using tide charts, the hydroperiod (inundation duration and frequency) of each elevation point was then calculated using the methodology described in Table 2-1. 2-10 Figure 2-4: The technique used to measure the level of tidal inundation across experimental transects. Measurement of tidal inundation at a site: (a) dyed tape is attached to stakes; (b) water washes dye out of the tape at high tide; (c) height reached by the high tide is measured. The photo shows the author measuring the level reached by the high tide; and (d) the point on the cotton tape below which the red water soluble coloured dye was washed away by the rising tide. Source (English et al., 1997). d Table 2-1: Definitions of hydroperiod terminology used in the current chapter. Focus area Parameter Definition Calculation method Hydroperiod Elevation (cm) Elevation of points at 5 m intervals along each transect above chart datum. Elevation = Height of tide above chart datum (from tide tables) – value measured from mark on cotton tape to the ground level. Duration of inundation (% of time per month) Average length of time in a month that a site is inundated expressed as a percentage. Calculated by reviewing the yearly tide tables for Ranong port (2002) and determining the amount of time in each day that the tide was above each elevation point. This value was then averaged over a twelve-month period. Frequency of inundation (number of inundations per month) Number of times in a month that each site is inundated. Calculated by reviewing the yearly tide tables for Ranong port (2002) and determining the number of times that the tide reached each elevation point. 2-11 Data obtained from the inundation study was combined with the forest structure data (discussed in Section 2.2.7.2) to create vegetation profiles diagrams which provide a visual representation of the change in forest structure across the tidal gradient. Forest structure Two lower intertidal mangrove forest communities were chosen for the surveys (Figure 2-5), selected as they are considered to be representative of the lower intertidal forest communities present in the Ngao river. The “Aegiceras forest” (Figure 2-6 a,b) is located approximately 1 km from the mouth of the Ngao river and the “Sonneratia-Avicennia forest” (Figure 2-6 c,d) is approximately 300 m upstream from the “Aegiceras forest”. Figure 2-5: Aerial photo of the Kraburi river estuary showing paired transects in the two forest types (source: Google Earth). Sonneratia-Avicennia forest (FT2) Aegiceras forest (FT1) 2-12 Figure 2-6: View of the “Aegiceras forest” (a) looking into the forest from the Ngao river at low tide, (b) looking towards the Ngao river from within the forest at high tide and the “SonneratiaAvicennia forest”, (c) from within the forest at high tide looking inland from the Ngao river and (d) within the forest looking towards the Ngao river at low tide (a) (b) (c) 2-13 Figure 2-6 continued (d) 2.2.7.1 Definition of univariate forest structure parameters Many ways exist to define the composition and structure of mangrove forests. For the purpose of this study, univariate forest structure parameters have been selected to describe the characteristics of individual mangrove trees and mangrove stands both independently and in relation to other species found in the forest community. These parameters are described in detail in other studies including those by Cintron and Schaeffer Novelli (1984) and Mueller-Dombois and Ellenberg (1974). A detailed description of each selected parameter is included in Table 2-2. Table 2-2: Definitions of univariate forest structure terminology used in the current chapter. Focus area Parameter Definition Calculation method Individual Tree height A measurement of the height of the Direct measurement of height of all trees within tree size (m) tallest stem of each tree from soil experimental quadrat (≥2.5cm Stem diameter) from soil surface to top of canopy. surface to the top of canopy. Stem Diameter of the stem at 1.3m above soil Girth at Breast Height (GBH) measurements made in the diameter surface (where possible and where not field (adjusted for difference in tree morphology as (cm) possible using method described in described in Section 2.2.7.2). Section 2.2.7.2). Conversion to stem 2-14 Stem diameter (d) was calculated using the mathematical Table 2-2 continued Focus area Parameter Definition Calculation method diameter through use of mathematical relationship between the circumference of a circle and its equation. diameter (d = GBH /). Stem Stem diameter of mangrove stems Physical count of all trees in each 100m2 quadrat falling diameter size broken down into size classes. Size into each diameter size class. Density of each species in class (cm) classes categorised as follows: <2.5cm, each size class (no. ha-1) = no. x 10,000 m2 100 m-2. ≥2.5-4cm, 4-6cm, 6-8cm, 8-10cm, 1012cm, 12-14cm, 14-16cm, 16-18cm, 1820cm and ≥20cm. Mangrove Total stand The total basal area of each species and Girth at Breast Height (GBH) measurements made in the stand basal area as a total found per unit area (hectare) field (adjusted for difference in tree morphology as (m2 ha-1) described in Section 2.2.7.2). Stem diameter (d) was calculated using the mathematical relationship between the circumference of a circle and its diameter (d = GBH /). Basal area of the tree stems (g) (cm2) calculated using the formula, =/4*(d)2). Basal area of tree stems per ha=sum of individual basal areas 10,000 m2 100 m-2. Total stem The total number of stems of each density species and as a total found per unit (Number of area (hectare) Physical count of all trees in each 100m2 quadrat. Density of each species (no. ha-1)= no. x 10,000 m2 100 m2. stems ha-1) Relative to Relative Basal area of a species expressed as Relative dominance = (total basal area of a species / other species dominance percentage of total basal area basal area of all species) x 100. in the stand (%) Relative Number of individuals of a given species Relative density = (number of individuals of each species/ density (%) per unit area expressed as a percentage total number of individuals of all species) x 100. of total number of individuals of all species per unit area Frequency The percentage of plots occupied by an Frequency = Number of occupied plots)/(total number of (%) individual species plots) × 100. Relative Frequency of a species divided by sum Relative frequency = (Frequency of an individual species/ frequency (%) of frequencies of all species, expressed total forest frequency of all species) x 100. as a percent Importance The importance value indicates the Importance value of a species = (Relative density + Value (%) structural importance of a species Relative dominance + Relative frequency)/300%. within a stand of mixed species (Curtis 1959). 2-15 2.2.7.2 Survey method The transect line method was used to study the structure of these two forest types following the methodology of English et al. (1997). In each forest type, two 50 m transects were established perpendicular to the river bank, running from the banks of the Ngao river to the extremities of the mangrove vegetation inland. Along each transect, five 10 m x 10 m quadrats were established aligned end to end along the length of the transects and tagged for ease of identification (Figure 27 and Figure 2-8). The total area sampled in the two forest types was 2000m2. No information is available on the total area of the two forest types within the Ngao river estuary but these forest types are not the dominant forest types in the estuary and limited in distribution to the lowest intertidal zones of the forest close to the mouth of the Ngao river. As a result this level of sampling effort was deemed sufficient. Figure 2-7: Experimental quadrats (labelled 1-5) used to measure forest structure within each forest type and transect. Numbers within each quadrat show the distance from the landward to seaward edge (m). Forest type Experimental quadrats and distance from the Ngao river (m) Transect “Aegiceras forest” “Sonneratia-Avicennia forest” 1 2 3 4 5 1 0-10 10-20 20-30 30-40 40-50 2 0-10 10-20 20-30 30-40 40-50 1 0-10 10-20 20-30 30-40 40-50 2 0-10 10-20 20-30 30-40 40-50 In each of the 10 m x 10 m quadrats, mature trees (girth at breast height (where stem diameter ≥ 2.5 cm) and saplings (where stem diameter < 2.5 cm and height ≥ 1 m) were identified and their GBH (approximately 1.3 m from ground level) measured using a measuring tape. Density of seedlings (height < 1 m) was estimated by conducting physical counts within 2 m x 2 m subplots in the left hand corner of each 10 m x 10 m quadrat. Due to the coppiced form of many of the mangrove trees found along each of the experimental transects, decisions had to be made whether stems present in the quadrats constituted separate trees or were simply branches. 2-16 Figure 2-8: Layout of 50 m experimental transects in the two forest types showing 10 m x 10 m quadrats To ensure consistency in decision making, the methodology of English et al. (1997) was followed;  When a stem forks below breast height, or sprouts from a single base close to the ground or above it, measure each branch as a separate stem;  When a stem forms at breast height or slightly above, measure the diameter at breast height or just below the swelling caused by the form;  When the stem has prop roots or fluted lower trunk, measure the stem diameter above them; and  When the stem has swellings, branches or abnormalities at the point of measurement, take the stem diameter slightly above or below the irregularity where it stops affecting normal form. 2-17 To obtain a visual representation of the two forest types, vegetation profiles were drawn of each of the four experimental transects. To draw the profiles, the author walked along the right hand edge of each transect (the opposite side to the 2 m x 2 m permanent quadrats), from the edge of the Ngao river inland and progressively measured the heights and GBH of trees, saplings and seedlings present within 2m of the transect line. The height of trees and saplings was measured using a telescopic measuring pole and GBH measured using a measuring tape. Profile drawings were made in the field using graph paper and then transferred into digital format using Microsoft Excel. Data analysis Exploratory data analysis was conducted on forest structure parameters described in Table 2-2 to explore for any errors and outliers in the data and prepare summary tables of central tendency, variability, and measures of distribution. The Shapiro Wilk test was used to determine whether the populations were normally distributed and therefore whether parametric or non-parametric inferential statistical tests could be used. Graphical methods of analysis were also adopted including scatter plots and frequency histograms. Forest structure data was analysed using both multivariate and univariate statistical tests using SPSS v.22 for Windows (Univariate analysis) and PRIMER 6 (Plymouth Routines In Multivariate Ecological Research) (Clarke and Warwick, 2001)(Multivariate analysis). Non-parametric multivariate techniques contained in PRIMER 6, as described in Clarke and Ainsworth (1993) were used to explore for patterns in mangrove community structure across the two forest types. Similarity matrices were constructed using the Bray–Curtis similarity measure on non-standardized square-root transformed total stem density data. Transformations carried out on the total stem density data served to reduce the impact of the most commonly recorded mangrove tree species. Similarities between mangrove forest communities were displayed using non-metric 2dimensional nMDS plots (Clarke and Green, 1988, Kruskal and Wish, 1978) based on the rank order of dissimilarities. Stress coefficients on each nMDS ordination plot indicate the extent to which the rank order of distances between samples in the nMDS ordination agrees with the rank order from the similarity matrices. The higher the stress coefficient, the higher the disagreement; for example, for 2D ordinations stress <0.05 gives an excellent representation while stress > 0.3 2-18 indicates the placement of points are close to being arbitrarily placed in the ordination space (Clarke and Warwick, 2001). To confirm the adequacy of the nMDS ordination, hierarchical agglomerative clustering using the group average method of the CLUSTER procedure in PRIMER 6 was performed on the Bray-Curtis similarity matrix prepared from total stem density data from each of the twenty experimental quadrats (Clarke and Warwick, 2001). To test for differences in community structure between and within forest types, formal significance tests for differences on the similarity matrices were performed using the PRIMER 6 ANOSIM permutation test (Clarke and Green, 1988, Clarke and Ainsworth, 1993). The mangrove species contributing most to the dissimilarities between and within forest types were then investigated using the similarities procedure (SIMPER) (Clarke and Ainsworth, 1993, Macintosh et al., 2002, Clarke and Warwick, 2001).The sum of the average dissimilarity divided by the standard deviation (Av. Diss./SD) generated by the SIMPER analysis was used as a measure of the consistency and power of the contribution of each mangrove species (Barlow et al., 2007, Clarke and Warwick, 2001). Due to the non-normal nature of the forest structure data, differences between univariate forest structure parameters (See Table 2-2) of groups established through nMDS ordination, were assessed using the Non parametric Mann Whitney U test (between Forest type differences) and Kruskal Wallis test (within Forest type differences). Multiple comparisons among experimental quadrats were Bonferoni adjusted. Correlations between forest structure parameters and hydroperiod parameters (inundation frequency and duration) were carried out using Spearman’s rank co-efficient. The Spearman’s rank co-efficient was also used to analyse the correlation between the axes of the nMDS ordination for within forest type groups and hydroperiod parameters (inundation frequency and duration). 2.3 Results This section of the chapter presents the results of the field studies of hydroperiod and forest structure. 2-19 Surface elevation and hydroperiod In the following subsection, results of surface elevation measurements and hydroperiod calculations are summarized in Table 2-3 and compared to international guidelines established by Watson (1928) where: 'low intertidal' represents areas inundated by medium high tides and flooded more than forty-five times a month (Watson classes 1 and 2), 'mid intertidal' represents areas inundated by normal high tides and flooded from twenty to forty-five times a month (Watson class 3), and 'high intertidal' represents areas inundated less than twenty times a month (Watson classes 4 & 5) (Duke et al., 1998). 2.3.1.1 Variation in hydroperiod between forest types Analysis of tidal inundation data indicates a clear difference in the surface elevation and hydroperiod between the two forest types. Experimental quadrats in the “Aegiceras forest” (average surface elevation 2.14m) were inundated for an average of 57% of time (13.68 hours per day) over a twelve-month period (Table 2-3). This compares to an average of 45.84% of time (11 hours per day) across a twelve month period in the “Sonneratia-Avicennia forest” (average elevation 2.45m above chart datum) (Mann Whitney U test, χ2(18)=4, p<0.001). Inundation frequency was not significantly different between the “Aegiceras forest” (59.30 inundations per month) to the “Sonneratia-Avicennia forest” (58.76 inundations per month) (Mann Whitney U test, χ 2(18)= 27.5, p>0.05). Table 2-3: Variation in duration and frequency of tidal inundation between and within the two “Aegiceras forest” Watson's class* Inundation frequency (No. month-1) Inundation duration (%) Mean surface elevation (m) above chart datum Distance from the Ngao river (m) Quadrat Transect Forest type forest types. 1 1 0-10 1.93 65.25 61.39 1/2 1 2 10-20 2.06 59.94 59.17 1/2 1 3 20-30 2.15 56.41 58.92 1/2 1 4 30-40 2.21 54.27 58.75 1/2 2-20 Watson's class* Inundation frequency (No. month-1) 1 5 40-50 2.31 50.82 58.75 1/2 2 1 0-10 1.98 63.13 60.72 1/2 2 2 10-20 2.08 59.08 59.08 1/2 2 3 20-30 2.16 56.04 58.75 1/2 2 4 30-40 2.22 53.97 58.75 1/2 2 5 40-50 2.3 51.12 58.75 1/2 2.14 57.00 59.30 Mean values for the “Aegiceras forest” “Sonneratia-Avicennia forest” Inundation duration (%) Mean surface elevation (m) above chart datum Distance from the Ngao river (m) Quadrat Transect Forest type Table 2-3 continued 1 1 0-10 2.28 51.86 58.78 1/2 1 2 10-20 2.39 47.96 58.75 1/2 1 3 20-30 2.49 44.22 58.75 1/2 1 4 30-40 2.56 41.74 58.75 1/2 1 5 40-50 2.51 43.42 58.75 1/2 2 1 0-10 2.28 51.86 58.78 1/2 2 2 10-20 2.39 47.96 58. 75 1/2 2 3 20-30 2.49 44.22 58.75 1/2 2 4 30-40 2.56 41.74 58.75 1/2 2 5 40-50 2.51 43.42 58.75 1/2 2.45 45.84 58.76 Mean values for the “Sonneratia-Avicennia forest” Note: *Watson’s class 1, 56-62 inundations per month, Class 2, 45-59 inundations per month. 2.3.1.2 Variation in hydroperiod within forest types Surface elevation and hydroperiod within the two forest types showed similar spatial trends. Both inundation duration and frequency decreased with distance inland from the Ngao river and corresponding increases in surface elevation. In the “Aegiceras forest”, transect 1, for example, the quadrat closest to the river (0-10m from the Ngao river) was inundated for an average of 2-21 65.25% of time each month while the most inland quadrat (40-50m from the Ngao river) was inundated on average only 51.12% of time each month. A significant negative correlation was observed between surface elevation and inundation duration and frequency in the “Aegiceras forest”. Both inundation duration and inundation frequency decreased with an increase in surface elevation along the tidal gradient (Spearmans rank co-efficient rs(9)= -1, p<0.001 and Spearmans rank co-efficient rs(9)= -0.963, p<0.001) respectively). In the “Sonneratia-Avicennia forest”, the quadrats closest to the river (0-10m) were inundated on 51.86% of time each month, while the quadrat 40m from the river was inundated only 41.74% of time each month. In contrast to the “Aegiceras forest”, however, inundation frequency in the “Sonneratia-Avicennia forest” was largely constant across the tidal gradient in all experimental quadrats which were flooded by all high tides (mean inundation frequency of 58.78 times per month) (Table 2-3, Figure 2-9). In the “Sonneratia-Avicennia forest”, a correlation was observed between surface elevation and inundation duration (Spearmans rank co-efficient rs(9)=-1, p<0.001) but not between surface elevation and inundation frequency (Spearmans rank co-efficient rs (9)= -0.471, p>0.05). 2.3.1.3 Seasonal variation in hydroperiod The highest and lowest spring tides in the two forest types coincided with the winter and summer equinoxes in April (High: 4.31 m / Low: 0.34m) and September (High: 4.34 m / Low: 0.37m) (Figure 2-10). In the “Aegiceras forest”, there was only slight variation in mean inundation duration (from a low of 52.3% of time in February to 59.36% of time in October) and inundation frequency (from a low of 54.27 inundations per month in February to 61.5 inundations per month in August) across the 2002 calendar year (Table 2-4 and Figure 2-10). Similarly in the “Sonneratia-Avicennia forest” there was only slight variation in inundation duration (from a low of 42.63% of time in February to 47.64% of time in December) and inundation frequency (from a low of 54 inundations per month in February to 60 inundations per month in January, March, May, October and December) (Table 2-4 and Figure 2-9). 2-22 Figure 2-9: (a) Higher surface elevation, (b) Higher mean inundation duration and (c) Constant Inundation frequency with distance from the Ngao river across transects in the “Aegiceras forest” relative to the “Sonneratia-Avicennia forest”: Note: MHWN (Mean High Water Neap), MSL (Mean Sea level) and MLWN (Mean Low Water Neaps). Aegiceras forest Transect 1 Transect 2 Sonneratia-Avicennia forest Transect 3 and 4 (a) (b) (c) 2-23 Table 2-4: Predicted inundation frequency and duration in each forest type showing variation across the calendar year. Month March April May June July August September October November December “SonneratiaAvicennia forest” February “Aegiceras forest” Hydroperiod parameter January Forest type Mean inundation duration (% of time) 57.43 52.30 58.14 56.69 58.42 56.30 57.89 58.23 57.14 59.36 56.36 57.94 Mean Inundation frequency (No. month-1) 60.09 54.27 60.95 59.45 61.05 58.45 60.18 61.50 59.23 60.59 58.68 60.23 Mean inundation duration (% of time) 47.44 42.63 47.14 45.44 47.02 45.93 47.48 47.18 44.83 46.97 45.92 47.64 Mean Inundation frequency (No. month-1) 60.00 54.00 60.00 58.00 60.00 58.00 60.00 59.36 57.73 60.00 58.00 60.00 Forest structure In the following subsection, variation in spatial patterns of forest structure is initially described between the “Aegiceras forest” and the “Sonneratia-Avicennia forest”, followed by a description of within forest variation in each of the two forest types. 2.3.2.1 Variation in forest structure between forest types Results from forest surveys indicate distinctive spatial patterns of forest structure in the two forest types. Forest composition In total 2313 mangrove trees and saplings were recorded in the twenty experimental quadrats sampled in the two forest types comprised of 1085 trees (Stem diameter≥2.5cm) and 1,228 saplings (stem diameter <2.5cm). 2-24 Figure 2-10: (a) Predicted tidal level over the 2002 calendar year (the shaded window represents the range of mean surface elevations across the two forest types), (b) mean inundation duration and (c) mean inundation frequency of experimental quadrats in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Aegiceras forest Sonneratia-Avicennia forest (a) (b) (c) Month 2-25 Trees and saplings were unevenly distributed across the two forest types, with 2131 stems (996 trees, 1135 saplings) in the “Aegiceras forest” and 182 stems (89 trees, 93 saplings) in the “Sonneratia-Avicennia forest”. Six mangrove species were recorded in total across the two forest types, representing four families and five genera (Table 2-5). Species composition was similar in the two forest types with the only difference in composition being the distribution of B. parviflora which was restricted to the “Aegiceras forest” and R. mucronata which was restricted to the “Sonneratia-Avicennia forest”. In addition to mature trees and saplings, significant numbers of seedlings were found which will form the basis of the subsequent chapter of the thesis (Chapter 3). Table 2-5: The six mangrove tree species recorded in the two forest types in the study area Family Species “Aegiceras forest” “Sonneratia-Avicennia forest” Sonneratiaceae Sonneratia alba Y Y Myrsinaceae Aegiceras corniculatum Y Y Rhizophoraceae Rhizophora apiculata Y Y Rhizophoraceae Bruguiera parviflora Y N Avicenniaceae Avicennia alba Y Y Rhizophoraceae Rhizophora mucronata N Y Comparison of community structure in the two forest types Analysis of the non-metric MDS (nMDS) ordination of square-root transformed mangrove tree stem density data in two dimensions indicated differences in community structure between the two forest types. The nMDS ordination (Figure 2-11a) showed a low stress value (0.11) indicating that the mangrove community is well represented by the ordination. The nMDS ordination showed that the ten experimental quadrats in the “Aegiceras forest” were clustered to the left of the nMDS plot, while the remaining ten quadrats from the “Sonneratia-Avicennia forest” were clustered to the right of the plot. The dendrogram prepared for the tree stem density data (Figure 2-11b) also showed strong spatial variation consistent with the nMDS ordination. The major groupings in the nMDS ordinations were supported by one-way ANOSIM results (Global R = 0.904, p<0.01) (Table 2-6). 2-26 Table 2-6: ANOSIM comparisons and SIMPER analysis of mangrove community structure between the “Aegiceras forest” and the “Sonneratia-Avicennia forest” confirming differences in community structure in the two forest types and highlighting the mangrove species responsible for the majority of the difference between the forest types. ANOSIM SIMPER Comparison R statistic p Dissimilarity Species Dissimilarity (DS)/ Standard deviation (SD) ratio (%) Mangrove trees in the “Aegiceras forest” and the 0.904 <0.01 74.62 A.corniculatum 3.74 (60.96%) S.alba 1.62 (14.06%) R.apiculata 0.95 (9.6%) A.alba 1.13 (7.83%) “Sonneratia- Avicennia forest" The SIMPER procedure also supported the nMDS ordination, indicating that mangrove tree communities in the “Aegiceras forest” were 74.62% dissimilar to those in the “SonneratiaAvicennia forest”. The SIMPER analysis for mangrove trees also indicated that more than 92.46% of the dissimilarity between the two forest types was due to differences in stem density of four species; A. corniculatum (60.96%, Diss/SD ratio of 3.74), S.alba (14.06%, Diss/SD ratio of 1.62), R. apiculata (9.6%, Diss/SD ratio of 0.95) and A.alba (7.83%, Diss/SD ratio of 1.13)(Table 2-6). Comparison of univariate forest structure parameters between forest types The significant differences in community structure between the two forest types identified through the nMDS ordination are in agreement with results of analysis of univariate forest structure parameters. These include size (height and stem diameter) of individual mangrove trees, stand characteristics (total basal area and total stem density) and relative forest structure values (relative density, dominance, frequency and importance values) between the two forest types. The structural attributes of the two forest types are given at Table 2-7 and Figure 2-12 and discussed below in relation to each of the nMDS groups, the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. 2-27 Figure 2-11: Association of experimental quadrats in the two forest types based on mangrove tree stem density (no. ha-1). (a) nMDS ordination from Bray Curtis similarities on square root transformed stem density data highlighting dissimilarity in community structure of the two forest types. The dissimilarity between forest types of 74.6% is shown with the coloured circle; (b) Dendrogram of the twenty experimental quadrats using group average clustering from Bray Curtis similarities. The two forest types separated at the 74.6% dissimilarity threshold (dotted line). (a) (b) 2-28 Significantly different results between the two forest types are highlighted in Table 2-7 as annotations and box plots in Figure 2-13 and Figure 2-14. When all species were pooled, individual mangrove trees were on average larger in stem diameter in the “Sonneratia-Avicennia forest” (9.54+/-1.62cm) than in the “Aegiceras forest” (6.93+/1.63cm) (Mann Whitney U test, χ2(18)=232, p<0.01). Of the individual species, A.corniculatum trees had larger mean stem diameter in the “Aegiceras forest” (3.80+/-0.15cm) than the “SonneratiaAvicennia forest” (2.85 +/- 0.20cm) (Mann Whitney U test, χ2(11)=0, p<0.01). In contrast, A. alba trees had larger mean stem diameter in the “Sonneratia-Avicennia forest” (17.15+/-6.2cm) than the “Aegiceras forest” (5.35+/-1.22cm) (Mann Whitney U test, χ2(2)= 1, p<0.05). The mean stem height of A.corniculatum trees was also greater in the “Aegiceras forest” (3.32+/-0.3cm) than the “Sonneratia-Avicennia forest” (2.17+/-0.29cm) (Mann Whitney U test, χ2(15)= 8, p<0.01). Mean total basal area of mangrove trees in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” was 17.87+/-3.06m2ha-1 and 10.10+/-3.25m2ha-1 respectively. When all species were pooled, no significant differences were apparent in total stand basal area between the two forest types (Mann Whitney U test, χ2(18)= 3026.5, p>0.05). Of the individual species, mean total basal area of A. corniculatum trees was however greater in the “Aegiceras forest” (10.46+/-2.76 m2ha-1) (representing 57.08% of total basal area) than in the “Sonneratia-Avicennia forest” (0.03+/-0.01 m2ha-1)(representing only 0.41% of total basal area) (Mann Whitney U test, χ2(18)=55, p<0.001). Mean total stem density in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” was 9,960+/-2,413 stems ha-1 and 890+/-145 stems ha-1 respectively. When all species were pooled, no significant differences were apparent in total stem density between the two forest types (Mann Whitney U test, χ2(18)=3026.5, p>0.05). Of the individual species, total tree density of A. corniculatum was significantly higher in the “Aegiceras forest” (mean 8,530+/-2214 stems ha1 )(representing on average 81.51% of stems), a forest type characterised by the presence of dense stands of small shrubby A. corniculatum trees, than in the “Sonneratia-Avicennia forest” where the species was sparsely distributed (40+/-22.11 stems ha-1) (3.97% of stems)(Mann Whitney U test, χ2(18)=6.75, p<0.001). Significant differences between the two forest types were also recorded in the relative density of S.alba trees which represented 56.44+/-9.96% of the total number of stems in the “Sonneratia-Avicennia forest” compared to only 11.93+/-4.28% in the “Aegiceras forest” (Mann Whitney U test, χ2(18)=8, p<0.001). 2-29 The proportion of stems falling into different size classes was significantly different in the two forest types (Table 2-8). A.alba had a greater proportion of trees with stem diameter larger than 10cm in the “Sonneratia-Avicennia forest” (4.35+/-2.08%) than the “Aegiceras forest” (0.83+/0.83% (Mann Whitney U test, χ2(18)=1621.5, p<0.05). In contrast, A.corniculatum had a greater proportion of smaller (stem diameter<10cm) stems in the “Aegiceras forest” (20+/-2.98%) than the “Sonneratia-Avicennia forest” (16+/-4.92%) (Mann Whitney U test, χ2(18)=675.5, p<0.001) (Table 2-8, Figure 2-15). Visual examination of the break down in size classes (Figure 2-15a) also indicates differences in size class distributions with stem diameter size classes of S. alba heavily right skewed with the population characterised by trees with large basal areas with few saplings present while A. corniculatum, was left skewed. In the “Sonneratia-Avicennia forest” (Figure 2-15b) the distributions of S. alba, stem diameter size classes are right skewed with the trees with large basal areas with few saplings present while A.corniculatum had a left skewed distribution. Further significant differences between the two forest types were recorded for A. corniculatum which was the most broadly distributed species in the “Aegiceras forest”, found in 100% of experimental quadrats in the forest while restricted to only 30% of quadrats in the “SonneratiaAvicennia forest” (Mann Whitney U test, χ2(18)=8, p<0.001). None of the other species found were significantly different in their frequency of occurrence between the two forest types. Consistent with the high total basal area, density and frequency values, the mean importance values (sum of relative dominance, density and frequency divided by 300) of A. corniculatum trees in the “Aegiceras forest” (58.70+/-3.74%) were also significantly higher (Mann Whitney U test, χ2(18)= 55, p<0.001) than in the “Sonneratia-Avicennia forest” (4.35+/-2.31%). In contrast, S. alba trees had a higher mean importance value in the “Sonneratia-Avicennia forest” (Mean Importance value, 52.87+/-9.44%) than the “Aegiceras forest” (24.47 +/- 5.48%) (Mann Whitney U test, χ2(18)= 21, p<0.05). The Importance values of the other species recorded did not differ between forest types. 2-30 Table 2-7: Summary of forest structure parameters in the two forest types for mangrove trees (stem diameter≥2.5cm). Parameter Forest type Statist ic Species S.alba Mean Mean Total Basal Area (m2 ha-1) Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest Total B.parviflora R. apiculata A. alba R. mucronata 8.19 (1.22) 6.4(0.50) A.corniculatu m 3.32(0.3)** 2.17(0.29) 1.25(0) NP 3.30(0) 3.55(1.21) 3.70(2.2) 9.00(1.29) NP 4.43(1.15) 4.53 (0.65) 5.21 (0.54) Mean Mean 12.77 (5.56) 8.87(0.97) 3.8(0.15) ** 2.85(0.20) 4.75(0.19) NP 6.11(0.71) 9.91(2.5) 5.35(1.22) * 17.15(6.2) NP 5.54(0.74) 6.93 (1.63) * 9.54 (1.62) Max Max 51.57 26.10 9.23 3.24 6.68 NP 14.32 32.50 10.82 44.88 NP 8.28 Mean 6.33(1.89) 0.08(0.04) 0.87(0.45) 0.14(0.11) NP 17.87 (3.06) Mean 4.01(1.26) NP 1.59(1.44) 4.2(1.84) 0.27(0.14) 10.10 (3.25) Relative dominance (%) SonneratiaAvicennia forest Aegiceras forest 10.46(2.76) *** 0.03(0.01) Mean 32.32(9.09) 0.26(0.14) 9.53(4.68) 0.8(0.48) NP - Mean 55.18(10.96) NP 5.77(4.54) 32.13(9.56) 6.52(3.8) - Total Stem density (No. ha-1) SonneratiaAvicennia forest Aegiceras forest 57.08(7.93)** * 0.41(0.27) Mean 1070(385.88) 40(22.11) 280(136.46) 40(22.11) NP Mean 460(115.66) NP 100(78.88) 200(74.54) 90(37.86) 9,960 (2,413) 890 (145) Relative density (%) SonneratiaAvicennia forest Aegiceras forest 8530(2214) *** 40(22.11) Mean 11.93(4.28)* 0.57(0.43) 4.73(2.29) 1.25(0.87) NP - Mean 56.44(9.96) NP 8.78(6.2) 19.48(6.97) 11.33(5.13) - Mean Mean 80.00 (13.33) 100.00 (0) 100.00 (0)** 30.00 (15.28) 30.00 (15.28) NP 50.00 (16.77) 30.00 (15.28) 30.00 (15.28) 60.00 (16.33) NP 50.00 (16.67) 36.25 5.41) 33.75 (5.32) Relative frequency (%) SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest 81.51(3.73)** * 3.97(2.15) Mean 29.17 (5.86) 8.33 (4.30) 16.67 (5.96) 8.33 (4.30) NP - Mean 47.00 (9.11) NP 10.33(5.72) 18.67(5.24) 15.33(5.26) - Importance value (IV)(%) SonneratiaAvicennia forest Aegiceras forest 37.50 (3.57)*** 8.67(4.56) Mean 24.47(5.48)* 58.70(3.74)** * 4.35(2.31) 3.06(1.57) 10.31(4.04) 3.46(1.79) NP - 8.29(4.74) 23.43(6.89) 11.06(4.39) - Stem Height (m) Stem diameter (cm) Stem diameter (cm) Frequency SonneratiaMean 52.87(9.44) NP Avicennia forest Note: *P<0.05, **P<0.01, ***P<0.001, NP, Not present in forest type, All mean values are presented (+/- 1 S.E). 2-31 Figure 2-12: Difference in selected mangrove forest structure parameters of mature mangrove trees of different mangrove species in the “Aegiceras forest” (orange) and the “SonneratiaAvicennia forest” (green). A.corniculatum B.parviflora R.apiculata A.alba R.mucronata Mean basal area (m2 ha-1) Mean stem diameter (cm) S.alba Forest type 2-32 -1 Mean total stem density (No. ha ) Figure 2.12: continued 2-33 Figure 2-13: Significantly higher mean stem height, stem diameter, total basal area, total stem density and Importance value of A.corniculatum trees in the “Aegiceras forest” compared to the -1 Mean total basal area (m2 ha ) Aegiceras forest (18) =55, p<0.001 -1 2 χ Mean stem density (No. ha ) Mean stem diameter (cm) “Sonneratia-Avicennia forest”. Sonneratia-Avicennia forest Forest type 2-34 Figure 2-14: Significantly higher mean stem diameter of (a) mature trees (when all species are pooled)(b) A.alba trees and (c) relative density and Importance value of S.alba trees in the “Sonneratia-Avicennia forest” compared to the “Aegiceras forest”. (a) Relative density (%) Importance Value (IV) (%) Mean stem diameter (cm) Mean stem diameter (cm) (b) Forest type 2-35 Table 2-8: Stem size distribution of mangrove saplings and trees of the six mangrove species found in the two forest types. Data shown are the mean of experimental quadrats (n=20). Values are mean values (+/- S.E) and mean percentage of trees of that species in the forest type that fall into that stem size category (in brackets) Species S.alba A. corniculatum B.parviflora R.apiculata A.alba R.mucronata Subtotal Total Forest Stem size (cm) <2.5 (Saplings) (C0) 2.5-4 (C1) 4-6 (C2) 6-8 (C3) 8-10 (C4) 10-12 (C5) 12-14 (C6) 14-16 (C7) 16-18 (C8) 18-20 (C9) 20+ (C10) Total >10cm (%) <10cm (%) Aegiceras forest 200+/-86.93 (8.95) 130+/39.59 (15.98) 70+/-30 (8.55) 4.63+/2.34 10+/-10 (9.68) 0+/-0 (4.01) 580 14.76+ /-2.73 4.37+/1.21 0+/0(0.86) 40+/16.33 (10.82) 0+/-0 (0.55) 20+/13.34 (3.51) 20+/13.34 (6.59) 0+/-0 (0) 12.44+ /-2.52 19,490 570+/195.54 (75.17) 10+/-10 (10) 40+/-22.12 (12.8) 0+/0(2.75) 0+/-0(0) 0+/-0(0) 0+/0(0) 0+/0(0) 0+/0(0) 0+/-0(0) 0+/-0(0) 610 20.00+ /2.98** * 16.00+ /-4.92 0+/-0 SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest 2660+/782.48 *** (29.69) *** 0+/0(4.84) 50+/16.67 (19.14) 40+/16.33 (23.14) 0+/-0 (2.23) 1270 5500+/1383.16 *** (4.39) *** 60+/26.67 (4.97) 50+/26.88 (4.53) 0+/-0 (2.68) 0+/-0 (2.75) 10960+/3109.1 ***(52.93)* 190+/65.75 (4.11) 110+/34.81 (5.37) 340+/111.76 (1.6)** 0+/-0 (7.72) 120+/-35.91 (19.99) 360+/158.61 (10.44) 50+/-30.74 (11.3) 10+/-10 (9.87) SonneratiaAvicennia forest Aegiceras forest 250+/153.66 (4.03) 70+/-33.5 (5.08) 20+/-13.34 (5) NP 10+/10(5) NP 0+/-0(20) 0+/-0 (13.34) NP 0+/0(5) NP 0+/0(5) NP 0+/0(0) NP 0+/-0(0) 0+/-0(0) 50 NP NP 8+/3.60 0+/-0 0+/-0 NP 10+/-10 (10) NP 110+/-69.05 (11.94) 70+/-36.67 (6.03) 10+/-10 (4.15) 0+/-0(0) 390 9.61+/2.70 0.33+/0.24 0+/0(9.94) 10+/-10 (1.25) 0+/0(1.25) 0+/0(10.72 ) 0+/0(3.75) 0+/0(9.95) 10+/-10(10) 20+/13.34 (12.49) 30+/-30 (11.25) 10+/10(6.2) 10+/-10(10) 50+/40.14 (3.94) 10+/-10 (1.25) 0+/0(5.17) SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest 120+/59.26(6.72 ) 20+/-13.34 (1.25) 0+/0(3.75) 20+/20(0) 110 5.5+/2.88 2.08+/1.71 70+/-42.3 (28) 130+/-66.75 (16.29) 10+/-10 (14.37) 40+/-30.56 (6.84) 20+/-13.34 (2) 30+/-15.28 (3.56) 0+/-0(2) 0+/-0(15) 0+/0(0) 10+/10(3.4) 0+/0(0) 0+/0(2.92) 0+/-0(5) 110 0+/0(1.97) 330 0.83+/0.83* 4.35+/2.08 NP NP NP NP NP NP 30+/15.28 (2.02) NP 9+/3.79 6.78+/1.86 NP 20+/13.34 (5.77) NP 0+/0(0) 10+/-10 (5.36) 0+/-0(0) 40+/22.12 (2.5) NP 10+/-10 (10.68) 20+/13.34 (3.53) NP 0+/-0 0+/-0 100+/-39.45 (39.72) 20+/-13.34 (12.52) 20+/-13.34 (7) 10+/-10 (11.43) 0+/0(9.95) 0+/0(1.43) 0+/0(1.43) 0+/-0(0) 190 14.00+ /-3.92 0+/-0 5840 3180 180 70 0+/0(10.43 ) 60 0+/0(5.98) 11350 40+/22.12 (5.18) 590 20 0 0 20 21,310 9.84 0.97 930 180 120 200 130 70 60 20 40 0 70 1,820 9.51 1.80 12280 6020 3300 790 310 140 120 40 40 0 90 23,130 19.35 2.77 30+/15.28 (14.62) NP Note: *P<0.05, **P<0.01, ***P<0.001, NP, Not present in forest type. 2-36 0+/-0 (0.37) 0+/-0 0+/-0 Figure 2-15: Difference in stem size class distributions of mangrove species in (a) the “Aegiceras forest” and (b) the “Sonneratia-Avicennia forest”. Data shown are the mean of experimental quadrats (n=20) S.alba A.corniculatum B.parviflora R.apiculata A.alba R.mucronata (a) (b) 2.3.2.2 Variation in forest composition and structure within forest types Aegiceras forest Comparison of Community structure within groups in the “Aegiceras forest” The “Aegiceras forest” is characterised by open mudflats and then dense stands of S. alba and A.corniculatum saplings and trees growing in a narrow (50-60 m) belt, seaward of a larger area of forest dominated by a mixed community comprised of trees of Rhizophora and Bruguiera spp. trees at higher elevations and large individual S. alba trees interspersed by some unvegetated or lightly vegetated areas. Analysis of the non-metric MDS (nMDS) ordination of square-root transformed total stem density data from the “Aegiceras forest” in two dimensions, confirmed these observations of within-forest type differences in forest community structure. The nMDS ordination prepared 2-37 for the “Aegiceras forest” showed low stress values (0.05), indicating that the mangrove communities in the forest type are well represented by the ordination. In the nMDS ordination for the “Aegiceras forest”, (Figure 2-16a), four groups were established largely with distance from the Ngao river, group 1 (comprised of quadrats T1Q2, T2Q2, T1Q3, and T2Q3), group 2 (T1Q1, T1Q2), group 3 (T1Q5) and group 4 (T1Q4, T2Q4 and T2Q5). A dendrogram prepared for the “Aegiceras forest” stem density data (Figure 2-16b), also showed strong spatial variation in within forest data, consistent with the nMDS ordination. The major groupings in the nMDS ordination are supported by one way ANOSIM results (Global R = 0.663, p<0.05) (Table 2-9). Pairwise comparisons using ANOSIM indicated that there were only significant differences between the groups 1 and 2 (Global R = 0.796, p<0.05) whereas all other group combinations were not significantly different (Table 2-9). Similarly, the SIMPER procedure supported the nMDS ordination, indicating that the four groups could be discriminated from 23.27% to 52.85% dissimilarity (Table 2-9). Group 1 and 3 have the highest dissimilarity (52.85%) while the lowest species dissimilarity is 23.27% between groups 2 and 3. Group 1 and 2 had a dissimilarity of 45.44% with Group 1 discriminated from Group 2 by A.corniculatum which was found at high abundance in Group 1 compared to Group 2 (Diss/SD ratio, 3.34). The lack of a significant difference between the other groups is likely to be due to the fact that all groups shared a species in common (A.corniculatum). Table 2-9: ANOSIM comparisons and results of SIMPER analysis of mangrove tree community structure amongst groups within the “Aegiceras forest” confirming differences in community structure in nMDS groups and highlighting the mangrove species responsible for the majority of the difference between groups (where p<0.05). ANOSIM SIMPER Comparison R statistic p Dissimilarity Species Dissimilarity (DS)/ Standard deviation (SD) ratio (%) With distance in the “Aegiceras forest” 0.663 <0.05 - - - Groups 1 and 2 0.796 <0.05 45.44 A.corniculatum 3.34 (24.75) Groups 1 and 3 1 >0.05 52.85 - - Groups 1 and 4 1 >0.05 38.19 - - Groups 2 and 3 -0.778 >0.05 23.27 - - Groups 2 and 4 0 >0.05 35.70 - - Groups 3 and 4 1 >0.05 42.70 - - 2-38 Figure 2-16: Association of experimental quadrats in the “Aegiceras forest” based on total tree stem density (no. ha-1). (a) nMDS ordination from Bray Curtis similarities on square root transformed stem density data highlighting dissimilarity in community structure of the nMDS groups. (b) Dendrogram of the ten experimental quadrats using group average clustering from Bray Curtis similarities. (a) Axis 1 (b) Axis 2 2-39 Comparison of univariate forest structure parameters within groups in the “Aegiceras forest” The differences in forest structure within groups in the “Aegiceras forest” are in agreement with the results of statistical tests between univariate forest structure parameters between the four groups given at Table 2-10 and displayed in Figure 2-17. Significantly different results are highlighted in Table 2-10 as annotations and displayed as box plots in Figure 2-18. In particular the mean density and stem diameter of A.corniculatum trees were significantly different between nMDS groups (Kruskal Wallis test, χ2(3)=8.182, p<0.05) and (Kruskal Wallis test, χ2(3)=7.891, p<0.05) respectively. Group 1 includes four quadrats, located 20-40m from the Ngao river in both transects (T1Q2, T1Q3, T2Q2 and T2Q3). The group is characterised by the highest (although not significantly different after Bonferroni correction) density of A. corniculatum trees (16,400 +/- 1277.37 stems ha-1) amongst the four groups in the forest type. S.alba was also present in each quadrat in the group overtopping the smaller A.corniculatum trees while R.apiculata and B.parviflora were each present in two of the quadrats in the group. Group 2 includes the two quadrats closest to the Ngao river (T1Q1, T2Q1), 0-10m along the two experimental transects. The group is characterised by the presence of A.corniculatum and S.alba trees and areas of unvegetated mudflats. A.corniculatum trees in the group have the largest (although not significantly different after Bonferroni correction) mean stem diameter (4.58+/-0.07cm) amongst the four groups. Group 3 includes 1 quadrat (T1Q5) located furthermost inland quadrat of transect 1 (40-50m from the Ngao river) and at the highest surface elevation in the forest type. The group is characterised by the presence of four tree species, A.corniculatum, S.alba, B.parviflora and A.alba. A.corniculatum trees in the quadrat had the lowest (although not significantly different after Bonferroni correction) total tree density (1900 stems ha-1) and smallest (although not significantly different after Bonferroni correction) stem diameter (3.40+/-0cm) of A.corniculatum trees respectively of the four groups. Group 4 includes three quadrats (T1Q4, T2Q4 and T2Q5) located 30-50m inland from the Ngao river. Group 4 is characterised by the presence of three species, A.corniculatum, R.apiculata and A.alba trees and the passage of a small drainage channel running parallel to the Ngao river. 2-40 Table 2-10: Forest structure parameters of the five species found in nMDS ordination groups in the “Aegiceras forest” Species Parameter Group 1 Group 2 Group 3 Group 4 S.alba Mean Tree Height (m) 7.29+/-0.89 6.75+/-0 12.9+/-0 0+/-0 Mean Stem diameter (cm) 7.04+/-0.9 7.77+/-0.84 51.57+/-0 6.97+/-0 Mean Total Basal area (m2ha-1) 5.55+/-1.43 7.25+/-0.18 20.89+/-0 1.91+/-1.91 Mean Relative dominance 20.19+/-3.43 60.21+/-4.37 87.21+/-0 11.62+/-11.62 Mean Density of stems (no. ha-1) 1725+/-825 1350+/-250 100+/-0 333.34+/-333.34 Mean Relative density (%) 8.55+/-3.46 34.15+/-7.96 4.35+/-0 4.17+/-4.17 Mean Frequency (%) 100+/-0 100+/-0 100+/-0 33.34+/-33.34 Mean Relative frequency (%) 35.42+/-5.25 50+/-0 25+/-0 8.34+/-8.34 Mean Importance value (IV) (%) 21.39+/-3.05 48.12+/-4.11 38.86+/-0 8.04+/-8.04 Mean Tree Height (m) 3.74+/-0.36 4.28+/-0.51 2.65+/-0 2.37+/-0.25 Mean Stem diameter (cm) 3.81+/-0.07 4.58+/-0.07 3.4+/-0 3.43+/-0.11 Mean Total Basal area (m2ha-1) 20.19+/-1.9 4.88+/-1 1.83+/-0 4.12+/-0.98 Mean Relative dominance 77.98+/-3.01 39.8+/-4.37 7.63+/-0 57.24+/-11 Mean Density of stems (no. ha-1) 16400+/-1277.37 2650+/-450 1900+/-0 4166.67+/-731.06 Mean Relative density (%) 90.68+/-3.46 65.86+/-7.96 82.61+/-0 79.37+/-4.83 Mean Frequency (%) 100+/-0 100+/-0 100+/-0 100+/-0 Mean Relative frequency (%) 35.42+/-5.25 50+/-0 25+/-0 36.12+/-7.35 Mean Importance value (IV) (%) 68.03+/-2.78 51.89+/-4.11 38.41+/-0 57.58+/-6.8 Mean Tree Height (m) 1.25+/-0 0+/-0 0+/-0 NP Mean Stem diameter (cm) 4.9+/-0.2 0+/-0 4.46+/-0 NP 0.16+/-0.1 0+/-0 0.16+/-0 NP Mean Relative dominance 0.5+/-0.32 0+/-0 0.66+/-0 NP Mean Density of stems (no. ha-1) 75+/-47.88 0+/-0 100+/-0 NP Mean Relative density (%) 0.35+/-0.22 0+/-0 4.35+/-0 NP Mean Frequency (%) 50+/-28.87 0+/-0 100+/-0 NP Mean Relative frequency (%) 14.59+/-8.59 0+/-0 25+/-0 NP Mean Importance value (IV) (%) 5.15+/-3.04 0+/-0 10+/-0 NP Mean Tree Height (m) 0+/-0 0+/-0 0+/-0 3.3+/-0 Mean Stem diameter (cm) 7.39+/-1.25 0+/-0 0+/-0 5.26+/-0.51 0.3+/-0.18 0+/-0 0+/-0 2.5+/-1.03 Mean Relative dominance 1.34+/-0.82 0+/-0 0+/-0 29.98+/-5.3 Mean Density of stems (no. ha-1) 75+/-47.88 0+/-0 0+/-0 833.34+/-233.34 Mean Relative density (%) 0.44+/-0.26 0+/-0 0+/-0 15.2+/-0.67 Mean Frequency (%) 50+/-28.87 0+/-0 0+/-0 100+/-0 Mean Relative frequency (%) 14.59+/-8.59 0+/-0 0+/-0 36.12+/-7.35 Mean Importance value (IV) (%) 5.46+/-3.22 0+/-0 0+/-0 27.1+/-3.84 Mean Tree Height (m) 1.5+/-0 0+/-0 0+/-0 5.9+/-0 Mean Stem diameter (cm) 0+/-0 0+/-0 7.69+/-0 4.19+/-0.63 0+/-0 0+/-0 1.09+/-0 0.1+/-0.06 Mean Relative dominance 0+/-0 0+/-0 4.53+/-0 1.18+/-0.7 Mean Density of stems (no. ha-1) 0+/-0 0+/-0 200+/-0 66.67+/-33.34 Mean Relative density (%) 0+/-0 0+/-0 8.7+/-0 1.28+/-0.75 Mean Frequency (%) 0+/-0 0+/-0 100+/-0 66.67+/-33.34 Mean Relative frequency (%) 0+/-0 0+/-0 25+/-0 19.45+/-10.02 Mean Importance value (IV) (%) 0+/-0 0+/-0 12.74+/-0 7.3+/-3.8 A. corniculatum B. parviflora - Mean Total Basal area R. apiculata Mean Total Basal area A. alba Mean Total Basal area (m2ha-1) (m2ha-1) (m2ha-1) Note: *p<0.05, NP, not present, All mean values are presented (+/- 1 S.E) 2-41 p * * Figure 2-17: (a) Vegetation profiles of experimental transects in the “Aegiceras forest” showing visual representation of the mangrove forest from the banks of the Ngao river to 50 m inland, the elevation of the transects at 5 m intervals, degree of inundation at the spring tide and location of nMDS ordination groups in relation to each experimental transect. (b) Comparison of forest structure parameters of the five species found in nMDS ordination groups in the “Aegiceras forest” S.alba A.corniculatum B.parviflora R.apiculata A.alba R.mucronata (a) Mean stem diameter (cm) Mean stem height (m) (b) 2-42 Figure 2-17 Continued Forest type 2-43 Figure 2-18: Variation in stem diameter and total stem density of A.corniculatum trees between nMDS ordination groups in the “Aegiceras forest”. nMDS Group Relationship between community structure and hydroperiod parameters in the “Aegiceras forest” Rank correlation of hydroperiod parameters with the first two nMDS ordination axes for the “Aegiceras forest” showed a significant correlation between both inundation duration and frequency (Spearman’s rank co-efficient, rs (9)=-0.758, p<0.05 and Spearman’s rank co-efficient, rs(9)=0.666, p<0.05 respectively) and Axis 2 of the ordination (Table 2-11, Figure 2-19). No correlation was found between these hydroperiod parameters and Axis 1 of the nMDS ordination (Spearman’s rank co-efficient, rs(9)=0.091, p>0.05) (Spearman’s rank co-efficient rs(9)=0.252, p>0.05). Bubble plots and biplots (Figure 2-20) allow inundation frequency and duration to be overlaid on nMDS plots to illustrate their correlation with the nMDS ordination axis. Table 2-11 Spearman’s rank correlation coefficient of hydroperiod parameters along two nMDS ordination axes for mangrove vegetation. Significance levels are shown as (*): p<0.05; Correlations without annotations are not significant. Hydroperiod parameters Axis 1 Axis 2 Inundation duration (% of time per month) 0.091 0.758* Inundation frequency (No. times month -1) 0.252 0.666* 2-44 Figure 2-19: Significant correlation between nMDS ordination axis 2 and Inundation duration and frequency in the “Aegiceras forest” Inundation duration (% of time per month) Comparison of univariate forest structure parameters and hydroperiod in the “Aegiceras forest” Rank correlation of hydroperiod parameters with univariate forest structure parameters for the “Aegiceras forest” showed a significant positive correlation between inundation duration and S.alba tree density and relative density, A.corniculatum stem diameter and stem height and R.apiculata stem diameter, with all parameters increasing as a result of increases in inundation duration. In contrast, the basal area, relative dominance, stem density, relative density and Importance value of A.alba trees showed a significant negative correlation, decreasing as a result of increases in inundation duration (Figure 2-21). 2-45 Figure 2-20: Bubble plots and biplots overlaid on the nMDS plots of mangrove tree composition based on stem density in the “Aegiceras forest”. Green circles indicate (a) the extent of inundation duration and (b) frequency in each of the groups established by the nMDS ordination in the “Aegiceras forest” with larger circles signifying greater inundation duration or frequency. Axis 1 Axis 2 2-46 Figure 2-21: Significant correlations between forest structure parameters and Inundation duration in the “Aegiceras forest”. Inundation duration (% of time per month) 2-47 Figure 2-21 continued Inundation duration (% of time per month) 2-48 The “Sonneratia-Avicennia forest” Comparison of community structure within groups in the “Sonneratia-Avicennia forest” The “Sonneratia-Avicennia forest” is dominated by S. alba and A. alba trees and takes on a more open low density woodland forest with an understory of A.corniculatum saplings and trees. Analysis of the nMDS ordination of square-root transformed total stem density data from the “Sonneratia-Avicennia forest” in two dimensions, also indicated within-forest type differences in forest community structure. The nMDS ordination prepared showed low stress value (0.11), indicating that the mangrove communities in the forest types are well represented by the ordination. In the nMDS ordination for the “Sonneratia-Avicennia forest”, (Figure 2-22a), four groups were established; group 1 (T4Q5, T4Q3, T4Q2), group 2 (T3Q4, T4Q4), group 3 (T4Q1 and T3Q1) and group 4 (T3Q2, T3Q3 and T3Q5). The dendrogram prepared for the “Sonneratia-Avicennia forest” stem density data (Figure 2-22b), was consistent with the nMDS ordination. The major groupings in the nMDS ordination were supported by one way ANOSIM results (Global R=0.919, p<0.05). Pairwise comparisons using ANOSIM indicated that there were no significant differences between groups in the forest type (Table 2-12) due to the fact that all groups shared a species in common (S.alba). Similarly, the SIMPER procedure supported the nMDS ordination, indicating that the four groups could be discriminated from 60.18% to 40.74% dissimilarity. Table 2-12: ANOSIM comparisons of mangrove tree community structure amongst groups within the “Sonneratia-Avicennia forest” confirming differences in community structure in nMDS groups. ANOSIM SIMPER Species Dissimilarity (DS)/ Standard deviation (SD) ratio (%) - - - >0.05 60.18 - - 0.917 >0.05 48.33 - - Group 1 and 4 1 >0.05 56.79 - - Group 2 and 3 1 >0.05 55.51 - - Group 2 and 4 0.75 >0.05 42.31 - - Group 3 and 4 0.917 >0.05 40.74 - - Dissimilarity Comparison R statistic p With group in the “Sonneratia-Avicennia forest” 0.919 <0.05 Group 1 and 2 1 Group 1 and 3 2-49 Figure 2-22: Association of experimental quadrats in the “Sonneratia-Avicennia forest” based on stem density (no. ha-1). (a) nMDS ordination from Bray Curtis similarities on square root transformed stem density data highlighting dissimilarity in community structure of the nMDS groups. (b) Dendrogram of the ten experimental quadrats using group average clustering from Bray Curtis similarities on square root transformed stem density data. (a) a Axis 1 New data set Group Axis 2 average (b) b Transform: Square root Resemblance: S17 Bray Curtis similarity 60 Dissimilarity 40 F2T4Q1 F2T3Q1 F2T3Q5 F2T3Q2 F2T3Q3 F2T4Q4 F2T3Q4 F2T4Q5 F2T4Q3 0 F2T4Q2 20 Samples Group 1 Group 2 Group 4 2-50 Group 3 Comparison of univariate forest structure parameters within groups in the “SonneratiaAvicennia forest” The differences in forest structure within the “Sonneratia-Avicennia forest” are in partial agreement with the results of statistical tests between univariate forest structure parameters which are summarized in Table 2-13 and displayed in Figure 2.23. Significantly different results are highlighted in Table 2-13 as annotations and displayed as box plots in Figure 2-24. Group 1 includes three quadrats (T4Q2, T4Q3 and T4Q5) located 20-50m from the Ngao river in transect 4. The group is characterised by the presence of S.alba trees with a small number of R.apiculata trees in quadrat T4Q3 (20-30m from the Ngao river). Group 2 includes two quadrats (T3Q4 and T4Q4) located 30-40m from the Ngao river in both transects. These quadrats represented the highest surface elevation in the forest type. The group is characterised by mixed forest consisting of large individual S.alba trees and small numbers of R. mucronata and A.corniculatum trees. The mean density of R. mucronata trees in the group (200 stems ha-1) was the highest (although not significantly different after Bonferroni correction) amongst all groups in the “Sonneratia-Avicennia forest”. A.alba and the higher intertidal species R.apiculata were also present in the group in quadrat T4Q4. The mean relative density of A.alba in the group (3.85+/- 3.85% of all stems in the group) was the lowest (although not significantly different after Bonferroni correction) amongst the four groups in the forest. Group 3 includes two quadrats (T3Q1 and T4Q1) located closest to the Ngao river, 0-10m along the 2 experimental transects and the lowest surface elevation in the forest type. The group is characterised by the presence of four species, S.alba, A.alba, R.apiculata and A.corniculatum. The group is differentiated by the restricted presence of R.apiculata to quadrat T3Q1 and A.corniculatum to quadrat T4Q1. Group 4 includes three quadrats (T3Q2, T3Q3 and T3Q5) located 20-50m from the Ngao river in transect 3. The group is characterised by the presence of S.alba, A.alba and R.mucronata trees in all quadrats. The mean relative density of A.alba in the group (47.95 +/- 4.13%) of all stems in quadrats in the group) was the highest (although not significantly different Bonferroni correction) amongst the four groups in the forest. Similarly, the mean relative dominance (% of total basal area) of A.alba in the group (65.07 +/- 3.41%) of the total basal area of quadrats in the group) was the highest (although not significantly different after Bonferroni correction) amongst the four groups in the forest type. 2-51 Table 2-13: Forest structure parameters of the five species found in nMDS ordination groups in the “Sonneratia-Avicennia forest” Species Parameter Group 1 Group 2 Group 3 Group 4 S.alba Mean Tree Height (m) 6.65+/-1.4 6.4+/-0.76 6.31+/-1.02 6.23+/-1.14 Mean Stem diameter (cm) 10.39+/-3.08 8.69+/-0.2 9.76+/-0.5 6.91+/-1.07 Mean Total Basal area (m2ha-1) 5.3+/-2.61 1.34+/-0.06 8.9+/-2.35 1.27+/-0.45 Mean Relative dominance 96.13+/-3.88 32.13+/-27.77 60.66+/-13.69 25.95+/-3.75 Mean Density of stems (no. A. corniculatum R. apiculata 466.67+/-176.39 200+/-0 1000+/-300 266.67+/-33.34 Mean Relative density (%) 93.34+/-6.67 24.36+/-8.98 67.55+/-13.71 33.53+/-5.28 Mean Frequency (%) 100+/-0 100+/-0 100+/-0 100+/-0 Mean Relative frequency (%) 83.34+/-16.67 26.67+/-6.67 33.34+/-0 33.34+/-0 Mean Importance value (IV) (%) 90.94+/-9.07 27.72+/-14.47 53.85+/-0.01 30.94+/-3.01 Mean Tree Height (m) 1.55+/-0 2.14+/-0.47 1.82+/-0.27 2.88+/-0.83 Mean Stem diameter (cm) 0+/-0 2.93+/-0.32 2.71+/-0 0+/-0 Mean Total Basal area (m2ha-1) 0+/-0 0.07+/-0.02 0.06+/-0.06 0+/-0 Mean Relative dominance 0+/-0 1.38+/-1.12 0.66+/-0.66 0+/-0 Mean Density of stems (no. ha-1) 0+/-0 100+/-0 100+/-100 0+/-0 Mean Relative density (%) 0+/-0 12.18+/-4.49 7.7+/-7.7 0+/-0 Mean Frequency (%) 0+/-0 100+/-0 50+/-50 0+/-0 Mean Relative frequency (%) 0+/-0 26.67+/-6.67 16.67+/-16.67 0+/-0 Mean Importance value (IV) (%) 0+/-0 13.41+/-4.1 8.34+/-8.34 0+/-0 Mean Tree Height (m) 5.45+/-0 3.9+/-0 0+/-0 1.3+/-0 Mean Stem diameter (cm) 12.8+/-0 12.02+/-0 4.94+/-0 0+/-0 Mean Total Basal area (m2ha-1) 0.43+/-0.43 7.24+/-7.24 0.1+/-0.1 0+/-0 Mean Relative dominance 3.88+/-3.88 22.63+/-22.63 0.4+/-0.4 0+/-0 Mean Density of stems (no. A. alba R.mucrona ta ha-1) ha-1) p * 33.34+/-33.34 400+/-400 50+/-50 0+/-0 Mean Relative density (%) 6.67+/-6.67 30.77+/-30.77 3.13+/-3.13 0+/-0 Mean Frequency (%) 33.34+/-33.34 50+/-50 50+/-50 0+/-0 Mean Relative frequency (%) 16.67+/-16.67 10+/-10 16.67+/-16.67 0+/-0 Mean Importance value (IV) (%) 9.07+/-9.07 21.14+/-21.14 6.74+/-6.74 0+/-0 Mean Tree Height (m) 0+/-0 12.85+/-0 6.7+/-0 8.49+/-1.45 Mean Stem diameter (cm) 0+/-0 44.89+/-0 16.39+/-8.67 8.42+/-0.85 Mean Total Basal area (m2ha-1) 0+/-0 7.92+/-7.92 7.32+/-5.18 3.84+/-2.11 Mean Relative dominance) 0+/-0 24.76+/-24.76 38.3+/-13.94 65.08+/-3.42 * Mean Density of stems (no. ha-1) 0+/-0 50+/-50 300+/-100 433.34+/-145.3 * Mean Relative density (%) 0+/-0 3.85+/-3.85 21.64+/-9.14 47.95+/-4.13 Mean Frequency (%) 0+/-0 50+/-50 100+/-0 100+/-0 * Mean Relative frequency (%) 0+/-0 10+/-10 33.34+/-0 33.34+/-0 * Mean Importance value (IV) (%) 0+/-0 12.87+/-12.87 31.09+/-1.61 48.79+/-1.85 Mean Tree Height (m) 2.85+/-0.15 3.25+/-0 0+/-0 5.89+/-2.11 Mean Stem diameter (cm) 0+/-0 5.35+/-0.35 0+/-0 5.67+/-1.34 0+/-0 0.51+/-0.31 0+/-0 0.57+/-0.37 0+/-0 19.12+/-18.51 0+/-0 8.99+/-3.32 Mean Total Basal area (m2ha-1) Mean Relative dominance Mean Density of stems (no. ha-1) 0+/-0 200+/-100 0+/-0 166.67+/-66.67 Mean Relative density (%) 0+/-0 28.85+/-21.16 0+/-0 18.53+/-3.15 Mean Frequency (%) 0+/-0 100+/-0 0+/-0 100+/-0 * Mean Relative frequency (%) 0+/-0 26.67+/-6.67 0+/-0 33.34+/-0 * Mean Importance value (IV) (%) 0+/-0 24.88+/-15.45 0+/-0 20.29+/-1.17 Note: *p<0.05, All mean values are presented (+/- 1 S.E) 2-52 * Figure 2-23: (a) Vegetation profiles of experimental transects in the “Sonneratia-Avicennia forest” showing visual representation of the mangrove forest from the banks of the Ngao river to 50 m inland, the elevation of the transects at 5 m intervals, degree of inundation at the spring tide and location of nMDS ordination groups in relation to each experimental transect. (b) Comparison of forest structure parameters of the five species found in nMDS ordination groups in the “SonneratiaAvicennia forest”. a S.alba A.corniculatum B.parviflora R.apiculata A.alba R.mucronata (a) Mean Relative dominance (%) Mean stem height (m) (b) Mean stem diameter (cm) b nMDS Group 2-53 Mean stem density (no. ha-1) Figure 2-23 continued nMDS Group 2-54 Figure 2-24: Comparison of significantly different forest structure characteristics of S.alba, A.alba and R.mucronata trees between nMDS ordination groups in the “Sonneratia-Avicennia forest”. All values Mean +/- 2 S.E. nMDS Group 2-55 Figure 2-24 continued. Relationship between community structure and hydroperiod parameters in the “SonneratiaAvicennia forest” Rank correlation of hydroperiod parameters with the first two nMDS ordination axes for the “Sonneratia-Avicennia forest” showed no significant correlations between hydroperiod and either axes of the nMDS ordination (Table 2-14). Table 2-14 Spearman’s rank correlation coefficient of hydroperiod parameters along two nMDS ordination axes for mangrove vegetation confirming no significant correlations between hydroperiod parameters and nMDS axes. Hydroperiod parameters Axis 1 Axis 2 Inundation duration (% of time per month) 0.074 0.482 Inundation frequency (No. times month -1) -0.406 0.058 2-56 Comparison of univariate forest structure parameters and hydroperiod in the “SonneratiaAvicennia forest” Rank correlation of hydroperiod parameters with univariate forest structure parameters for the “Sonneratia-Avicennia forest” showed no significant correlations between inundation duration and measured parameters of all species recorded in the forest type. 2.4 Discussion The lower intertidal forests in southern Thailand have been rarely studied and little quantitative data is available about the existence of spatial patterns of mangrove trees within these forests and the environmental factors responsible for these patterns. The findings of the current chapter of the thesis quantify the existence of spatial patterns within lower intertidal forests in the Ngao river, Ranong province, southern Thailand. The results show clear variation in forest composition, species distribution and forest structure as well as a clear relationship between forest structure and hydroperiod in the two forest types. These points will be expanded upon below with further discussion focusing initially on the spatial and seasonal variation in hydroperiod in the two forest types followed by discussion of variation in species composition and spatial patterns of forest structure in the two forest types and environmental parameters responsible for observed patterns of community structure. Spatial and seasonal variation in hydroperiod in the two forest types Surface elevation of lower intertidal communities in the Ngao river ranged from 1.93 to 2.56m above chart datum. 60% of experimental quadrats in the forest types were located below Mean Sea Level (MSL) of the estuary (2.34m) and the remaining 40% of quadrats were above MSL but below Mean Low Water Neaps (MHWN). Previous studies in the Ngao river estuary have not measured hydroperiod in a way which is comparable to the results of the current study instead stating that the S.alba-A.alba community is “around mean sea level” (Mochida et al., 1999) or using relative differences between transects (Komiyama et al., 1996). Compared to the wider mangrove literature, however, the range of surface elevations and the inundation duration (5165% of time per month) reported in the current study, place these forest communities at tidal elevations on the lower end of that usually considered to be mangrove habitat and in the lowest 2-57 two classes in Watson’s classification scheme (Watson, 1928). Literature values for mangrove inundation levels do however vary greatly. In a comprehensive study of the effect of hydroperiod on forest structure in Florida, for example, a range of inundation durations ranging from 3541 hrs per year (9.7% of time) to 8653 hrs per year (23.7% of time) were reported (Castañeda-Moya et al., 2013). Similarly, a study in Louisiana reported that inundation duration at experimental sites varied from 14.9% to 32.6% (Alleman and Hester, 2011), while Zhang et al. (1997) observed that inundation durations of 2.9- 47.5% were suitable for mangrove growth in China. Importantly, results of the study have also shown that mean surface elevations were significantly lower and inundation duration was higher in the “Aegiceras forest” than the “SonneratiaAvicennia forest”. A significant correlation also existed between elevation and inundation duration in the two forest types. Despite variations in tidal level during across the calendar year there was no significant correlation between season and tidal level or variation in tidal duration or frequency with season. This result contrasts with the findings of Castañeda-Moya et al. (2013) who reported significant differences in inundation frequency with season, with some locations only flooded by tides during the wet season. Variation in species composition and spatial patterns of forest structure in the two forest types The two forest types which formed the focus of the current study were low in mangrove tree species diversity with only six species found in total. The number and general composition of mangrove tree and sapling species in lower intertidal communities in the Ngao river is in the range of those found in other studies in the vicinity of the study area. Imai et al. (2006), for example, in their Ngao river study reported that species composition in the lower intertidal zone comprised a similar group of six species, with S.alba, A. alba, A. officinalis, R. mucronata, A.corniculatum and R.apiculata. The higher intertidal species, B. parviflora, which was only found in small numbers in the current study, was the only species recorded not recorded in the study by Imai et al. (2006). Other authors have confirmed that the lower intertidal zone is often the most species poor in mangrove environments, perhaps due to the challenging environmental conditions faced in these forest communities including exposure to erosion due to wind and waves during storms (Salmo et 2-58 al., 2013). The species found in the current study are consistent with those accepted as lower intertidal specialist species as defined by Duke et al. (1998) and Tomlinson (1987). nMDS ordination and cluster analysis of stem density values of mangrove vegetation in the two forest types enabled the identification of two clear groups. Significant differences were also recorded in many other forest structure parameters confirming the results of the ordination. In particular clear differences in both the size of individual A.corniculatum and A.alba trees, and stand characteristics of A.corniculatum and S.alba trees in the two forest types. Adding another dimension of complexity are the significant differences recorded in the size class distribution of mangrove stems between the two forest types suggesting that different patterns of recruitment exist in the two forest types. The size of individual trees of some species in experimental quadrats in the two forest communities was significantly different between the two forest types and within groups across the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. This suggests a possible preference of these species to environmental conditions provided by the different forest types and groups. A.corniculatum trees, for example, were on average larger in stem diameter and stem height in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, A.alba trees had larger mean stem diameter in the “Sonneratia-Avicennia forest”. Within the two forest types, the size of individual A.corniculatum trees also differed across groups across the forest types, highest in group 2, the group with the lowest density of A.corniculatum stems, suggesting that competition for resources in this group is potentially restricting the growth of mature A.corniculatum trees. In their study in Florida, Chen and Twilley (1998) also found a relationship between tree density and tree stem diameter, with tree density reported to decrease exponentially with increases in stem diameter between sites in the lower to upper estuary. The maximum stem diameter of individual trees in the two forest communities (51.57 cm) was consistent with values reported by (Imai et al., 2006) in the literature for their lower intertidal zone (65.2-87.1cm). For individual species, the maximum stem diameter of A. corniculatum trees in current study (maximum stem diameter 9.23cm) are consistent with other global studies where the species is typically described as taking a shrubby habit (Chapman, 1976) as did a study in Indonesia which observed that A.corniculatum has “a shrubby habit with many thin stems” (Hinrichs et al., 2008) (although the values in the current were smaller than the maximum stem 2-59 diameter recorded of 22.6cm in that study). The maximum stem diameter of individual S.alba and A.alba trees (51.57 cm and 44.88 cm respectively) is also consistent with previous studies in the Ngao river which reported finding trees of these species with stem diameters in excess of 34cm (Macintosh et al., 2002). Studies in Indonesia reported lower maximum stem diameters for S.alba and A.alba (27.69 and 27.85cm respectively) (Hinrichs et al., 2008), possibly due to the large degree of human use of the mangroves in the vicinity of the Indonesian case study. Total stem density in the two forests reported in this study (9,960 stems ha-1 in the “Aegiceras forest” and 890 stems ha-1 in the “Sonneratia-Avicennia forest”) was much higher than that reported in other studies in the Ngao river by Imai et al. (2006) who recorded total stem densities of only 187.3-191.1 stems ha-1 for the lower intertidal forest component of their study. Total stem density of individual species in the “Aegiceras forest” was also much higher than those reported by Imai et al. (2006)(1-38.1 stems ha-1) largely due to the high total tree density of A.corniculatum trees (8,530 stems ha-1) while total tree density of A.corniculatum trees in the “SonneratiaAvicennia forest” (40 stems ha-1) was closer to the values reported by Imai et al. (2006). Total tree density of other species found by Imai et al. (2006) (S.alba from 44.4-111.1 stems ha-1, A.alba from 31.1-46.0 stems ha-1 and R.mucronata from 11.1-52.4 stems ha-1) were also in the range found in the current study. In comparison with other studies presented in the literature, total stem density values were similar to those recorded in Indonesia by Hinrichs et al. (2008) (0.80 individuals m-2) (equating to 8000 stems ha-1) and other regional studies reported in Hinrichs et al. (2008) (ranging between 70 and 7000 stems ha-1) in Thailand (Macintosh et al., 2002, Chasang, 1984), Papua New Guinea (Robertson et al., 1991) and Belize (Murray et al., 2003). Values in the current study were relatively low in comparison to studies in India and Malaysia which presented very high densities (4,700–44,500 stems ha-1 (Ashton and Macintosh, 2002, Satyanarayana et al., 2002). Within the two forest types, high densities of A.corniculatum trees were observed in the middle quadrats of the forest and low densities of A.corniculatum trees in quadrats closest to the Ngao river. The values for total stem density in both forest types were lower than those in higher intertidal zones of the forest communities dominated by R. apiculata (1776.7-2072.2 stems ha-1) and mixed Rhizophora-Bruguiera forests (1571.7 stems ha-1) recorded by Imai et al. (2006) in the Ngao river. Again this result is consistent with previously reported increases in tree density and tree stem diameter between sites in the lower to upper estuary (Chen and Twilley (1998). Jiménez and 2-60 Sauter (1991) also reported higher density and basal area of one species (Avicennia bicolor) in landward plots and higher density (and basal area) of Rhizophora racemosa in the seaward stands. The total stem basal areas of mangrove trees in the “Aegiceras forest” and the “SonneratiaAvicennia forest” were 17.87m2 ha-1 and 10.10m2 ha-1 respectively. The difference reflects the high densities of relatively small stem diameter A.corniculatum trees in the “Aegiceras forest” (10.46 m2 ha-1) compared to the “Sonneratia-Avicennia forest” (0.03m2 ha-1). Total stem basal areas were in the range of values found by (Hinrichs et al., 2008) in Indonesia (9.86 m2 ha-1) and the lower intertidal zone of Imai et al. (2006) (11.3-13m2 ha-1). Total stem basal area in the current study was lower than that found in other studies (again previously reported by Hinrichs et al. (2008) (13.3–67 m2ha-1) from studies in Malaysia, Thailand, Papua New Guinea and Belize). Change dynamics Observations of the stem diameter size classes of individuals of S. alba, A. corniculatum and A. alba in the two forest types, as well as in field observations, indicate that there are significant differences in the size distributions of these species between the two forest types which provides some interesting information about change and species replacement over time in these forests. In the “Aegiceras forest” for example, the population structure of the two dominant species in this forest type show that for S. alba, stem diameter size classes are heavily right skewed, with the population characterised by trees with large stem diameters with few saplings present. This is indicative of a mature population of trees which is not regenerating. The distribution of stem diameter size classes of A. corniculatum, in contrast, had a left skewed population indicating that this species is regenerating and maintaining itself at this location. This data, as well as visual observations made by the author of isolated tall S. alba trees growing throughout the middle zones of the “Aegiceras forest” dominated by A. corniculatum suggest that that “secondary succession” has taken place and A. corniculatum has replaced S. alba throughout much of this forest type with S. alba only continuing to dominate sites at the extremities of the lower intertidal community. In the “Sonneratia-Avicennia forest”, a similar demographic pattern exists although at an earlier stage than the “Aegiceras forest”. In this forest type, S. alba stem diameter size classes are heavily right skewed with the population characterised by trees with large basal areas with few saplings 2-61 present. In contrast A.corniculatum trees had a left skewed population (although overall density of this species were much lower than in the “Aegiceras forest”) indicating that this species is regenerating and maintaining itself at this location and possible expanding its range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” (Figure 2-25). These patterns of distribution in the two forests could be used to explain future changes in species composition in the two forest types where species are replaced as the elevation of the substrate increases due to sediment deposition, and competition between different species or vegetation units. This concept will be tested in subsequent chapters of the study which focus on recruitment processes, particularly propagule supply and seedling establishment and growth. Figure 2-25: Observed and predicted patterns of secondary succession in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. a) Aegiceras forest showing possible past forest profile and existing forest, b) Sonneratia-Avicennia forest showing existing and predicted forest profile. Key to colours in figure: Black- S.alba, Blue-A.corniculatum and Grey, A.alba. 2-62 Previous studies have described similar patterns of demographic change in mangrove populations over time. Studies in Florida for example describe competition between two species of mangrove, R. mangle and Laguncularia racemosa in a mangrove estuary as a result of variation in soil fertility allowing R. mangle to successfully outcompete L.racemosa at sites with lower nutrient levels (Castañeda-Moya et al., 2013). Another study described the colonization of high intertidal habitats in Florida by R.mangle and L.racemosa where it was hypothesized (through observations of living and dead tree densities and densities of saplings and seedlings) that Laguncularia was being replaced by Rhizophora in lower intertidal areas (with greater degree of inundation) but able to compete effectively in the higher intertidal areas (with reduced degrees of inundation) (Ball, 1980). Environmental parameters responsible for observed patterns of community structure The clear difference in community and forest structure both between and within the two forest types raises the question as to what factors are responsible for these distinct spatial patterns. The distribution of groups across the two forest types appears initially to suggest that hydroperiod plays the foremost role in determining the community structure in the two forests, with forest communities in the lower surface elevation “Aegiceras forest” possibly more tolerant to greater rates of tidal inundation than those in the higher surface elevation “Sonneratia-Avicennia forest”. Closer analysis of within forest structure in the two forest types in groups with distance from the Ngao river in the “Aegiceras forest”, suggests that hydroperiod plays an important role in determining the community structure within this forest type with S.alba and A.alba trees dominating groups closest to the Ngao river in the “Aegiceras forest” while those groups dominated by A.corniculatum trees were found at more inland sites towards the middle of the forest. The presence of S.alba and A.alba trees at the sites with lowest surface elevations and highest tidal inundation durations in the “Aegiceras forest” suggests by extension that hydroperiod doesn’t play an important role in determining the overall difference in community structure in the two forests. Instead it appears that all three dominant species may have a broad tolerance to hydroperiod in the two forest types and that within forest spatial patterns have arisen due to other environmental factors such as variation in physical effects at exposed locations in sites close to the Ngao river compared to more sheltered inland sites. At the more exposed sites, it 2-63 appears that S.alba and A.alba trees are able to outcompete A.corniculatum trees while at inland sheltered sites, the opposite pattern is occurring with A.corniculatum the better competitor. Ball (1980) reported a similar interaction between hydroperiod and salinity in further studies of mangrove communities in Florida (Ball, 1980). This concept will be tested in subsequent chapters of the study which focus on propagule dispersal and development and seedling establishment and growth between and within the two forest types. Extensive evidence for the relationship between hydroperiod and forest structure exists in the mangrove literature (Krauss et al., 2006) and duration of tidal inundation is considered by many researchers to be the “most obvious parameter which varies across the tidal zone and is most often cited as a cause of zonation” (Smith, 1992) and described by some authors as a “proxy for the wide variety of environmental conditions that affect plant growth including soil salinity, redox potential and water logging” (Ellison, 2001). As to the environmental factors responsible for observed spatial patterns, tidal gradients such as those observed at the study sites are known to introduce physiological effects on mangrove plants through waterlogging (Giglioli and Thornton, 1965, Clarke and Hannon, 1967), and reduced soil oxygen levels (Hutchings and Saenger, 1987), increased soil redox values and resulting in transformation of chemical elements essential for plant growth (Clough, 1992). Through their study in the Ngao river estuary, Imai et al. (2006) also identified a clear pattern of forest structure (composition, density and basal area) across the tidal gradient from the S.alba dominated community (which is similar in composition and structure to the “Sonneratia-Avicennia forest” of the current study), to forests dominated by Rhizophora, Bruguiera and Ceriops higher in the tidal gradient abutting terrestrial forests. Unfortunately no detailed analysis was carried out of forest structure within the “S.alba dominated community” and A.corniculatum forests don’t seem to have been included in Imai et al. (2006)’s surveys. In addition no measurements were made of hydroperiod to enable comparison with the current study’s results. Previous studies have also compared tolerances to tidal inundation of some of the species found in the current study. In southern Australia, for example, distributions of Avicennia marina and A. corniculatum were differentiated along gradients of soil moisture, and A.corniculatum observed to 2-64 be less prevalent in saturated conditions due to reduced tolerance for waterlogged conditions than A.marina (Saintilan, 1998). In contrast, S. alba has been described by Ball and Pidsley (1995) as commonly found in perennially waterlogged, muddy banks at the lowest elevations colonized by mangroves suggesting a high tolerance to reduced soil conditions. In a further recent study of the relationship between duration of inundation and forest structure in Northern Australia, S. alba dominated areas inundated for more than 50% of the time, R. stylosa in areas inundated between 10 and 40%, and the higher intertidal species Ceriops tagal was reported to have the greatest probability of dominance at locations where duration of inundation was less than 10% (Crase et al., 2013). These results are consistent with the observed role of S.alba in the current study as the species found at the extremities of the lower intertidal. Studies in Florida have also shown that duration of inundation is the dominant factor controlling distribution of mangrove species with R. mangle typically occupying the lower intertidal and A. germinans occupying less frequently inundated sites (McKee, 1993b, Lewis, 2005). A study by Koch (1997), further quantified the hydroperiod differences between mangrove communities in Florida noting that actual inundation durations of between 50 and 75% of time often result in populations of R.mangle, while L. racemosa and A. germinans are generally restricted to sites with inundation durations less than 50%. This relationship between duration of tidal inundation and spatial patterns is known to extend to adjacent ecosystems with surface elevation reductions as small as 5cm reported to be responsible for an increase of 15% time flooded resulting in changes from the Avicennia to Spartina dominated communities (Carlson et al., 1983). Evidence for the relationship between physical effects on mangroves and forest structure is also well documented in the mangrove literature. Physical effects on mangroves include the impact of wave action which can result in broken stems (Aksornkoae, 1993), undermining and washing away of seedlings (Clarke, 1995, Balke et al., 2013) sedimentation of pneumatophores, seeds and seedlings (Ellison, 1999, Oliver, 1982, Thampanya et al., 2002) and impacts to mangrove seedlings by floating debris. These impacts are likely to be more pronounced in the lower intertidal forest areas than those higher up in the tidal gradient due to the comparatively stronger mechanical effects of tides and wave currents in these areas that are inimical to propagule rooting and hence seedling establishment (Clarke and Allaway, 1993, Clarke and Myerscough, 1993). Variation in flooding tolerance of different mangrove species such as that observed in the current study, is well 2-65 accepted in the literature due to variation in root structure and lenticel density amongst species (Youssef and Saenger, 1996). In this regard, S.alba and A.alba trees, for example, have pneumatophores and cable roots while A.corniculatum trees have none of these adaptations (Hutchings and Saenger, 1987) which again suggests an enhanced tolerance of these 2 species to exposed locations. The above inferences about the environmental factors leading to the current observed spatial patterns across the two forest types are limited as it is difficult to prove causality from observations of mature mangrove trees as a result of: (1) physiological tolerances of seedlings may be much narrower than those of adults (Ball, 1988, Mckee et al., 1988); (2) changes in environmental conditions at a site over time such that adults persist but seedlings can no longer become established (Smith, 1992); and (3) potential interactions between these factors and other factors which also change across the tidal gradient such as dispersal, interspecific competition, predation and salinity making it difficult to separate the influence of a single factor from another (Smith, 1992, McKee, 1993a). The observations are however useful as a base on which to develop hypotheses for further controlled experiments in subsequent chapters to strengthen the inferences underlying causal mechanisms (Platt, 1964). The application of information developed in this chapter to mangrove restoration is discussed in detail in Chapter 5 of this study. 2.5 Conclusions The objective of the chapter was to answer the questions, “What is the hydroperiod in the two forest types and how does this vary spatially and seasonally across the sites?”; “What are the natural spatial patterns of forest structure within lower intertidal communities in the Ngao river?”; and “How do spatial patterns of forest structure in the two forest communities relate to hydroperiod in these communities?” Measurement of hydroperiod parameters confirmed that surface elevation was lower and hydroperiod parameters higher in the “Aegiceras forest” than that in the “Sonneratia-Avicennia forest”. Inundation duration and frequency both decreased with distance inland from the Ngao river and corresponding increases in surface elevation in both forest types and there was only slight seasonal variation in hydroperiod. 2-66 Results of the study show that two forest types were both low in mangrove tree species diversity and assemblages of mature trees differed between the two forest types. Differences between the two forest types were also apparent in overall tree size (total stem diameter) significantly higher in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. For individual species, forest structure parameters (mean stem diameter, total stand basal area and Importance value) of the dominant species, A.corniculatum were greater in the “Aegiceras forest” than the “SonneratiaAvicennia forest”. In contrast S.alba and A. alba trees had larger average stem diameters and higher Importance values in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. Analysis of stem diameter size classes also suggests differences in the demographics of mature trees in the two forest types. The “Aegiceras forest” was characterised by a greater proportion of smaller A.corniculatum trees compared to the “Sonneratia-Avicennia forest” while the “Sonneratia-Avicennia forest” was characterised by a greater proportion of larger A.alba trees (Figure 2-26a). Within the “Aegiceras forest”, assemblages of mature trees differed with distance from the Ngao river with experimental quadrats grouped into zones along the tidal gradient. A.corniculatum tree densities and mean stem diameter were significantly different between “zones” . The duration of tidal inundation was significantly correlated with assemblages of mature trees, mean density of S.alba trees and mean stem diameter of A.corniculatum trees. These correlations and the differences in natural assemblages of mature mangrove trees between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest” suggest that assemblages of mature trees are related to hydroperiod and duration of inundation, but that forest communities in both the “Aegiceras forest and “Sonneratia-Avicennia forest” could have a broad tolerance to hydroperiod characteristics present within the two forest types. Within the two forest types, the physical effects associated with more exposed locations close to the Ngao river appear to determine within forest spatial patterns, with A.corniculatum trees seemingly less tolerant of these more exposed locations than A.alba and S.alba trees but a better competitor than A.alba and S.alba trees in inland more sheltered sites. 2-67 Figure 2-26: Visual summary of differences in univariate forest structure parameters (a) between the two forest types and (b) within each of forest types a) Evidence “Aegiceras forest” “Sonneratia Avicennia forest” -Mature trees Hydroperiod < > Composition = = ≠ Assemblages ≠ Tree stem diameter < > A.corniculatum > < S.alba = = A.alba < > Tree height = = A.corniculatum > < S.alba = = A.alba = = Total Stem density = = A.corniculatum > < S.alba < > A.alba = = Total Basal area = = A.corniculatum > < S.alba = = A.alba = = A.corniculatum > < S.alba < > A.alba = = Importance value 2-68 Figure 2-26 continued. b) Evidence Aegiceras forest Sonneratia Avicennia forest -Mature trees Group # G2 G1 G1 G4 G4 G3 G3 G1 G4 G1 G4 G2 G1 G4 Distance (m) 0-10 Hydroperiod > 10-20 20-30 30-40 40-50 40-50 0-10 10-20 10-20 20-30 20-30 30-40 40-50 40-50 >> >>> >>>> >>>>> >>>>> > >> >>>> >>> >>> Assemblages ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ ≠ >>> Stem diameter >>>> >>> >>> >> >> >> = = = = = = = = S.alba = = = = = = = = = = = = = = A.alba = = = = = = = = = = = = = = A.corniculatum = = = = = = = = = = = = = = S.alba = = = = = = = = = = = = = = A.alba = = = = = = = = = = = = = = A.corniculatum >> >>>> >>>> >>> >>> >> = = = = = = = = S.alba = = = = = = = = = = = = = = A.alba = = = = = = >>> > >>>> > >>>> >> > >>>> A.corniculatum = = = = = = = = = = = = = = S.alba = = = = = = = = = = = = = = A.alba = = = = = = >>> > >>>> > >>>> >> > >>>> A.corniculatum Tree height Stand density Stand Basal area 2-69 Information on forest structure obtained in Chapter 2 suggests that secondary succession of A.corniculatum trees as possibly taken place within the “Aegiceras forest” outcompeting S.alba trees in all quadrats except for the outer and most inner quadrats. In the “Sonneratia-Avicennia forest”, forest structure data suggests that the species A.corniculatum is expanding its range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” . The above inferences about factors leading to the current observed spatial patterns across the two forest types are useful as a base on which to develop hypotheses for further controlled experiments in subsequent chapters of the thesis. The specific application of information developed in this chapter to mangrove restoration is discussed in detail in Chapter 5 of this thesis. 2.6 References AKSORNKOAE, S. 1993. The Ecology of Mangroves, IUCN. ALLEMAN, L. K. & HESTER, M. W. 2011. Reproductive Ecology of Black Mangrove (Avicennia germinans) Along the Louisiana Coast: Propagule Production Cycles, Dispersal Limitations, and Establishment Elevations. Estuaries and Coasts, 34, 1068-1077. ALONGI, D. 2002. Environmental Conservation, 29, 331. ASHTON, E. C., MACINTOSH, D. J. & HOGARTH, P. J. 2003. A baseline study of the diversity and community ecology of crab and molluscan macrofauna in the Sematan mangrove forest, Sarawak, Malaysia. Journal of Tropical Ecology, 19. ASHTON, L. & MACINTOSH, D. J. 2002. Preliminary assessment of the plant diversity and community ecology of the Semantan mangrove. Forest Ecology and Management, 166, 111-129. BALKE, T., WEBB, E. L., VAN DEN ELZEN, E., GALLI, D., HERMAN, P. M. & BOUMA, T. J. 2013. Seedling establishment in a dynamic sedimentary environment: a conceptual framework using mangroves. J Appl Ecol, 50, 740-747. BALL, M. 1980. Patterns of secondary succession in a mangrove forest of Southern Florida. Oecologia, 44, 226-235. BALL, M. C. 1988. Salinity Tolerance in the Mangroves Aegiceras corniculatum and Avicennia marina I. Water Use in Relation to Growth, Carbon Partitioning, and Salt Balance. Aust. J. Plant Physiol., 15, , 447-64. BALL, M. C. & PIDSLEY, S. M. 1995. Growth Responses to Salinity in Relation to Distribution of Two Mangrove Species, Sonneratia alba and S. lanceolata, in Northern Australia. Functional Ecology, 9, 77-85. BARLOW, J., GARDNER, T. A., ARAUJO, I. S., ÁVILA-PIRES, T. C., BONALDO, A. B., COSTA, J. E., ESPOSITO, M. C., FERREIRA, L. V., HAWES, J., HERNANDEZ, M. I. M., HOOGMOED, M. S., LEITE, R. N., LO-MAN-HUNG, N. F., MALCOLM, J. R., MARTINS, M. B., MESTRE, L. A. M., MIRANDA-SANTOS, R., NUNES-GUTJAHR, A. L., OVERAL, W. L., PARRY, L., PETERS, S. L., 2-70 RIBEIRO-JUNIOR, M. A., DA SILVA, M. N. F., DA SILVA MOTTA, C. & PERES, C. A. 2007. Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences, 104, 18555-18560. BUNT, J. S. 1996. Mangrove Zonation: An Examination of Data from Seventeen Riverine Estuaries in Tropical Australia. Annals of Botany, 78, 333-341. CARLSON, P. R., YARBRO, L. A. & ZIMMERMAN C.F. MONTGOMERY, J. R. 1983. Pore water chemistry of an overwash mangrove island. Florida Science, 46, 239–249. CASTAÑEDA-MOYA, E., TWILLEY, R. R. & RIVERA-MONROY, V. H. 2013. Allocation of biomass and net primary productivity of mangrove forests along environmental gradients in the Florida Coastal Everglades, USA. Forest Ecology and Management, 307, 226-241. CHAPMAN, V. J. 1976. Mangrove vegetation, J. Cramer. CHASANG, H. 1984. Structure of a mangrove forest at Ko Yao Yai, Southern Thailand. In: SOEPADMO, E., RAO, A. & MACINTOSH, D. (eds.) Asian symposium on mangrove environment research and management. Kuala Lumpur University of Malaya and UNESCO. CHEN, R. & TWILLEY, R. R. 1998. A gap dynamic model of mangrove forest development along gradients of soil salinity and nutrient resources. Journal of Ecology, 86, 37-51. CHUNKAO, K., NIYOM, W. & TANTANAWARIT, S. 1985. Sediment and water flow affected to estuarine zone and its productivity. International Seminar on Environmental Factors in Agricultural Production, Thailand. CINTRON, G. & SCHAEFFER NOVELLI, Y. 1984. Methods for studying mangrove structure. In: SNEDAKER, S. & SNEDAKER, J. (eds.) The mangrove ecosystem: research methods. Paris: UNESCO. CLARKE, K. R. & AINSWORTH, M. 1993. A method of linking multivariate community structure to environmental variables. MARINE ECOLOGY PROGRESS SERIES, 92, 205-219. CLARKE, K. R. & GREEN, R. H. 1988. Statistical design and analysis for a 'biological effects' study. Marine Ecology Progress Series, 46, 213-226. CLARKE, K. R. & WARWICK, R. M. 2001. Change in marine communities: an approach to statistical analysis and interpretation, , Plymouth, PRIMER-E. CLARKE, L. D. & HANNON, N. J. 1967. The Mangrove Swamp and Salt Marsh Communities of the Sydney District: I. Vegetation, Soils and Climate. Journal of Ecology, 55, 753-771. CLARKE, L. D. & HANNON, N. J. 1970. The Mangrove Swamp and Salt Marsh Communities of the Sydney District: III. Plant Growth in Relation to Salinity and Waterlogging. Journal of Ecology, 58, 351-369. CLARKE, P. 1995. The population dynamics of the mangrove Avicennia marina; demographic synthesis and predictive modelling. Hydrobiologia, 295, 83-88. CLARKE, P. & MYERSCOUGH, P. 1993. The intertidal distribution of the grey mangrove {Avicennia marina) in southeastern Australia: The effects of physical conditions, interspecific competition, and predation on propagule establishment and survival. Australian Journal of Ecology 18, 307-315. CLARKE, P. J. & ALLAWAY, W. G. 1993. The regeneration niche of the grey mangrove (Avicennia marina): effects of salinity, light and sediment factors on establishment, growth and survival in the field. Oecologia, 93, 548-556. CLOUGH, B. F. 1992. Primary Productivity and Growth of Mangrove forests. In: ROBERTSON, A. I. & ALONGI, D. M. (eds.) Tropical Mangrove Ecosytems. Washington, DC: American Geophysical Union. 2-71 CRASE, B., LIEDLOFF, A., VESK, P. A., BURGMAN, M. A. & WINTLE, B. A. 2013. Hydroperiod is the main driver of the spatial pattern of dominance in mangrove communities. Global Ecology and Biogeography, 22, 806-817. DUKE, N., BALL, M. & ELLISON, J. 1998. Factors influencing biodiversity and distributional gradients in mangroves. Global Ecology & Biogeography Letters, 7, 27-47. DUKE, N. C. 2001. Gap Creation and regenerative processes driving diversity and structure of mangrove ecosystems. Wetlands Ecology and Management, 9, 257-269. ELLISON, A., MUKHERJEE, B. & KARIM, A. 2000. Testing patterns of zonation in mangroves: scale dependence and environmental correlates in the Sundarbans of Bangladesh. Journal of Ecology 88, , 813-824. ELLISON, A. M. 2001. Macroecology of mangroves: large-scale patterns and processes in tropical coastal forests. Trees, 16, 181-194. ELLISON, J. C. 1999. Impacts of Sediment Burial on Mangroves. Marine Pollution Bulletin, 37, 420426. ENGLISH, S., WILKINSON, C. & BAKER, V. 1997. Survey Manual for Tropical Marine Resources, 2nd Edition. (Townsville, Australian Institute of Marine Science. FIELD, C. D. 1998. Rehabilitation of Mangrove Ecosystems-An Overview. Marine Pollution Bulletin 37, 383-392. GIGLIOLI, M. E. C. & THORNTON, I. 1965. The mangrove swamps of Keneba, lower Gambia river basin. I. Descriptive notes on the climate, the mangrove swamps, and the physical condition of their soils. . Journal of Applied Ecology, , 2, 81-103. HINRICHS, S., NORDHAUS, I. & GEIST, S. J. 2008. Status, diversity and distribution patterns of mangrove vegetation in the Segara Anakan lagoon, Java, Indonesia. Regional Environmental Change, 9, 275-289. HUTCHINGS, P. & SAENGER, P. 1987. Ecology of Mangroves., St Lucia, University of Queensland Press. IMAI, N., TAKYU, M., NAKAMURA, Y. & NAKAMURA, T. 2006. Gap formation and regeneration of tropical mangrove forests in Ranong, Thailand. Plant Ecology, 186, 37-46. JIMÉNEZ, J. & SAUTER, K. 1991. Structure and dynamics of mangrove forests along a flooding gradient. Estuaries, 14, 49-56. KHEMNARK, C. 1995. Ecology and Management of Mangrove Restoration and Regeneration in East and Southeast Asia. Proceedings of the Ecotone IV Surat Thani, Thailand: Amarin Co. Ltd. KOCH, M. S. 1997. Rhizophora mangle L. Seedling Development into the Sapling Stage across Resource and Stress Gradients in Subtropical Florida1. Biotropica, 29, 427-439. KOMIYAMA, A., KONG SANGCHAI, J., PATANAPONPAIBOON, P., AKSORNKOAE, K. & OGINO, K. 1992. Socio-Ecosystem Studies on Mangrove Forests Charcoal Industry and Primary Productivity of Secondary Stands. TROPICS, 1, 233-242. KOMIYAMA, A., SANTIEAN, T., HIGO, M., PATANAPONPAIBOON, P., KONGSANGCHAI, J. & OGINO, K. 1996. Microtopography, soil hardness and survival of mangrove (Rhizophora apiculata BL) seedlings planted in an abandoned tin-mining area. . Forest Ecology and Management 81, 243–248. KRAUSS, K. W., DOYLE, T. W., TWILLEY, R. R., RIVERA-MONROY, V. H. & SULLIVAN, J. K. 2006. Evaluating the relative contributions of hydroperiod and soil fertility on growth of south Florida mangroves. Hydrobiologia, 569, 311-324. KRUSKAL, J. B. & WISH, M. 1978. Multidimensional scaling, Sage Publications, Beverly Hills. LEWIS, R. R., III. 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecological Engineering, 24, 403-418. 2-72 MACINTOSH, D. J., AKSORNKOAE, S., VANUCCI, M., FIELD, C. D., CLOUGH, B. F., KJERFVE, B., PAPHAVASIT, N. & WATTAYAKORN, G. 1991. Final Report of the Integrated Multidisciplinary Survey and Research Programme of the Ranong Mangrove Ecosystem. . UNDP/UNESCO regional project: research and its application in the management of the mangroves of Asia and the Pacific (RAS/86/120). Bangkok, Thailand: Funny Publishing Limited Partnership,. MACINTOSH, D. J., ASHTON, E. C. & HAVANON, S. 2002. Mangrove Rehabilitation and Intertidal Biodiversity: a Study in the Ranong Mangrove Ecosystem, Thailand. Estuarine, Coastal and Shelf Science, 55, 331-345. MACNAE, W. 1966. Mangroves in eastern and southern Australia. Australian Journal of Botany, 14, 67-104. MACNAE, W. 1969. A General Account of the Fauna and Flora of Mangrove Swamps and Forests in the Indo-West-Pacific Region. In: FREDERICK, S. R. & MAURICE, Y. (eds.) Advances in Marine Biology. Academic Press. MATSUI, N., SUEKUNI, J., NOGAMI, M., HAVANOND, S. & SALIKUL, P. 2010. Mangrove rehabilitation dynamics and soil organic carbon changes as a result of full hydraulic restoration and re-grading of a previously intensively managed shrimp pond. Wetlands Ecology and Management, 18, 233-242. MCKEE, K. L. 1993a. Soil Physicochemical Patterns and Mangrove Species Distribution--Reciprocal Effects? Journal of Ecology, 8, 477-487. MCKEE, K. L. 1993b. Soil Physicochemical Patterns and Mangrove Species Distribution--Reciprocal Effects? Journal of Ecology, 81, 477-487. MCKEE, K. L. 1995. Seedling recruitment patterns in a Belizean mangrove forest: effects of establishment ability and physico-chemical factors. Oecologia, 101, 448-460. MCKEE, K. L., MENDELSSOHN, I. A. & HESTER, M. W. 1988. Reexamination of Pore Water Sulfide Concentrations and Redox Potentials near the Aerial Roots of Rhizophora-Mangle and Avicennia-Germinans. American Journal of Botany, 75, 1352-1359. MIYAWAKI, A., OKUDA S, SUZUKI K & K, F. 1985. Phytosociological studies of mangrove vegetation in Thailand. In: MIYAWAKI (ed.) Ecological studies on the vegetation of mangrove forests in Thailand. Yokohama National University. MOCHIDA, Y., K, F., T, M., S, I., T, M., T, K. & P, P. 1999. A phytosociological study of the mangrove vegetation in the Malay Peninsula. Tropics, 8, 207–220. MUELLER-DOMBOIS, D. & ELLENBERG, H. 1974. Aims and Methods of Vegetation Ecology, John Wiley & Sons. MURRAY, M. R., ZISMAN, S. A., FURLEY, P. A., MUNRO, D. M., GIBSON, J., RATTER, J., BRIDGEWATER, S., MINTY, C. D. & PLACE, C. J. 2003. The mangroves of Belize: Part 1. distribution, composition and classification. Forest Ecology and Management, 174, 265279. NICKERSON, N. H. & THIBODEAU, F. R. 1985. Association between Pore Water Sulfide Concentrations and the Distribution of Mangroves. Biogeochemistry, 1, 183-192. OLIVER, J. 1982. The Geographic and Environmental aspects of Mangrove communities: Climate. In: CLOUGH, B. F. (ed.) Mangrove Ecosystems in Australia, structure function and management. Australian Institute of Marine Science, Cape Ferguson: Australian National University Press. PLATHONG, J. 1998. Status of Mangrove Forests in Southern Thailand. Wetlands International Thailand programme/ PSU. 2-73 PLATT, J. R. 1964. Strong Inference, Certain systematic methods of scientific thinking may produce much more rapid progress than others. Science, 146, 347-353. PUTZ, F. E. & CHAN, H. T. 1986. Tree growth, dynamics, and productivity in a mature mangrove forest in Malaysia. Forest Ecology and Management, 17, 211-230. RABINOWITZ, D. 1978b. Early Growth of Mangrove Seedlings in Panama, and an Hypothesis Concerning the Relationship of Dispersal and Zonation. Journal of Biogeography, 5, 113133. ROBERTSON, A. I., A., D. P. & P., D. 1991. Mangrove forest structure and productivity in the Fly River estuary, Papua New Guinea. Marine Biology, 111,, 147-155. SAINTILAN, N. 1998. Relationships between height and girth of mangroves and soil-water conditions in the Mary and Hawkesbury River estuaries, eastern Australia. Australian Journal of Ecology, 23, 322-328. SALMO, S. G., LOVELOCK, C. E. & DUKE, N. C. 2013. Assessment of vegetation and soil conditions in restored mangroves interrupted by severe tropical typhoon ‘Chan-hom’ in the Philippines. Hydrobiologia, 733, 85-102. SANTISUK, T. 1983. Taxonomy and Distribution of Terrestrial Trees and Shrubs in the Mangrove Formations in Thailand. . Natural History Bulletin, Siam Society. , 31 63-91. SATYANARAYANA, B., RAMAN, A. V., DEHAIRS, F., KALAVATI, C. & CHANDRAMOHAN, P. 2002. Mangrove floristic and zonation patterns of Coringa, Kakinada Bay, East Coast of India. Wetlands Ecology and Management, 10, 25-37. SEMENIUK, V. 1983. Mangrove distribution in Northwestern Australia in relationship to regional and local freshwater seepage. Vegetatio, 53, 11-31. SMITH, T. I. 1987. Effects of seed predators and light level on the distribution of Avicennia marina (Forsk.) Vierh. in tropical, tidal forests. Estuarine, Coastal and Shelf Science, 25, 43-51. SMITH, T. J. 1992. Forest structure. In: ROBERTSON, A. I. & ALONGI, D. M. (eds.) Tropical Mangrove Ecosytems. Washington, DC. : American Geophysical Union SMITH, T. J., III, CHAN, H. T., MCIVOR, C. C. & ROBBLEE, M. B. 1989. Comparisons of Seed Predation in Tropical, Tidal Forests From Three Continents. Ecology, 70, 146-151. SNEDAKER, S. C. 1982. Mangrove species zonation: why? In: SEN, D. & RAJPUROHIT, K. (eds.) Contributions to the ecology of halophytes. Springer Netherlands. THAMPANYA, U., VERMAAT, J. E. & DUARTE, C. M. 2002. Mar. Ecol. Prog. Ser., 237, 111. THULSTRUP, H. D. 1998. Man and the Biosphere: Southeast Asian experiences in coastal zone protected area management. . In: CENTRE FOR TROPICAL ECOSYSTEMS RESEARCH, U. O. A., DENMARK (ed.) Proceedings of the TCE-Project Workshop No. II Coastal Environmental Improvement in Mangrove/Wetland Ecosystems. Funny Publishing, . TOMLINSON, P. 1987. The Botany of Mangroves. . Brittonia, 39, 10-10. WATSON 1928. Mangrove forests of the Malay Peninsula, Singapore., Fraser and Neave. WOLANSKI, E., MAZDA, Y. & RIDD, P. 2013. Mangrove Hydrodynamics. Tropical Mangrove Ecosystems. American Geophysical Union. YOUSSEF, T. & SAENGER, P. 1996. Anatomical Adaptive Strategies to Flooding and Rhizosphere Oxidation in Mangrove Seedlings. Australian Journal of Botany, 44, 297-313. ZHANG, M., USTIN, L., REIMANKOVA, E. & SANDERSON, E. W. 1997. Monitoring Pacific Coast Salt Marshes Using Remote Sensing. Ecological Applications, 7, 1039-1053. 2-74 3. Dispersal and early development of mangrove propagules and seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand Abstract The dispersal and early development stages of the mangrove tree life cycle play an important role in determining eventual spatial patterns of mature trees within mangrove forests. Unfortunately, information on these early stages of the mangrove tree life cycle is lacking. Particular gaps in knowledge exist regarding the influence of environmental factors on these stages of the mangrove tree life cycle and patterns of dispersal and development of propagules and seedlings in lower intertidal forests. A combined twelve month observational and experimental propagule release study was conducted to better understand patterns of dispersal and development of mangrove propagules and seedlings in relation to environmental parameters within two lower intertidal mangrove communities in the Ngao river estuary, southern Thailand. Non metric MDS ordination showed that assemblages of propagules and developing seedlings differed between forest types largely due to the high density of A. corniculatum propagules and seedlings in the “Aegiceras forest” and high density of A. alba propagules and seedlings in the “Sonneratia-Avicennia forest”. Univariate analysis also showed that the density of A. corniculatum propagules and seedlings was significantly higher in the “Aegiceras forest” than the “Sonneratia-Avicennia forest” while, densities of A. alba propagules and seedlings and S. alba seedlings were higher in the “Sonneratia-Avicennia forest”. Propagule and seedling assemblages also differed with distance from the Ngao river in the “Aegiceras forest”. The greatest density of A.corniculatum propagules was found in the middle zone of the forest and the lowest density found in the most inland quadrats. A.alba propagules and seedlings in contrast were confined to quadrats closest to the Ngao river. There was clear seasonal variation in propagule and seedling abundance types in both forest types. A clear peak in production of A. corniculatum and A. alba propagules was evident in August and September followed by a peak in seedling density in subsequent months. Seasonal differences were also apparent between forest types. A. corniculatum, A. alba and S.alba seedlings tended to persist across the year in the “Sonneratia-Avicennia forest” but not in the “Aegiceras forest”. Peaks in production coincided with equinox periods and the lowest neap 3-1 tides suggest a preference for peak production of propagules in periods of lowest low water and relatively calm wind and high rainfall. Assemblages of propagules and seedlings in the forest types appear to be related to mature forest composition and the Importance value of conspecific trees in the canopy. Within the two forest types, a secondary role of duration of inundation appears to exist. Physical effects associated with more exposed locations close to the Ngao river appear to determine within forest propagule and seedling distribution patterns. A.corniculatum propagules and seedlings were seemingly less tolerant to these more exposed locations than those of A.alba and S.alba. A higher proportion of A. corniculatum propagules released into enclosures in the propagule release component of the study developed into mature seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, a higher proportion of A.alba propagules developed into mature seedlings in the “Sonneratia-Avicennia forest”, than the “Aegiceras forest”. This was largely due to the higher proportion of “failed seedlings” which germinated but were unable to establish into seedlings in the “Aegiceras forest”. In addition, a higher proportion of A.corniculatum propagules developed into mature seedlings 20-30m from the Ngao river and least 30-40m from the Ngao river. In contrast, a higher proportion of A.alba propagules developed into mature seedlings further away from the Ngao river (30-40m) and a lower proportion closest to the Ngao river, again largely due to the high proportion of undeveloped or “failed seedlings” in the quadrats closest to the Ngao river. The proportion of A.corniculatum propagules developing into mature seedlings was not significantly correlated with the duration of inundation in the “Aegiceras forest”. The proportion of A.alba propagules developing into mature seedlings was significantly negatively correlated with duration of inundation. Information on propagule and seedling populations obtained in Chapter 3 also suggest an apparent difference in reproductive strategy amongst lower intertidal species in the Ngao river. A.corniculatum appears to produce large quantities of propagules irrespective of mature tree dominance possibly to increase the chance of recruitment to the sapling stage and repress colonisation by other mangrove species. In the “Sonneratia-Avicennia forest”, it appears that this strategy is allowing A.corniculatum to expand its range and take over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” where large S.alba trees overtop a dense understory of A.corniculatum trees. These results add further weight to the theory that secondary succession is possibly taking place within these lower intertidal forests. 3-2 3.1 Introduction As mangroves primarily colonise both new areas, and repair damaged areas by dispersal and seedling recruitment (Duke, 2001), the dispersal and early development stages of the mangrove tree life cycle play an important role in determining eventual spatial patterns within mangrove forests (Tomlinson, 1987, Duke, 1996). These early stages of the mangrove tree life cycle are also usually characterised by the highest mortality (Patterson et al., 1997, Harper, 1977) as a result of biological (e.g. predators) and physical (e.g. tides, waves and currents) influences which ultimately affect recruitment to the subsequent seedling stage (Smith, 1992). Despite its importance, the dispersal and early development stages of the mangrove tree life cycle is not well studied and characteristics and tolerances of propagules and seedlings of the individual species that comprise lower intertidal forest communities in southern Thailand to environmental factors are not well established. These gaps in knowledge impede management and monitoring of these important mangrove communities and successful restoration of degraded lower intertidal forests. Improvements to our understanding of the dispersal and early establishment of mangrove propagules and seeds has direct application to mangrove restoration and help answer questions regarding for example the sufficiency of propagule supply for natural regeneration (Sousa et al., 2007) and understand the bottlenecks to seedling establishment (Friess et al., 2012, Balke et al., 2013). In this chapter, I describe a combined observational propagule dispersal and experimental propagule transplant study using two lower intertidal forest communities along the Ngao river, in the Kraburi river estuary, Ranong province, southern Thailand as case studies. The aim of the chapter was to describe how naturally dispersed propagules arrive and develop into seedlings in relation to variation in environmental parameters present in these communities. The experimental propagule transplant study assesses the role of predation and physical environmental factors on development of propagules of the three dominant species, A. corniculatum, A. alba and S.alba, transplanted to experimental enclosures installed in the different forest types. Specifically, in this chapter I propose to answer the following questions:  “What are the spatial and seasonal patterns of propagule and seedling assemblages in the two lower intertidal forest types?”; 3-3  “How are observed spatial patterns of propagule and seedling assemblages related to variation in environmental parameters between forest types and with distance from the Ngao river within each forest type?”; and  “Does protection from predation and physical disturbance influence the development of transplanted propagules in the two forest types?”. I answer these questions in the results section of the chapter by first describing environmental characteristics of these forest types (3.3.1), followed by descriptions of the propagule and seedling assemblages with forest type (3.3.2), within forest type, relative to distance from the Ngao river (3.3.3) and with month of monitoring (3.3.4). Section 3.3.5 goes on to examine the relationship between observed mangrove propagule and seedling assemblages and environmental parameters. Section 3.3.6 describes the results of enclosure experiments and variation in development of transplanted propagules with forest type, within forest type relative to distance from the Ngao river and the relationship between propagule and environmental parameters. 3.2 Methods 3.2.1 Study sites The study was conducted in two lower intertidal mangrove forests of the Ngao river, Ranong province, southern Thailand. The first forest type, the “Aegiceras forest” is characterised by open mudflats and then dense stands of A.corniculatum and S.alba saplings and trees growing in a narrow (50-60 m) belt, seaward of a larger area of forest dominated by a mixed community comprised of trees of Rhizophora and Bruguiera spp. trees at higher elevations, and large individual S. alba trees interspersed by some unvegetated or lightly vegetated areas. The second forest type, the “Sonneratia-Avicennia forest”, is an open woodland community comprised of large individual S. alba and A. alba trees with an understory of A.corniculatum saplings and trees and a less pronounced zonation pattern across the tidal gradient. 3.2.2 3.2.2.1 Experimental method Monitoring of 2 x 2m quadrats The ability of mangrove propagules to arrive and germinate and establish as seedlings was tested by establishing a series of twenty, 2 m x 2 m quadrats with distance from the Ngao river in the two forest types (Figure 3-1-Figure 3-4). 3-4 Experimental quadrats were distributed within 10 m x 10 m quadrats used previously to survey mature forest structure along the two 50 m transects in each forest type. Monitoring of quadrats took place on a monthly basis over a twelve-month period and involved a physical search for the presence of mangrove reproductive materials (propagules, seedlings and fruits) for 10 minutes and classification of materials by species and reproductive category. Categories used to classify reproductive material were: (1) mature but ungerminated propagules; (2) germinated propagules whose roots had not yet entered the soil; (3) immature seedlings (propagules whose roots had entered the soil and still had their cotyledons attached); and (4) mature seedlings. For ease of analysis, categories one and two were combined into the category “propagules” and categories three and four combined into the category “seedlings”. 3.2.2.2 Enclosure experiments The ability of propagules of the three dominant lower intertidal species, A.corniculatum, S.alba and A. alba to germinate and establish in the two forest types away from the influence of predators and the disruptive physical influence of tides and currents, was tested by installing a series of sixty enclosures across the two forest types within 10 m x 10 m quadrats. This quadrat size was used previously to survey forest structure along the two 50 m transects in each forest type (Figure 3-5). Three enclosures were distributed randomly in each 10 m x 10 m quadrat. Enclosures were constructed of wooden stakes of dimensions 70 cm x 50 cm x 50 cm and totally enclosed by mesh fishing netting (Figure 3-5) and dug into the soil to a depth of 10 cm. The mesh netting allowed unimpeded exposure to tides while also protecting propagules from predation by crabs and snails and ensuring that propagules could not move away from the site with currents and tidal movement. Twenty A. corniculatum and A. alba propagules and ten S.alba seeds were placed in the 60 enclosures (1,200, 1,200 and 600 in total for each species respectively), over different, but overlapping periods during the study period coinciding with the time that ripe propagules were available in the natural forest. Ripe propagules of A. corniculatum and A. alba were collected from the natural forest along the Ngao river by gently shaking trees allowing propagules to fall to the ground where they were collected by hand and inspected to ensure that they had had no signs of insect attack. Fruit of S. alba were collected when considered ripe, when the outer skin cracked and the main body of the fruit separated from the calyx (Ball and Pidsley, 1995). To obtain individual S.alba seeds, fruit were placed in a container of sea water for 48 hours, allowing the fruit to disintegrate and release the seeds. 3-5 Figure 3-1: Distribution of experimental 2 m x 2 m quadrats across experimental transects in the two forest types. Aegiceras forest (FT1) Sonneratia-Avicennia forest (FT2) Figure 3-2: Layout of 50m experimental transects in the two forest types showing 2 m x 2 m quadrats Aegiceras forest Sonneratia-Avicennia forest 3-6 Figure 3-3: Example of 2 m x 2 m quadrats located (a) Adjacent to the Ngao river in the “Aegiceras forest”, and (b) In the middle zone of the “Sonneratia-Avicennia forest” looking towards the Ngao river. (a) (b) 3-7 Figure 3-4: Soil surface of one quadrat containing germinating propagules of A. corniculatum. The proportion of propagules reaching each stage of development (using the same categories as for the observational study described in Section 3.2.2.1) was recorded on a monthly basis for four months but results presented in this chapter focus on development at the two-month stage as this was the period where the majority of propagules reached maturity. A summary of the experimental design for the two studies is provided in Table 3-1. Table 3-1: Summary table highlighting links between overall study objectives and survey design Focus of experiment Hypothesis tested Understanding Composition of of Experimental design propagule and Five 2 x 2m Data collected experimental established with Propagules and seedlings patterns of arrival of seedling assemblages is independent quadrats propagules and early of forest type, distance from the Ngao distance from the Ngao river in categorized into one of development of river and other environmental factors two 50m transects in each of two four development stages seedlings in lower which vary across the two forest types lower intertidal forest types (total intertidal mangrove (Duration of inundation, Importance of values of mature trees and stem monitored each month for 12 diameter of mature trees). months. forests 20 3-8 quadrats). Quadrats identified and Table 3.1 continued Focus of experiment Hypothesis tested Understanding of Physical factors development of influence the propagules of lower intertidal Experimental design Data collected 60 enclosures established with Counts of of distance from the Ngao river in and seedlings propagules of lower intertidal species. the two transects in two lower different intertidal stages. and predation development mangrove species under forest propagules of types. A.alba 20 propagules of development and conditions protected A.corniculatum, 10 S.alba seeds from predation and released into each enclosures physical removal by (3,000 in total). Development of tides. propagules and seedlings in each enclosure monitored each month for 4 months. Figure 3-5: Experimental enclosure used in the study showing (a). dimensions of experimental enclosure, (b). actual enclosure in the field in the “Sonneratia-Avicennia forest”. (b) (a) 3.2.3 3.2.3.1 Measurement parameters Environmental parameters Previous studies of the two lower intertidal communities conducted as part of Chapter 2 of this thesis identified hydroperiod (duration and frequency of tidal inundation) as the most 3-9 important environmental parameter which has potentially led to observed spatial patterns of mature mangrove trees within the two lower intertidal mangrove forests. For the present study, these two hydroperiod parameters have been combined with the Importance value of mature trees (combined relative density, relative dominance and relative frequency of mature trees) and total stem diameter, factors which are known from other studies (Krauss et al., 2008, Peterson and Bell, 2012) to play an important role in dispersal of mangrove propagules and seeds within lower intertidal mangrove forests and facilitation of seedling establishment respectively. A full description of the selected parameters is included in Table 3-2. Table 3-2: Definitions of environmental parameters as applied to the current chapter. Focus Parameter Definition Calculation method Hydroperiod Inundation duration Percentage time that per month Average length of time in a day that a site is (% time inundated that each site is inundated by inundated expressed as a percentage. per month) tides. Inundation Frequency frequency (No. of times inundated, per of (number of inundation inundations per Number of times in a month that each site is inundated. month) month) Importance of Importance value of The Importance value indicates Importance value of a species = (Relative density mature trees of mature trees of each the structural importance of an + different species individual species (%) individual frequency)/300% mangrove in the canopy as within a quadrat in relation to the other source of propagules Facilitation each species Relative dominance + Relative experimental species in the quadrat. of Total stem diameter Provides an index of horizontal Girth at Breast Height (GBH) measurements seedling of mature trees in structures in the forest which made in the field (adjusted for difference in tree establishment the serve experimental quadrat (cm) to trap retain morphology as recommended by English et al. facilitate (1997). Stem diameter (d) was then calculated establishment (Peterson and Bell, using the mathematical relationship between 2012). the circumference of a circle and its diameter (d propagules and and = GBH /). 3.2.3.2 Univariate Propagule and Seedling distribution parameters Univariate mangrove propagule and seedling parameters were selected to describe the characteristics of propagules and seedlings of each species both individually and relative to other mangrove species recorded in the experimental quadrat. A detailed description of each selected parameter is included in Table 3-3. 3-10 Table 3-3: Definitions of univariate propagule and seedling parameters used in the current chapter. Focus Parameter Definition Calculation method Propagule Propagule The total number of propagules of each Physical count of all propagules in each 4m2 assemblage density (No. ha- species as a total found per unit area quadrat (What is 1) (hectare) ungerminated seeds/ propagules or 2) germinated arriving in which are either 1) mature but seeds/ propagules whose roots had not yet each entered the soil. Density of propagules species for experimental each species (no./ha-1)= no. x 2,500 m2 / 4 m2 quadrat?) Relative density Number of propagules of a given Relative density = (number of individuals of each (%) species per unit area expressed as a species/ total number of individuals of all species) percentage x 100 of total number of individuals of all species per unit area Relative Frequency of propagules or seedlings Relative frequency = (frequency of an individual frequency (%) of a species divided by sum of species/ total frequency of all species) x 100 frequencies of all species, expressed as a percent Seedling Seedling density The total number of seedlings of each Physical count of all seedlings in each 4m2 quadrat assemblage (No. ha-1) species as a total found per unit area which (hectare) (propagules whose roots had entered the soil and (What is are either: (1) immature seedlings developing still had their cotyledons attached); and (2) in each mature seedlings that had lost their cotyledonary experimental leaves. Density of seedlings for each species in quadrat) each size class (no. ha-1)= no. x 2,500 m2 4 m-2 Relative density Number of seedlings of a given species Relative density = (number of individuals of each (%) per species/ total number of individuals of all species) unit percentage area of expressed total as number a of x 100 individuals of all species per unit area Relative Frequency of seedlings or seedlings of Relative frequency = (frequency of an individual frequency (%) a species divided by sum of frequencies species/ total frequency of all species) x 100. of all species, expressed as a percentage 3.2.3.3 Enclosure parameters Parameters for the enclosure experiments were selected to describe the development of propagules released in experimental enclosures. A detailed description of each parameter is included in Table 3-4. 3-11 Table 3-4: Definitions of parameters used to describe propagule development in experimental enclosures in the current chapter. Parameter Definition Proportion The of Physical count of propagules of each species which were at one of 5 development of propagules released of stages: (1) ungerminated seeds/ propagules; (2) germinated seeds/ propagules whose propagules each species classified roots had not yet entered the soil; (3). Immature seedlings (propagules whose roots released according had entered the soil and still had their cotyledons attached); (4) mature seedlings that recorded at development stage. proportion Calculation method to had lost their cotyledonary leaves, and (5) Dead propagules. each development stage (%) 3.2.4 Data analysis Exploratory data analysis was conducted on environmental and forest structure parameters described in Tables 3-2 to 3-4 to explore for any errors and outliers in the data. The Shapiro Wilk test was used to determine whether the population was normally distributed and therefore whether parametric or non-parametric inferential statistical tests could be used. Graphical methods of analysis were also adopted including scatter plots, draftsman plots and frequency histograms to detect whether distributions of data for any variable was skewed and would benefit from transformation and view pairwise correlations. For the environmental datasets, data were normalized in PRIMER before being analysed using Principal component analysis (PCA) to identify which variables contribute to the variability between sites (Clarke and Warwick 2001). Mangrove propagule and seedling abundance data was analysed using both multivariate and univariate statistical tests using SPSS v.22 for Windows (Univariate analysis) and PRIMER 6 (Clarke and Warwick, 2001)(Multivariate analysis). Non-parametric multivariate techniques were used to explore for patterns in the structure of mangrove propagule and seedling assemblages across and within the two forest types. Similarity matrices were constructed using the Bray–Curtis similarity measure on nonstandardized square-root transformed total propagule and total seedling density data. Square root transformations carried out on the propagule and seedling density data reduced the impact of the most commonly recorded mangrove tree species. Similarities in mangrove propagule and seedling assemblages between and within the two mangrove forests and within mangrove communities over the twelve month monitoring period were displayed using non-metric 2-dimensional nMDS plots (Kruskal and Wish, 1978, Clarke and Green, 1988) based on the rank order of dissimilarities. Stress coefficients on each nMDS 3-12 ordination plot indicate the extent to which the rank order of distances between samples in the nMDS ordination agrees with the rank order from the similarity matrices. The higher the stress coefficient, the higher the disagreement; for example, for 2D ordinations stress <0.05 gives an excellent representation while stress>0.3 indicates the placement of points are close to being arbitrarily placed in the ordination space (Clarke and Warwick, 2001). To test for differences in propagule and seedling assemblages between and with distance from the Ngao river in both forest types, formal significance tests for differences on the similarity matrices were performed using the PRIMER ANOSIM permutation test (Clarke and Green, 1988, Clarke and Ainsworth, 1993). The ANOSIM test generates a value of R between -1 and +1 as an indication of how separate a priori defined groups are relative to each other. The mangrove species contributing most to the dissimilarities between groups were then investigated using the similarities procedure of PRIMER (SIMPER); (Clarke and Ainsworth, 1993, Macintosh et al., 2002, Clarke and Warwick, 2001), Clarke and Warwick, 2001). The sum of the average dissimilarity divided by the standard deviation (Av. Diss./SD) generated by the SIMPER analysis was reported as a measure of the consistency and power of the contribution of each mangrove species to the dissimilarity of the a priori defined group (Barlow et al., 2007, Clarke and Warwick, 2001). Due to the non-normal nature of the mangrove propagule and seedling abundance data, differences between univariate parameters of groups established through nMDS ordination were assessed using the Non parametric Mann Whitney U test (between forest type differences (n=2)) and the Kruskall Wallis test (within forest type with distance (n=5)) and with month of monitoring (n=12). Multiple comparisons among experimental quadrats were adjusted using the Bonferoni correction. The resemblance matrices of the mean propagule and density abundance data were compared to matrices of data using the RELATE procedure of PRIMER procedure (Clarke and Warwick, 1994). The BIOENV procedure of PRIMER was used to ascertain the reason for observed patterns of mangrove propagule and seedling assemblages (Clarke and Ainsworth, 1993, Clarke and Warwick, 2001). BIOENV links the multivariate biotic patterns to suites of environmental variables to determine the best correlations between the biotic and abiotic factors and identify which environmental variables are contributing the most to shifts in propagule and seedling assemblages in the two forest types (Clarke and Warwick, 2001). 3-13 Correlations between the most important environmental parameters identified through BIOENV and univariate propagule and seedling parameters was then assessed by calculating the Spearman’s Correlation co-efficient for each parameter individually. Logistic regression analysis was used to determine if there was a relationship between forest type and distance from the Ngao river and the proportion of propagules found at different development stages two months after release. 3.3 Results 3.3.1 Environmental parameters The environmental parameters described in Table 3-2 were subjected to Principal Components Analysis (PCA) to examine their spatial distribution across the two forest types as the basis for subsequent comparisons with observed patterns of mangrove propagule and seedling assemblages. Results of the PCA demonstrated that 76.7% of the total amount of variation in the environmental data collected for the study was determined by principal component axes one and two (Table 3-5). Within the PCA plot (Figure 3-6), the direction of the eigenvectors on the PCA plot showed that the Importance value of mature A.corniculatum trees (%), Inundation duration (%) and total tree stem diameter (cm) were well correlated with PC axis 1 (PC1) (0.525, -0.503 and -0.461 respectively) (Table 3-6). Experimental quadrats in the two forest types were grouped across PC1 with quadrats in the “Aegiceras forest” located to the left of the plot (signifying high inundation duration and importance of A.corniculatum trees and high total tree stem diameter values) and quadrats in the “Sonneratia-Avicennia forest” to the right of the plot (low inundation duration, low Importance values of A.corniculatum trees and low mature tree stem diameter values)(Figure 3-6). PC axis 2 (PC2), in contrast, was well correlated with the Importance value of S.alba (0.671) and inundation frequency (0.557) (Figure 3-6, Table 3-6). Further grouping of the experimental quadrats in the “Aegiceras forest” occurred across PC2 with quadrats located close to the Ngao river (quadrat 1) located at top of the plot (signifying high inundation frequency and Importance value of S.alba trees) and quadrats located further away from the Ngao river (quadrats 4 and 5) located at the bottom of the plot (signifying low inundation duration and low importance of S.alba trees) (Figure 3-6, Table 3-5). The experimental quadrats in the right hand side of the plot (“Sonneratia-Avicennia forest”) were closer together on the PCA plot and split based on quadrats with higher S.alba Importance values and those with higher A.alba Importance values (Figure 3-6). 3-14 Table 3-5: Eigenvalues for six environmental variables in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” showing that a large proportion of variation in the environmental variables was accounted for by PC1 and PC2. PC Eigenvalues % variation Cum% variation PC1 3.16 52.6 52.6 PC2 1.45 24.1 76.7 PC3 0.76 12.7 89.4 PC4 0.447 7.5 96.8 PC5 0.111 1.8 98.7 Table 3-6: Eigenvectors for six environmental variables in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” highlighting the direction of the correlation between environmental variables and the axes of the PCA. Variable PC1 PC2 PC3 PC4 PC5 Inundation duration (%) -0.503 0.191 -0.273 0.303 0.512 Inundation frequency (no. of times month-1) -0.297 0.557 -0.539 -0.240 -0.478 Total mature tree stem diameter -0.461 -0.242 0.224 0.621 -0.536 Importance value of mature S.alba trees 0.235 0.671 0.262 0.488 0.220 Importance value of mature A.corniculatum trees -0.525 -0.196 0.022 -0.203 0.408 Importance value of mature A.alba trees 0.340 -0.327 -0.718 0.431 0.090 3.3.2 Variation in composition and distribution of propagules and seedling assemblages with forest type 3.3.2.1 Composition In total 932 mangrove propagules (Table 3-7) and 1607 seedlings (Table 3-8) belonging to seven lower intertidal mangrove species were recorded in the 20 experimental 2 x 2 m quadrats sampled in the two forest types over the twelve-month period. A total of 887 propagules (95.17%) and 931 (57.93%) seedlings were recorded from the “Aegiceras forest” and 45 (4.83%) propagules and 677 (42.07%) seedlings from the “Sonneratia-Avicennia forest” (Table 3-7 and Table 3-8). 3-15 Figure 3-6: Principal Components Analysis (PCA) plot showing variation in environmental variables (a) across the two forest types, and (b) with distance from the Ngao river. Principal component axes 1 and 2 cumulatively account for 76.7% of the total variation present. The annotations adjacent to symbols in plot b indicate the distance of the experimental quadrat to the Ngao river (m). (a) (b) 3-16 Across both forest types A. corniculatum propagules were most abundant (95.92%), followed by A.alba (2.9%) and A.ilicifolius (0.75%)(Table 3-7). Propagules of the other threespecies were found in relatively small quantities (<1%). Similarly, across both forest types A. corniculatum seedlings were the most abundant (79.53%), followed by A.alba (15.74%) and A. ilicifolius (3.36%)(Table 3-8). Seedlings of the other three species were found in relatively small quantities (<1%). S.alba was only found at the seedling stage (<1% of total seedling quantities) while R.apiculata was only found as a propagule and restricted to the “Sonneratia-Avicennia forest”. B.parviflora was found only at the propagule stage in the “Aegiceras forest” and as a seedling in the “Sonneratia-Avicennia forest”. Both N.fructicans propagules and seedlings were restricted to the “Aegiceras forest”. Table 3-7: Composition of mangrove propagules of the six species in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” Forest type “Aegiceras forest” “Sonneratia-Avicennia forest” Species Count Cum. count % Cum. % % of both forest s comb ined Count Cum. count % Cum. % % of both forest s comb ined A. corniculatum 872 872 98.31 98.31 97.54 22 22 48.89 48.89 2.46 894 95.9 A.alba 7 879 0.79 99.10 25.93 20 42 44.44 93.33 74.07 27 2.90 A.ilicifolius 5 884 0.56 99.66 71.43 2 44 4.44 97.78 28.57 7 0.75 N.fructicans 2 886 0.23 99.89 100 0 44 0 97.78 0 2 0.21 B.parviflora 1 887 0.11 100 0 0 44 0 97.78 100 1 0.11 R.apiculata 0 887 0 100 100 1 45 2.22 100 0 1 0.11 Total 887 887 100 100 95.17 45 45 100 100 4.83 932 100 Total % Table 3-8: Composition of mangrove seedlings of the six species in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” Forest type Species “Aegiceras forest” “Sonneratia-Avicennia forest” Total % Count Cum. count % Cum % % of both Count Cumcount % Cum% % of both A.corniculatum 880 880 94.52 94.52 68.86 398 398 58.88 58.88 31.14 1278 79.5 A.alba 23 903 2.47 96.99 9.09 230 628 34.02 92.90 90.91 253 15.7 3-17 Table 3.8 continued Forest type Species “Aegiceras forest” “Sonneratia-Avicennia forest” Total % Count Cum. count % Cum % % of both Count Cumcount % Cum% % of both A.ilicifolius 19 922 2.04 99.03 35.19 35 663 5.18 98.08 64.81 54 3.4 N.fructicans 5 927 0.54 99.57 100 0 663 0 98.08 64.81 5 0.3 S.alba 4 931 0.43 100.0 25 12 676 1.78 99.85 75 16 0.06 B.parviflora 0 931 0 99.57 0 1 664 0.15 100 100 1 0 Total 931 931 100 100 57.93 676 676 100 100 42.07 1607 100 3.3.2.2 nMDS ordination of community structure in propagule and seedling assemblages in the two forest types Analysis of non-metric MDS (nMDS) ordinations of propagule and seedling density data indicated significant differences in propagule and seedling assemblages between the two forest types. The nMDS ordinations (Figure 3-7a and b) both showed low stress values (0.09 and 0.08 respectively) indicating that the community structure was well represented by the ordinations. For propagule density, the nMDS ordination, Figure 3-7a showed that the ten experimental quadrats in the “Aegiceras forest” were generally clustered to the right of the nMDS plot, while the remaining ten quadrats from the “Sonneratia-Avicennia forest” were clustered to the left of the plot. For seedling density, the two forest types were also located separately on the nMDS plot with experimental quadrats in the “Aegiceras forest” were located at the bottom of the plot and those from the “Sonneratia-Avicennia forest” at the top of the plot. The grouping of experimental quadrats by forest type in the nMDS ordinations for mangrove propagules and seedlings are supported by one way ANOSIM results (Global R=0.291, p<0.05 and Global R=0.413, p<0.01 for propagules and seedlings respectively (Table 3-9). The SIMPER procedure also supported the nMDS ordinations for propagules and seedlings, indicating that propagule distribution in the “Aegiceras forest” was 64.97% dissimilar to that in the “Sonneratia-Avicennia forest” while seedlings were 43.08% dissimilar (Table 3-9). The SIMPER analysis for mangrove propagules indicated that 92.5% of the dissimilarity between the two forest types was due to differences in propagule density of two species; A. corniculatum (74.18%, Diss/SD ratio of 1.66) and A.alba (18.31%, Diss/SD ratio of 0.99) (Table 3-9). 3-18 Figure 3-7: nMDS ordination showing clear differences in community structure between the two forest types based on: (a) propagule and (b) seedling density data. Plot (a) shows experimental quadrats in the “Aegiceras forest” to the right of the plot and “SonneratiaAvicennia forest” to the left of the plot while plot (b) shows experimental quadrats in the “Aegiceras forest” to the top of the plot and “Sonneratia-Avicennia forest” to the bottom of the plot (a) Axis 1 (b) Axis 2 3-19 Table 3-9: Summary of ANOSIM and SIMPER results showing significant differences in community structure in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” based on propagule and seedling density data. ANOSIM SIMPER Comparison R statistic p Dissimilarity Mangrove propagule density in the “Aegiceras forest” and the “SonneratiaAvicennia forest” 0.291 <0.05 64.97 Mangrove seedling density in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” 0.413 <0.01 Species responsible dissimilarity most for Diss/ SD ratio (%) A.corniculatum 1.66 (74.18) A.alba 0.99 (18.31) A.alba 1.82 (38.48) A.corniculatum 1.6 (37.33) Acanthus ilicifolius 1.27 (13.69) S.alba 1.17 (7.82) 43.08 The SIMPER analysis for mangrove seedlings indicated that 97.32% of the dissimilarity between the two forest types was due to differences in seedling density of four species; A. alba (38.48%, Diss/SD ratio of 1.82, A.corniculatum (37.33%, Diss/SD ratio of 1.60), A.ilicifolius (13.69%, Diss/SD ratio of 1.27) and S.alba (7.82%, Diss/SD ratio 1.17) (Table 3-9). 3.3.2.3 Comparison of univariate forest structure parameters between forest types The significant differences in mangrove propagule and seedling assemblages between the two forest types identified through the nMDS ordinations are in general agreement with results of analysis of univariate propagule and seedling density parameters. The propagule and seedling parameters for the two forest types are shown in Figure 3-8 and discussed below in relation to the forest groups established in the nMDS ordinations. Significantly different results between the two forest types are highlighted in box plots in Figure 3-9. Propagules When all species were pooled, there were no significant differences apparent in propagule density between the two forest types (Mann Whitney U test, χ2(18)= 3167, p>0.05). 3-20 For individual species, the mean and relative density of A. corniculatum propagules was significantly higher in the “Aegiceras forest” (mean density, 18,166.67 propagules +/- 8,125.82 ha-1) (mean relative density 95.81 +/-2.66% of all propagules in the forest type) than the “Sonneratia-Avicennia forest” where the species was sparsely distributed (mean density, 458.34 +/- 106.69 propagules ha-1) (mean relative density 47.11 +/-10.84%)(Mann Whitney U test, χ2(18)= 11, p<0.01 and Mann Whitney U test, χ2(18)=8, p<0.001 for mean propagule density and relative density respectively) (Figure 3-9a). The mean density of A. alba propagules was not significantly different between the two forest types (Mann Whitney U test, χ2(18) =31, p>0.05) however the relative density of A.alba propagules was greater in the “Sonneratia-Avicennia forest” (mean relative density 38.17 +/11.25%), than the “Aegiceras forest” (mean relative density 3.95 +/-2.69 %) (Mann Whitney U test, χ2(18)= 20, p<0.05) (Figure 3-9b). The mean and relative density of the other species of propagules was not significantly different between the two forest types. Seedlings When all species were pooled, the mean seedling density in the “Aegiceras forest” (2,424.48+/-958.99 seedlings ha-1) was significantly greater than that in the “SonneratiaAvicennia forest” (1,760.42+/-362.70 seedlings ha-1) (Mann Whitney U test, χ2(18)=2694.5, p<0.05). For individual species, the mean density of A. corniculatum seedlings was not significantly different between the two forest types (Mann Whitney U test, χ2(18) =45, p>0.05) however the relative density of A. corniculatum seedlings was significantly higher in the “Aegiceras forest” (mean relative density, 92.02 +/-4.02% of all seedlings in the forest type) than the “SonneratiaAvicennia forest” (mean relative density, 58.21 +/-3.95%) (Mann Whitney U test, χ2(18) =4, p<0.001) (Figure 3-9a). The mean and relative density of A. alba seedlings was also significantly higher in the “Sonneratia-Avicennia forest” (mean density, 4791.67 seedlings +/- 897.42 ha-1)(mean relative density, 34.12+/-3.96% of seedlings), than the “Aegiceras forest” (mean density, 479.17 +/232.51 seedlings ha-1) (mean relative density, 4.74+/-2.96% of seedlings)(Mann Whitney U test, χ2(18) =1, p<0.001 and χ2(18)=3, p<0.001 respectively). 3-21 The mean density of S. alba seedlings, although low in absolute quantities in both forest types, showed a contrasting spatial pattern, significantly higher in the “Sonneratia-Avicennia forest” (mean 250+/-60.55 seedlings ha-1)(mean density, 2.46+/-0.99% of seedlings), as compared to the “Aegiceras forest” (mean density, 83.34 +/- 34.03 seedlings ha-1) (Mann Whitney U test, χ2(18) =25, p<0.05)(Figure 3-9c). The mean and relative density of seedlings of the remaining species was not significantly different between the two forest types. Further significant differences between the two forest types were recorded in the relative frequency of occurrence of A. alba seedlings, which was 28.67+/-1.63 % in the “SonneratiaAvicennia forest”, compared to only 11.67+/-4.84% in the “Aegiceras forest” (Mann Whitney U test, χ2(18)=21, p<0.05) (Figure 3-9b). Figure 3-8: Variation in mean density of propagules and seedlings of the lower intertidal species found in the “Aegiceras forest” (orange colour) and the “Sonneratia-Avicennia forest” (green colour) over the twelve-month monitoring period. S.alba A.corniculatum B.parviflora Nypa fructicans R.apiculata A.alba R.mucronata Acanthus ilicifolius Forest type 3-22 Figure 3-9: Significant differences in mean density, relative density and relative frequency of (a) A.corniculatum, (b) S.alba and (c) A.alba propagules and seedlings between the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Of note are the mean, relative density and relative frequency of A.corniculatum propagules and seedlings which were greater in the “Aegiceras forest” (orange colour) than the “Sonneratia-Avicennia forest” (green colour) and relative density and frequency of A.alba propagules and seedlings which were greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. (a) Forest type 3-23 Figure 3-9 continued (b) A.alba (c) Forest type 3-24 3.3.3 3.3.3.1 Variation in propagule and seedling assemblages with month of monitoring Comparison of univariate propagule and seedling density parameters The “Aegiceras forest” In the “Aegiceras forest”, the mean density of A. corniculatum propagules varied significantly with month of monitoring, peaking in September (181,750+/-80,226.28 propagules per ha-1) (mean relative density, 79% of propagules in each experimental quadrat in that monitoring period) (Kruskal Wallis test, χ2(11)=56.653, p<0.001)(Figure 3-10a). Subsequent to this peak, propagule density quickly declined so that by January, propagules were not recorded in any of the experimental quadrats for the following four months of the year (Figure 3-10a). In contrast, the density of A. corniculatum seedlings peaked in the subsequent month of monitoring (October) (mean seedling density, 81,000+/-29,716.71 seedlings per ha-1)(mean relative density 72%) (Kruskal Wallis test, χ2(11)= 78.408, p<0.001)(Figure 3-10a). b). In contrast to the populations of propagules, the mean density of A.corniculatum seedlings experienced a more gradual decline with seedlings recorded in relatively small quantities for 11 of the 12 of the monitoring periods (Figure 3-10b). Patterns of abundance of A. alba propagules also varied across the 12 month monitoring period in the “Aegiceras forest” with a non-significant peak evident in propagule density (mean density, 1000+/-763.76 propagules per ha-1) (Kruskal Wallis test, χ2(11)=17.216, p>0.05) during the month of November (Figure 3-10a). The mean density of A.alba seedlings also peaked in November (mean seedling density, 3,750+/-1796.98 seedlings per ha-1) (Kruskal Wallis test, χ2(11)= 28.040, p<0.01) (mean relative density, 8.68%) (Figure 3-10b). Patterns of abundance of A.ilicifolius seedling abundance peaked in October (Mean density, 2750+/-1314.97 seedlings per ha-1) (Mean relative density 4.98%) (Kruskal Wallis test, χ2(11)= 39.010, p<0.001) (Figure 3-10b). S.alba seedlings were only recorded in the “Aegiceras forest” during four months of the year (April, July, August and October), and mean seedling density was not significantly different between month of monitoring (Kruskal Wallis test, χ2(11)= 8.207, p>0.05) (Figure 3-10b). The mean density of B.parviflora propagules and N.fructicans propagules and seedlings recorded in the forest type was not significantly different across the monitoring periods (Kruskal Wallis test, χ2(11)=11, p>0.05 and χ2(11)=11, p>0.05 respectively) (Figure 3-10a and b). 3-25 Figure 3-10: Comparison of the mean density of: (a) propagules and (b) seedlings of the lower intertidal mangrove species in the “Aegiceras forest” in each month of monitoring. Of note are the significantly different mean densities of A.corniculatum and A.ilicifolius propagules and seedlings and A.alba seedlings across the twelve month monitoring period. N=12, error bars are mean +/-2 S.E. (a) Month 3-26 Figure 3-10 continued (b) Month 3-27 The “Sonneratia-Avicennia forest” In the “Sonneratia-Avicennia forest” as well as differing in absolute quantities, the pattern of abundance of propagules and seedlings over the twelve month monitoring period differed substantially from the “Aegiceras forest” for the three dominant species. A. corniculatum propagules were only found during one month of the year (August)(mean density 5,500+/-1,280 propagules ha-1) (Kruskal Wallis test, χ2(11)=93.365, p<0.001) (Figure 3-11a). In contrast, A. corniculatum seedlings were recorded in all monitoring periods with a small peak from September to November (mean density 5,500+/-1,280 ha-1 to mean density 12.25+/-1765.8ha-1) (Kruskal Wallis test, χ2(11)= 32.54, p<0.001)(Figure 3-11b). Of the other species, A. alba propagules were only found in October (mean density 5000+/2204.79 propagules ha-1) (Kruskal Wallis test, χ2(11) =81.010, p<0.001) (Figure 3-11a). Similar to those of A. corniculatum, A.alba seedlings had a broad distribution, found in all monitoring periods with a peak in November (mean density 16,250+/-3768.47 seedlings ha-1) (Kruskal Wallis test, χ2(11) =68.707, p<0.001) (Figure 3-11b). No S.alba propagules were found in the forest type and S.alba seedlings had an insignificant peak in abundance in the month of October (12,250+/-559.01ha-1) and small quantities were found intermittently for another five months of the year (Kruskal Wallis test, χ2(11) = 18.134, p>0.05) (Figure 3-11b). A.ilicifolius propagules found in the forest type did not vary significantly with month of monitoring (Kruskal Wallis test, χ2(11)= 10.085, p>0.05) (Figure 3-11a). A.ilicifolius seedling density peaked in September (mean density, 4,500+/-1740.05 seedlings per ha-1) (mean relative density 22.6%) (Kruskal Wallis test, χ2(11)= 55.710, p<0.001) (Figure 3-11b). The mean density of R.apiculata propagules and B.parviflora recorded in the forest type seedlings (Figure 3-11a and b) did not vary significantly (Kruskal Wallis test, χ2(11)= 11, p>0.05 and χ2(11)= 11, p>0.05 respectively) across the monitoring periods. 3-28 Figure 3-11: Comparison of mean density of: (a) propagules and (b) seedlings of the 6 lower intertidal specialist species in the “Sonneratia-Avicennia forest” with month of monitoring. Of note are the significantly different mean densities of A.corniculatum and A.alba propagules and seedlings and A.ilicifolius seedlings over the 12 month monitoring period. N=12, error bars are mean +/-2 S.E). (a) Month 3-29 Figure 3-11 continued (b) Month 3-30 3.3.4 Variation in propagule and seedling assemblages with distance from the Ngao river within forest types 3.3.4.1 nMDS ordination of community structure in propagule and seedling assemblages with distance from the Ngao river in the two forest types Analysis of the nMDS ordinations for propagule and seedling density data separately for each forest type, indicated within forest type differences in propagule and seedling assemblages with distance from the Ngao river. Separate nMDS ordinations prepared for each forest type showed low stress values (0.01 for both propagules and seedlings in the “Aegiceras forest” (Figure 3-12) and 0.04 and 0.08 for propagules and seedlings respectively in the “Sonneratia-Avicennia forest”)(Figure not shown). The low stress numbers indicate that the propagule and seedling assemblages in the two forest types are well represented by the ordinations. The “Aegiceras forest” In the nMDS ordination for propagules in the “Aegiceras forest”, (Figure 3-12a), four groups were established along across the plot with the “a priori” established factor “distance” (from the Ngao river); group 1 (comprised of quadrats T1Q2, T2Q2, located 10-20m from the Ngao river), group 2 (T1Q3, T2Q3, T1Q4, T2Q4, located 20-30m from the Ngao river), group 3 (T1Q1 and T2Q1, 0-10m from the Ngao river) and group 4 (T1Q5 and T2Q5 (40-50m from the Ngao river). The distance groupings in the nMDS ordination were supported by one way ANOSIM results (Global R = 0.82, p<0.01)(Table 3-10). Pairwise comparisons using ANOSIM, however, indicated that that there were no significant differences between each of the ten group combinations (p>0.05). The SIMPER procedure supported the nMDS ordination, indicating that the four groups could be discriminated from 15.04% to 84.99% dissimilarity (Table 3-10). Group 2 and 5 have the highest dissimilarity (84.99%) while the lowest species dissimilarity was between groups 3 and 4 (15.04%). Dissimilarity between groups in the forest type was largely due to differences in abundance of A.alba propagules (Groups 1 and 3, 1 and 4, and 1 and 5) and A.corniculatum (Groups 1 and 2, 2 and 4, 2 and 5, 3 and 5 and 4 and 5). The nMDS ordination for seedlings in the “Aegiceras forest”, was similar to that for propagules with the exception that experimental quadrats in group 3 (T1Q3, T2Q3 and T1Q4, T2Q4) were spaced further apart in the seedling nMDS plot as compared to quadrats in the propagule nMDS 3-31 Figure 3-12: nMDS ordinations of:(a) propagule and(b) seedling density data showing association of experimental quadrats in the “Aegiceras forest” with distance from the Ngao river. Statistic shown is the Global R statistic. (a) Axis 1 (b) Axis 1 Axis 2 3-32 plot. The distance groupings in the seedling nMDS plot which demonstrated that the composition of distance groups in the seedling nMDS (Global R = 0.53, p<0.05) were less similar to the composition of groups in the propagule nMDS plot (Global R =0.82, p<0.01) (Table 3-10). Pairwise comparisons using ANOSIM indicated that there were no significant differences between each of the ten group combinations (p>0.05)(Table 3-10). The SIMPER procedure supported the nMDS ordination, indicating that the distance groups could be discriminated from 35.73% to 64.93% dissimilarity. Group 2 and 5 had the highest dissimilarity (64.93%) while groups 2 and 4 had the lowest species dissimilarity (35.73%). Dissimilarity between groups 1 and 2, 3 and 4, 3 and 5 and 4 and 5 was due to the presence/ absence of A.corniculatum seedlings while dissimilarity between groups 1 and 4, 1 and 5, 2 and 3, 2 and 4 and 2 and 5 was due to presence/ absence of A.alba seedlings. Table 3-10: ANOSIM and SIMPER comparisons of mangrove propagule and seedling assemblages with distance from the Ngao river in the “Aegiceras forest”. ANOSIM SIMPER Dissimilarity Species responsible dissimilarity most for R statistic p With distance group in the “Aegiceras forest” 0.82 <0.01 1,2 1 >0.05 74.70 A.corniculatum 7.02 (80.22%) 1,3 1 >0.05 52.52 A.alba 4.14 (20.80%) 1,4 1 >0.05 52.98 A.alba 10.43 (20.40%) 1,5 0.75 >0.05 39.09 A.alba 14.66 (55.62%) 2,3 1 >0.05 48.50 A.ilicifolius 3.99 (11.58%) 2,4 1 >0.05 49.25 A.corniculatum 74.26 (74.43%) 2,5 1 >0.05 84.99 A.corniculatum 11.40 (81.84%) 3,4 0 >0.05 15.04 A.corniculatum 51.84 (100%) 3,5 1 >0.05 60.76 A.corniculatum 4.91 (100) 4,5 1 >0.05 61.68 A.corniculatum 10.45 (100) With distance group in the “Aegiceras forest” 0.53 <0.05 1,2 1 >0.05 45.34 A.corniculatum 8.67 (74.31%) 1,3 0 >0.05 51.18 S.alba 2.88 (10.46%) Comparison Diss/ SD ratio (%) Propagules Seedlings 3-33 Table 3-10 continued ANOSIM SIMPER Dissimilarity Species responsible dissimilarity most for Comparison R statistic p 1,4 1 >0.05 44.67 A.alba 8.72 (32.61%) 1,5 0.5 >0.05 40.33 A.alba 7.74 (48.30%) 2,3 0.5 >0.05 45.11 A.alba 3.45 (13.84%) 2,4 1 >0.05 35.73 A.alba 5.76 (16.71%) 2,5 1 >0.05 64.93 A.alba 5.95 (11.17%) 3,4 0 >0.05 36.31 A.corniculatum 1.56 (75.49%) 3,5 -0.25 >0.05 37.96 A.corniculatum 1.4 (69.45%) 4,5 0.5 >0.05 39.16 A.corniculatum 2.61 (91.30%) Diss/ SD ratio (%) The “Sonneratia-Avicennia forest” The nMDS ordination for propagules and seedlings in the “Sonneratia-Avicennia forest”, (Figure not shown), showed no clear pattern of distribution of propagule and seedling assemblages with distance from the Ngao river. The absence of patterns in the nMDS were supported by results of the one way ANOSIM (R=0.035, p>0.05 and R=0.24, p>0.05 for propagules and seedlings respectively) (Table 3-11). Pairwise comparisons between groups were also not significantly different. 3.3.4.2 Comparison of univariate propagule and seedling density parameters with distance from the Ngao river The significant differences in the distribution of mangrove propagule and seedling assemblages with distance from the Ngao river identified through the nMDS ordinations for the “Aegiceras forest” are in partial agreement with results of analysis of univariate propagule and seedling density parameters between distance groups within the forest type. The propagule and seedling parameters for the two forest types are shown at Figure 3-13b and Figure 3-14b and discussed below in relation to the distance groups established in the nMDS ordinations. Significantly different results between quadrats within each of the two forest types are highlighted on Figure 313c and Figure 3-14c as annotations. 3-34 Table 3-11: ANOSIM and SIMPER comparisons of mangrove propagule and seedling assemblages in the “Sonneratia-Avicennia forest”. ANOSIM SIMPER Dissimilarity Species responsible dissimilarity >0.05 34.81 A.alba 1.27 (38.25%) 0 >0.05 31.80 A.alba 1.35 (48.02%) 1,4 0.25 >0.05 40.97 A.corniculatum 1.14 (54.71%) 1,5 0.125 >0.05 67.88 A.alba 3.12 (38.79%) 2,3 0.25 >0.05 42.13 A.alba 2.48 (79.87%) 2,4 -0.5 >0.05 44.66 A.corniculatum 1.18 (69.06%) 2,5 0 >0.05 71.09 A.corniculatum 1.01 (60.34%) 3,4 0.25 >0.05 46.06 A.alba 2.41 (58.91%) 3,5 0.25 >0.05 79.74 A.alba 2.63 (62.39%) 4,5 0.375 >0.05 86.05 A.alba 1.27 (54.14%) With distance group in the “SonneratiaAvicennia forest” 0.24 >0.05 1,2 0.5 >0.05 16.83 A.ilicifolius 1.83 (43.27%) 1,3 -0.5 >0.05 12.83 A.ilicifolius 1.36 (23.46%) 1,4 -0.5 >0.05 11.13 A.corniculatum 3.40 (14.52%) 1,5 0.5 >0.05 27.47 A.ilicifolius 2.12 (29.94%) 2,3 0.25 >0.05 15.28 A.ilicifolius 1.68 (30.44%) 2,4 1 >0.05 23.95 S.alba 9.71 (22.81%) 2,5 0 >0.05 21.14 A.alba 1.59 (42.70%) 3,4 0.5 >0.05 17.48 S.alba 4.94 (23.66%) 3,5 0.5 >0.05 27.89 A.ilicifolius 1.99 (18.41%) 4,5 0.5 >0.05 28.74 S.alba 3.50 (19.16%) R statistic P With distance group in the “SonneratiaAvicennia forest” 0.035 >0.05 1,2 -0.5 1,3 Comparison most for Diss/ SD ratio (%) Propagules Seedlings 3-35 The “Aegiceras forest” In the “Aegiceras forest”, the density of A. corniculatum propagules was significantly different between distance groups (Kruskal Wallis test, χ2(4)= 10.713, p<0.05)(Figure 3-13c. Propagules were found in the greatest (although not significantly different after Bonferroni correction) quantities 10-20m from the Ngao river (distance group 2) (65,104.16+/-7,395.83 propagules ha-1) and the lowest quantities in the innermost zone of the forest, 40-50m from the Ngao river (distance group 5) (625 +/-205.33 propagules ha-1) and closest to the Ngao river (distance group 1)(1,666.66+/1041.66 propagules ha-1) (Figure 3-13c). A. corniculatum seedlings followed a similar pattern, found in the greatest quantities 20-30m from the Ngao river (distance group 2)(46,770.83+/-4,479 seedlings ha-1) and the lowest in the innermost zone (distance group 5) (3854.16+/-1770.83 seedlings ha-1) and closest to the Ngao river (distance group 1 )(5,416.66 +/-208.33 seedlings ha-1). Differences in density of A.corniculatum seedlings between distance groups were not statistically significant (Kruskall Wallis test, χ2(4)= 5.217, p>0.05). A. alba propagules were found in low quantities in the “Aegiceras forest” and confined to the zone closest to the Ngao river (distance group 1 and 2)(312.5+/-104.1 propagules ha-1 and 416.66+/416.66 propagules ha-1 respectively although these differences were not statistically significant (Kruskall Wallis test, χ2(4)= 8.225, p>0.05). A. alba seedlings followed a similar pattern with the mean density of 1,770.83+/-104.16 and 625+/-208.33 seedlings ha-1 in distance group 1 and 2 respectively (Kruskall Wallis test, χ2(4)= 20.246, p<0.001) (Figure 3-13c). Mean density of A. ilicifolius propagules, also differed significantly, confined to the zone of the forest 10-20m from the Ngao river (distance group 2) (520.83+/-104.16 propagules ha-1) (Kruskall Wallis test, χ2(4=16.41, p<0.01) (Figure 3-13c). A. ilicifolius seedlings had a broader distribution across the forest (0-40m)(distance groups 1-3) (520.83+/-104.16, 1354.16+/-1145.83 and 104.16+/-104.16 respectively) and the mean density of this species was not significantly different across distance groups (Kruskall Wallis test, χ2(4) = 8.89, p>0.05). 3-36 Figure 3-13: (a) Vegetation profiles of experimental transects in the “Aegiceras forest”, (b) bar graphs showing variation in mean propagule and seedling density with distance from the Ngao river and (c) box plots showing significant differences in the density and frequency of propagules and seedlings of lower intertidal species with distance from the Ngao river S.alba A.corniculatum B.parviflora Nypa fructicans R.apiculata A.alba R.mucronata Acanthus ilicifolius (a) (b) Distance from the Ngao river (m) 3-37 Figure 3-13 continued (c) Distance from the Ngao river (m) No S. alba propagules were found in the “Aegiceras forest” over the twelve-month period. S.alba seedlings were distributed in small quantities in the lowest intertidal zone, 0-10m from the Ngao river (distance group 1) (208.33 seedlings ha-1) and the innermost zone of the forest (distance groups 4 and 5) (104.16 +/-104.16) in both quadrats) and not significantly different across distance groups (Kruskall Wallis test, χ2(4)=3.591, p>0.05). Of the other species, N. fructicans propagules were found at low densities (208.33+/-208.33 propagules ha-1), restricted to the zone 10-20m from the Ngao river (distance group 2) and not significantly different across distance groups (Kruskall Wallis test, χ2(4)=4, p>0.05). Similarly, N. 3-38 fructicans seedlings were found at low densities (520.83+/-104.16 seedlings ha-1) restricted to the zone 20-30m from the Ngao river (distance group 3) and not significantly different across distance groups (Kruskall Wallis test, χ2(4)=4, p>0.05). B.parviflora propagules were also found in very small quantities (104.16+/- 104.16 propagules ha1 ) and restricted to one quadrat of the forest (distance group 2, 10-20m from the Ngao river) and not significantly different across distance groups (Kruskall Wallis test, χ2(4)=4, p>0.05). The “Sonneratia-Avicennia forest” In the “Sonneratia-Avicennia forest”, A. corniculatum propagules were distributed across the forest type with no clear pattern apparent with distance from the Ngao river with a range of 208.33 propagules ha-1+/-208.33 to 729.16+/-104.16) (Kruskall Wallis test, χ2(4)= 0.831, p>0.05). In contrast, the density of A.corniculatum seedlings was significantly different across distance groups ranging from 10,520.83 +/-1,562.5 seedlings ha-1 in distance group 3 (20-30m from the Ngao river) to 4,270.83 +/-2,812.5 seedlings ha-1 40-50m from the Ngao river (distance group 5)(Kruskall Wallis test, χ2(4)= 31.474 p<0.001) (Figure 3-14c). There was a strong trend (although non-significant) towards greater density of A. alba propagules and seedlings in the middle zone of the forest (20-40m from the Ngao river) (distance group 3) (6,979.16+/-3020.83 for propagules and 1,354.16+/-520.83 seedlings ha-1 for seedlings) and 1020m from the Ngao river (distance group 2) (6,875+/-2,708.33 (propagules)(Kruskall Wallis test, χ2(4)=8.678, p>0.05). No S.alba propagules were found in the forest type while S.alba seedlings were found in low quantities (208.33 -416.66 seedlings ha-1) in all zones of the forest with the exception of the zone 30-40m from the Ngao river (distance group 4) (Kruskall Wallis test, χ2(4)=4.277, p>0.05). A.ilicifolius seedlings had a broad but not significantly different distribution in the “SonneratiaAvicennia forest” found in all distance groups with the highest density of seedlings 0-10m from the Ngao river (distance group 1) (Kruskall Wallis test, χ2(4)=5.658, p>0.05). No trends with distance from the Ngao river were apparent for the other species of propagules/ seedlings were found in this forest type. 3-39 Figure 3-14: (a) Vegetation profiles of experimental transects in the “Sonneratia-Avicennia forest”, (b) bar graphs showing variation in mean propagule and seedling density with distance from the Ngao river and (c) box plots showing significant differences in the density of A.corniculatum propagules and seedlings with distance from the Ngao river S.alba A.corniculatum B.parviflora Nypa fructicans R.apiculata A.alba R.mucronata Acanthus ilicifolius (a) (b) (c) Distance from the Ngao river (m) 3-40 3.3.5 Relationship between mangrove propagule and seedling assemblages and environmental parameters 3.3.5.1 Relationship between nMDS ordinations of community structure in propagule and seedling assemblages and environmental parameters The RELATE procedure of PRIMER confirmed that the assemblages of both mangrove propagules and seedlings in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” were related to the environmental parameters present in these forest types (R=0.212, p<0.05 and R=0.251, p<0.01) for propagules and mangrove seedlings respectively. This relationship can be demonstrated through biplots where the environmental data for each of the 20 experimental quadrats is overlaid with data on the assemblages of propagules and seedlings (Figure 3-15). The biplots clearly show that the assemblages of propagules and seedlings in the two forest types are separated based on a combination of the effect of inundation duration, total tree stem diameter and Importance value of A.corniculatum trees all of which have greater values in the “Aegiceras forest” compared to the “Sonneratia-Avicennia forest”. Experimental quadrats within the “Aegiceras forest” were split largely due to duration inundation and the opposite gradient of Importance value of A.alba and S.alba trees. The location of the quadrats experiencing the highest inundation duration (F1T1Q1 and F1T2Q1) are the exceptions to this rule, appearing to be influenced by the relatively high importance role of S.alba in these quadrats. For mangrove propagules, within the “Aegiceras forest”, quadrat T1Q5 separates from the other quadrats from left to right, based on the relatively high importance of A.alba in the quadrat and its location at the highest inundation duration. Application of the PRIMER BIOENV procedure to each forest type individually, allowed assessment of the most important environmental parameters in explaining observed spatial patterns of propagule and seedling assemblages. BIOENV results showed that the resemblance matrix for mangrove propagule density was most highly correlated (although not significantly) (R=0.44, p>0.05) with inundation duration and total tree stem diameter (Table 3-12). In comparison, the resemblance matrices for mangrove seedling density in the “Aegiceras forest” was most highly correlated (although not significantly) with Inundation duration (R = 0.314, p >0.05)(Table 3-12). 3-41 In the “Sonneratia-Avicennia forest” correlations between environmental variables and propagule and seedling data in were weaker than in the “Aegiceras forest” suggesting a weaker relationship between selected environmental variables (total stem diameter and Importance value of A.alba and R.mucronata for propagules and total tree stem diameter and Importance value of S.alba, A.corniculatum, and A.alba for seedlings) and resemblance matrices for both propagule (R=0.162, p>0.05) and seedling assemblages (R=0.275, p>0.05) (Table 3-12). Table 3-12: Summary of BIOENV statistical analysis showing correlation between mangrove propagule and seedling assemblages in the two forest types and environmental parameters Parameter R p Environmental variables included Mean density of Mangrove propagules 0.44 >0.05 Inundation duration (%), Total tree stem diameter Mean density of Mangrove seedlings 0.314 >0.05 Inundation duration (%), Total tree stem diameter Mean density of Mangrove propagules 0.162 >0.05 Importance value, A.alba, Importance value, R.mucronata Mean density of Mangrove seedlings 0.275 >0.05 Total Tree stem diameter, Importance Value S.alba, Importance Value A.corniculatum, Importance Value A.alba The “Aegiceras forest” The “Sonneratia-Avicennia forest” 3.3.5.2 Relationship between univariate propagule and seedling parameters and environmental parameters Duration of inundation In the “Aegiceras forest”, the relative density of A. corniculatum propagules and seedlings found within experimental quadrats over the twelve month monitoring period was significantly negatively correlated with the duration of tidal inundation (Propagules: Spearmans rank coefficient Rs(9)= -0.874, p<0.001 and Seedlings: Spearman’s rank coefficient Rs(9)= -0.681, p<0.05) respectively). Relative density of propagules and seedlings of this species increasied as the proportion of the day in which the site was inundated decreased (Figure 3-16). 3-42 Figure 3-15: Biplots of (a) propagules and (b) seedlings in the two forest types showing clearly that the assemblages of propagules and seedlings are separated based on a combination of the effect of Inundation duration, total tree stem diameter and Importance value of mature A.corniculatum trees (all higher in the Aegiceras forest) and the Importance value of mature S.alba and A.alba trees (higher in the Sonneratia-Avicennia forest). The blue circle in relation to the biplot vectors indicates a correlation of 1. a(a) b(b) 3-43 In contrast, the mean density and relative density of A.alba propagules and seedlings in the “Aegiceras forest”, was significantly positively correlated (Mean density of propagules, Spearman’s rank coefficient Rs(9)=0.768, p<0.01; Relative density of propagules, Spearman’s rank coefficient Rs(9)=0.798, p<0.05, Mean density of propagules, Spearman’s rank coefficient Rs(9)=0.888, p<0.01; Relative density of seedlings, Spearman’s rank coefficient Rs(9)=0.888, p<0.01). Each of these parameters increased significantly as the proportion of the day in which the site was inundated increased (Figure 3-16). There were no significant correlations apparent between propagule or seedling densities and duration of inundation for the other species recorded in the forest type. In the “Sonneratia-Avicennia forest”, the mean density of A.alba seedlings found within experimental quadrats over the 12 month monitoring period was also significantly positively correlated (Spearman’s rank coefficient Rs(9)=0.681, p<0.05) with duration of inundation, increasing significantly as duration of tidal inundation increased (Figure 3-16). There were no significant correlations apparent between propagule or seedling densities and duration of inundation for the other species recorded in the forest type. Total tree stem diameter In the “Aegiceras forest”, the relative density of S.alba seedlings found within experimental quadrats over the twelve month monitoring period was significantly negatively correlated (Spearman’s rank coefficient Rs(9)= -0.651, p<0.05) with the total stem diameter of mature trees in the experimental quadrats, decreasing as the stem diameter of mature trees decreased (Figure 317). In contrast, the mean density of A. corniculatum propagules found within experimental quadrats over the twelve month monitoring period was significantly positively correlated (Spearman’s rank coefficient Rs(9)= 0.77, p<0.01), with the total stem diameter of mature trees in the experimental quadrats, increasing significantly as the total stem diameter of mature trees increased (Figure 317). There were no significant correlations apparent between propagule or seedling densities and total stem diameter of mature trees for the other species recorded in the forest type. 3-44 Figure 3-16: Correlations between duration of tidal inundation and the density and relative density of propagule and seedlings recorded in experimental quadrats over the twelve-month monitoring period. Inundation duration (%) 3-45 Figure 3-16 continued Inundation duration (%) In the “Sonneratia-Avicennia forest”, the mean and relative density of A.ilicifolius seedlings found within experimental quadrats over the 12 month monitoring period was significantly positively correlated (Spearman’s rank coefficient Rs(9)= 0.713, p<0.05) with total stem diameter of mature trees, increasing significantly as the total stem diameter increased (Figure 3-17). There were no correlations apparent between propagule or seedling densities and total stem diameter for the other species recorded in the forest type. 3-46 Importance values of mature trees In the “Aegiceras forest”, the density and relative density of S.alba seedlings found within experimental quadrats over the twelve month monitoring period correlated significantly (Spearman’s rank coefficient Rs(9) = 0.713, p<0.05 and Spearman’s rank coefficient Rs(9) = 0.749, p<0.05) with the presence of mature S.alba trees in the canopy, increasing as the Importance value of S.alba mature trees increased (Figure 3-18). There were no correlations apparent between propagule or seedling densities and duration of inundation for the other species recorded in the forest type. In the “Sonneratia-Avicennia forest”, the relative density of S.alba seedlings found within experimental quadrats over the twelve month monitoring period correlated significantly (Spearman’s rank coefficient Rs(9)= 0.665, p<0.05) with the Importance value of S.alba trees in the canopy, increasing as the Importance value of S.alba mature trees increased (Figure 3-18). The mean density of A.alba seedlings found within experimental quadrats over the 12 month monitoring period correlated significantly (Spearman’s rank coefficient Rs(9)= 0.632, p<0.05) with the Importance value of A.alba trees in the canopy, increasing as the Importance value of A.alba mature trees increased (Figure 3-18). There were no correlations apparent between propagule or seedling densities and duration of inundation for the other species recorded in the forest type. 3-47 Figure 3-17: Correlations between the total stem diameter of mature trees and the density and relative density of propagule and seedlings recorded in experimental quadrats in the two forest types over the twelve month monitoring period. The graphs highlight lower relative densities of S.alba seedlings and higher A.corniculatum propagule density in quadrats with high total stem diameter in the “Aegiceras forest”. The graphs also highlight higher mean density of A.ilicifolius seedlings in quadrats with high total tree total stem diameter in the “Sonneratia-Avicennia forest”. Total mature tree stem diameter (cm) 3-48 Figure 3-18: Correlations between the Importance value of mature trees and the density and relative density of propagule and seedlings recorded in experimental quadrats in the two forest types over the twelve-month monitoring period. The graphs highlight higher relative densities of S.alba seedlings in quadrats with high Importance values of mature S.alba trees in the “Aegiceras forest”. The graphs also highlight higher relative density of S.alba seedlings and mean densities of A.alba seedlings in quadrats with high Importance values of mature S.alba and A.alba trees in the “Sonneratia-Avicennia forest” respectively. Importance value of mature trees (%) 3-49 3.3.6 3.3.6.1 Enclosure experiments Variation in development of released propagules in the two forest types A greater proportion of A. corniculatum propagules released into experimental enclosures germinated into mature seedlings two after release into enclosures in the “Aegiceras forest” (0.2383) than in the “Sonneratia-Avicennia forest” (0.105) (Logistic regression, χ2(1)=35.755, p<0.001)(Table 3-13 and 3-14). In contrast, a greater proportion of A. alba propagules released into experimental enclosures germinated into mature seedlings in the “Sonneratia-Avicennia forest” (0.6433) than the “Aegiceras forest” (0.1233) (Logistic regression, χ2(1)=286.976, p<0.001) (Table 3-13 and 3-14). The proportion of S. alba propagules released into experimental enclosures germinating into mature seedlings in the “Aegiceras forest” was the same as in the “SonneratiaAvicennia forest” (0.0033) (Logistic regression, χ2(1)=0, p>0.05) (Table 3-13 and 3-14). The proportion of A.alba propagules germinating but remaining as “failed seedlings” was higher in the “Aegiceras forest” (0.1483) than the “Sonneratia-Avicennia forest” (0.003) (Logistic regression, χ2(1)= 42.392, p<0.001). No A.corniculatum or S.alba propagules were recorded at this development stage in either forest type (Table 3-13, Table 3-14 and Figure 3-19. The proportion of A.corniculatum propagules remaining at the immature seedling stage two months after release in the “Aegiceras forest” (0.0066) was not significantly different to the proportion in the “Sonneratia-Avicennia forest” (0) (Logistic regression, χ2(1)=0, p>0.05). Similarly, the proportion of A.alba propagules remaining at the immature seedling stage two months after release in the “Sonneratia- Avicennia forest” (0.0083) was not significantly different to the proportion in the “Aegiceras forest” (0.001) (Logistic regression, χ2(1)=0.091, p>0.05) (Table 3-13, Table 3-14 and Figure 3-19. 3-50 Table 3-13: The proportion of propagules of A.corniculatum, A.alba and S.alba propagules released into enclosures reaching each development stage 2 months after release. Species of propagule released/ Forest type Development stage Aegiceras corniculatum Avicennia alba Sonneratia alba “Aegiceras forest” “SonneratiaAvicennia forest” “Aegiceras forest” “SonneratiaAvicennia forest” “Aegiceras forest” “SonneratiaAvicennia forest” Ungerminated (1) 0 0 0 0 0 0 Germinated propagules not yet in soil (Failed seedlings) (2) 0 0 0.1483 0.03 0 0 Immature seedlings (3) 0.0066 0 0.01 0.0083 0 0 Mature seedlings (4) 0.2383 0.105 0.1233 0.6433 0.0033 0.0033 Dead (5) 0.755 0.895 0.7183 0.3183 0.9966 0.01 Table 3-14: Results of logistic regression analysis of the proportion of propagules released into enclosures at different development stages two months after release in the two forest types. Development stage/ Species of propagule released B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Mature seedling Forest A.corniculatum -0.981 0.164 35.755 1 <0.001 0.375 0.272 0.517 A.alba -2.551 0.151 286.976 1 <0.001 12.821 9.544 17.223 S.alba 0 1.417 0 1 >0.05 1 0.062 16.062 A.corniculatum 16.199 1640.871 0 1 >0.05 0 0 . A.alba -0.184 0.608 0.091 1 >0.05 0.832 0253 2.741 S.alba - - - - - - - - A.corniculatum - - - - - - - - A.alba -1.728 0.265 42.392 1 <0.001 0.178 0.106 0.299 S.alba - - - - - - - - Immature seedlings Forest Germinated propagules not yet in soil (Failed seedlings) Forest 3-51 Figure 3-19: Proportion of propagules of A.corniculatum, A.alba and S.alba recorded at each development stage two months after release into experimental enclosures in the “Aegiceras forest” (orange) and “Sonneratia-Avicennia forest” (green). Of note are the significantly greater proportion of A.corniculatum propagules developing into mature seedlings in the “Aegiceras forest” compared to the “Sonneratia-Avicennia forest” and the significantly greater proportion of A.alba propagules developing into mature seedlings in the “Sonneratia-Avicennia forest” compared to the “Aegiceras forest”. Ungerminated Immature seedlings Germinated propagules (Not yet in soil) Mature seedlings Forest type 3-52 Dead 3.3.6.2 Variation in development of propagules and seedlings relative to distance from the river The “Aegiceras forest” The proportion of A.corniculatum propagules released into experimental enclosures that germinated into mature seedlings two months after release differed significantly with distance from the Ngao river (Logistic regression, χ2(4)=59.928, p<0.001)(Table 3-16). The highest proportion (0.43) was located 20-30m from the Ngao river while the lowest (0.075) was located 30-40m from the Ngao river (Table 3-15, Table 3-16 and Figure 3-20). In contrast, the proportion of A. corniculatum propagules released into experimental enclosures that germinated into immature seedlings did not differ significantly with distance from the Ngao river (Logistic regression, χ2(4)= 0, p>0.05). Immature seedlings were only found 0-10m from the Ngao river and represented only 0.0066 of propagules released into enclosures in this location (Table 3-15, Table 3-16). The proportion of A.alba propagules that germinated into mature seedlings differed significantly with distance from the Ngao river (Logistic regression, χ2(4)=30.283, p <0.001). The highest proportion (0.258) was found 30-40m from the Ngao river and the lowest (0.05) found 0-10m from the Ngao river (Table 3-15, Table 3-16 and Figure 3-20). However, the proportion of A.alba propagules germinating into immature seedlings did not vary significantly with distance from the Ngao river (Logistic regression, χ2(4)= 0.944, p>0.05) (Table 3-15, Table 3-16). The proportion of A.alba propagules released into experimental enclosures germinating but not establishing as seedlings “failed seedlings”, differed significantly with distance from the Ngao river. The highest proportion 0.36) was found closest to the Ngao river and the lowest was found 3040m from the Ngao river (0.056) (Logistic regression, χ2(4)= 48.368, p<0.001) (Table 3-15, Table 316). The single S. alba propagule that developed into a mature seedling in this forest type did so in the zone closest to the Ngao river (Logistic regression, χ2(4)=0, p>0.05) (Table 3-15, Table 3-16 and Figure 3-20). 3-53 Table 3-15: Proportion of propagules released into enclosures at each development stage two months after release Forest/ Distance group Species Development stage A.cornicu -latum A.alba S. alba “The Avicennia forest” The “Sonneratia-Avicennia forest” 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 Ungerminated (1) 0 0 0 0 0 0 0 0 0 0 Germinated propagules not yet in soil (failed seedlings)(2) 0 0 0 0 0 0 0 0 0 0 Immature seedlings (3) 0.033 0 0 0 0 0 0 0 0 0 Mature seedlings (4) 0.117 0.183 0.433 0.075 0.383 0.125 0.108 0.075 0.158 0.058 Dead (5) 0.85 0.817 0.567 0.925 0.617 0.875 0.892 0.925 0.842 0.942 Ungerminated (1) 0 0 0 0 0 0 0 0 0 0 Germinated propagules not yet in soil (failed seedlings)(2) 0.358 0.142 0.067 0.058 0.117 0.05 0.1 0 0 0 Immature seedlings (3) 0 0 0.025 0.008 0.017 0.008 0 0.008 0.017 0.008 Mature seedlings (4) 0.05 0.075 0.067 0.258 0.167 0.633 0.683 0.558 0.7 0.642 Dead (5) 0.592 0.783 0.842 0.675 0.7 0.308 0.217 0.433 0.283 0.35 Ungerminated (1) 0 0 0 0 0 0 0 0 0 0 Germinated propagules not yet in soil (failed seedlings)(2) 0 0 0 0 0 0 0 0 0 0 Immature seedlings (3) 0 0 0 0 0 0 0 0 0 0 Mature seedlings (4) 0.017 0 0 0 0 0 0 0.017 0 0 Dead (5) 0 0 0 0 0 0 0 0 0 0 3-54 Table 3-16: Results of logistic regression analysis of the proportion of propagules released into enclosures at different development stages two months after release based on distance from the Ngao river in the two forest types. Development stage/ propagule released Species of B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Mature seedling The “Aegiceras forest” A.corniculatum - - 59.928 4 <0.001 - - - A.alba - - 30.283 4 <0.001 - - - S.alba - - 0 4 >0.05 - - - A.corniculatum - - 7.723 4 >0.05 - - - A.alba - - 6.348 4 >0.05 - - - S.alba - - 0 4 >0.05 - - - A.corniculatum - - 0 4 >0.05 - - - A.alba - - 0.944 4 >0.05 - - - S.alba - - - - - - - - A.corniculatum - - - - - - - - A.alba - - 0.583 4 >0.05 - - - S.alba - - - - - - - - A.corniculatum - - - - - - - - A.alba - - 48.368 4 <0.001 - - - S.alba - - - - - - - - A.corniculatum - - - - - - - - A.alba - - 2.083 4 >0.05 - - - S.alba - - - - - - - - The “Sonneratia-Avicennia forest” Immature seedlings The “Aegiceras forest” The “Sonneratia-Avicennia forest” Germinated propagules not yet in soil (Failed seedlings) The “Aegiceras forest” The “Sonneratia-Avicennia forest” 3-55 Figure 3-20: Proportion of released propagules developing as mature seedlings two months after release into experimental cages with distance away from the river in the “Aegiceras forest”(Orange); and “Sonneratia-Avicennia forest” (Green). Of note is the significantly higher proportion of A.corniculatum propagules that germinated into mature seedlings in the middle zone of the “Aegiceras forest” and lowest 30-40m from the Ngao river while a significantly higher proportion of A.alba propagules germinated into mature seedlings 30-40m from the Ngao river and the lowest proportion closest to the Ngao river. A.alba A.corniculatum S.alba Distance from the Ngao river (m) 3-56 The “Sonneratia-Avicennia forest” In the “Sonneratia-Avicennia forest”, the trend in development of propagules with distance from the Ngao river was less clear with no significant differences found with distance from the Ngao river for any of the three species and development stages (Table 3-15, Table 3-16). 3.3.6.3 Variation in development of propagules and seedlings with duration of tidal inundation The “Aegiceras forest” In the “Aegiceras forest”, the proportion of A. corniculatum propagules germinating into mature seedlings within enclosures two months after release was not significantly correlated with duration of inundation (Spearman’s rank coefficient Rs(29)= -0.280, p>0.05). In contrast, the proportion of A. corniculatum propagules developing into immature seedlings within enclosures at month two was significantly positively correlated with duration of inundation (Spearman’s rank coefficient Rs(29)= 0.419, p<0.05) with more seedlings developing into immature seedlings at locations with higher inundation durations (Figure 3-21a). In contrast, the proportion of A. alba propagules that germinated into mature seedlings was significantly negatively correlated with duration of inundation (Spearman’s rank coefficient Rs(29)= 0.561, p<0.001). Fewer seedlings germinated into mature seedlings in locations with the highest inundation durations (Figure 3-21a). In these locations there was a non-significant trend towards a higher proportion of A.alba propagules germinating into immature seedlings (Spearman’s rank coefficient Rs(29)= -0.318, p<0.01). No significant correlations were recorded for S.alba propagules or seedlings in this forest type. The “Sonneratia-Avicennia forest” The proportion of A. alba propagules germinating but not developing into seedlings (failed seedlings) in the “Sonneratia-Avicennia forest” differed significantly with duration of inundation (Spearman’s rank coefficient Rs(29)= 0.5, p<0.001) (Figure 3-21b). 3-57 No significant differences were recorded for A.corniculatum and S.alba propagules or seedlings in this forest type. Figure 3-21: Correlation between the duration of inundation (%) and the proportion of propagules of the same species developing into different development stages two months after release into experimental cages in the: (a) “Aegiceras forest” (Orange); and (b) “Sonneratia-Avicennia forest” (Green). In the “Aegiceras forest”, of note is the significant positive correlation between the proportion of A.corniculatum propagules developing into immature seedlings and duration of inundation and significant negative correlation between proportion of A.corniculatum propagules developing into mature seedlings and duration of inundation. In the “Sonneratia-Avicennia forest”, of note is the significant negative correlation between the proportion of A.alba propagules germinating into failed seedlings and duration of inundation. (a) (b) Duration of inundation (% of time) 3-58 3.3.6.4 Variation in development of propagules and seedlings with Importance value of mature trees The “Aegiceras forest” In the “Aegiceras forest”, the proportion of A. corniculatum propagules germinating into mature or immature seedlings within enclosures two months after release was not significantly correlated with the Importance value of mature A.corniculatum trees (Spearman’s rank coefficient Rs(9)= 0.109, p>0.05 and Rs(9)= 0.290, p>0.05 respectively). In contrast, the proportion of A. alba propagules germinating into immature seedlings was significantly positively correlated with the Importance value of mature A.alba trees (Spearman’s rank coefficient Rs(9)=0.640, p<0.05). A greater proportion of seedlings developed into immature seedlings in locations with higher A.alba mature trees Importance values. No significant differences were recorded for S.alba seedlings in this forest type (Figure 3-22a). The “Sonneratia-Avicennia forest” In the “Sonneratia-Avicennia forest”, the proportion of A. corniculatum propagules developing into mature seedlings within enclosures was significantly positively correlated with the Importance value of mature A.corniculatum trees (Spearman’s rank coefficient Rs(9)= 0.681, p<0.05 (Figure 322b). In contrast, the proportion of A. alba propagules developing into mature seedlings within enclosures at month two was significantly negatively correlated with the Importance value of mature A.alba trees (Spearman’s rank coefficient Rs(9)= -0.893, p<0.001 (Figure 3-22b). No significant differences were recorded for S.alba seedlings in this forest type. 3-59 Figure 3-22: Correlation between Importance values of mature trees and the proportion of propagules developing into different development stages two months after release into experimental cages in the: (a) “Aegiceras forest” (Orange); and (b) “Sonneratia-Avicennia forest”(Green). In the “Aegiceras forest”, of note is the correlation between the proportion of A.alba propagules developing into immature seedlings and the Importance value of mature A.alba trees. In the “Sonneratia-Avicennia forest”, of note is the correlation between the proportion of A.corniculatum propagules developing into mature seedlings and the Importance value of mature A.corniculatum trees and correlation between the proportion of A.alba propagules developing into mature seedlings and the Importance value of mature A.alba trees. (a) (b) Importance value (IV) of mature trees of the same species 3-60 3.4 Discussion Mangroves have apparently little capacity for vegetative reproduction and spread (Tomlinson, 1987) and as such the dispersal of propagules and early establishment of seedlings plays a crucial role in determining spatial patterns of different mangrove species within mangrove forests (McKee, 1995b, Duke, 2001). Spatial and seasonal patterns in the distribution of mangrove propagules and early establishment of seedlings in the lower intertidal mangrove forests of southern Thailand have been rarely studied and as such little quantitative data is available about these distributions and environmental factors responsible for observed patterns. The findings of the current chapter of the thesis on the early growth stages of lower intertidal forest communities in the Ngao river estuary, provide some useful new information about spatial and seasonal patterns in the distribution of mangrove propagules and early development of seedlings. Specifically, the results of the study show clear variation in composition and spatial distribution of propagules and seedlings between and within two different lower intertidal forest communities, clear patterns of variation with season across an annual period and a clear correlation between observed spatial patterns and environmental conditions. These points will be expanded upon below with further discussion focusing initially on variation in the composition and distribution of assemblages of propagules and seedlings with forest type, distance from the Ngao river and season. The discussion then focuses on development of experimentally released propagules and environmental factors responsible for observed spatial patterns. 3.4.1 3.4.1.1 Composition and distribution of propagules and seedling assemblages Composition The composition of propagules and seedlings recorded in the two forest types was low in diversity with only seven species found in total, one more than that for the number of mature trees recorded at the site in Chapter 2 of this thesis, with the addition of the mangrove associate N.fructicans and understory species A.ilicifolius, which were not observed as mature trees, and the absence of R.mucronata. The presence of N.fructicans and A.ilicifolius as seedlings suggests that seedlings and mature trees of these species have differing tolerances to environmental conditions 3-61 present in these two forests. This phenomena was previously described in A.germinans populations in Louisiana where very young seedlings (less than 12 months old) were more tolerant to surface elevations than older seedlings and mature trees (Alleman and Hester, 2011). A similar observation of tolerance limits was made in A.marina populations in Australia (Clarke and Myerscough, 1993). The absence of R. mucronata is not unexpected as only small quantities of mature trees of this species were found and these were restricted to the higher elevations of the “SonneratiaAvicennia forest”. Of the other members of the Rhizophoracea family, B.parviflora was only found as propagules and restricted to one forest type (the “Aegiceras forest”) while very small quantities of R.apiculata propagules were found in in the “Aegiceras forest” and similar quantities of seedlings in the “Sonneratia-Avicennia forest”. The absence and limited distribution of propagules and seedlings of species of the Rhizophoracea family is consistent with the limited contribution that these higher intertidal species were observed to play in the forest structure of these two lower intertidal forests in the previous chapter of the thesis. Other conspicuous absences from the propagule assemblages in the two forest types were seeds of S.alba, largely due to their small size and difficulty in recording them on the forest floor during monitoring surveys. The general composition of propagules and seedlings in lower intertidal communities in the Ngao river agrees well with previous studies conducted in the Ngao river in the vicinity of the study area and the immediate geographic region. De Haan (1931) for example describes a similar “AvicenniaSonneratia alba” assemblage of species in Indonesia, which develops in the low intertidal zone on young unconsolidated soils comprised of a tree layer of S. alba, A. marina with an undergrowth of Acanthus ilicifolius and saplings of A. corniculatum. The low diversity of this community is also consistent with other studies in southern Thailand which have examined colonisation of sites within the lower intertidal (Matsui et al., 2010, Stevenson, 1997, Panapitukkul et al., 1998). A comparison of propagule and seedling densities measured in other lower intertidal stands in Thailand and south-east Asia showed that the mean density values across all species recorded through the current study (2309.9+/-1181.19 propagules ha-1 and 2424.48 +/-958.99 seedlings ha-1 in the “Aegiceras forest” and 117.19 +/-33.12 propagules ha-1 and 1760.42 +/-362.71 seedlings hain the “Sonneratia-Avicennia forest”) were in the same range as that reported in other studies of 1 lower intertidal communities and comparable to figures of 2,500 seedlings ha-1 given by Srivastava and Bal (1984), as a minimum number to ensure adequate regeneration. 3-62 3.4.1.2 Differences in assemblages of propagules and seedlings between forest types and with distance from the Ngao river The current study found significant differences in the community structure of assemblages of propagules and seedlings in the two different forest types, largely due to differences in the density of propagules and seedlings of A. alba, A.corniculatum, A.ilicifolius, and S.alba between the two forest types. Three of these species, A. alba, A.corniculatum, and S.alba also played an important role in separating the “Aegiceras forest” and the “Avicennia-Sonneratia forest” in the ordination of mature trees carried out in the same forest types in Chapter 2 of the thesis. When species were analysed individually, the mean density of A.corniculatum propagules was greater in the “Aegiceras forest” than the “Sonneratia-Avicennia forest” as was the relative density of A. corniculatum seedlings. This finding is consistent with the dominant role of this species in the mature forest canopy in the forest type described in Chapter 2 of this thesis, where individual A. corniculatum trees were on average larger in stem diameter, and taller while total A. corniculatum stands were higher in total basal area and stem density, frequency and Importance value than in the “Sonneratia-Avicennia forest”. In contrast, propagules and seedlings of the two other dominant lower intertidal species, A.alba and S.alba, generally had the opposite pattern of distribution within the two forest types, with higher mean density of propagules (A.alba) and seedlings (A.alba and S.alba) in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. The results for A.alba and S.alba were also consistent with the dominant role of these two species in the mature forest canopy in the “Sonneratia-Avicennia forest” as compared to the “Aegiceras forest” as described in Chapter 2 of the thesis. Surprisingly, seedlings of S.alba were found in disproportionately low quantities in both the “Aegiceras forest” and the “Sonneratia-Avicennia forest” compared to the significant contribution that mature S.alba trees makes to the latter forest type. Other previous studies have also reported low seedling densities for S.alba. Bosire et al. (2006) for example reported seedling density of 496+/-119 ha-1 from studies of natural S.alba forests in Kenya. Similarly, Mohammed (2009) who also reported very low regeneration rates of S.alba from studies of natural S.alba forests in Kenya. The low quantities of S.alba seedlings can possibly be explained by year to year variation in fruit production of this species as previously documented by Clarke and Myerscough (1993). Inter-annual variation in propagule production in Avicennia germinans trees has been described in Louisiana where trees of this species had 3-63 alternating years of relatively higher reproductive growth with years of relatively higher vegetative growth (Alleman and Hester, 2011). The high reported overall propagule and seedling density in the study area in both forest types was largely due to the high density of A.corniculatum propagules and seedlings in the “Aegiceras forest”. Density of reproductive material of this species was significantly higher in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”, and significantly higher than the density of propagules and seedlings of A.alba and S.alba and the other four species found in the forest type. The values for S.alba in the current study were in the range of those recorded by Imai et al. (2006) of 18.4-218.2 seedlings ha-1 also in the Ngao river estuary. No other studies quoting the densities of A.corniculatum, A.alba and Acanthus propagules or seedlings were found to enable comparison of the current results The large quantities of mangrove propagules observed in the study can be explained in ecological terms. Large quantities of propagules for example increase the chances that a minimum number will develop into adult individuals in an ecosystem where edaphic factors are unfavourable for seedlings during their early stages. They also serve to repress colonisation by other mangrove species (Proisy et al., 2009, Delgado et al., 2001, Baldwin et al., 2001, Din et al., 2002, Sousa and Mitchell, 1999). Different mangrove species have different dispersal strategies. Ceriops tagal populations in Kenya, for example, is similar to A.corniculatum in the current study, releasing a large number of propagules which locally saturate faunal predators and disperse easily due to their slender morphology and small size (De Ryck et al., 2012). This strategy contrasts with that of Rhizophora mangle in the same study which only disperses a small number of propagules which are more tolerant to predation due to larger size and more suited to long distance dispersal. Of the species found in the current study, A.corniculatum and A.alba appear to belong to the first group and S.alba the latter (De Ryck et al., 2012). In addition to differences in assemblages of propagules and seedlings between the forests, the current study found significant differences in the structure of assemblages of propagules and seedlings with distance from the Ngao river in the “Aegiceras forest” but observed no clear pattern of difference in the “Sonneratia-Avicennia forest”. In the “Aegiceras forest”, the difference in assemblages of propagules with distance from the Ngao river was largely due to differences in density of abundance of A.alba and A.corniculatum. Differences in assemblages of seedlings, in comparison, were due to presence and absence of S.alba and A.alba seedlings. These three 3-64 species and R.apiculata also played an important role in separating groups with distance from the Ngao river in the mature tree nMDS ordination carried out in the same forest types in Chapter 2 of this thesis. Similarly, univariate statistical analysis confirmed that there were significantly higher density seedlings of all species in the middle zone of the forest as compared to the quadrats closest to the Ngao river and further most inland. This trend continued for A.corniculatum propagules and seedlings which were found in high densities in the middle zone of the forest dominated by mature A. corniculatum trees. Propagules and seedlings of A. alba, in contrast, were concentrated in the low intertidal quadrats one and two (0-20m from the Ngao river) also inhabited by mature A.alba trees. Propagules and seedlings of S. alba were found in such low quantities that no trend was discernable with distance from the Ngao river. Variation in spatial patterns of mangrove propagule and seedling density with distance from the shoreline or river bank is well documented in the literature. In an earlier study also in southern Thailand, A.alba dominated the vegetation at the progressive edge of the mangrove, followed by S.caseolaris and in lower abundances S.alba and R.apiculata, with density of A.alba and S.caseolaris increasing from minimum values under the canopy to maximum values of between 50 to 150m from the forest edge (Panapitukkul et al., 1998). Other regional and global studies of propagule and seedling distributions have observed similar trends in variation of propagule and seedling abundance according to position along the tidal gradient and hydroperiod. In a survey of A. marina seedling density in mangrove forests in New Zealand, “the lowest number of seedlings was found in the low seaward edges while the highest occurred in the upper (landward) parts of the forest” (Osunkoya and Creese, 1997). Similarly, studies of A. marina seedlings in southern Australia have shown that the highest quantities of seedlings in all transects were found at the upper mangrove level while the least were found at the mudflat level (Clarke and Myerscough, 1993). Consistent with the apparent preference of the dominant lower intertidal species for different degrees of inundation duration in the current study, studies in Belize have added further information, showing that the relative densities of seedlings of R. mangle, A. germinans and L. racemosa varied across the intertidal zone. Of these three species R. mangle was found at highest densities at the lowest intertidal elevation and density of this species decreased with distance towards the shoreline. Of the other species, A. germinans had the highest densities in the middle of the forest and decreased in density towards the shoreline while L. racemosa grew at the highest density close to the shoreline (McKee, 1995b). 3-65 3.4.1.3 Seasonal patterns of propagule and seedling distribution in the two forest types Analysis of seasonal patterns of propagule and seedling distributions through univariate statistical analysis indicates that the populations of propagules and seedlings differed significantly with time across the two forest types. A distinct peak occurred in the month of June in populations of A. corniculatum in the “Aegiceras forest” and a lesser peak evident in the “Sonneratia-Avicennia forest”. The populations of A. alba also peaked but two months later in the month of August in both forest types. These peaks in production coincide with equinox periods and the lowest neap tides and suggest a preference for peak production of propagules in periods of lowest low water and relatively calm wind and high rainfall and therefore environmental conditions which enable maximum exposed area and the most favourable conditions for establishment. The duration of the recruitment phase of propagules and seedlings of all species differed considerably between the two forest types. In the “Aegiceras forest” for example, although a large cohort of propagules was present, no seedlings recruited permanently to the population. This contrasts to the “Sonneratia-Avicennia forest”, where small increases in quantities of A. corniculatum and A. alba seedlings was recorded over the twelve month monitoring period. These patterns provide further insight into the change taking place in the population structure of mangroves in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” as discussed in the previous section. They also provide further evidence that the “Aegiceras forest” is stable in its species composition while the “Sonneratia-Avicennia forest” appears to be undergoing change with an increasing population of seedlings of A. corniculatum adding to the already young population of this species in this forest type. These results also suggest that competition for light resources may play an important role limiting the presence of a persistent A.corniculatum seedling underlayer in the “Aegiceras forest”. The peaked nature of propagule production has been previously reported in populations of A. marina in Australia where this pattern of production is considered as beneficial to the recruitment of seedlings to any available regeneration niches in the forest (Clarke and Allaway, 1993). Similarly, these results are consistent with those reported in other studies in southern Thailand where A.alba produced fruit in peaks from August, September and R.apiculata from July to August including sites from Ranong province (Thampanya et al., 2002). Observed peaks in propagule production have been attributed by previous authors to: conditions provided by the wet season such as the greater dispersal potential due to increased river flow (Troup, 1921, Coupland et al., 3-66 2005, Harun-or-Rashid et al., 2009, Duke et al., 1998); above average high tide events providing opportunities for propagule transport (Peterson and Bell, 2012) and periods of lowest neap tides or “low water establishment windows” which allow for mud banks to emerge allowing establishment of propagules (Proisy et al., 2009, Alleman and Hester, 2011). 3.4.1.4 Change dynamics Significant differences in absolute propagule and seedling quantities between the two forest types and within the “Aegiceras forest” with distance from the Ngao river, suggest that different patterns of recruitment are occurring in the two forest types amongst the dominant lower intertidal species; A.corniculatum, A.alba and S.alba. A.corniculatum propagules and seedlings were widely distributed in both forest types apparently irrespective of the density of mature trees in the forest canopy. In the “Aegiceras forest” for example, the large cohort of A.corniculatum propagules is consistent with the dominance of mature A.corniculatum trees in the canopy as described in Chapter 2 of this thesis. For the “Sonneratia-Avicennia forest”, which doesn’t have a dominant mature forest canopy of A.corniculatum trees, the high density of A.corniculatum propagules and seedlings suggests that this species is expanding its range, with a high density of seedlings observed throughout the forest which persisted in the forest type for most of the year. The two other dominant species in the lower intertidal communities, A.alba and S.alba in contrast had distinct preference for the “Sonneratia-Avicennia forest” over the “Aegiceras forest” consistent with the dominant role that mature trees of these species play in this forest type. Significant differences in absolute propagule and seedling quantities of the three lower intertidal species between the two forest types may also indicate that secondary succession is taking place within these lower intertidal forests. These patterns are particularly evident in the “SonneratiaAvicennia forest”, where it appears that A.corniculatum is expanding its range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” where large S.alba trees are still present in the canopy above a dense understory of A.corniculatum trees. Replacement of lower intertidal species due to interspecific competition has previously been reported in high intertidal habitats in Florida through observations of living and dead tree densities and densities of saplings and seedlings of two mangrove species, R.mangle and 3-67 Laguncularia racemosa, allowing the author to conclude that Laguncularia was being replaced by Rhizophora (Ball, 1980). 3.4.2 3.4.2.1 Development of experimentally released propagules Observed differences in germination in enclosures between and within forest types Overall amongst species, rates of germination into mature seedlings of propagules released into enclosures ranged from less than 1% for S.alba to 64% for A.alba in the “Sonneratia-Avicennia forest”. These results are comparable to a previous cage experiment in Australia where the average rate of propagule establishment within enclosures was increased by 61% within enclosures (Clarke and Myerscough, 1993). Analysis of the proportion of propagules released into enclosures developing to mature seedlings shows that significantly more A. corniculatum propagules developed into seedlings in enclosures in the “Aegiceras forest” than the “Sonneratia-Avicennia forest. In contrast for A.alba the opposite trend was true, while there was no difference between forest types for S.alba seeds. In addition the proportion of A.alba propagules germinating but remaining as “failed seedlings was higher in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. Analysis of the proportion of propagules in enclosures developing into seedlings with distance from the Ngao river provides further insights. Within the “Aegiceras forest”, the highest proportion of A.corniculatum propagules developed to the seedling stage in the middle zone of the forest. This middle zone coincides with the zone dominated by A.corniculatum trees, as recorded in Chapter 2 of the thesis, with a general trend of increasing germination with distance from the Ngao river where it appears that environmental conditions preclude successful development. Propagules of A.alba displayed a similar pattern of increased germination away from the Ngao river, with a higher proportion of propagules germinating at higher elevations and a greater proportion of propagules germinating but remained as “failed seedlings”, unable to progress to the seedling stage in the quadrats closest to the Ngao river. 3.4.3 Environmental factors responsible for observed patterns in natural propagule and seedling assemblages and experimentally released propagules The significant differences in the composition of natural assemblages of mangrove propagules and seedlings, differences in development of propagules in enclosures between the two forest types 3-68 and further significant differences with distance from the Ngao river in the “Aegiceras forest” raises the question as to what environmental factors are responsible for these distinct spatial patterns. Clear differences in assemblages of propagules and seedlings between the two forest types appears to be related to the mature forest composition and the Importance value of conspecific trees in the canopy. Evidence for the important role of conspecific trees in controlling propagule and seedling assemblages in the two forest types comes from the results of the correlation analysis between univariate parameters of propagule and seedling densities of individual species and Importance values of mature trees of the same species. For example the density and relative density of S.alba seedlings in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” correlates significantly with the presence of mature S.alba trees, increasing as the Importance value of S.alba mature trees increased. This suggests that fruit of this species are not distributed far from the parent tree where they rot, distribute their seeds and germinate. A similar pattern was observed for A.alba seedlings in the “Sonneratia-Avicennia forest”, which increased as the Importance value of A.alba mature trees increased. A.corniculatum propagules and seedlings on the other hand did not correlate significantly with the presence of mature A.corniculatum trees in the canopy, suggesting an ability of this species to develop in a wider range of habitats than A.alba and S.alba propagules and seedlings. Additional evidence for the role of conspecific trees in determining spatial patterns of propagule and seedlings comes from the correlation of rates of germination of propagules released into enclosures with the Importance value of mature trees of the same species, especially in the “Sonneratia-Avicennia forest”. The trend of establishment of seedlings close to conspecific mature trees is well described in the mangrove literature. In Thailand, for example, the majority of artificially dispersed R.mucronata propagules in the Ngao river were recaptured within 300m downstream one month after release (Komiyama et al., 1992). Other regional and global studies of propagule and seedling distributions have observed similar trends and reported variation of propagule and seedling abundance according to presence or absence of adult trees of the same species (McKee, 1995a, Clarke and Myerscough, 1993, McGuinness, 1997). In a comprehensive study in Belize, for example, peaks in relative densities of each species coincided with the region dominated by conspecific adults. This was attributed to the fact that the propagules were not dispersed far from adults. In addition, establishment rates of each species varied spatially depending on the buoyancy characteristics of 3-69 its propagules and tidal fluctuations relative to the soil surface and/ or factors contribute to seedling mortality were observed to vary spatially and were less intense near conspecific adults (McKee, 1995b). In a further study in Florida, A. germinans and R. mangle seedlings also regenerated naturally along transects according to the spatial distribution of conspecific adults (Castañeda-Moya et al., 2013). Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to determine within forest distribution patterns. This relationship is especially clear in the “Aegiceras forest” where a negative correlation between the relative density of A.corniculatum propagules, seedlings and the duration of inundation exists, decreasing as inundation increased. Similarly, in the enclosures experiment in the “Aegiceras forest”, a greater proportion of A.corniculatum released propagules germinated at sites with lower inundation durations than those at higher inundations. In the same forest type, A.alba seedlings had the opposite pattern, found at higher densities in quadrats of higher inundation duration adjacent to the Ngao river. At the same time there was also a correlation between the proportion of “failed A. alba seedlings” and inundation duration in the forest type. These results suggest that at low intertidal sites, A.alba seedlings are influenced by the duration of tidal inundation which impedes establishment of seedlings which are “buoyed away from contact with the soil surface”. This provides further evidence of a physical effect associated with the lowest intertidal sites, suggesting that that the buoyancy and obligate dispersal time for this species to root precluded a large proportion of propagules of this species from developing into mature seedlings at these locations. Interestingly, this result contrasts with the observed habitat of mature A.alba trees as described in Chapter 2 of the study which were found in greater densities at lower elevation sites. This suggests that despite difficulties in establishing at lower elevations, this habitat is still the preferred habitat for A.alba perhaps as a result of other factors influencing development of A.alba seedlings in other parts of the forest such as predation and interspecific competition. A similar phenomena has been reported in studies of A. germinans in Louisiana where large proportions of propagules (>90%) rooted but failed to establish (Delgado et al., 1999). Failure to strand before loss of viability was also a major cause of propagule mortality as “tidal action limited contact with the soil surface” (McKee, 1995b). Other physical factors reported to influence spatial patterns in lower intertidal mangrove forests (including studies in the Ngao river) include wave 3-70 action resulting undermining and washing away of seedlings (Houwing et al., 1999, Clarke, 1995, Tamai and Iampa, 1988) and impacts to seedlings due to transport of floating debris in the form of algae, seagrass and other materials which break or smother mangrove stems. Physical impacts on seedlings have been reported to be greater at lower intertidal areas, areas which received more tidal and wave buffering than at higher tidal elevations and are considered to be an important factor leading to low number of seedlings found on mudflats (Clarke and Myerscough, 1993). Mangrove species are reported to vary in tolerance to physical effects such as those present in more exposed lower intertidal areas. In southern Thailand for example the tolerance to flooding was reported to decrease in order from R. mucronata, S. alba, R. apiculata, A. officinalis, C. tagal, B. cylindrica and X.granatum (Kitaya et al., 2002). Kitaya’s attribution of high flooding tolerance to R.mucronata and R.apiculata contrasts to the results of the current study where these two species had higher densities in quadrats of lower inundation duration. In a further study in southern Thailand, lower intertidal colonisers such as A. alba and S. alba had greater survival rates at exposed sites with high hydrodynamic energy and higher degree of sediment disturbance than Rhizophora sp. (Thampanya et al., 2002). In the same study, amongst the lower intertidal species, the faster-growing A.alba showed less mortality than the slower-growing S. alba. Other studies have also reported that A.alba appears to be well adapted to establish on dynamic bare mudflats due to its ability to anchor rapidly, resist hydrodynamic forces from waves and currents within a few days, and then resist sediment movements of the upper sediment layer (Balke et al., 2011). Numerous authors have also recognised variation in the morphological characteristics of the diaspores of mangrove species such as buoyancy, period of obligate dispersal, longevity and period of establishment which could influence their susceptibility to the influence of tides, currents and winds (Rabinowitz, 1978b). Rabinowitz (1978b), for example, proposed a ‘tidal sorting’ hypothesis in which consistently buoyant propagules, for example, could only become established at higher positions in the intertidal zone, whereas larger, heavier propagules would be restricted to lower elevations. These results were corroborated in studies of A. germinans, R. mangle, and L. racemosa in Florida, where the relative importance of factors determining seedling success differed amongst the three species and between high and low tidal elevations. A. germinans propagules, which were small and consistently buoyant could establish at higher elevations in the intertidal zone, but failed to take root in the low intertidal zone where tidal action limited contact with the soil surface. In contrast, establishment of R. mangle seedlings was 3-71 not differentially affected by site conditions (McKee, 1995a). Rabinowitz’s hypothesis has been refuted by numerous authors who claim that her tidal sorting hypothesis “has never been demonstrated and does not occur in all mangrove forests. In the mangrove forests in northern Australia species such as Avicennia with small diaspores occupy the lower tidal levels and species such as Rhizophora with large propagules occupy the higher levels” (Smith, 1992). The observations in the current study however are also inconsistent with Rabinowitz’s hypothesis with mangrove genera with the smallest propagules (Sonneratia) abundant in the lowest intertidal areas, Avicennia and Aegiceras, with similar size propagules found at sites slightly higher in surface elevation. Related to the effect of physical environmental factors on propagule and seedling distribution in the lowest intertidal quadrats, is the role that structures within the forest play in facilitating trapping and retention of propagules and seeds and protect germinating seedlings from wave action. Results of the correlation analysis between propagule and seedling density of individual species and total tree diameter of mature trees were however mixed, suggesting that this factor may be less important in the two forest types than the Importance value of mature trees and duration of inundation in determining propagule and seedling distribution. For example, the relative density of S.alba seedlings had a significant negative correlation with total stem diameter of mature trees, increasing significantly as the total stem diameter of mature trees decreased. Similarly, there were no correlations apparent between A.alba propagules or other species or seedlings and total stem diameter of mature trees which would be expected if facilitation was taking place in these forests. The mean density of A. corniculatum propagules correlated significantly with the total stem diameter of mature trees, increasing significantly as this parameter increased. This result could suggest a dependence on propagules of A.corniculatum for structures within the forest which they can use for added protection during their initial establishment stages. Equally, the significant correlation in could have been due to the confounding effect of the high Importance value of mature A.corniculatum in the canopy. The facilitation role played by physical structures encountered by dispersing stages of plants has been described by other authors and recognised to greatly influence plant recruitment (Schneider and Sharitz, 1988, Sousa et al., 2007). In a comparison of natural regeneration in planted and unplanted sites in Kenya, mangrove regeneration has modified site conditions in a way that facilitates settling and establishment of propagules by altering hydrodynamics (Bosire et al., 2003). 3-72 Similarly in Thailand, the lack of shelter is an important factor controlling establishment and some species (e.g R.mucronata) appear to prefer sheltered sites while others i.e Avicennia and Sonneratia prefer sites of low shelter consistent with the observed distributions of these species in the current study. The final environmental factor which has the potential to influence the distribution of propagules and seedlings across the lower intertidal zone is that of predation. Although not explicitly examined in the current study, evidence for the role of predation in the two forest types comes partially from a comparison of the relatively high level of germination of A.corniculatum and A.alba propagules in enclosures across all locations in the “Aegiceras forest” to the observations made of naturally arriving propagules in experimental quadrats. The role of predation in determining the distribution of propagules and seedlings cannot be discounted, as post dispersal predation is a major factor affecting seedling establishment in Australian mangrove forests (Smith, 1987) although not yet documented in the lower intertidal mangrove forests of southern Thailand. In other studies, predation rates have been related to location in relation to tidal inundation. Smith et al., 1989, for example, recorded variation in the predator population across different tidal levels resulting in predation rates that were inversely correlated with species dominance in some mangrove forests (Smith et al., 1989). Vulnerability of different species of propagules to predation has also been reported, with some predators noted to prefer propagules of particular species over another. Some authors also reported high propagule predation rates of smaller-sized propagules (e.g. those of Ceriops, Bruguiera, Avicennia and Aegiceras spp.) compared to larger propagules (e.g. those of Rhizophora spp.) (Smith, 1987a, McKee, 1995a, McGuinness, 1997, Sousa and Mitchell, 1999). Propagules of Avicennia, for example, are regarded as being particular susceptible to predation compared to other species (Smith, 1987a). In further experiments in north Queensland, in a forest dominated by Avicennia marina Smith, (1988) found that 95% of Avicennia marina propagules were consumed, compared to only 25% of propagules of the other species. The preference for smaller propagules by crab predators has been attributed to easier facilitation of burial of smaller propagules in burrows, and the high nutritive value and low concentration of inhibiting chemicals (e.g. tannins) of smaller propagules (Smith 1987, McKee 1995a, Clarke & Kerrigan 2002). This preference, in the long term, significantly influences forest structure in areas where propagule predators are abundant (Bosire et al., 2005). Some authors have recorded an “inverse relationship between the dominance of the 3-73 mangrove species in the canopy and the amount of predation on its propagules”, observing that “predation was higher where the species was absent from the canopy and lower where conspecific adults were present” (Smith, 1987a). This finding was corroborated by Bosire et al. (2006) who found that propagule predation was the most likely factor influencing seedling establishment. Subsequent studies by Smith (1992) in Panama noted that predation on propagules may effectively preclude establishment of Avicennia germinans and Lumnitzera racemosa in forests dominated by Rhizophora mangle and Pellicieria rhizophorae. The reverse situation however was observed not to be true i.e. predation can account for some of the species distribution patterns in Panama but not all. The above inferences about the environmental factors leading to the current observed spatial patterns of propagules and seedlings across the two forest types are useful as a base on which to develop further controlled experiments in the subsequent chapter of this thesis (Chapter 4), focusing on the post establishment phase and further work to strengthen our understanding of the mechanisms influencing the important pre-establishment phase of the mangrove tree life cycle. The application of information developed in this chapter to mangrove restoration is discussed in detail in Chapter 5 of this thesis. 3.5 Conclusions The objective of the chapter was to answer the questions, “What are the spatial and seasonal patterns of propagule and seedling assemblages in the two lower intertidal forest types?”, “Does protection from predation and physical disturbance influence the development of transplanted propagules in the two forest types?”, and “How are these spatial patterns related to variation in environmental parameters between forest types and with distance from the Ngao river within each forest type?” Results of the study have clearly answered these questions as follows: nMDS ordination analysis showed that assemblages of propagules and developing seedlings are different in the two forest types largely due to the high density of A. corniculatum propagules and seedlings in the “Aegiceras forest” and high density of A.alba propagules and seedlings in the “Sonneratia-Avicennia forest”. Additional univariate analysis supported this result, showing higher 3-74 densities of A. corniculatum propagules and seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest” and higher densities of A. alba propagules and seedlings and S. alba seedlings in the “Sonneratia-Avicennia forest” (Figure 3-23a). Within each forest type, propagule and seedling assemblages were also different with distance from the Ngao river in the “Aegiceras forest” (Figure 3-23b). Figure 3-23: Visual summary of differences in univariate forest structure parameters (a) between the two forest types and (b) within each of forest types (a) Evidence “Aegiceras forest” “Sonneratia Avicennia forest” 1. Hydroperiod < > 2. Composition of propagules and seedlings = = 3. Correlation with duration of inundation Propagules -ve A.corniculatum (Relative density) +ve A.alba (Relative density) Seedlings -ve A.corniculatum (Relative density) +ve A.alba (Relative density) 4. Correlation with total stem diameter – mature trees Propagules A.corniculatum (Rel.density) +ve S.alba (Rel.density) -ve 5. Correlation with Imp.values of conspecific trees Seedlings +ve S.alba (rel.density) +ve +ve A.alba (rel.density) Assemblages Propagules ≠ = Seedlings ≠ = 6. Propagule density = = A.corniculatum > < S.alba = = A.alba < > 7. Seedling density < > A.corniculatum (rel.density) > < S.alba < > A.alba < > 3-75 Figure 3-23 continued (b) Evidence Aegiceras forest -Mature trees Group # 0-10 10-20 Sonneratia Avicennia forest 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 > >> >>> >>>> >>>>> > >> >>> >>>> >>> Propagules = = = = = = = = = = Seedlings = = = = = = = = = = >> >>>>> >>>> >>> > > >>>>> >>> >> >>>>> >>>> > Hydroperiod Assemblages Propagule density A.corniculatum S.alba A.alba Seedling density A.corniculatum S.alba A.alba Seasonal patterns in propagule and seedling densities were also reported with a clear peak in production of A. corniculatum and A. alba propagules in August and September followed by a peak in seedling density in subsequent months. Some seasonal differences were also apparent between forest types with A. corniculatum, A. alba and S.alba seedlings for example, able to persist across the year in the “Sonneratia-Avicennia forest” as compared to the “Aegiceras forest” where they were only recorded for one to two months of the year. Peaks in production coinciding with equinox periods and the lowest neap tides suggest a preference for peak production of propagules in periods of lowest low water and relatively calm wind and high rainfall and therefore environmental conditions which enable maximum exposed area and the most favourable conditions for establishment. The propagule release experiment showed that a higher proportion of A. corniculatum propagules developed into mature seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, more A.alba propagules developed into mature seedlings in enclosures in the “Sonneratia-Avicennia forest”, largely due to the higher proportion of “failed seedlings” which germinated but were unable to establish into seedlings in the “Aegiceras forest”. Significant 3-76 differences in development of released propagules were also observed with distance from the Ngao river. Most A.corniculatum propagules developed into mature seedlings 20-30m from the Ngao river and the least 30-40m from the Ngao river. In contrast, more A.alba propagules developed into mature seedlings further away from the Ngao river and the least closest to the Ngao river, again largely due to the high proportion of “failed seedlings” closest to the Ngao river. There was a significant negative correlation between the proportion of A.alba propagules developing into mature seedlings and duration of inundation. These observations suggest differences in reproductive strategy amongst lower intertidal species in the Ngao river. A.corniculatum trees produced producing large quantities of propagules potentially to increase the chance of recruitment to the sapling stage and repress colonisation by other mangrove species. In contrast, the other two species of propagules, S.alba and A.alba were found in lower quantities suggesting a different reproductive strategy. Significant differences in absolute propagule and seedling quantities between the two forest types also indicate that secondary succession is taking place within these lower intertidal forests. A.corniculatum propagules and seedlings had a wide distribution across both forest types, irrespective of mature tree dominance, consistent with the large sapling density in the previous study of the mature forest structure (Chapter 2 of this thesis). In the “Sonneratia-Avicennia forest”, it appears that A.corniculatum is expanding its range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” where large S.alba trees are still present in the canopy above a dense understory of A.corniculatum trees. The differences in natural assemblages of mangrove propagules and seedlings and development of propagules in enclosures between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest” suggest that assemblages of propagules and seedlings between the two forest types are related to the mature forest composition and the Importance value of conspecific trees in the canopy. These factors in turn are driven by the variation in duration of inundation between the two forest types as reported in Chapter 2 of this thesis. Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to determine within forest distribution patterns, with A.corniculatum propagules and seedlings seemingly less tolerant to these more exposed locations closer to the Ngao river than A.alba and S.alba. 3-77 The above inferences about the environmental factors leading to the current observed spatial patterns of propagules and seedlings across the two forest types are useful as a base on which to develop hypotheses for further controlled experiments in the subsequent chapter of this study and further work external to the study to strengthen our understanding of the mechanisms influencing the important pre-establishment phase of the mangrove tree life cycle. The application of information developed in this chapter to mangrove restoration is discussed in detail in Chapter 5 of this thesis. 3-78 3.6 Appendices Table A3-1: Mean density (+/- S.E) of mangrove propagules and seedlings of mangrove species recorded from experimental plots in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Parameter Forest type Total propagule density (No. ha-1) Relative density propagules (%) of Total seedling density (No. ha-1) Relative density of seedlings (%) Frequency of propagules Frequency of seedlings (%) Relative frequency propagules (%) of Relative frequency seedlings (%) of Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Aegiceras forest SonneratiaAvicennia forest Sonneratia alba NP Aegiceras corniculatum 18166.67 (8125.82) * Bruguiera parviflora 458.34 (106.69) 95.81(2.66) * 20.84 (20.84) Rhizophora apiculata NS Avicennia alba Nypa fructicans NS 41.67 (41.67) NS Acanthus ilicifolius 104.17 (71.16) NS Total 2309.9 (1181.19) NS 0(0) 145.84 (88.12) NP 20.84 (20.84) 416.67 (183.74) NP 41.67 (27.78) 117.19 (33.12) 0.03(0.03) 0(0) 3.95(2.69) *** 0.06(0.06) NS 0.17(0.12) NS 12.51(3.58) NS NP 1.12(1.12) 38.17(11.25) NP) 3.62(2.63) 11.25(2.8) NP 0(0) NS 479.17 (232.51) 104.17 (104.17) NS 395.84 (247.39) NS 2424.48 (958.99)* NP NP NP 47.12(10.84) NS 83.34 (34.03)* 18333.34 (5682.24) *** 250 (60.55) 8291.67 (932.63) 20.84(20.84) 0(0) 4791.67 (897.42) NP 729.17 (239.06) 1760.42 (362.71) 0.98 (0.45)NS 92.02(4.02) * NP 0(0) NS 4.74(2.96) *** 0.32(0.32) NS 1.97(1.11) NS 12.5(3.45) NS 2.47(0.99) 58.21(3.95) 0.12(0.12) 0(0) 34.12(3.96) NP 5.11(1.56) 12.5(2.41) NP 100(0) NS 10(10) 0(0) NS 30(15.28) NS 10(10) NS 20(13.34) NS 21.25(4.61) NS 80(13.34) NP 10(10) 70(15.28) NP 20(13.34) 22.5(4.7) 40(16.33) NS 100(0) NS NP 0(0) 40(16.33) *** 10(10) NS 50(16.67) NS 30(5.16) NS 70(15.28) 100(0) 10(10) 0(0) 100(0) NP 80(13.34) 45(5.6) 77(9.79) * 2(2) 0(0) 12(6.64) *** 2(2) NS 7(5.18) NS 12.5(3.2) NS 42.5(9.17) NP 2.5(2.5) NS 37.5(10.04) NP 7.5(5.34) 11.25(2.59) 15(6.67) NS 55(10.19) * NP 0(0) 11.67(4.85) *** 3.34(3.34) NS 15(5.1) NS 12.5(2.6) NS 18.67(4.27) 28.67(1.63) 2(2) 0(0) 28.67(1.63) NP 22(3.94) 12.5(1.6) NP NP NP Note: *P<0.05, **P<0.01, ***P<0.001, NS p >0.05, NP, Not present in forest type, All mean values are presented (+/- 1 S.E). 3-79 Table A3-2: Propagule and seedling data of the 7 species with season (mean +/- S.E) Month Spec ies S.alb a A. corni culat um B.pa rviflo ra Parameter The “Aegiceras forest” Mean Propagule density (No. ha-1) NP Mean Propagule Relative Density (%)NP Mean propagule frequency (%)NP Mean Propagule relative frequency (%)NP Mean Seedling Density No. ha-1) NS Mean Seedling Relative Density (%)NS Mean Seedling Frequency (%)NS Mean Relative Frequency (%)NS Mean Propagule density (No. ha-1)*** Mean Propagule Relative Density (%)*** Mean propagule frequency (%)*** Mean Propagule relative frequency (%)*** Mean Seedling Density No. ha-1) *** Mean Seedling Relative Density (%)*** Mean Seedling Frequency (%)*** Mean Relative Frequency (%)*** Mean Propagule density (No. ha-1) NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1) NP Mean Seedling Relative Density (%)NP 1 (May) 2 (June) 3 (Jul) 4 (Aug) 5 (Sep) 6 (Oct) 7 (Nov) 8 (Dec) 9 (Jan) 10 (Feb) 11 (Mar) 12 (Apr) Mean 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250 (250) 0(0) 0(0) 10(10) 10(10) 0(0) 0(0) 250 (250) 0.29 (0.29) 0(0) 0(0) 0(0) 250 (250) 0(0) 0(0) 0(0) 250 (250) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 10(10) 10(10) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 1000 (763.76) 20( 13.33) 20 (13.33) 20 (13.33) 10(10) 2500 (1707.8) 40 (16.32) 40 (16.32) 40 (16.32) 10(10) 181750 (80226) 79.81 (13.3) 80 (13.33) 67.5 (13.96) 5(5) 8250 (4014.7) 52.27 (15.07) 60 (16.32) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 250 (250) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 10(10) 10(10) 0(0) 100(0) 68.33 (8.76) 100(0) 83.33 (8.6) 0(0) 2500 (1236.03 ) 50 (16.66) 50 (16.66) 50 (16.66) 0(0) 1250 (671.85) 30 (15.27) 30 (15.27) 30 (15.27) 750 (533.59) 20 (13.33) 20(13.33 ) 20(13.33 ) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 500 (333.33) 15 (10.67) 20 (13.33) 15 (10.67) 250 (250) 0.03 (0.03) 0(0) 7500 (1825.74 ) 76.66 (13.19) 80 (13.33) 75 (13.43) 0(0) 1250 (768.29) 27.5 (14.16) 30 (15.27) 25 (13.43) 50(14.9) 75000 (26116.7 2) 90.15 (5.65) 0(0) 1250 (768.29) 25 (13.43) 30 (15.27) 25(13.43 ) 49000 (12528.8 5) 95.96 (2.83) 0(0) 1250 (768.29) 27.5 (14.16) 30 (15.27) 25 (13.43) 0(0) 23000 (12532.) 67.35 (14.78) 70 (15.27) 53.33 (14.01) 81000 (29716.7 1) 72.78 (12.79) 80 (13.33) 49.99 (12.17) 0(0) 83.33 (41.13) 2.52 (1.43) 3.33 (1.64) 2.91 (1.48) 18166.6 (7903.5) 24.53 (3.85) 25.83 (4.01) 22.15 (3.62) 18333.33 (4281.98 ) 42.13 (4.34) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0.83 (0.83) 0(0) 0(0) 0(0) 2.5(2.5) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0.2(0.2) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 3-80 0(0) 0(0) 0(0) 0(0) 45(4.56) 36.8 (4.06) 20.83 (20.83) Month Spec ies R.api culat a A. alba N. fruct icans Parameter Mean Seedling Frequency (%)NP Mean Relative Frequency (%)NP Mean Propagule density (No. ha-1) NP Mean Propagule Relative Density (%)NP Mean propagule frequency (%)NP Mean Propagule relative frequency (%)NP Mean Seedling Density No. ha-1) NP Mean Seedling Relative Density (%)NP Mean Seedling Frequency (%)NP Mean Relative Frequency (%)NP Mean Propagule density (No. ha-1)NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1)** Mean Seedling Relative Density (%)** Mean Seedling Frequency (%)** Mean Relative Frequency (%)** Mean Propagule density (No. ha-1) NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1) NS Mean Seedling Relative Density (%)NS 1 (May) 2 (June) 3 (Jul) 4 (Aug) 5 (Sep) 6 (Oct) 7 (Nov) 8 (Dec) 9 (Jan) 10 (Feb) 11 (Mar) 12 (Apr) Mean 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250(250) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 5(5) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 1000 (763.76) 7.72 (5.42) 20(13.33 ) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 500 (333.33) 1.89 (1.65) 20 (13.33) 8.33 (5.69) 5(5) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 500 (500) 0.06 (0.06) 500 (333.33) 13.33 (10.18) 20(13.33 ) 15(10.67 ) 0(0) 0(0) 750 (533.59) 3.11 (2.84) 20(13.33 ) 8.33 (5.69) 0(0) 0(0) 750 (381.88) 2.22 (1.39) 30 (15.27) 9.99 (5.09) 10(6.66) 3750 (1796.98 ) 8.68 (5.77) 40(16.32 ) 18.33 (7.63) 0(0) 145.83 (74.23) 1.21 (0.63) 4.16 (1.83) 1.94 (0.86) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250 (250) 0(0) 250 (250) 0(0) 250 (250) 2.5(2.5) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250(250) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 2.5(2.5) 2.5(2.5) 5(5) 0(0) 0(0) 250 (250) 0.16 (0.16) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 3-81 479.16 (180.68) 2.27 (1.03) 9.16 (2.64) 4.3(1.34) 41.66 (41.66) 0(0) 0.83 (0.83) 0.2(0.2) 104.16 (45.79) 1.68 (0.96) Month Spec ies Parameter Mean Seedling Frequency (%)NS Mean Relative Frequency (%)NS Mean Propagule density (No. ha-1)* Mean Propagule Relative Density (%)* Mean propagule frequency (%)* Mean Propagule relative frequency (%)* S. alba Mean Seedling Density No. ha-1) *** Mean Seedling Relative Density (%)*** Mean Seedling Frequency (%)*** Mean Relative Frequency (%)*** The “SonneratiaAvicennia forest” Mean Propagule density (No. ha-1) NP Mean Propagule Relative Density (%)NP Mean propagule frequency (%)NP Mean Propagule relative frequency (%)NP Mean Seedling Density No. ha-1) NS Mean Seedling Relative Density (%)NS Mean Seedling Frequency (%)NS Mean Relative Frequency (%)NS A. corni culat um Mean Propagule density (No. ha-1)*** Mean Propagule Relative Density (%)*** Mean propagule frequency (%)*** Mean Propagule relative frequency (%)*** Mean Seedling Density A. ilicif olius 1 (May) 2 (June) 3 (Jul) 4 (Aug) 5 (Sep) 6 (Oct) 7 (Nov) 8 (Dec) 9 (Jan) 10 (Feb) 11 (Mar) 12 (Apr) Mean 10(10) 10(10) 10(10) 10(10) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 4.16 (1.83) 10(10) 5(5) 5(5) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 750 (533.59) 0.75 (0.54) 20 (13.33) 8.33 (5.69) 2750 (1314.97 ) 5(5) 0(0) 5(5) 500 (333.33) 0.07 (0.05) 20 (13.33) 7.5 (5.33) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 3.33 (3.33) 0(0) 0(0) 4.98 (3) 50 (16.66) 19.99 (6.93) 500 (333.33) 0.92 (0.68) 20(13.33 ) 8.33 (5.69) 0(0) 0(0) 1500 (1500) 0.69 (0.69) 0(0) 0(0) 0(0) 0(0) 0(0) 395.83 (177.5) 0.55 (0.28) 6.66 (2.28) 2.63 (0.92) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250 (250) 0(0) 250 (250) 1.66 (1.66) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250(250) 0(0) 0(0) 0(0) 0(0) 250 (250) 0(0) 0(0) 2.5(2.5) 0(0) 0(0) 0(0) 2.5 (2.5) 10(10) 3.33 (3.33) 10(10) 3.33 (3.33) 10(10) 3.33 (3.33) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 250 (74.84) 1.74 (0.54) 9.16 (2.64) 0(0) 10(10) 3.33 (3.33) 0(0) 500 (333.33) 5.83 (3.93) 20 (13.33) 0(0) 0(0) 0(0) 5(5) 10(6.66) 3.4(1) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 458.33 (172.47) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 6250 0(0) 6000 0(0) 6000 0(0) 5500 (1280.19 ) 78 (13.14) 80 (13.33) 75 (13.43) 6000 0(0) 1250 (559.01) 5.03 (2.31) 40(16.32 ) 12.49 (5.15) 0(0) 2(2) 0(0) 250 (250) 1.42 (1.42) 0(0) 11250 0(0) 11250 0(0) 12250 0(0) 10500 0(0) 8500 0(0) 8250 0(0) 6750 0(0) 6500 3-82 2.5(1.16) 104.16 (54.52) 0.06 (0.04) 3.33 (1.64) 1.31 (0.67) 6.5(2.23) 6.66 (2.28) 6.25 (2.17) 8291.66 Month Spec ies Parameter 1 (May) 2 (June) 3 (Jul) 4 (Aug) 5 (Sep) 6 (Oct) 7 (Nov) 8 (Dec) 9 (Jan) 10 (Feb) 11 (Mar) 12 (Apr) Mean No. ha-1) *** (1070.43 ) 69.5 (11.6) (1000) (1000) (1000) (1356.56 ) (1765.8) 69.66 (11.49) 72.5 (11.72) 39.9(3.9) 45.01 (5.81) (1280.19 ) 51.98 (5.99) (1247.21 ) 69.15 (6.37) (1293.68 ) 94.16 (3.93) (1057.38 ) 92.5 (4.01) (1130.38 ) 92.5 (4.01) (403.48) 67.92 (11.63) (1547.84 ) 59.21 (7.84) 90(10) 63.33 (11.05) 90(10) 63.33 (11.05) 90(10) 63.33 (11.05) 90(10) 70 (11.05) 90(10) 100(0) 46.66 (2.22) 100(0) 100(0) 100(0) 50(0) 100(0) 56.66 (7.53) 100(0) 38.33(5) 100(0) 35.83 (2.5) 90 (6.66) 85 (7.63) 85(7.63) 68.67 (2.71) 95.83 (1.83) 62.29 (2.67) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 20.83 (20.83) 0(0) 0(0) 0(0) 0(0) 0(0) 250 (250) 1.25 (1.25) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 3.33 (3.33) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250 (250) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 0(0) 0.1(0.1) 0.83 (0.83) 0.27 (0.27) 20.83 (20.83) 0.83 (0.83) 0.83 (0.83) 0.83 (0.83) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 5000 (2204.79 ) 70(15.27 ) 70(15.27 ) 70(15.27 ) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) R. apic ulat a Mean Seedling Relative Density (%)*** Mean Seedling Frequency (%)*** Mean Relative Frequency (%)NS Mean Propagule density (No. ha-1) NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1) NP Mean Seedling Relative Density (%)NP Mean Seedling Frequency (%)NP Mean Relative Frequency (%)NP Mean Propagule density (No. ha-1) NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1) NP Mean Seedling Relative Density (%)NP Mean Seedling Frequency (%)NP Mean Relative Frequency (%)NP A.al ba Mean Propagule density (No. ha-1)*** Mean Propagule Relative Density (%)*** Mean propagule frequency (%)*** Mean Propagule relative frequency (%)*** B. parvi flora 3-83 416.66 (216.06) 5.83 (2.14) 5.83 (2.14) 5.83 (2.14) Month Spec ies N. fruct icans A.ilic ifoliu s Parameter Mean Seedling Density No. ha-1) *** Mean Seedling Relative Density (%)*** Mean Seedling Frequency (%)*** Mean Relative Frequency (%) NP Mean Propagule density (No. ha-1) NP Mean Propagule Relative Density (%)NP Mean propagule frequency (%)NP Mean Propagule relative frequency (%)NP Mean Seedling Density No. ha-1) NP Mean Seedling Relative Density (%)NP Mean Seedling Frequency (%)NP Mean Relative Frequency (%)NP Mean Propagule density (No. ha-1)NS Mean Propagule Relative Density (%)NS Mean propagule frequency (%)NS Mean Propagule relative frequency (%)NS Mean Seedling Density No. ha-1)*** Mean Seedling Relative Density (%)*** Mean Seedling Frequency (%)*** Mean Relative Frequency (%)*** 1 (May) 2 (June) 3 (Jul) 4 (Aug) 5 (Sep) 6 (Oct) 7 (Nov) 8 (Dec) 9 (Jan) 2250 (1017.21 ) 18.5 (8.16) 50 (16.66) 23.33 (7.93) 2750 (1261.06 ) 20.64 (8.14) 50 (16.66) 23.33 (7.93) 2250 (1017.21 ) 18.66 (7.53) 50 (16.66) 23.33 (7.93) 2000 (1040.83 ) 17.5 (8.73) 40(16.32 ) 12750 (2456.56 ) 44.49 (4.98) 16250 (3768.47 ) 51.76 (5.13) 11500 (2793.84 ) 48.01 (5.99) 100(0) 35.83 (2.5) 100(0) 46.66 (2.22) 100(0) 50(0) 3750 (1003.46 ) 28.3 (5.8) 80 (13.33) 36.66 (6.47) 20(8.16) 2500 (912.87) 16.92 (5.43) 70(15.27 ) 29.99 (6.93) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10 (Feb) 11 (Mar) 12 (Apr) Mean 750 (533.59) 5.83 (3.93) 20 (13.33) 500 (333.33) 10 (6.66) 10 (6.66) 5(5) 4791.66 (670.49) 23.1 (2.23) 57.5 (4.53) 26.18 (2.12) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250(250) 0(0) 250(250) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 41.66 (29.33) 0(0) 0(0) 0(0) 2(2) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 0(0) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 5(5) 10(10) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 4500 (1740.05 ) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 250(250) 0.71 (0.71) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 10(10) 3.33 (3.33) 0(0) 0(0) 0(0) 22.6(6.7) 70(15.27 ) 28.33 (6.59) 3500 (1354) 10.57 (3.9) 50(16.66 ) 15.83 (5.33) 0(0) 0(0) 500 (333.33) 2.53 (1.7) 20 (13.33) 6.66 (4.44) Note: *P<0.05, **P<0.01, ***P<0.001, NS p >0.05, NP, Not present in forest type, All mean values are presented (+/- 1 S.E). 3-84 5 (3.55) 20 (13.33) 250(250) 1.66 (1.66) 10(10) 1(0.84) 1.66 (1.17) 1.25 (0.92) 729.16 (224.14) 3.03 (0.87) 12.5 (3.03) 4.51 (1.11) Table A3-3 Propagule and seedling data of the 7 species with distance from the Ngao river (Mean +/- S.E). “The Aegiceras forest” The “Sonneratia-Avicennia forest” Species Parameter /Distance 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 p S. alba Mean Propagule density (No. ha-1) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) NS Mean Propagule relative density (%) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) NS Mean Propagule relative frequency (%) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) NS Mean Seedling density (No. ha-1) 208.3 3(0) 0(0) 0(0) 104.16 (104.16) 104.16 (104.16) NS 208.33 (208.33) 416.66(0) 312.5 (104.16) 0(0) 312.5 (104.16) NS Mean Seedling relative density (%) 8.33( 0) 0(0) 0(0) 4.16(4.16) 4.16(4.16) NS 4.16(4.16) 8.33(0) 8.33(0) 0(0) 8.33(0) NS Mean Seedling relative frequency (%) 25(0) 0(0) 0(0) 25(25) 25(25) NS 12.5(12.5) 29.16(4.16) 22.5(2.5) 0(0) 29.16(4.16) NS Mean Propagule density (No. ha-1) 1666. 66 (1041 .66) 65104.16 (7395.83) 12708.33 (6875) 10729.16 (104.16) 625(208.33 ) * 729.16 (104.16) 520.83 (312.5) 520.83 (312.5) 208.33 (208.33) 312.5 (312.5) NS Mean Propagule relative density (%) 8.33( 0) 8.33(0) 8.33(0) 8.33(0) 8.33(0) NS 8.33(0) 8.33(0) 8.33(0) 4.16(4.16) 4.16 (4.16) NS Mean Propagule relative frequency (%) 50(0) 35(15) 100(0) 100(0) 100(0) NS 37.5(12.5) 75(25) 50(0) 25(25) 25(25) NS Mean Seedling density (No. ha-1) 5416. 66 (208. 33) 46770.83 (4479.16) 17500 (14791.66) 18125 (5000) 3854.16 (1770.83) NS 9479.16 (312.5) 7604.16 (1562.5) 10520.83 (1562.5) 9583.33 (1250) 4270.83 (2812.5) *** Mean Seedling relative density (%) 8.33( 0) 8.33(0) 8.33(0) 8.33(0) 8.33(0) NS 8.33(0) 8.33(0) 8.33(0) 8.33(0) 8.33(0) NS Mean Seedling relative frequency (%) 25(0) 33.33(0) 66.66 (33.33) 75(25) 75(25) NS 29.16(4.16) 29.16(4.16) 22.5(2.5) 33.33(0) 29.16(4.16) NS Mean Propagule density (No. ha-1) 0(0) 104.16 (104.16) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Propagule relative density (%) 0(0) 4.16(4.16) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Propagule relative frequency (%) 0(0) 10(10) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Seedling density 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 104.16 0(0) 0(0) A.corniculatum B. parviflora 3-85 p NS Species Parameter /Distance “The Aegiceras forest” 0-10 (No. R.apiculata A. alba N. fructicans 10-20 The “Sonneratia-Avicennia forest” 20-30 30-40 40-50 p 0-10 10-20 ha-1) 20-30 30-40 40-50 p (104.16) Mean Seedling relative density (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 4.16 (4.16) 0(0) 0(0) NS Mean Seedling relative frequency (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 10(10) 0(0) 0(0) NS Mean Propagule density (No. ha-1) 0(0) 0(0) 0(0) 0(0) 0(0) NS 104.16 (104.16) 0(0) 0(0) 0(0) 0(0) NS Mean Propagule relative density (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 4.16(4.16) 0(0) 0(0) 0(0) 0(0) NS Mean Propagule relative frequency (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 12.5(12.5) 0(0) 0(0) 0(0) 0(0) NS Mean Seedling density (No. ha-1) 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Seedling relative density (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Seedling relative frequency (%) 0(0) 0(0) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Propagule density (No. ha-1) 312.5 (104. 16) 416.66 (416.66) 0(0) 0(0) 0(0) *** 416.66 (208.33) 104.16 (104.16) 1354.16 (520.83) 208.33(0) 0(0) NS Mean Propagule relative density (%) 8.33( 0) 4.16(4.16) 0(0) 0(0) 0(0) NS 8.33(0) 4.16(4.16) 8.33(0) 8.33(0) 0(0) NS Mean Propagule relative frequency (%) 50(0) 10(10) 0(0) 0(0) 0(0) NS 37.5(12.5) 25(25) 50(0) 75(25) 0(0) NS Mean Seedling density (No. ha-1) 1770. 83 (104. 16) 625 (208.33) 0(0) 0(0) 0(0) NS 4687.5 (729.16) 6875 (2708.33) 6979.16 (3020.83) 2812.5 (937.5) 2604.16 (729.16) NS Mean Seedling relative density (%) 8.33( 0) 8.33(0) 0(0) 0(0) 0(0) NS 8.33(0) 8.33(0) 8.33(0) 8.33(0) 8.33(0) NS Mean Seedling relative frequency (%) 25(0) 33.33(0) 0(0) 0(0) 0(0) NS 29.16(4.16) 29.16(4.16) 22.5(2.5) 33.33(0) 29.16(4.16) NS Mean Propagule density (No. ha-1) 0(0) 208.33 (208.33) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Propagule relative density (%) 0(0) 4.16(4.16) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean 0(0) 10(10) 0(0) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Propagule 3-86 Species Parameter /Distance “The Aegiceras forest” The “Sonneratia-Avicennia forest” 0-10 10-20 20-30 30-40 40-50 p 0-10 10-20 20-30 30-40 40-50 p Mean Seedling density (No. ha-1) 0(0) 0(0) 520.83 (520.83) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Seedling relative density (%) 0(0) 0(0) 4.16 (4.16) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Seedling relative frequency (%) 0(0) 0(0) 16.66 (16.66) 0(0) 0(0) NS 0(0) 0(0) 0(0) 0(0) 0(0) Mean Propagule density (No. ha-1) 0(0) 520.83 (104.16) 0(0) 0(0) 0(0) ** 104.16 (104.16) 0(0) 0(0) 0(0) 104.16 (104.16) NS Mean Propagule relative density (%) 0(0) 8.33(0) 0(0) 0(0) 0(0) NS 4.16(4.16) 0(0) 0(0) 0(0) 4.16 (4.16) NS Mean Propagule relative frequency (%) 0(0) 35(15) 0(0) 0(0) 0(0) NS 12.5(12.5) 0(0) 0(0) 0(0) 25(25) NS Mean Seedling density (No. ha-1) 520.8 3 (312. 5) 1354.16(1 145.83) 104.16 (104.16) 0(0) 0(0) NS 1458.33 (833.33) 104.16 (104.16) 729.16 (312.5) 1250 (416.66) 104.16 (104.16) NS Mean Seedling relative density (%) 8.33( 0) 8.33(0) 4.16 (4.16) 0(0) 0(0) NS 8.33(0) 4.16(4.16) 8.33(0) 8.33(0) 4.16 (4.16) NS Mean Seedling relative frequency (%) 25(0) 33.33(0) 16.66 (16.66 0(0) 0(0) NS 29.16(4.16) 12.5(12.5) 22.5(2.5) 33.33(0) 12.5 (12.5) NS relative frequency (%) A.ilicifolius 3-87 3.7 References ALLEMAN, L. K. & HESTER, M. W. 2011. Reproductive Ecology of Black Mangrove (Avicennia germinans) Along the Louisiana Coast: Propagule Production Cycles, Dispersal Limitations, and Establishment Elevations. Estuaries and Coasts, 34, 1068-1077. BALDWIN, A., EGNOTOVICH, M., FORD, M. & PLATT, W. 2001. Regeneration in fringe mangrove forests damaged by Hurricane Andrew. Plant Ecology, 157, 151-164. BALKE, T., BOUMA, T. J., HERMAN, P. M. J., HORSTMAN, E. M., SUDTONGKONG, C. & WEBB, E. L. 2013. Cross-shore gradients of physical disturbance in mangroves: implications for seedling establishment. Biogeosciences, 10, 5411-5419. BALKE, T., BOUMA, T. J., HORSTMAN, E. M., WEBB, E. L., ERFTEMEIJER, P. L. A. & HERMAN, P. M. J. 2011. Windows of opportunity: thresholds to mangrove seedling establishment on tidal flats. Marine Ecology Progress Series, 440, 1-9. BALL, M. 1980. Patterns of secondary succession in a mangrove forest of Southern Florida. Oecologia, 44, 226-235. BALL, M. C. & PIDSLEY, S. M. 1995. Growth Responses to Salinity in Relation to Distribution of Two Mangrove Species, Sonneratia alba and S. lanceolata, in Northern Australia. Functional Ecology, 9, 77-85. BARLOW, J., GARDNER, T. A., ARAUJO, I. S., ÁVILA-PIRES, T. C., BONALDO, A. B., COSTA, J. E., ESPOSITO, M. C., FERREIRA, L. V., HAWES, J., HERNANDEZ, M. I. M., HOOGMOED, M. S., LEITE, R. N., LO-MAN-HUNG, N. F., MALCOLM, J. R., MARTINS, M. B., MESTRE, L. A. M., MIRANDA-SANTOS, R., NUNES-GUTJAHR, A. L., OVERAL, W. L., PARRY, L., PETERS, S. L., RIBEIRO-JUNIOR, M. A., DA SILVA, M. N. F., DA SILVA MOTTA, C. & PERES, C. A. 2007. Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences, 104, 18555-18560. BOSIRE, J. O., DAHDOUH-GUEBAS, F., KAIRO, J. G. & KOEDAM, N. 2003. Colonization of nonplanted mangrove species into restored mangrove stands in Gazi Bay, Kenya. Aquatic Botany, 76, 267-279. BOSIRE, J. O., DAHDOUH-GUEBAS, F., KAIRO, J. G., WARTEL, S., KAZUNGU, J. & KOEDAM, N. 2006. Success rates of recruited tree species and their contribution to the structural development of reforested mangrove stands. Marine Ecology Progress Series, 325, 85-91. BOSIRE, J. O., KAIRO, J. G., KAZUNGU, J., KOEDAM, N. & DAHDOUH-GUEBAS, F. 2005. Predation on propagules regulates regeneration in a high-density reforested mangrove plantation. Marine Ecology Progress Series, 299, 149-155. CASTAÑEDA-MOYA, E., TWILLEY, R. R. & RIVERA-MONROY, V. H. 2013. Allocation of biomass and net primary productivity of mangrove forests along environmental gradients in the Florida Coastal Everglades, USA. Forest Ecology and Management, 307, 226-241. CLARKE, K. R. & AINSWORTH, M. 1993. A method of linking multivariate community structure to environmental variables. MARINE ECOLOGY PROGRESS SERIES, 92, 205-219. 3-88 CLARKE, K. R. & GREEN, R. H. 1988. Statistical design and analysis for a 'biological effects' study. Marine Ecology Progress Series, 46, 213-226. CLARKE, K. R. & WARWICK, R. M. 1994. Similarity-based testing for community pattern: the twoway layout with no replication. Marine Biology, 118, 167-176. CLARKE, K. R. & WARWICK, R. M. 2001. Change in marine communities: an approach to statistical analysis and interpretation, , Plymouth, PRIMER-E. CLARKE, P. 1995. The population dynamics of the mangrove Avicennia marina; demographic synthesis and predictive modelling. Hydrobiologia, 295, 83-88. CLARKE, P. & MYERSCOUGH, P. 1993. The intertidal distribution of the grey mangrove {Avicennia marina) in southeastern Australia: The effects of physical conditions, interspecific competition, and predation on propagule establishment and survival. Australian Journal of Ecology 18, 307-315. CLARKE, P. J. & ALLAWAY, W. G. 1993. The regeneration niche of the grey mangrove (Avicennia marina): effects of salinity, light and sediment factors on establishment, growth and survival in the field. Oecologia, 93, 548-556. COUPLAND, G. T., PALING, E. I. & MCGUINNESS, K. A. 2005. Vegetative and reproductive phenologies of four mangrove species from northern Australia. Australian Journal of Botany, 53, 109. DE HAAN, J. H. 1931. Het een en ander over de Tijlatjapsche vloedbosschen. . Tectona 24, , 39-75. DE RYCK, D. J. R., ROBERT, E. M. R., SCHMITZ, N., VAN DER STOCKEN, T., DI NITTO, D., DAHDOUHGUEBAS, F. & KOEDAM, N. 2012. Size does matter, but not only size: Two alternative dispersal strategies for viviparous mangrove propagules. Aquatic Botany, 103, 66-73. DELGADO, P., HENSEL, P. F., JIMÉNEZ, J. A. & DAY, J. W. 2001. The importance of propagule establishment and physical factors in mangrove distributional patterns in a Costa Rican estuary. Aquatic Botany, 71, 157-178. DELGADO, P., JIMÉNEZ, J. A. & JUSTIC, D. 1999. Population dynamics of mangrove Avicennia bicolor on the Pacific coast of Costa Rica. Wetlands Ecology and Management, 7, 113-120. DIN, N., PRISO, R., KENNE, M., NGOLLO, D. & BLASCO, F. 2002. Early growth stages and natural regeneration of Avicennia germinans (L.) Stearn in the Wouri estuarine mangroves (Douala-Cameroon). Wetlands Ecology and Management, 10, 461-472. DUKE, N., BALL, M. & ELLISON, J. 1998. Factors influencing biodiversity and distributional gradients in mangroves. Global Ecology & Biogeography Letters, 7, 27-47. DUKE, N. C. 1996. Mangrove reforestation in Panama: an evaluation of planting in areas deforested by a large oil spill. In: FIELD, C. D. (ed.) Restoration of Mangrove ecosystems. Okinawa, Japan: International Society for Mangrove Ecosystems (ISME). DUKE, N. C. 2001. Gap Creation and regenerative processes driving diversity and structure of mangrove ecosystems. Wetlands Ecology and Management, 9, 257-269. 3-89 ENGLISH, S., WILKINSON, C. & BAKER, V. 1997. Survey Manual for Tropical Marine Resources, 2nd Edition. (Townsville, Australian Institute of Marine Science. FRIESS, D. A., KRAUSS, K. W., HORSTMAN, E. M., BALKE, T., BOUMA, T. J., GALLI, D. & WEBB, E. L. 2012. Are all intertidal wetlands naturally created equal? Bottlenecks, thresholds and knowledge gaps to mangrove and saltmarsh ecosystems. Biological Reviews, 87, 346-366. HARPER, J. L. U. 1977. Population biology of plants, Academic Press. HARUN-OR-RASHID, S., BISWAS, S. R., BÖCKER, R. & KRUSE, M. 2009. Mangrove community recovery potential after catastrophic disturbances in Bangladesh. Forest Ecology and Management, 257, 923-930. HOUWING, E. J., HOUWING, E.-J., VAN DUIN, W., WAAIJ, Y.-V., DIJKEMA, K. & TERWINDT, J. J. 1999. Biological and abiotic factors influencing the settlement and survival of Salicornia dolichostachya in the intertidal pioneer zone. Mangroves and Salt Marshes, 3, 197-206. IMAI, N., TAKYU, M., NAKAMURA, Y. & NAKAMURA, T. 2006. Gap formation and regeneration of tropical mangrove forests in Ranong, Thailand. Plant Ecology, 186, 37-46. KITAYA, Y., JINTANA, V., PIRIYAYOTHA, S., JAIJING, D., YABUKI, K., IZUTANI, S., NISHIMIYA, A. & IWASAKI, M. 2002. Early growth of seven mangrove species planted at different elevations in a Thai estuary. Trees, 16, 150-154. KOMIYAMA, A., KONG SANGCHAI, J., PATANAPONPAIBOON, P., AKSORNKOAE, K. & OGINO, K. 1992. Socio-Ecosystem Studies on Mangrove Forests Charcoal Industry and Primary Productivity of Secondary Stands. TROPICS, 1, 233-242. KRAUSS, K. W., LOVELOCK, C. E., MCKEE, K. L., LÓPEZ-HOFFMAN, L., EWE, S. M. L. & SOUSA, W. P. 2008. Environmental drivers in mangrove establishment and early development: A review. Aquatic Botany, 89, 105-127. KRUSKAL, J. B. & WISH, M. 1978. Multidimensional scaling, Sage Publications, Beverly Hills. MACINTOSH, D. J., ASHTON, E. & TANSAKUL, V. 2002. Utilisation and Knowledge of Biodiversity in the Ranong Biosphere Reserve, Thailand. MATSUI, N., SUEKUNI, J., NOGAMI, M., HAVANOND, S. & SALIKUL, P. 2010. Mangrove rehabilitation dynamics and soil organic carbon changes as a result of full hydraulic restoration and re-grading of a previously intensively managed shrimp pond. Wetlands Ecology and Management, 18, 233-242. MCGUINNESS, K. A. 1997. Seed predation in a tropical mangrove forest: a test of the dominancepredation model in northern Australia. Journal of Tropical Ecology, 13, 293-302. MCKEE, K. L. 1995a. Seedling recruitment patterns in a Belizean mangrove forest: effects of establishment ability and physico-chemical factors. Oecologia, 101, 448-460. MCKEE, K. L. 1995b. Seedling recruitment patterns in a Belizean mangrove forest: effects of establishment ability and physico-chemical factors. Oecologia 101, 448-460. 3-90 OSUNKOYA, O. & CREESE, R. G. 1997. Population Structure, Spatial Pattern and Seedling Establishment of the Grey Mangrove, Avicennia marina var.australasica, in New Zealand. Australian Journal of Botany, 45, 707-725. PANAPITUKKUL, N., DUARTE, C. M., THAMPANYA, U., KHEOWVONGSRI, P., SRICHAI, N., GEERTZHANSEN, O., TERRADOS, J. & BOROMTHANARATH, S. 1998. Mangrove Colonization: Mangrove Progression Over the Growing Pak Phanang (SE Thailand) Mud Flat. Estuarine, Coastal and Shelf Science, 47, 51-61. PATTERSON, S., MCKEE, K. & MENDELSSOHN, I. 1997. Effects of tidal inundation and predation on Avicennia germinans seedling establishment and survival in a sub-tropical mangal/salt marsh community. Mangroves and Salt Marshes, 1, 103-111. PETERSON, J. M. & BELL, S. S. 2012. Tidal events and salt-marsh structure influence black mangrove (Avicennia germinans) recruitment across an ecotone. Ecology, 93, 1648-1658. PROISY, C., GRATIOT, N., ANTHONY, E. J., GARDEL, A., FROMARD, F. & HEURET, P. 2009. Mud bank colonization by opportunistic mangroves: A case study from French Guiana using lidar data. Continental Shelf Research, 29, 632-641. RABINOWITZ, D. 1978b. Early Growth of Mangrove Seedlings in Panama, and an Hypothesis Concerning the Relationship of Dispersal and Zonation. Journal of Biogeography, 5, 113133. SCHNEIDER, R. L. & SHARITZ, R. R. 1988. Hydrochory and Regeneration in A Bald Cypress-Water Tupelo Swamp Forest. Ecology, 69, 1055-1063. SMITH, T. J. 1987. Seed Predation in Relation to Tree Dominance and Distribution in Mangrove Forests. Ecology, 68, 266-273. SMITH, T. J. 1987a. Seed Predation in Relation to Tree Dominance and Distribution in Mangrove Forests. Ecology, 68, 266-273. SMITH, T. J. 1992. Forest structure. In: ROBERTSON, A. I. & ALONGI, D. M. (eds.) Tropical Mangrove Ecosytems. Washington, DC. : American Geophysical Union SMITH, T. J., III, CHAN, H. T., MCIVOR, C. C. & ROBBLEE, M. B. 1989. Comparisons of Seed Predation in Tropical, Tidal Forests From Three Continents. Ecology, 70, 146-151. SOUSA, W. P., KENNEDY, P. G., MITCHELL, B. J. & ORDÓÑEZ L, B. M. 2007. SUPPLY-SIDE ECOLOGY IN MANGROVES: DO PROPAGULE DISPERSAL AND SEEDLING ESTABLISHMENT EXPLAIN FOREST STRUCTURE? Ecological Monographs, 77, 53-76. SOUSA, W. P. & MITCHELL, B. J. 1999. The Effect of Seed Predators on Plant Distributions: Is There a General Pattern in Mangroves? Oikos, 86, 55-66. SRIVASTAVA, P. & BAL, H. 1984. Composition and distribution pattern of natural regeneration after second thining in Matang mangrove reserve, Perak, Malaysia. In: SOEPADIMO, E., RAO, A. & MACINTOSH, D. (eds.) Asian Symposium on Mangrove Environments and Resource Managment Kuala Lumpur. 3-91 STEVENSON, N. J. 1997. Disused shrimp ponds: Options for redevelopment of mangroves. Coastal Management, 25, 425-435. TAMAI, S. & IAMPA, P. 1988. Establishment and growth of mangrove seedling in mangrove forests of southern Thailand. Ecological Research, 3, 227-238. THAMPANYA, U., VERMAAT, J. E. & DUARTE, C. M. 2002. Colonization success of common Thai mangrove species as a function of shelter from water movement. Marine Ecology Progress Series, 237, 111-120. TOMLINSON, P. 1987. The Botany of Mangroves. . Brittonia, 39, 10-10. TROUP, R. S. 1921. The Silviculture of Indian Trees., Oxford, UK., Clarendon Press. 3-92 Survival and growth of transplanted mangrove seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand Abstract The post-establishment phase of the mangrove tree life cycle is dependent upon individual species tolerances (autoecology) to the many, sometimes interacting environmental factors prevalent in mangrove ecosystems. Gaps in knowledge about key processes hamper successful restoration of lower intertidal forests. An experimental seedling transplant study was conducted to understand the relative ability of mature mangrove seedlings of two lower intertidal specialist species (S.alba and A.corniculatum) to survive and grow in relation to environmental parameters within two lower intertidal mangrove communities in the Ngao river estuary, southern Thailand. Experimental seedling transplantation experiments reveal significant differences in the survival and growth of transplanted seedlings between the two forest types as well as with distance from the Ngao river and canopy type and as a result provide some insights into specific tolerances of S.alba and A. corniculatum seedlings to experimental treatments. Survival of S.alba and A. corniculatum seedlings was greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” while across the two species, survival of A.corniculatum seedlings was greater than S.alba seedlings in both forest types. The growth of transplanted seedlings of both species was also consistently greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” and growth increments of S.alba seedlings were also greater than those of A.corniculatum in both forest types. There was also a significant interaction between forest type and species of seedling planted on growth, with the increased growth associated with transplanting to the “Sonneratia-Avicennia forest” less pronounced for A.corniculatum seedlings. Within the two forest types, survival of both species increased with distance from the Ngao river. A greater proportion of A.corniculatum seedlings survived in quadrats located further away from the Ngao river than S.alba seedlings. Growth of mangrove seedlings also differed 4-1 with distance from the Ngao river, significantly higher for both species transplanted furthest from the Ngao river. Results of the final treatment in Chapter 4 showed that the proportion of transplanted seedlings of both species surviving differed with canopy type, with a greater proportion of seedlings surviving in light gaps than under the forest canopy. There was also a trend towards higher growth increments in light gaps but differences were not statistically significant. There was a significant interaction between canopy type and forest, canopy type and distance and canopy type and species and seedling survival. Canopy type had less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest”. In addition, canopy type had less of an effect on survival of transplanted seedlings closest to the Ngao river than it did further inland. Canopy type also had less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings. The main driver for the differing rates of survival and growth of transplanted seedlings in the two forest types appears to be the significantly different levels of light availability in the two forest types as a result of the differing forest structure of these two lower intertidal communities, the denser “Aegiceras forest” and the more open canopy “Sonneratia-Avicennia forest” as discussed in Chapter 2 of the thesis. Forest type in turn is driven by variation in the duration of inundation between the two forest types. Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to determine rates of survival and growth of transplanted seedlings, both of which were significantly greater in plots away from the Ngao river than plots close to it with A.corniculatum seedlings were seemingly less tolerant to more exposed locations than S.alba seedlings. The effects of light level and physical disturbance on seedling establishment, although not unequivocal are consistent with results from the literature and also provide further insights into tolerances of the two lower intertidal species to environmental conditions typically observed at restoration sites. Knowledge of these species specific tolerances to tidal inundation, physical environmental factors and light will serve to improve current and future restoration efforts. 4-2 4.1 Introduction Subsequent to mangrove propagules arriving at a site, germinating and establishing as mature seedlings, their continued survival and eventual recruitment to the sapling stage through the post-establishment phase of the mangrove tree life-cycle is dependent upon their individual species tolerances (autoecology) to the many, sometimes interacting environmental factors prevalent in these habitats. These include hydroperiod and interspecific competition (including light availability)(Krauss et al., 2008, Osunkoya and Creese, 1997). Gaps in knowledge about the influence of key environmental factors on seedling progression to the sapling stage ultimately hamper successful restoration of lower intertidal forests. In this chapter, I describe an experimental study of the relative abilities of seedlings of three lower intertidal specialist species; Aegiceras corniculatum, Avicennia alba, and Sonneratia alba to survive and grow in sites exhibiting variation in tidal elevation, physical and light condition, using two lower intertidal forest communities along the Ngao river, in the Kraburi river estuary, Ranong province, southern Thailand as case studies. The aim of the chapter is to develop an understanding of the factors affecting the post establishment stage of the mangrove tree life cycle through analysis of the ability of mature mangrove seedlings to survive and grow under different light, tidal and physical environmental conditions. Specifically, in this chapter I propose to answer the following questions:  “Does growth and survival of transplanted mangrove seedlings vary across the two different lower intertidal communities with treatments of forest type, distance from the Ngao river and canopy type?”; and  “What is the relationship between hydroperiod and light availability and survival and growth of transplanted seedlings?” I answer these questions in the results section of the chapter by describing seedling survival (4.3.1) and growth (4.3.2) with forest type, within forest type relative to distance from the Ngao river and relative to canopy type within experimental quadrats. Section 4.3.3 goes on to examine the relationship between mangrove seedling survival and growth and hydroperiod. 4-3 4.2 Methods 4.2.1 Study sites The study was conducted in two lower intertidal mangrove forests of the Ngao river, Ranong province, southern Thailand. The first forest type, the “Aegiceras forest”, is characterised by open mudflats and then dense stands of A.corniculatum and S.alba saplings and trees growing in a narrow (50-60 m) belt, seaward of a larger area of forest dominated by a mixed community comprised of trees of Rhizophora and Bruguiera spp. trees at higher elevations and large individual S. alba trees interspersed by some unvegetated or lightly vegetated areas. The second forest type, the “Sonneratia-Avicennia forest”, is an open woodland community comprised of large individual S. alba and A. alba trees with an understory of A.corniculatum saplings and trees and a less pronounced zonation pattern across the tidal gradient. 4.2.2 Species description The current chapter of the thesis focuses on three lower intertidal mangrove species: A.corniculatum, A. alba and S.alba. Aegiceras corniculatum, family Myrsinaceae, is a low evergreen tree or shrub with a wide distribution from Ceylon to South China, through the Malay Archipelago to Polynesia and north-eastern Australia (Tomlinson, 1987). A. corniculatum trees grow on the banks of tidal rivers subject to daily inundation, but not very strong tides, occasionally on more or less sandy places in the outer mangrove fringe (Santisuk et al., 1988). Propagules of A.corniculatum are crypto-viviparous containing a precociously developed embryo which often ruptures the seed coat whilst still attached to the parent tree (Clarke, 1995b, Tomlinson, 1987)(Figure 4-1). Avicennia alba, family Acanthaceae is a medium to large sized tree, distributed from India to Indochina, through the Malay archipelago to the Philippines, New Guinea and New Britain (Santisuk et al., 1988, Tomlinson, 1987). A. alba is a pioneer species which typically colonises newly formed, well drained mud flats often on the sea-face, but usually within the influence of rivers near the sea. It is generally an indicator of rich soil (Santisuk et al., 1988). Propagules of A.alba are also crypto-viviparous and the precociously developed embryo occupies the fruit cavity (Duke, 1991, Tomlinson, 1987) (Figure 4-1). Sonneratia alba, family Lythraceae is a medium-sized tree, with a broad distribution from east Africa and Madagascar to Southeast Asia, the Malay Archipelago to the Philippines and tropical 4-4 Australia to Micronesia, the New Hebrides and New Caledonia (Santisuk et al., 1988, Tomlinson, 1987). S. alba is one of the common pioneer species growing in seaward areas which are inundated daily, and observed to grow well on soft, deep, unconsolidated soils (Santisuk, 1983, Santisuk et al., 1988). S.alba may be associated with Avicennia spp. in the lower intertidal zone (Thampanya et al., 2002), although it has also been reported to be found further seaward of Avicennia due to the ability of Sonneratia spp. to spread vegetatively especially on muddy substrata (Friess et al., 2012). Unlike A.corniculatum and A.alba, S.alba have large (3-4cm high, 4-5cm across) fruit containing 100-150 individual seeds (Santisuk et al., 1988) (Figure 4-1). 4.2.3 Experimental method Collection of plant material Propagules and fruit of A. alba, S. alba and A. corniculatum were collected from mature shrubs and trees growing along the Ngao river during the fruiting season of each of these respective species. Propagules of A. corniculatum, and A. alba were collected by gently shaking the branches of fruit laden trees to facilitate the release of ripe propagules (Figure 4-2 a and b). Fruit of S. alba fruit obtained by first examining individual fruit to assess their ripeness; the fruit considered ripe when the outer skin was cracked and the main body of the fruit had separated from the calyx (Ball and Pidsley, 1995, Duke and Jackes, 1987)(Figure 4-2c). Nursery pre-treatment Propagules and fruit collected in the field were transported to an experimental nursery located at the Ranong Mangrove Forest Research Centre on the banks of the Ngao river (Figure 4-3). To prepare for the arrival of the propagules of each species, seedling enclosures within the nursery had been lined on the bottom and sides with fine mesh to ensure that crabs could not enter and disturb the germinating seedlings. To protect the developing seedlings from desiccation, the enclosures were also covered at the top by shade cloth which restricted 50 % sunlight from reaching the seedlings. Upon arrival at the experimental nursery, propagules/ fruit were pre-treated according to the following regime: 4-5  Propagules of A. corniculatum required no pre-treatment and individual propagules were simply placed on the surface of the soil of seedling bags (18 cm x 10 cm) to allow them to germinate naturally (Figure 4-3a). Figure 4-1: Illustrations of (i) S. alba (a) habit, (b) flowers and immature fruit, (c) Mature fruit with typical reflexed sepals; (ii) A. corniculatum (a) branchlet with cluster of fruits, (b)fruit, (c) flower (d) longitudinal section of flower (e) stamen and (f) stems, and (iii) A. alba (a) habit, (b) branch with flowers and fruit, (c) flower seen from top and side (d) fruit and (e) buds (Giesen et al., 2006). (i) (ii) (iii) 4-6 Figure 4-2: Ripe propagules and fruit of (a) A. corniculatum, (b) A. alba and (c) S. alba prior to collection and transport to the experimental nursery. (a) (b) 4-7 Figure 4-2 continued (c)  A.alba propagules were soaked in a bucket of saline water for 48 hours after collection from the forest to facilitate the shedding of the external pericarp from the enclosed precociously developed seedling. Seedlings were then transplanted to seedling bags where they were placed on the soil surface and allowed to germinate naturally (Figure 43b) similar to A. corniculatum. During the period when A. alba seedlings were growing in the experimental nursery, the seedlings were attacked by monkeys, resulted in the death of the majority of the seedlings. Replacement natural seedlings (still with cotyledons attached) were sourced from mudflats in the natural mangrove forest surrounding the nursery. These seedlings were obtained by digging a core of mud surrounding the seedlings from mudflat areas and transplanting these seedlings to the experimental nursery.  S. alba fruit capsules were soaked in a bucket of saline water for a three-day period allowing the fruit to disintegrate and the seeds to be released. Seeds were then transferred to a plastic container containing soft muddy soil and scattered on the surface of the soil to facilitate germination (Figure 4-3c).  Newly germinated seedlings of S. alba were kept in this container for two weeks and then transferred to the seedling bags similar to the other two species. 4-8 Figure 4-3: Seedlings of (a) A. alba, (b) S. alba and (c) A. corniculatum seedlings in the experimental nursery prior to transplanting. (a) (b) 4-9 Figure 4-3 continued (c) b Growth of seedlings in the nursery Following pre-treatment, seedlings of the three species were grown in the experimental nursery for three months, after which healthy seedlings of uniform size were selected and tagged with vinyl numbered tags containing consecutive numbers attached by gardening wire. Of the three species, 610 individuals (each) of A. corniculatum and A. alba were selected and 305 S. alba seedlings. Fewer S. alba seedlings were selected due to the high mortality of seedlings in the nursery. 4-10 Transplanting of mature mangrove seedlings Once tagged and measured, seedlings of the three species were transplanted to the field sites by boat (Figure 4-4). Seedlings were planted according to a design where in each forest type, seedlings of one species were planted in one of the two transects (Figure 4-5). Within each forest type, the design for A. corniculatum and A. alba involved planting seedlings in ten m x ten m quadrats along the 50m transects. Within each quadrat seedlings were planted in six blocks of ten, three blocks in light gaps and three blocks under the canopy, making a total of 60 seedlings in each quadrat/ 300 seedlings in each forest type for each species (Figure 4-5). The design for S. alba differed from the other two species in that reduced numbers of seedlings were used, with seedlings planted in blocks of five instead of ten, making a total of 150 seedlings in each forest type. Field monitoring After transplanting, initial measurements were made of: (1) Seedling height from the soil surface to the apical tip; (2) Stem diameter at the soil surface; and (3) Number of leaves. Seedlings were then grown under experimental conditions in the field for twelve months. At the three, six and twelve month stages, seedling survival was recorded (seedlings classified as either alive or dead) and repeat measurements (from time of initial measurement in the nursery) made of the above seedling growth parameters. 4.2.4 Measurement parameters A detailed description of parameters selected to describe survival and growth of transplanted mangrove seedlings is included in Table 4-1. Table 4-1: Survival and growth parameters used during field monitoring Parameter Method Frequency Seedling survival Physical count of individual seedlings alive at each monitoring time and calculation of the proportion of transplanted seedlings still alive. Time 0, 3, 6, 12 months Change in seedling height to apical meristem In field monitoring after transplanting Time 3, 6, 12 months Height increment 3,6,12 months = Height at 3,6,12 months – Height 0 4-11 Figure 4-4: Tagged and measured A. corniculatum seedlings ready to be transported to the field. 4.2.5 Data analysis Survival To determine if there was a relationship between species of seedling transplanted, forest type, distance from the Ngao river and canopy type and the survival of transplanted mangrove seedlings three, six and twelve months after release, logistic regression analysis was used. Correlations between survival and environmental variables were calculated using the Pearsons correlation co-efficient. 4-12 Figure 4-5: Distribution of blocks of transplanted (a) A. corniculatum, (b) A. alba, and (c) S. alba seedlings in the two forest types. Figure 4-6: Block of transplanted seedlings at field site. 4-13 Growth A five-factor repeated measures ANOVA was performed to compare growth of transplanted seedlings growth data with the following factors: Within subjects: time (three months, six months and twelve months); Between subjects; species (A.corniculatum and S.alba), forest type (the “Aegiceras forest” and “Sonneratia-Avicennia forest”), distance from the Ngao river (0-10, 10-20, 20-30, 30-40 and 40-50m), canopy type (light gaps and under the canopy). Assumptions of the ANOVA that there are no outliers in any group (or overall) in the growth data were met by plotting the dataset using box plots and removing any outliers (defined as 1.5 times the interquartile range). Assumptions of the Homogeneity of variances, was assessed by Levene's test of homogeneity of variance and growth data were subsequently square-root-transformed to stabilise variances. Homogeneity of covariances, was assessed by Box's test of equality of covariance matrices while the normality of growth parameters for all interventions across all monitoring periods was assessed using the Kolomogorov Smirnov test. The assumption of sphericity was checked using Mauchly’s test. The Bonferroni method was used to perform pairwise comparisons of the distance factor following the significant test result. Correlations among growth and environmental variables were calculated using the Pearsons correlation co-efficient. 4.3 Results 4.3.1 Survival Forest and species All A.alba seedlings died within three months of transplanting to the field in both forest types. Three months The proportion of transplanted S. alba and A.corniculatum seedlings surviving to the three month monitoring period was significantly higher in the “Sonneratia-Avicennia forest” (0.87) than the “Aegiceras forest” (0.43) (Logistic regression, χ2(1) = 22.147, p<0.001) (Table 4-2, Table 4-3 and Figure 4-7). 4-14 A greater proportion of A. corniculatum seedlings transplanted into the two forest types survived after three months (0.79) than S. alba seedlings (0.37)(Logistic regression χ2(1) 21.949, p<0.001) (Table 4-2 and Table 4-3). There was no significant interaction between forest type and species of seedling planted on the proportion of transplanted seedlings surviving at three months (Logistic regression, χ2(1) = 3.003, p>0.05) indicating a consistent effect of species on seedling survival across the two forest types (Table 4-3). Six months The proportion of transplanted S. alba and A.corniculatum seedlings surviving to the six month monitoring period in the “Sonneratia-Avicennia forest” (0.74) was significantly higher than the “Aegiceras forest” (0.23) (Logistic regression, χ2(1) =32.954, p<0.001) (Table 4-2 and Table 4-3). A greater proportion of A. corniculatum seedlings transplanted into the two forest types survived to the six month monitoring period (0.59) than S. alba seedlings (0.27)(Logistic regression, χ2(1) = 18.436, p<0.001) (Table 4-2, Table 4-3 and Figure 4-7). There was a significant interaction between forest type and species of seedling planted on survival of seedlings at six months (Logistic regression, χ2(1) = 8.954, p < 0.01, indicating that the trend of varying seedling survival with forest type was not the same across the two species. Comparison of seedling survival of the two species with forest type indicates that S.alba had a stronger trend towards survival in the “Sonneratia-Avicennia forest”, than A.corniculatum seedlings (Table 4-2 and Figure 4-7). For example in the “Sonneratia-Avicennia forest”, the proportion of S.alba seedlings surviving to the six month monitoring period (0.51) was much greater than in the “Aegiceras forest” (0.027). In contrast in the “Sonneratia-Avicennia forest” the proportion of transplanted A.corniculatum seedlings surviving (0.85) was greater than that in the “Aegiceras forest” (0.34) (Table 4-2 and Figure 4-7). Twelve months At the twelve month period the proportion of A.corniculatum seedlings surviving (0.47) was greater than that of S.alba seedlings, none of which survived to this period (Table 4-2 and Figure 4-7). 4-15 Table 4-2: Variation in the proportion of transplanted (a) A. corniculatum and (b) S. alba seedlings surviving three, six and twelve months after transplanting. Values are proportion +/95% C.I. Forest type “Aegiceras forest” “SonneratiaAvicennia forest” Total Species of seedling transplanted No. of Months since transplanting 0 3 6 12 Count Proportion Count Proportion Count Proportion Count Proportion A.corniculatum 300 1 183 0.61 101 0.34 0 0.00 S.alba 150 1 12 0.08 4 0.027 0 0.00 Subtotal 450 1 195 0.43 105 0.23 0 0.00 A.corniculatum 300 1 293 0.98 255 0.85 140 0.47 S.alba 150 1 98 0.65 76 0.51 0 0.00 Subtotal 450 1 391 0.87 331 0.74 140 0.31 A.corniculatum 600 1 476 0.79 356 0.59 140 0.23 S.alba 300 1 110 0.37 80 0.27 0 0.00 Combined 900 1 586 0.65 436 0.48 140 0.16 Table 4-3: Results of logistic regression analysis predicting the probability of survival of transplanted seedlings at month three, six and twelve based on species and forest type. Factor B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Seedling survival 3 months Species 9.345 1.995 21.949 1 <0.001 11444.433 229.436 570857.667 Forest 9.127 1.94 22.147 1 <0.001 9203.704 205.611 411982.08 Forest * Species -1.157 0.668 3.003 1 >0.05 0.314 0.085 1.164 Seedling survival 6 months Species 8.809 2.052 18.436 1 <0.001 6696.135 120.071 373431.294 Forest 11.942 2.08 32.954 1 <0.001 153631.574 2604.357 9062760.454 Forest * Species -2.207 0.737 8.954 1 <0.01 0.11 0.026 0.467 Note: As seedlings of only one species (A.corniculatum in one forest type “Sonneratia-Avicennia forest” were alive at twelve months, “species” and “forest” factors and the interaction term “forest *species” are not included in the analysis at this time period. 4-16 Figure 4-7: Variation in the proportion of transplanted (a) A.corniculatum and (b) S.alba seedlings surviving in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” at three, six and twelve months. Figures show the significantly higher proportion of seedlings of both species in the “Sonneratia-Avicennia forest” and significantly higher survival of A.corniculatum seedlings compared to S.alba. Data are pooled across canopy and distance treatments. N=2, values are proportion +/-95% C.I. (a) (b) Distance from the Ngao river Three months The proportion of transplanted seedlings of both species surviving to the three month monitoring period differed significantly with distance from the Ngao, apart from an unexplained dip in seedling survival in quadrats located 20-30m from the Ngao river (0.33+/4-17 0.12) (Logistic regression, χ2(4) = 20.757, p<0.05) (Table 4-4, Table 4-5 and Figure 4-8). There was no significant interaction between species and distance (Logistic regression, χ2(4)=8.655, p>0.05) or between forest and distance (Logistic regression, χ2(4) =7.926, p>0.05 at the three month stage indicating a consistent distance effect on seedling survival across the two species and two forest types. Six months The proportion of transplanted seedlings of both species surviving to the six month monitoring period differed significantly with distance from the Ngao river with a general trend towards greater survival in quadrats further from the Ngao river (Logistic regression, χ2(4)=13.593, p<0.01) (Table 4-4, Table 4-5 and Figure 4-8). There was a significant interaction between species and distance at the six month stage (Logistic regression, χ2(4)=28.235, p<0.001), indicating that the trend of varying seedling survival with distance from the Ngao river was not the same across the two species. Comparison of seedling survival of the two species with distance at the six month monitoring period indicates that A.corniculatum seedlings had a stronger trend of survival in quadrats located further away from the Ngao river than S.alba seedlings (Table 4-4 and Figure 4-8). For example in the “Sonneratia-Avicennia forest”, the proportion of A.corniculatum seedlings surviving was the lowest (0.56+/-0.07) in the quadrat closest to the Ngao river and highest (0.96+/-0.02) in the most inland quadrat (40-50m from the Ngao river). There was no significant interaction between forest type and distance at the six month stage (Logistic regression, χ2(4)=6.976, p>0.05 indicating a consistent distance effect on seedling survival across the two forest types (Table 4-5). Twelve months The proportion of transplanted seedlings of both species surviving to the twelve month monitoring period were not statistically different (Logistic regression, χ2(4)=2.153, p>0.05), survival of transplanted A.corniculatum seedlings in the “Sonneratia-Avicennia forest” tended to increase with distance from the Ngao river, with the lowest survival recorded in the quadrat next to the river (0-10m) (0.11) and the highest in the quadrat furthest inland (0.65) (Tables 44 and Table 4-5, Figure 4-8). 4-18 Table 4-4: Variation in the proportion of transplanted (a) A. corniculatum and (b) S. alba seedlings surviving with distance from the Ngao river in the two forest types three, six and twelve months after transplanting. Data are pooled across canopy treatments. Values are proportion +/-95% C.I. Species of seedling transplanted Months since transplanting S.alba A.corniculatum Distance from the Ngao river (m) 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 “Aegiceras forest” 3 0.03 (0.03) 0(0) 0.1 (0.1) 0.1 (0.1) 0.16 (0.09) 0.55 (0.17) 0.43 (0.17) 0.51 (0.09) 0.81 (0.07) 0.73 (0.08) 6 0(0) 0(0) 0(0) 0.1 (0.1) 0.03 (0.03) 0.03 (0.03) 0.26 (0.15) 0.28 (0.1) 0.55 (0.1) 0.55 (0.06) 12 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) “Sonneratia-Avicennia forest” 3 0.63 (0.15) 0.86 (0.06) 0.33 (0.12) 0.56 (0.14) 0.86 (0.08) 0.93 (0.03) 0.98 (0.01) 1(0) 1(0) 0.96 (0.02) 6 0.53 (0.17) 0.63 (0.17) 0.1 (0.04) 0.5 (0.16) 0.76 (0.13) 0.56 (0.07) 0.88 (0.04) 0.88 (0.06) 0.95 (0.02) 0.96 (0.02) 12 0(0) 0(0) 0(0) 0(0) 0(0) 0.11 (0.04) 0.55 (0.04) 0.45 (0.05) 0.56 (0.08) 0.65 (0.02) Table 4-5: Results of logistic regression analysis predicting the probability of survival of transplanted seedlings at month three, six and twelve based on distance from the Ngao river Factor B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Seedling survival three months Distance 20.757 4 <0.001 Distance * Species 8.655 4 >0.05 Distance * Forest 7.926 4 >0.05 Distance 13.593 4 <0.01 Distance * Species 28.235 4 <0.001 Distance * Forest 6.976 4 >0.05 2.153 4 >0.05 Seedling survival 6 months Seedling survival 12 months Distance Note: As seedlings of only one species (A.corniculatum in one forest type “Sonneratia-Avicennia forest” were alive at twelve months, “the interaction terms distance*species” and “distance *forest” are not included in the analysis at this time period. 4-19 Figure 4-8: Variation in proportion of transplanted (a) S.alba and (b) A.corniculatum seedlings surviving with distance from the Ngao river in the “Aegiceras forest” and the “SonneratiaAvicennia forest”. Figures show the significantly higher proportion of seedlings of both species surviving in plots further away from the Ngao river, a trend more pronounced for A.corniculatum seedlings compared to S.alba. Data are pooled across canopy treatments. N=5. Values are proportion +/-95% C.I. Aegiceras forest Sonneratia-Avicennia forest (a) No. of months since transplanting 4-20 Figure 4.8 continued Aegiceras forest Sonneratia-Avicennia forest (b) No. of months since transplanting 4-21 Canopy type Three months The proportion of transplanted seedlings of both species surviving to the three-month stage differed significantly with canopy type, with a greater proportion of seedlings surviving in light gaps (0.74+/-0.02) than under the forest canopy (0.56+/-0.02) (Logistic regression, χ2(1) = 20.944, p<0.001) (Table 4-6, Table 4-7 and Figure 4-9). There was also a significant interaction between species and canopy type (Logistic regression, χ2(1) = 11.234, p<0.001), with canopy type having less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings (Table 4-6, Table 4-7 and Figure 4-9). Similarly, significant interactions were recorded between forest and canopy type (Logistic regression, χ2(1) =8.558, p<0.01) with canopy type having less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest” (Table 4-7). The final significant interaction was between distance and canopy type (Logistic regression, χ2(4)=27.16, p<0.001) with canopy type having less of an effect on survival of transplanted seedlings in the quadrats closest to the Ngao river than it did further inland (Table 4-7). Six months The proportion of transplanted seedlings of both species surviving to the six month monitoring period differed significantly with canopy type with a greater proportion of seedlings surviving in light gaps (0.57+/-0.02) than under the forest canopy (0.40+/-0.02) (Logistic regression, χ2(1)=25.12, p<0.001) (Table 4-6 and Table 4-7). There was a significant interaction between species and canopy type (Logistic regression, χ2(1)=20.805, p<0.001), with canopy type continuing to have less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings (Table 4-7). The significant interaction between forest and canopy also continued at the six month stage (Logistic regression, χ2(1)=12.722, p<0.05) with canopy type having less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest” (Table 4-7). In addition, the significant interaction between distance and canopy reported at the three month stage also continued to the six month stage (Logistic regression, χ2(4)= 13.036, p<0.001) with canopy type having less of an effect on survival of transplanted seedlings in the quadrats closest to the Ngao river than it did further inland (Table 4-7). 4-22 Twelve months The proportion of transplanted A.corniculatum seedlings surviving to the twelve month monitoring in light gaps (0.49+/-0.04) than under the forest canopy (0.45+/-0.04 although the differences were not statistically significant (Logistic regression, χ2(1)=0.148, p>0.05) (Table 4-6, Table 4-7 and Figure 4-9). There was no significant interaction between distance and canopy on survival of A.corniculatum seedlings at the twelve month stage (Logistic regression, χ2 (1)=0.875, p >0.05) (Table 4-7). Table 4-6: Variation in proportion of transplanted A. corniculatum and S. alba seedlings surviving in light gaps and under the forest canopy in the two forest types three, six and twelve months after transplanting. Values are proportion +/-95% C.I. Forest type Species No. of months since transplanting 3 “Aegiceras forest” “SonneratiaAvicennia forest” Total 6 12 Canopy Light gaps Canopy Light gaps Canopy Light gaps S. alba 0.01(0.01) 0.15(0.04) 0(0) 0.05(0) 0(0) 0(0) A. corniculatum 0.45(0.04) 0.77(0.05) 0.2(0) 0.46(0) 0(0) 0(0) Subtotal 0.30 (0.03) 0.56 (0.03) 0.14 (0.02) 0(0) 0(0) S. alba 0.48(0.06) 0.83(0.04) 0.29(0.05) 0.72(0.05) 0(0) 0(0) A. corniculatum 0.99(0.01) 0.97(0.01) 0.84(0.03) 0.86(0.03) 0.45(0.04) 0.49(0.04) Subtotal 0.82 (0.03) 0.66 (0.03) 0.81 (0.03) 0.30 (0.03) 0.32 (0.03) S. alba 0.25 (.04) 0.49 (.04) 0.15 (0.03) 0.39 (0.04) 0.04 (0) 0 (0) A. corniculatum 0.72 (0.03) 0.87 (0.02) 0.52 (0.03) 0.66 (0.03) 0.22 (0.02) 0.24 (0.02) Combined 0.56 (0.02) 0.74 (0.02) 0.40 (0.02) 0.57 (0.02) 0.15 (0.02) 0.16 (0.02) 0.92 (0.02) 0.33 (0.03) Table 4-7: Results of logistic regression predicting the probability of survival of transplanted seedlings at three, six and twelve months after transplanting based on canopy type. Development stage/ Species B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Seedling survival 3 months Canopy 10.533 2.302 20.944 1 <0.001 37532.052 412.398 3415766.232 Canopy * Species -2.636 0.787 11.234 1 <0.001 0.072 0.015 0.335 Canopy * Forest -2.245 0.767 8.558 1 <0.01 0.106 0.024 0.477 27.16 4 <0.001 Canopy * Distance 4-23 Table 4.7 continued continued Development stage/ Species B S.E Wald df p Odds ratio 95% C.I for odds ratio Lower Upper Seedling survival 6 months Canopy 10.17 2.029 25.12 1 <0.001 26104.068 489.256 1392771.405 Canopy * Species -2.843 0.623 20.805 1 <0.001 0.058 0.017 0.198 Canopy * Forest -2.076 0.582 12.722 1 <0.001 0.125 0.04 0.392 13.036 4 <0.05 0.148 1 >0.05 0.741 0.161 3.411 0.875 4 >0.05 Canopy * Distance Seedling survival 12 months Canopy -0.299 0.779 Canopy * Distance Note: As seedlings of only one species (A.corniculatum in one forest type “Sonneratia-Avicennia forest” were alive at twelve months, “the interaction terms canopy*species” and “canopy *forest” are not included in the analysis at this time period. 4.3.2 Growth Time Although growth of transplanted mangrove seedlings of both species increased over time, differences in mean height increment over time were not statistically significant (ANOVA, F(1,52)=1.097, p>0.05, partial η2=0.021. Similarly, there was no statistically significant interaction between month of monitoring and species, forest, distance or canopy (Table 4-8). Forest and species Comparisons of seedling growth are only possible between A.corniculatum and S.alba seedlings, as all A.alba seedlings died within three months of transplanting to the natural forest. In addition, as no S.alba seedlings survived to the twelve month monitoring period, no height comparisons can be made for this species for the twelve month period. Species The main effect of species also showed that there was a statistically significant difference in mean seedling height increment of species planted (ANOVA, F(1,52)=5.584, p<0.05, partial η2 = 0.097), with height increments greater for S.alba seedlings than A.corniculatum (Table 4.8, Figure 4-10). In the “Aegiceras forest”, for example the mean height increment of S.alba seedlings at the three and six month monitoring periods (0.7+/- 0.32cm and 2.13 +/- 0.43cm respectively) was higher than those of A.corniculatum (0.63 +/- 0.11 and 0.72+/- 0.13 respectively at the same period (Table 4-9). 4-24 Figure 4-9: Variation in proportion of transplanted (a) S.alba and (b) A.corniculatum seedlings surviving in light gaps and under the forest canopy in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Figures show the significantly higher proportion of seedlings of both species surviving in plots in light gaps than under the forest canopy, a trend more pronounced for S.alba seedlings compared to A.corniculatum and more pronounced in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. N=2. Values are proportion +/-95% C.I. (a) (b) No. of months since transplanting 4-25 Forest The main effect of forest type showed that there was a statistically significant difference in mean seedling height increment with forest type (ANOVA, F(1,52)=14.203, p <0.001, partial η2 = 0.215) with height increments greater for seedlings planted in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” (Table 4-8). For example when the two species were pooled, the mean height increment at the six month monitoring period in the “Aegiceras forest” was 0.85+/-0.15cm as compared to 3.67+/-0.45cm in the “Sonneratia-Avicennia forest” (Table 4-8 and Figure 4-10). Species * Forest There was a significant two-way interaction between forest type and species of seedling planted on mean seedling height increment, ANOVA, F(1,52)= 4.021, p<0.05, partial η2 =0.072, indicating that the trend of varying seedling growth with forest type was not the same across the two species. Comparison of seedling growth of the two species with forest type indicates although seedlings of both species had greater height increments in the “Sonneratia-Avicennia forest”, than the “Aegiceras forest”, the increased growth associated with transplanting to the “Sonneratia-Avicennia forest” was less pronounced for A.corniculatum seedlings. For example mean height increment A.corniculatum in the “Aegiceras forest” at six months was 0.72+/-0.13 compared to 1.8+/-0.23 in the “Sonneratia-Avicennia forest”. This compares the mean height increment of S.alba seedlings of 2.13 +/- 0.43 and 6.1+/-0.74 in the “Aegiceras forest” and the “Sonneratia-Avicennia forest” respectively (Table 4-9 and Figure 4-10). Table 4-8: Results of repeated measures ANOVA analysis comparing mean height increment of transplanted seedlings at month three and six based on time since planting, species, forest type, distance from the Ngao river and canopy type. Source Type III Sum of squares df Mean square F p Partial Eta Squared Time 0.054 1 0.054 1.097 >0.05 0.021 Time*species 0.108 1 0.108 2.197 >0.05 0.041 Time * forest 0.001 1 0.001 0.027 >0.05 0.001 Time*distance 0.674 4 0.168 3.442 >0.05 0.209 Time*canopy 0.065 4 0.016 0.332 >0.05 0.025 Within subjects effects 4-26 Table 4.8 continued continued Source Type III Sum of squares Error df Mean square 52 0.049 F p Partial Eta Squared Between subjects effects Species 1.884 1 1.884 5.584 <0.05 0.097 Forest 4.793 1 4.793 14.203 <0.001 0.215 Species*Forest 1.357 1 1.357 4.021 <0.05 0.072 Distance 4.292 4 1.073 3.180 <0.05 0.197 Species*Distance 0.838 4 0.209 0.621 >0.05 0.046 Forest*Distance 0.317 4 0.079 0.235 >0.05 0.018 Canopy 1.196 1 1.196 3.544 >0.05 0.064 Species*Canopy 0.157 1 0.157 0.466 >0.05 0.009 Forest*Canopy 0.059 1 0.059 0.175 >0.05 0.003 Distance*Canopy 0.617 4 0.154 0.457 >0.05 0.034 Error 17.548 52 0.337 Table 4-9: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings at three, six and twelve months after transplanting in the two forest types. Data are pooled across canopy and distance treatments. N=2. Values are mean +/-S.E. No. of months since transplanting Species of seedlings Forest type transplanted 3 6 12 “Aegiceras forest” Sonneratia alba 0.7 +/- 0.32 2.13 +/- 0.43 . Aegiceras corniculatum 0.63+/- 0.11 0.72+/- 0.13 . Total 0.64+/-0.10 0.85+/0.15 . 3.39+/- 0.42 6.1+0.74 . Aegiceras corniculatum 1.26+/- 0.15 1.8+/- 0.23 10.33+/- 1.16 Total 2.29+/-0.26 3.67+/-0.45 10.33+/-1.16 Sonneratia alba 2.91+/-0.39 5.78+/-0.71 . Aegiceras corniculatum 0.95+/-0.10 1.37+/-0.17 10.33+/-1.16 Total 1.67+/-0.18 2.84+/-0.35 10.33+/-1.16 “Sonneratia-Avicennia Sonneratia alba forest” Total 4-27 Figure 4-10: The mean height increment of (a) S. alba and (b) A. corniculatum seedlings three, six and twelve months after transplanting to the “Aegiceras forest” and the “SonneratiaAvicennia forest”. Figures show greater growth increments in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” which was more pronounced for S.alba seedlings than A.corniculatum. Figures also show higher greater mean height increments of S.alba seedlings as compared to A.corniculatum seedlings. Data are pooled across canopy and distance (a) Mean Height Increment (cm) treatments. N=2. Values are mean +/-S.E. Mean Height Increment (cm) (b) No. of months since transplanting 4-28 Distance from the Ngao river The main effect of distance showed that there was a statistically significant difference in mean seedling height increment with distance from the Ngao river (ANOVA, F(4,52)=3.180, p<0.05, partial η2 = 0.197)(Table 4-8). There was also a trend in both species towards greater height increment in quadrats further away from the Ngao river especially in the “SonneratiaAvicennia forest”. At the six month period, for example, the highest mean height increment of S.alba and A.corniculatum seedlings was found in the innermost quadrat (8.61+/-1.45cm, and 2.82+/- 0.35cm respectively). In contrast, the lowest mean height increment was recorded in the quadrat closest to the Ngao river (3.85 +/-0.93 and 0.68 +/-0.27 respectively)(Table 4-10 and Figure 4-11). There was no interaction between species of seedling planted and distance from the Ngao river (ANOVA, F(4,52) =0.621, p>0.05, partial η2 = 0.046) or interaction between forest type and distance from the Ngao river (ANOVA, F(4,52) =0.235, p>0.05, partial η2=0.018), indicating a consistent distance affect across these other treatments (Table 4-8). Table 4-10: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings three, six and twelve months after transplanting with distance from the Ngao river in the two forest types. Data are pooled across canopy treatments. N=5. Values are mean +/-S.E. “Aegiceras forest” Species S. alba A.corniculatum No. of months since transplanti ng 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 3 0(0) 0(0) 2.03(0) 1.03(0) 0.37 (0.24) 0.69 (0.38) 0.55 (0.19) 0.43 (0.15) 0.57 (0.18) 0.86 (0.22) 6 0(0) 0(0) 0(0) 2.56(0) 1.7(0) 0(0) 0.56 (0.17) 0.87 (0.29) 0.37 (0.17) 1.16 (0.26) 12 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 0(0) 20-30 30-40 40-50 “Sonneratia-Avicennia forest” Species S. alba No. of months since transplanti ng 0-10 A.corniculatum 10-20 20-30 30-40 40-50 4-29 0-10 10-20 Table 4.10 continued continued “Aegiceras forest” Species S. alba A.corniculatum No. of months since transplanti ng 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 3 2.18 (0.3) 2.42 (0.77) 2.17 (0.62) 4.37 (0.9) 5.37 (0.99) 0.49 (0.15) 1.03 (0.19) 1.15 (0.16) 1.53 (0.46) 2.06 (0.24) 6 3.85 (0.93) 4.5 (1.73) 4.7 (0.77) 7.3 (1.62) 8.61 (1.45) 0.68 (0.27) 1.39 (0.36) 1.66 (0.29) 2.47 (0.76) 2.82 (0.35) 12 0(0) 0(0) 0(0) 0(0) 0(0) 2.81 (2.2) 8.51 (2.32) 9.43 (0.94) 12.9 (2.66) 15.47 (1.94) Canopy type The main effect of canopy type showed that mean height increments of seedlings grown in light gaps were generally larger than those grown under the forest canopy but differences were not statistically significant (ANOVA, F(1,52)=3.544, p>0.05, partial η2=0.064) (Table 4-11 and Figure 4-12). There was no significant interaction between species of seedling transplanted and canopy type (ANOVA, F(1,52)=0.466, p>0.05, partial η2=0.009), between forest and canopy (ANOVA, F (1,52)=0.175, p>0.05, partial η2=0.003) or distance and canopy (ANOVA, F (1,52) = 0.457, p >0.05, partial η2 = 0.034) indicating a consistent canopy effect on height increment across these other treatments (Table 4-8). 4-30 Figure 4-11: Variation in the mean height increment of transplanted (a) S.alba and (b) A, corniculatum seedlings three, six and twelve months after transplanting at different distances from the Ngao river in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Figures show greater height increments in plots located away from the Ngao river. Data are pooled across canopy treatments. N=30. Values are mean +/-S.E. (a) (b) No. of months since transplanting 4-31 Figure 4-11 continued. No. of months since transplanting Table 4-11: Variation in the mean height increment (cm) of S. alba and A. corniculatum seedlings at three, six and twelve months after transplanting in different canopy types in the two forest types. N=2. Values are mean +/-S.E. Forest type Species No. of months since transplanting 6 3 “Aegiceras forest” “SonneratiaAvicennia forest” Total 12 Shade Sun Shade Sun Shade Sun S. alba 0.3(0) 0.78(0.38) 0(0) 2.13(0.43) 0(0) 0(0) A. corniculatum 0.59(0.19) 0.66(0.11) 0.69(0.31) 0.74(0.13) 0(0) 0(0) Subtotal 0.58(0.18) 0.69(0.12) 0.69(0.31) 0.92(0.18) 0(0) 0(0) S. alba 2.54(0.44) 4.24(0.65) 5.01(0.98) 6.79(1.02) 0(0) 0(0) A. corniculatum 1.01(0.18) 1.51(0.22) 1.4(0.23) 2.21(0.39) 7.48(1.09) 13.18(1.78) Subtotal 1.74(0.27) 2.83(0.42) 2.76(0.53) 4.42(0.68) 7.48(1.09) 13.18(1.78) S. alba 2.39(0.43) 3.33(0.6) 5.01(0.98) 6.21(0.97) 0(0) 0(0) A. corniculatum 0.81(0.14) 1.08(0.14) 1.18(0.19) 1.52(0.26) 7.48(1.09) 13.18(1.78) Combined 1.35(0.2) 1.95(0.29) 2.29(0.44) 3.23(0.51) 7.48(1.09) 13.18(1.78) 4-32 Figure 4-12: Variation in the mean height increment of (a) S. alba and (b) A. corniculatum seedlings at three, six and twelve months after transplanting to light gaps and under the forest canopy in the “Aegiceras forest” and the “Sonneratia-Avicennia forest”. Figures show no significant difference between mean height increments in plots located in light gaps compared to those under the forest canopy. N=2. Values are mean +/-S.E. (b) (a) No. of months since transplanting 4.4 Discussion Once established at a site, many factors are known to influence the post-establishment survival and growth of mangrove seedlings in the lower intertidal zone and their progression to the sapling stage of the mangrove tree life cycle (Krauss et al., 2008, Osunkoya and Creese, 1997). Uncertainties remain as to the particular tolerances of mangrove species to environmental conditions prevalent in these forest communities. The focus of the current chapter of the thesis has been to assess the influence of forest type, distance from the Ngao river and canopy type and associated variation in hydroperiod and 4-33 light availability, on survival and growth of mature seedlings of lower intertidal specialist species transplanted to sites along the Ngao river, southern Thailand. Information presented in the current chapter of the thesis contribute to our understanding of the functioning of these relatively unstudied communities and have direct application to enhancement of restoration of these forest communities and specifically matching of appropriate species with environmental conditions at restoration sites. These points will be expanded upon below with further discussion focusing initially on survival and growth of mangrove seedlings under different experimental treatments, followed by examination of the relationship of seedling survival and growth to hydroperiod to observed results. 4.4.1 Survival and growth of mangrove seedlings Overall rates of survival in the current study declined gradually over the monitoring period. Across the two species, a greater proportion of A.corniculatum seedlings survived than S.alba in all monitoring periods, although differences were only significant for the initial three month monitoring period, with no significant difference at six months. At twelve months only A.corniculatum seedlings survived. The higher survival rate of A.corniculatum seedlings is consistent with the higher density of seedlings of this species (as compared to S.alba and A.alba) in both forest types, as reported in Chapter 3 of this thesis and adds further weight to the argument that the greater supplies of cotyledonary reserves of A.corniculatum seedlings allow for it to establish and survive (at least initially) in a broad range of sites including those which are unsuitable for longer term survival. Interestingly, although S.alba with a small seed reserve and therefore lower cotyledonary reserves had a lower overall rate of survival across the two forests, where seedlings of this species did survive, they grew at a faster rate than A.corniculatum seedlings. This finding is consistent with the observation of higher numbers of naturally occurring S.alba seedlings in the “Sonneratia-Avicennia forest” as compared to the “Aegiceras forest” as described in Chapter 3 and also demonstrates the adaptive ability of this species to conditions in the low intertidal. Variation in survival and growth of different species of transplanted mangrove seedlings is well documented in the literature and is considered as one of the drivers of commonly observed zonation patterns as “physiological attributes of species determine how individual species respond and interact with other species” (Ball and Pidsley, 1995). For example, studies in China 4-34 found that A.corniculatum and A.marina stem elongation was higher in lower elevations while B.gymnorhiza and R.stylosa had higher survival rates and quicker stem elongation and more biomass accumulation at higher elevations (He et al., 2007). The proportion of seedlings surviving to twelve months averaged 47% for A.corniculatum and 0% for S.alba. These survival rates compare those from studies of transplanted Avicennia spp. from New Zealand and South Eastern Australia of 54%, 49.6%, 47.0% and 53.2% (Burns and Ogden, 1985, Clarke and Allaway, 1993, Clarke and Myerscough, 1993) cited in Osunkoya and Creese (1997) respectively). Seedlings planted in warmer climates were reported to have lower survival rates (Osunkoya and Creese, 1997) ranging from 8% survival of Avicennia seedlings in Panama (Rabinowitz, 1978b), 17% survival of Avicennia seedlings in northern Queensland (Smith, 1987b) and 27.6% and 33.0% survival of transplanted R. stylosa and B. cylindrica seedlings in southern Thailand (Tamai and Iampa, 1988). Similarly growth rates for transplanted seedlings were relatively slow as compared to previous transplant studies of S.alba seedlings from Thailand (Thampanya et al., 2002) and Avicennia seedlings from north Queensland (0.4 and 0.15cm per month and 0.14 cm and 0.06cm in New Zealand (Osunkoya and Creese, 1997). These results are consistent with previous studies focusing on lower intertidal communities which characterised the lower intertidal environment as a stressful physico-chemical environment. Differences in survival and growth between forest types In the current study a greater proportion of seedlings of both species transplanted into the “Sonneratia-Avicennia forest” survived than those transplanted to the “Aegiceras forest”. Growth rates of transplanted seedlings of both species were also higher in the “SonneratiaAvicennia forest” than in the “Aegiceras forest”. Comparison of the results of the current study with studies of the mature forest (Chapter 2) and propagule and immature seedling populations (Chapter 3) provides some interesting insights. For A.corniculatum seedlings, the findings are inconsistent with the results of Chapter 2, where A.corniculatum was the dominant mature tree species across the “Aegiceras forest”. The result is also inconsistent with the higher density of naturally dispersed A. corniculatum seedlings found in the “Aegiceras forest” compared to the “Sonneratia-Avicennia forest”, as reported in Chapter 3 of the thesis. The findings are however consistent with the lack of 4-35 persistence of A.corniculatum seedlings reported in the “Aegiceras forest” as compared to the “Sonneratia-Avicennia forest” also reported in Chapter 3 where although found at lower densities than the “Aegiceras forest”, A.corniculatum seedlings were able to persist across the forest type for most of the twelve month monitoring period. In contrast, the higher rate of survival and growth of S.alba seedlings in the “SonneratiaAvicennia forest”, is consistent with the dominant role that this species plays in the mature “Sonneratia-Avicennia forest” as compared to the “Aegiceras forest” as described in Chapter 2 of the thesis. The strong growth and initial high survival rates of S.alba seedlings however aren’t consistent with the low density of natural seedlings of this species as described in Chapter 3. Variation in survival and growth of mangrove seedlings with forest type and the presence of conspecific trees is well documented in the literature. Similar to the results for A.corniculatum in the current study, studies in Florida, for example, have shown that the greatest survival and growth mangrove seedlings was reported in areas not dominated by conspecifics (Rabinowitz, 1978b). Similarly, in studies of four species of mangrove seedlings in Australia, three of the species performed best in habitat where adult conspecifics were least abundant and all four species appeared to prefer the same zone (high intertidal) (Smith, 1987a). Differences in survival and growth with distance from the Ngao river The current study found that that survival of both species of seedlings tended to increase with distance of planting from the Ngao river, a result more pronounced for A.corniculatum seedlings which had a stronger trend towards increased survival and growth in sites away from the river than did S.alba seedlings. Seedling growth of both species was also significantly higher for both species in quadrats furthest from the Ngao river. The trend for increased survival and growth of A.corniculatum with distance is partially consistent with the result of Chapter 3 of the thesis where the peak in naturally arriving A.corniculatum seedlings was reported in the middle zone of the “Aegiceras forest”. The trend is however consistent with the greater densities of naturally arriving A.corniculatum seedlings observed away from the Ngao river in the “Sonneratia-Avicennia forest” also reported in Chapter 3 of the thesis. The reduced effect of distance to the Ngao river on survival of S.alba seedlings in the current thesis suggests a higher tolerance of seedlings of this species to environmental conditions 4-36 closer to the Ngao river than A.corniculatum seedlings. Higher growth of this species in sites away from the river suggests that although seedlings of the species can tolerate environmental conditions at lower elevations, they grow best at high elevations. Focusing on the two species included in the current thesis, previous studies typically show Sonneratia spp. such as S.alba typically being located at the lowest tidal elevations of all mangrove species and attributed with high tolerance for sites exhibiting or prolonged inundation, perennially waterlogged, muddy banks at the lowest elevations colonized by mangrove seedlings (Ball and Pidsley, 1995). A.corniculatum seedlings are seemingly less tolerant to these more exposed locations than S.alba seedlings with other studies describing the distributions of A. corniculatum and A. marina as being differentiated along gradients of soil moisture, with A. corniculatum less prevalent in saturated conditions (Saintilan, 1998). Differences in survival and growth with canopy type Analysis of survival and growth of transplanted seedlings according to whether seedlings were transplanted to light gaps or sites under the canopy provides further insights as to environmental factors causing the differences in survival and growth of transplanted mangrove seedlings across the two forest types. A greater proportion of mangrove seedlings transplanted inside light gaps survived to the three and six month stage than those planted under the forest canopy. At twelve months however (when only A.corniculatum seedlings planted in the “Sonneratia-Avicennia forest” survived), there was no significant difference in survival. Seedling height increments were also consistently higher in light gaps than under the forest canopy. Canopy type (and specifically whether the seedlings were planted in light gaps) also appeared to have a larger positive (but not significantly different) effect on survival of transplanted seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. The greater effect of light gaps in the denser “Aegiceras forest” is consistent with the denser nature of this forest type compared to the more open canopy “Sonneratia-Avicennia forest” as described in Chapter 2 of the thesis. Similarly, planting seedlings in light gaps had a larger positive (but not significantly different) effect on survival of S.alba seedlings than A.corniculatum seedlings. The results of the current study are consistent with the work of many authors who have recognised the importance of light gaps in recruitment from the seedling stage to the sapling stage (Clarke and Allaway, 1993, Clarke and Myerscough, 1993, Smith et al., 1994) and species 4-37 specific effects of light gaps on seedling survival and growth. Clarke (2004), for example, noted that five species survived and attained greater heights in large canopy gaps than in smaller gaps and no species survived under the canopy. Previous studies in the Ngao river by Tamai and Iampa (1988) also found that survival and growth of mangrove seedlings (Rhizophora and Bruguiera) in light gaps were enhanced as compared to those planted under the forest canopy. Previous studies of A. corniculatum seedlings in Australia, for example, noted that similar to the current study, this species survived and grew best on open unshaded areas (Smith, 1992, Osborne and Smith, 1990, O'Grady et al., 1996), while also being able to survive in the understory of other species (Clarke, 1995a). Previous studies of S.alba have noted that the species grows best in unshaded areas (Thampanya et al., 2002, Smith, 1992, Macnae, 1969). 4.4.2 Change dynamics Taken together, the significant differences in survival and growth of transplanted seedlings between and within the two forest types and between canopy types adds further to our understanding of the different patterns of recruitment occurring and demographic change potentially taking place in the two forest types. A.corniculatum seedlings, for example, appear to be demonstrating their ability to survive and grow in the “Sonneratia-Avicennia forest” a forest type not dominated by mature A.corniculatum trees, suggesting that the large numbers of propagules and seedlings documented to arrive and persist over time in this forest type (Chapter 3) will most likely rapidly move through to the sapling stage and may come to dominate the forest type over time. In contrast although S.alba seedlings could grow quickly in this forest type, long term survival was nil, suggesting that physicochemical parameters present in this forest such as increased competition associated with populations of shade tolerant A.corniculatum seedlings and saplings were not suitable for continued growth of this species once its embryonic reserves were exhausted. This result raises questions as to how this species can continue to dominate the “Sonneratia-Avicennia forest” if few seedlings can establish initially (as documented in Chapter 3) and then at the later stage of seedling development established seedlings can initially survive and grow but cannot proceed to the sapling stage. The highly competitive nature of A.corniculatum has been described in previous studies in Australia where the species is described as being able to outcompete populations of Avicennia by establishing and forming dense stands in the understorey of Avicennia, and then overtime if stable conditions remained, being able to inhibit the establishment and recruitment of Avicennia seedlings leading to the die back of Avicennia stands until the next 4-38 disturbance (Clarke, 1995b). The results for the “Aegiceras forest”, in contrast, confirm results from Chapter 3 that for both species survival and growth of seedlings was low, with not even light gaps being able to provide necessary resources to help them move to the sapling stage. This result raises questions as to how the mature trees in the “Aegiceras forest” can replace themselves if no A.corniculatum seedlings can persist in this forest type. 4.4.3 Environmental factors responsible for observed patterns of survival and growth The significant differences in survival and growth of transplanted seedlings between the two forest types and further significant differences with canopy condition and with distance from the Ngao river raises questions as to what environmental factors are responsible these observed differences. The main driver for the differing rates of survival and growth of transplanted seedlings in the two forest types appears to be the significantly different levels of light availability in the two forest types arising from the differing forest structure of these two lower intertidal communities, the denser “Aegiceras forest” and the more open canopy “Sonneratia-Avicennia forest” as described in Chapter 2 of the thesis. Forest type in turn is driven by variation in the duration of inundation between the two forest types as described in Chapter 2 of the thesis which is a major factor controlling the overall spatial patterns of distribution of these two communities. Evidence for the preeminent role of light availability comes initially from the significant differences in the greater proportion of seedlings of both species surviving in light gaps than under the forest canopy and the fact that these differences were more substantial in the denser “Aegiceras forest” than the more open “Sonneratia-Avicennia forest”. Both of these facts suggest a positive response to transplanted mangrove seedlings to sites of increased light availability. The known sensitivity of mangrove seedlings which have lost their cotyledons to light intensity and other variation in resource availability (Clarke and Allaway, 1993) supports this argument. Variation in survival and growth of mangrove seedlings according to light intensity is a wellaccepted theory in the mangrove literature where mangrove species are typically classified into two categories; shade tolerant and shade intolerant. Using this system of classification, S. alba has been classified by Macnae (1969), Wells (1982), Saenger (1982) (Smith (1992), Osborne and Smith (1990), and O'Grady et al. (1996) as a shade intolerant species, or species “with high photosynthetic capacity, requiring high levels of light intensity for maximum 4-39 photosynthesis and thus growth” (Clough, 1992). The shade tolerance status of A. corniculatum however appears to be in dispute as some authors (for example (Clarke and Hannon, 1971, Wells, 1982, Hinrichs et al., 2008) regard A. corniculatum as shade tolerant while others (Osborne and Smith, 1990, Macnae, 1969) noted that A. corniculatum seedlings survived and grew best on open unshaded areas. Clarke (1995a) described different recruitment strategies of A. corniculatum and Avicennia (marina) in Australia, noting that “Aegiceras does not appear to require disturbance for recruitment as evidenced by the presence of shrubs in the understorey of Avicennia stands”. This inconsistency can be explained by variation in tolerance to light intensity according to the stage of the development of the plant, with seedlings sometimes found to be shade tolerant where mature trees are not (Saenger, 1982). This appears to be the case with A. corniculatum which as a seedling has been categorized as shade tolerant but as an adult tree commonly observed growing as with dense canopies (Ball, 1998). Light tolerance was offered as an explanation for observed difference in the species composition of mangrove forest in Indonesia by Hinrichs et al. (2008), noting that B. gymnorrhiza, is less tolerant to shading than A. marina or A. corniculatum and only develops in the shade of other trees. This was given as the reason why this species does not occur on lightened sediment accretion banks and their seedlings might be in strong competition with understory species. Supporting the argument for the importance of light levels in seedling survival and growth in the two forest types, over a direct role of duration of inundation in differentiating survival and growth of transplanted seedlings, comes from the results of Chapter 2 of the thesis where in the “Aegiceras forest” (where surface elevations were the lowest), S.alba trees occupied the lowest surface elevations and A.corniculatum dominated all quadrats except those closest to the Ngao river. This suggests that both mangrove species are tolerant to levels of tidal inundation which characterise the two forest types. Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to determine rates of survival and growth of transplanted seedlings, both of which were significantly greater in plots away from the Ngao river than plots close to it. An important additional finding was that A.corniculatum seedlings appeared to be more sensitive to environmental conditions in sites close to the Ngao river than S.alba seedlings. Physical 4-40 environmental factors at locations closer to the Ngao river are likely to include increased wave action which can result in broken stems (Aksornkoae, 1993), currents which undermine and wash away seedlings (Clarke and Myerscough, 1993, Balke et al., 2013, Friess et al., 2012) and increased rates of sedimentation which impact on pneumatophores, seeds and seedlings (Ellison, 1999, Oliver, 1982, Thampanya et al., 2002). These impacts are likely to be more pronounced due to the comparatively stronger mechanical effects of tides and wave currents in the lower intertidal establishment (Clarke and Allaway, 1993, Clarke and Myerscough, 1993). Mangrove species are reported to vary in tolerance to physical effects such as those present in more exposed lower intertidal areas. In southern Thailand for example the tolerance to flooding was reported to decrease in order from R. mucronata, S. alba, R. apiculata, A. officinalis, C. tagal, B. cylindrica and X.granatum (Kitaya et al., 2002). Highly relevant to the current results are the findings of another study in southern Thailand, where lower intertidal colonizers such as A. alba and S. alba had greater survival rates at exposed sites with high hydrodynamic energy and higher degree of sediment disturbance than Rhizophora sp. (Thampanya et al., 2002). Reasons given for the difference in relative ability of these three species to withstand physical effects are the root system of Avicennia and Sonneratia including long star shaped cable roots which extend across the mudflat (Augustinus, 1995 cited in Friess et al. (2012). Young seedlings such as those used in the current study are considered to be particularly vulnerable to physical effects due to lack of aerial roots, their overall size which may result in oxygen deficiencies during submergence by the tides (McKee, 1993). 4.5 Conclusions The objective of the chapter was to answer the questions, “Does growth and survival of transplanted mangrove seedlings vary across the two different lower intertidal communities?; and “Does variation in environmental parameters between forest types, with distance from the Ngao river within each forest type and with canopy type influence the rate of survival and growth of transplanted seedlings?” Results of the study show that survival and growth of S.alba and A. corniculatum seedlings was greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” while across the two species, survival of A.corniculatum seedlings was greater than S.alba seedlings in both forest types. In contrast, growth increments of S.alba seedlings were greater than those of A.corniculatum in both forest types. There was also a significant interaction between forest type and species of seedling planted on growth, with the increased growth associated with 4-41 transplanting to the “Sonneratia-Avicennia forest” was less pronounced for A.corniculatum seedlings (Figure 4-13a). Results of the study added further that survival and growth of both species increased with distance from the Ngao river and that the effect of distance was stronger for A.corniculatum seedlings than S.alba seedlings (Figure 4-13b). Results of the final treatment in Chapter 4 showed that the proportion of transplanted seedlings of both species surviving differed with canopy type, with a greater proportion of seedlings surviving in light gaps than under the forest canopy. There was also a trend towards higher growth increments in light gaps but differences were not statistically significant. Canopy type had less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest” and also had less of an effect on survival of transplanted seedlings in the quadrats closest to the Ngao river than it did further inland. In addition canopy type also had less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings (Figure 4-13c). The main driver for the differing rates of survival and growth of transplanted seedlings in the two forest types appears to be the significantly different levels of light availability in the two forest types as a result of the differing forest structure of these two lower intertidal communities, the denser “Aegiceras forest” and the more open canopy “Sonneratia-Avicennia forest”. Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to determine rates of survival and growth of transplanted seedlings, both of which were significantly greater in plots away from the Ngao river than plots close to it. The effects of light level and physical disturbance on seedling establishment, although not unequivocal, are consistent with results from the literature and also provide further insights into tolerances of the two lower intertidal species to environmental conditions. Knowledge of these species specific tolerances to tidal inundation, physical environmental factors and light will serve to improve current and future restoration efforts. 4-42 Figure 4-13: Visual summary of differences in univariate forest structure parameters (a) between the two forest types, (b) with distance from the Ngao river and (c) with canopy type. (a) Evidence Transplanted seedlings Aegiceras forest Sonneratia Avicennia forest Forest Survival A.corniculatum S.alba < < > >> Growth A.corniculatum S.alba < < > > Species Survival A.corniculatum S.alba > < > < Growth A.corniculatum S.alba < > < > (b) Evidence -Mature trees Distance (m) Aegiceras forest Sonneratia-Avicennia forest 0-10 10-20 20-30 30-40 40-50 0-10 10-20 20-30 30-40 40-50 Hydroperiod > >> >>> >>>> >>>>> > >> >>> >>>> >>> A.corniculatum > >>> >>> >>>> >>>> >>> >>>> >>>> >>>>> >>>>> S.alba > > > >> >> >> >>> > >>>> >>>>> A.corniculatum > >> >>> > >>> >> >> >>> >>> >>> S.alba > > > >> >> >> >>> >>> >>>> >>>> Survival Growth (c) Evidence Transplanted seedlings Canopy Survival A.corniculatum S.alba Growth A.corniculatum S.alba Aegiceras forest Light gaps >> >>> = = Sonneratia Avicennia forest Canopy < < = = 4-43 Light gaps > >> = = Canopy < < = = 4.6 References AKSORNKOAE, S. 1993. The Ecology of Mangroves, IUCN. BALKE, T., BOUMA, T. J., HERMAN, P. M. J., HORSTMAN, E. M., SUDTONGKONG, C. & WEBB, E. L. 2013. Cross-shore gradients of physical disturbance in mangroves: implications for seedling establishment. Biogeosciences, 10, 5411-5419. BALL, M. C. 1998. Mangrove Species Richness in Relation to Salinity and Waterlogging: A Case Study Along the Adelaide river Floodplain, Northern Australia. Global Ecology and Biogeography Letters, 7, 73-82. BALL, M. C. & PIDSLEY, S. M. 1995. Growth Responses to Salinity in Relation to Distribution of Two Mangrove Species, Sonneratia alba and S. lanceolata, in Northern Australia. Functional Ecology, 9, 77-85. BURNS, B. R. & OGDEN, J. 1985. The demography of the temperate mangrove [Avicennia marina (Forsk.) Vierh.] at its southern limit in New Zealand. Australian Journal of Ecology, 10, 125-133. CHEN, Y.-P. & YE, Y. 2014. Early responses of Avicennia marina (Forsk.) Vierh. to intertidal elevation and light level. Aquatic Botany, 112, 33-40. CLARKE, L. D. & HANNON, N. J. 1971. The Mangrove Swamp and Salt Marsh Communities of the Sydney District: IV. The Significance of Species Interaction. Journal of Ecology, 59, 535-553. CLARKE, P. 1995a. The population dynamics of the mangrove Avicennia marina; demographic synthesis and predictive modelling. Hydrobiologia, 295, 83-88. CLARKE, P. & MYERSCOUGH, P. 1993. The intertidal distribution of the grey mangrove {Avicennia marina) in southeastern Australia: The effects of physical conditions, interspecific competition, and predation on propagule establishment and survival. Australian Journal of Ecology 18, 307-315. CLARKE, P. J. 1995b. The population dynamics of the mangrove shrub Aegiceras corniculatum: fecundity, dispersal, establishment and population structure. Proc. Linn. Soc. N.S.W, 115, 35-44. CLARKE, P. J. 2004. Effects of experimental canopy gaps on mangrove recruitment: lack of habitat partitioning may explain stand dominance. Journal of Ecology, 92, 203-213. CLARKE, P. J. & ALLAWAY, W. G. 1993. The regeneration niche of the grey mangrove (Avicennia marina): effects of salinity, light and sediment factors on establishment, growth and survival in the field. Oecologia, 93, 548-556. CLOUGH, B. F. 1992. Primary Productivity and Growth of Mangrove Forests. Tropical Mangrove Ecosystems. American Geophysical Union. DUKE, N. 1991. A Systematic Revision of the Mangrove Genus Avicennia (Avicenniaceae) in Australasia. Australian Systematic Botany, 4, 299-324. 4-44 DUKE, N. & JACKES, B. 1987. A systematic revision of the mangrove genus Sonneratia (Sonneratiaceae) in Australasia. Blumea, 32, 277-302. ELLISON, A. M. & FARNSWORTH, E. J. 1993. Seedling Survivorship, Growth, and Response to Disturbance in Belizean Mangal. American Journal of Botany, 80, 1137-1145. ELLISON, J. C. 1999. Impacts of Sediment Burial on Mangroves. Marine Pollution Bulletin, 37, 420-426. FRIESS, D. A., KRAUSS, K. W., HORSTMAN, E. M., BALKE, T., BOUMA, T. J., GALLI, D. & WEBB, E. L. 2012. Are all intertidal wetlands naturally created equal? Bottlenecks, thresholds and knowledge gaps to mangrove and saltmarsh ecosystems. Biological Reviews, 87, 346-366. GIESEN, W., WULFFRAAT, S., ZIEREN, M. & SCHOLTEN, L. 2006. MANGROVE GUIDEBOOK for Southeast Asia, Bangkok, Thailand, Food and Agriculture Organization of the United Nations, Regional Office for Asia and the Pacific. HE, B., LAI, T., FAN, H., WANG, W. & ZHENG, H. 2007. Comparison of flooding-tolerance in four mangrove species in a diurnal tidal zone in the Beibu Gulf. Estuarine, Coastal and Shelf Science, 74, 254-262. HINRICHS, S., NORDHAUS, I. & GEIST, S. J. 2008. Status, diversity and distribution patterns of mangrove vegetation in the Segara Anakan lagoon, Java, Indonesia. Regional Environmental Change, 9, 275-289. KITAYA, Y., JINTANA, V., PIRIYAYOTHA, S., JAIJING, D., YABUKI, K., IZUTANI, S., NISHIMIYA, A. & IWASAKI, M. 2002. Early growth of seven mangrove species planted at different elevations in a Thai estuary. Trees, 16, 150-154. KOMIYAMA, A., SANTIEAN, T., HIGO, M., PATANAPONPAIBOON, P., KONGSANGCHAI, J. & OGINO, K. 1996. Microtopography, soil hardness and survival of mangrove (Rhizophora apiculata BL) seedlings planted in an abandoned tin-mining area. . Forest Ecology and Management 81, 243–248. KRAUSS, K. W., LOVELOCK, C. E., MCKEE, K. L., LÓPEZ-HOFFMAN, L., EWE, S. M. L. & SOUSA, W. P. 2008. Environmental drivers in mangrove establishment and early development: A review. Aquatic Botany, 89, 105-127. MACNAE, W. 1969. A General Account of the Fauna and Flora of Mangrove Swamps and Forests in the Indo-West-Pacific Region. In: FREDERICK, S. R. & MAURICE, Y. (eds.) Advances in Marine Biology. Academic Press. MCKEE, K. L. 1993. Soil Physicochemical Patterns and Mangrove Species Distribution-Reciprocal Effects? Journal of Ecology, 81, 477-487. O'GRADY, A. P., MCGUINNESS, K. A. & EAMUS, D. 1996. The abundance and growth of Avicennia marina and Rhizophora stylosa in the low shore zone of Darwin Harbour, Northern Territory. Australian Journal of Ecology, 21, 272-279. OLIVER, J. 1982. The Geographic and Environmental aspects of Mangrove communities: Climate. In: CLOUGH, B. F. (ed.) Mangrove Ecosystems in Australia, structure function and management. Australian Institute of Marine Science. 4-45 OSBORNE, K. & SMITH, T. J. I. 1990. Differential Predation on Mangrove Propagules in Open and Closed Canopy Forest Habitats. Vegetatio, 89, 1-6. OSUNKOYA, O. & CREESE, R. G. 1997. Population Structure, Spatial Pattern and Seedling Establishment of the Grey Mangrove, Avicennia marina var.australasica, in New Zealand. Australian Journal of Botany, 45, 707-725. RABINOWITZ, D. 1978b. Early Growth of Mangrove Seedlings in Panama, and an Hypothesis Concerning the Relationship of Dispersal and Zonation. Journal of Biogeography, 5, 113-133. SAENGER, P. Morphological, anatomical and reproductive adaptations of Australian mangroves. In: CLOUGH, B. F., ed. Mangrove ecosystems in Australia : structure, function and management ; proceedings of the Australian National Mangrove Workshop, 18-20 April, 1982 Australian Institute of Marine Science in Association with ANU Press, 153-191. SAINTILAN, N. 1998. Relationships between height and girth of mangroves and soil-water conditions in the Mary and Hawkesbury river estuaries, eastern Australia. Australian Journal of Ecology, 23, 322-328. SANTISUK, T. 1983. Taxonomy and Distribution of Terrestrial Trees and Shrubs in the Mangrove Formations in Thailand. . Natural History Bulletin, Siam Society. , 31 63-91. SANTISUK, T., PATANAPONPAIBOON, P. & UBOLCHOLAKET, A. 1988. Floristics (trees and Shrubs) and Ecology of the Mangrove Vegetation of Thailand, National Research Council of Thailand. SMITH, T. J. 1987a. Seed Predation in Relation to Tree Dominance and Distribution in Mangrove Forests. Ecology, 68, 266-273. SMITH, T. J. 1987b. Effects of seed predators and light level on the distribution of Avicennia marina (Forsk.) Vierh. in tropical, tidal forests. Est, Coast and Shelf Sci., 25, 43-51. SMITH, T. J. 1992. Forest structure. In: ROBERTSON, A. I. & ALONGI, D. M. (eds.) Tropical Mangrove Ecosytems. Washington, DC. : American Geophysical Union SMITH, T. J., III, ROBBLEE, M. B., WANLESS, H. R. & DOYLE, T. W. 1994. Mangroves, Hurricanes, and Lightning Strikes. BioScience, 44, 256-262. TAMAI, S. & IAMPA, P. 1988. Establishment and growth of mangrove seedling in mangrove forests of southern Thailand. Ecological Research, 3, 227-238. THAMPANYA, U., VERMAAT, J. E. & DUARTE, C. M. 2002. Colonization success of common Thai mangrove species as a function of shelter from water movement. Marine Ecology Progress Series, 237, 111-120. TOMLINSON, P. B. 1987. The Botany of Mangroves. WELLS, A. G. 1982. Mangrove vegetation of northern Australia. In: CLOUGH, B. F. (ed.) Mangrove Ecosystems in Australia: Function and Management. Canberra.: Austr. Nat. Univ. Press. 4-46 5. Summary and implications for mangrove restoration 5.1 Introduction The objective of the thesis was to demonstrate an evidence based approach to the planning and implementation of mangrove restoration projects which incorporates as its basic tenet a clear understanding of the natural functioning of mangrove ecosystems including the natural spatial distribution patterns and the critical environmental factors underlying these (Ellison (2000). The study demonstrated a case for an evidence based approach to mangrove restoration through the implementation of case studies in two lower intertidal mangrove communities in the Ngao river estuary southern Thailand, with development of a series of recommendations for how to incorporate this improved evidence base into different phases of mangrove restoration programs. This general discussion summarises the main findings of each of the three technical chapters of the thesis (Section 5.2) which focus on case studies focusing on mature trees, propagules and immature seedlings and mature seedlings and then goes on to make recommendations for how to incorporate this improved evidence base into different phases of mangrove restoration programs (Section 5.3). 5.2 Summary of technical chapters 5.2.1 Chapter 2: Spatial patterns of forest structure in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand Chapter 2 of the thesis showed that the lower intertidal forests included in the study were low in mangrove tree species diversity with only six species found in total. S.alba, A. alba and A.corniculatum were the dominant species, typical of other lower intertidal communities in the region. The study quantified the differences between assemblages of mature trees in the two forest types based on species similarity, which were largely due to the differences in stem density of A.corniculatum in the “Aegiceras forest”. Further analysis of univariate forest structure parameters, showed that mean stem diameter was also significantly larger in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. Amongst individual species, A.corniculatum mean stem diameter and height, total stand basal area and Importance value were all greater in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, 5-1 S.alba and A. alba trees had larger average stem diameters and higher Importance values in the “Sonneratia-Avicennia forest”. Chapter 2 of the thesis also provided information on the demographics of mature trees in the two forest types, highlighting differences in the proportion of stems falling into different size classes between the two forest types. A.alba, for example, had a greater proportion of trees with stem diameters larger than 10cm in the “Sonneratia-Avicennia forest” than the “Aegiceras forest” while in contrast, A.corniculatum had a greater proportion of smaller stems (Stem diameter<10cm) in the “Aegiceras forest”. S. alba trees also had heavily right skewed stem diameters with large basal areas with few saplings present in the “Aegiceras forest”, indicative of a mature population of trees which is not regenerating. In addition to between forest variation, Chapter 2 of the thesis also highlighted within forest differences in assemblages of mature trees with distance from the Ngao river, confirming the presence of distinct spatial patterns or zones. Further analysis of univariate forest structure parameters showed significant differences in A.corniculatum tree densities and stem diameter in these “zones”. In the “Sonneratia-Avicennia forest”, distance from the river appeared to play a less important role in defining spatial patterns. Chapter 2 of the thesis also examined change in hydroperiod across the two forest types confirming that the duration of inundation in the “Aegiceras forest” was significantly higher than that in the “Sonneratia-Avicennia forest”. Correlation analysis showed that duration of inundation between the two forest types was significantly correlated with assemblages of mature trees as well as key forest structure parameters including the mean density of S.alba trees and mean stem diameter of A.corniculatum trees. These correlations and the differences in natural assemblages of mature mangrove trees between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest”, suggest that assemblages of mature trees are related to hydroperiod and duration of inundation but that forest communities in both the “Aegiceras forest” and “Sonneratia-Avicennia forest” could have a broad tolerance to hydroperiod characteristics of the two forest types. Within the two forest types, the physical effects associated with more exposed locations close to the Ngao river appear to determine within forest spatial patterns with A.corniculatum seemingly less tolerant of these more exposed 5-2 locations than A.alba and S.alba but a better competitor than A.alba and S.alba in inland more sheltered sites. Information on forest structure obtained in Chapter 2 of the thesis also suggests that secondary succession has taken place within the “Aegiceras forest” where A.corniculatum has invaded the forest at some stage, outcompeting S.alba in all quadrats except for the outer and most inner quadrats. This suggestion is supported by the heavily right skewed basal area size classes of S. alba, and trees with large basal areas with few saplings present, indicative of a mature population of trees which is not regenerating. The distribution of basal area size classes of A. corniculatum, in contrast, had a left skewed population indicating that this species is regenerating and maintaining itself in the forest type. In contrast, in the “SonneratiaAvicennia forest”, it appears that A.corniculatum is expanding its range and taking over from the dominant S.alba and A.alba trees, perhaps in the same manner as it has done in past in the “Aegiceras forest”. The above inferences about the environmental factors leading to the current observed patterns and demographic change are limited as it is difficult to prove causality from observations of mature mangrove trees as physiological tolerances of seedlings may be much narrower than those of adults (Ball, 1988). It is also possible that environmental conditions at the site may have changed over time so that seedlings can no longer become established (Smith, 1992). The observations are however useful as a base on which to develop hypotheses for experiments in Chapter 3 and 4 of the thesis. 5.2.2 Chapter 3: Dispersal and early development of mangrove propagules and seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand In Chapter 3 of the thesis, the focus moved to the dispersal and early development stages of the mangrove tree life cycle which are known to play an important role in determining eventual spatial patterns of mature trees within mangrove forests. Similar to the study of mature trees in Chapter 2 of the thesis, a propagule dispersal study carried out in Chapter 3 showed that assemblages of propagules and developing seedlings differed between forest types, largely due to the high density of A. corniculatum propagules and seedlings in the “Aegiceras forest” and high density of A.alba propagules and seedlings in the “SonneratiaAvicennia forest”. Univariate analysis of propagule and seedling density also confirmed the 5-3 significant difference in density of A. corniculatum, A.alba propagules and seedlings as well as S. alba seedlings between the two forest types. In addition to the between forest differences, similar to the study of mature trees in Chapter 2 of the thesis, Chapter 3 of the thesis also found that spatial patterns of propagules and seedlings in the “Aegiceras forest” were also related to the distance from the Ngao river, with the greatest density of A.corniculatum propagules and seedlings found in the middle zone of the forest (20-30m from the Ngao river) and the least in the zone immediately adjacent to the Ngao river and inland. A.alba propagules and seedlings in contrast were confined to quadrats closest to the Ngao river. In the “Sonneratia-Avicennia forest”, A.corniculatum seedling density was also related to the distance from the Ngao river, highest in the zone adjacent to the river and lowest inland. Chapter 3 of the thesis also identified clear seasonal variation in propagule and seedling abundance across the calendar year with a clear peak in production of A. corniculatum and A. alba propagules followed by a peak in seedling density in subsequent months. Some seasonal differences were also apparent between forest types with A. corniculatum, A. alba and S.alba seedlings tending to persist across the year in the “Sonneratia-Avicennia forest” but not in the “Aegiceras forest”. Results of a separate propagule release experiment reported in Chapter 3 of the thesis were largely consistent with the propagule dispersal study, showing differences between the two forest types in that a higher proportion of A. corniculatum propagules released into enclosures developed into mature seedlings in the “Aegiceras forest” than the “Sonneratia-Avicennia forest”. In contrast, more A.alba propagules developed into mature seedlings in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. The propagule release experiment also found that a greater proportion of released A.corniculatum propagules developed into mature seedlings 20-30m from the Ngao river and the lowest number 30-40m from the Ngao river. In contrast, a greater proportion of A.alba propagules developed into mature seedlings further away from the Ngao river and the lowest amount closest to the Ngao river, largely due to the high proportion of “failed seedlings” (seedlings germinating but not establishing as seedlings in the soil) in the quadrat closest to the Ngao river. 5-4 Chapter 3 of the thesis also examined the relationship between duration of inundation and the development of released propagules into mature seedlings, with correlation analysis showing that duration of inundation in the “Aegiceras forest” was significantly negatively correlated with development of mature seedlings of A.corniculatum while A.alba was significantly correlated with the duration of inundation in both the “Aegiceras forest” and “SonneratiaAvicennia forest” with fewer propagules developing into mature seedlings in sites with the highest inundation duration. The differences in natural assemblages of mangrove propagules and seedlings and development of propagules in enclosures between the two forest types and further significant differences with distance from the Ngao river in the “Aegiceras forest” highlighted in Chapter 3 of the thesis suggest that differences in assemblages of propagules and seedlings between the two forest types are closely related to the mature forest composition and the Importance value of conspecific trees in the canopy. Within the two forest types, the propagule and seedling dispersal data also supports the secondary role of duration of inundation described in Chapter 2 of the thesis, with physical effects associated with more exposed locations close to the Ngao river appearing to determine within forest distribution patterns of A.corniculatum propagules and seedlings which were seemingly less tolerant to these more exposed locations closer to the Ngao river than A.alba and S.alba. Comparison of the results for A.alba propagules and seedlings in the natural dispersal study and experimental release study suggest that although the propagules of this species has difficulty establishing close to the Ngao river, this area was still the preferred area for seedling establishment due to unsuitable environmental conditions higher in the intertidal. Chapter 3 of the thesis also suggests differences in reproductive strategy amongst lower intertidal species in the Ngao river with A.corniculatum trees producing large numbers of propagules to increase the chance of recruitment to the sapling stage and repress colonization by other mangrove species. In contrast, the other two dominant species of propagules were found in lower numbers suggesting a different reproductive strategy. The chapter also provides additional evidence that secondary succession is taking place within these lower intertidal forests with the wide distribution of A.corniculatum propagules and seedlings across both forest types, irrespective of mature tree dominance. This adds weight to the theory espoused in Chapter 2 of this thesis that in the “Sonneratia-Avicennia forest”, it appears that A.corniculatum is expanding its range and taking over from the dominant S.alba and A.alba 5-5 trees, perhaps in the same manner as it has done in past in the “Aegiceras forest” where large S.alba trees are still present in the canopy above a dense understorey of A.corniculatum trees. 5.2.3 Chapter 4: Survival and growth of transplanted mangrove seedlings in lower intertidal mangrove communities of the Kraburi river estuary, southern Thailand In Chapter 4, the final technical chapter of the thesis, the focus was the post establishment stage of the mangrove seedling life cycle. Results of the transplant study showed that survival and growth of transplanted seedlings of both species of seedling; A. corniculatum and S.alba, was greater in the “Sonneratia-Avicennia forest” than the “Aegiceras forest”. Survival of A.corniculatum seedlings was also greater than that of S.alba in both forest types while growth rates of S.alba seedlings were greater than those of A.corniculatum. There was a significant interaction between forest type and species of seedling planted on growth, with the increased growth associated with transplanting to the “Sonneratia-Avicennia forest” was less pronounced for A.corniculatum seedlings. Results of Chapter 4 of the thesis also showed that in addition to the between forest differences, survival and growth of both species also increased with distance from the Ngao river in both forest types. A.corniculatum seedlings had a stronger trend of survival in quadrats located further away from the Ngao river than S.alba seedlings. Results of the final treatment in Chapter 4 of the thesis showed that the proportion of transplanted seedlings of both species surviving differed with canopy type, with a greater proportion of seedlings surviving in light gaps than under the forest canopy. There was also a trend towards higher growth increments in light gaps but differences were not statistically significant. There was a significant interaction between canopy type and forest, canopy type and distance and canopy type and species on seedling survival. Canopy type had less of an effect on survival of transplanted seedlings in the “Sonneratia-Avicennia forest” than it did in the “Aegiceras forest”. Canopy type had less of an effect on survival of transplanted seedlings in the quadrats closest to the Ngao river than it did further inland. Canopy type also had less of an effect on survival of A.corniculatum seedlings than on S.alba seedlings. 5-6 In both forest types, the survival and growth of transplanted seedlings of both species was correlated with the percentage of time that they were inundated by tides, with survival and growth increasing as the proportion of the day in which the site was inundated decreased. Clear differences in survival and growth of the two species appears to be related to significantly different levels of light availability in the two forest types as a result of the differing forest structure of these two lower intertidal types, the denser “Aegiceras forest” and the more open canopy “Sonneratia-Avicennia forest” as reported in Chapter 2 of this thesis. Forest type in turn is driven by variation in the duration of inundation between the two forest types. Within the two forest types, a secondary role of duration of inundation appears to exist, with physical effects associated with more exposed locations close to the Ngao river appearing to influence seedling survival and growth patterns, both of which were significantly greater in plots away from the Ngao river than plots close to it with A.corniculatum seedlings seemingly less tolerant to these more exposed locations than those of S.alba. The effects of light level and physical disturbance on seedling establishment although not unequivocal are consistent with results from the literature and also provide further insights into tolerances of the two lower intertidal species to environmental conditions typically observed at restoration sites. The effects of forest type, light levels and physical disturbance on seedling establishment are consistent with results from the literature and also provide further insights into tolerances of the three lower intertidal species to environmental conditions typically observed at restoration sites. Knowledge of these species specific tolerances to tidal inundation, physical environmental factors and light will serve to improve current and future restoration efforts. 5.3 Implications for restoration An important objective of the thesis was to make recommendations for how to incorporate the improved evidence base developed through the three case studies into the planning and implementation of restoration projects focussing on lower intertidal communities in Thailand and the immediate region. The following subsection of the thesis incorporates 2 parts, the first is a table (Table 5-1) summarising the possible application of each type of information developed through the case studies for mangrove restoration and the second a concept for a decision support system for restoration of lower intertidal forests (S5.3.1). 5-7 Table 5-1: Summary of applications of knowledge developed through the thesis to restoration of degraded lower intertidal forests in the immediate region Type of knowledge generated Possible application to mangrove restoration Chapter 2: Spatial patterns of forest structure 2.1 Understanding of species  composition in the two forest types.  Knowledge of what mangrove species should be selected as the basis for restoration projects in the vicinity of the two forest types.  Use this information to develop an understanding of other values of these forest types (e.g community use, fisheries productivity) recognising the different ecosystem services provided by different mangrove types (Ewel et al., 1998) 2.2 Information on forest  structure of mature trees in the two forest types and variation across the two forest types Use this information to formulate restoration objectives defining the specific attributes of the natural standing forest that the restoration activity is trying to restore and in what time frame (McKee and Faulkner, 2000). These objectives will then form the basis of regular site monitoring to objectively evaluate success. Use this information to formulate restoration objectives focusing on tree species diversity. An example of an objective could be to restore the forest so that stand density of A.corniculatum trees is 4,000 stems ha-1 in 2018/ or that the mean stand basal area is 35m2 ha-1 in 5 years’ time.  Use forest structure data and data on assemblages of community structure as the “baseline” to compare to future monitoring results.  Use information on distance from the Ngao river to enable matching of mangrove species with preference for protected sites (away from the river) or unprotected sites.  Use mature tree density data as a guide to planting density of different species. E.g A.corniculatum is commonly found growing at high densities as compared to S.alba and A.alba. As such planting of A.corniculatum could be conducted at closer spacing between seedlings (e.g 50cm * 50cm) while S.alba and 5-8 Type of knowledge generated Possible application to mangrove restoration A.alba seedlings could be planted at higher spacing (e.g 2m * 2m).  Understand that mangrove forests are often not stable but possible undergoing secondary succession as environmental conditions in the forest change over time (Ball, 1980). As such it provides important information about change taking place in the forest (e.g species replacement and die back) so mangrove restoration activity doesn’t try to work against natural processes. E.g in the Sonneratia-Avicennia forest, it appears that secondary succession is taking place with A.corniculatum taking over from S.alba and A.alba. This suggests that S.alba seedlings should not be planted as they will be outcompeted by the naturally arriving A.corniculatum seedlings in this forest type. Similarly, in the “Aegiceras forest”, seedlings shouldn’t be planted as survival rates are likely to be low given high shade conditions in the forest.  2.4 Information on hydroperiod  in the two forest types and with distance from the Ngao river in  the two forest types and correlation between species forest structure parameters and  hydroperiod Use forest structure data and data on assemblages of community structure as the “baseline” to compare future restoration of the site. Understand the role that hydroperiod plays in determining forest structure in mangrove forest communities. Identify if there are any impediments to hydrologic flow through the site which might require specific action (e.g removal of dikes, regrading etc) before restoration activities can continue. Understand different tolerances to duration of inundation of each species of mature tree in mangrove forest communities to enable matching of site conditions with species autoecology. E.g site slope to be regraded to match the 40cm above chart datum surface elevation of the neighbouring reference forest.  Provides background information on the surface elevation of the sites including level in relation to chart datum and degree of slope 5-9 Type of knowledge generated Possible application to mangrove restoration Chapter 3: Dispersal and early development of mangrove propagules and seedlings 3.1. Understanding of propagule/  seedling species composition in the two forest types. Contribute to understanding of processes of secondary succession taking place in the two forest types (in addition to information from item 2.3) 3.2. Information on propagule  and seedling mean and relative density of the different species of mangroves in the two forest types Assessment of whether propagule numbers are sufficient across both forest types to enable natural restoration to take place without any interventions and whether all species in the forest are regenerating or whether forests are propagule limited (Sousa et al., 2007) 3.3. Information on the  abundance of propagules and seedlings of different mangrove species varies with each forest  type with distance from the Ngao river in each forest type Assessment of tolerance limits of seedlings of different species to differing levels of exposure (Thampanya et al., 2002) 3.4 Information on the abundance of propagules and seedlings of different mangrove species varies with season Assessment of potential bottleneck in early establishment of seedlings which can be addressed through planting of larger seedlings (Duke, 1996)  Understanding of whether natural recruitment is sufficient at the site (e.g are propagules arriving, germinating and then persisting in the forest) or whether a bottleneck to recruitment exists.  Used by restoration planners to help decide on the most cost effective planting unit required to achieve success (e.g propagule, newly established propagule, seedling or mature tree) (Alleman et al 2011).  Assessment of the availability of propagules as the basis for nursery production of seedlings and planting programs or direct seeding programs (Lewis et al., 2005)  Guide to the best time of year to be planting to ensure the best environmental conditions (e.g temperature, 5-10 Type of knowledge generated Possible application to mangrove restoration rainfall, degree of wave/ current action) to ensure maximum seedling survival.  Understand tolerances of natural seedlings to hydroperiod as a guide to suitable planting elevations for planting of mature seedlings  Identify bottlenecks in development of seedlings. E.g A.alba has difficulties establishing at lower elevations although mature trees of this species are found growing there. Use as the basis for planting of the species to bypass the bottleneck. 3.6 Correlation between mean and relative density of propagules and seedlings of different mangrove species and total tree DBH of mature trees.  Understand benefits of using artificial structures in planting sites to facilitate natural recruitment by trapping propagules. 3.7 Correlation between mean and relative density of propagules and seedlings of different mangrove species and Importance value of mature tree species  Identify whether distance from mature stands is an impediment to natural recruitment (e.g if most propagules strand close to their parent tree). If so use as the justification for direct seedling programs etc. 3.5 Correlation between mean and relative density of propagules and seedlings of different mangrove species and hydroperiod Chapter 4: Survival and growth of transplanted mangrove seedlings 4.1 Understanding of variation in survival and growth of transplanted seedlings between forest type  Enables comparison of survival rates of planted species in a range of environmental conditions as the basis for restoration planning, including the number of seedlings required in a certain area, requirements for maintenance planting (in the event of poor survival).  Enables assessment of likely restoration trajectory based on seedling growth rates and whether the 5-11 Type of knowledge generated Possible application to mangrove restoration restoration project is likely to meet performance standards. 4.2. Understanding of variation in survival and growth of transplanted seedlings with distance from the Ngao river in the two forest types 4.3 Understanding of variation in survival and growth of transplanted seedlings with canopy type in the two forest types  Overall provides an idea of what the likely survival rates are going to be across the planting site and areas where planting losses are likely to be higher (e.g exposed areas near the river).  Understanding of tolerances of different species to exposed and sheltered conditions.  Overall-gives an idea of what the likely survival rates are going to be across the planting site and areas where planting losses are likely to be higher (e.g shaded areas under the canopy).  Understand tolerances of different species to shaded and non shaded conditions. 5-12 5.3.1 A concept for a decision support system for restoration of lower intertidal forests 5.3.2 Introduction The decision support system concept recognises that lack of information or “evidence” is only part of the problem facing mangrove restoration and that mangrove restoration projects are also often constrained by the lack of integration of knowledge about the ecosystem with decisions on restoration. The decision support system is a set of simple selection criteria in the form a flow chart for lower intertidal field sites based on results of studies of forest structure, propagules and mature seedlings (Figure 5-1). The flow chart attempts to demonstrate how each piece of information could potentially be applied in practice to enable the best restoration decisions in an objective manner. This flow chart could be used by a mangrove restoration practitioner working in the vicinity of the case study sites to guide decision making, when planning a hypothetical restoration project for a degraded mangrove site before making key decisions about activities at the site. The flow chart could also be further developed and applied at other localities using site specific information developed by a pilot study at that site using a similar suite of studies as included in the case studies. Depending on the characteristics of the field sites, the flow chart could be expanded to incorporate additional parameters of more relevance to other field sites including physico-chemical such as salinity, soil nutrients and redox potentials. In this way the flow chart could form the front end of a guideline for integration of mangrove ecology into mangrove restoration which could include more detailed information on methodologies for field studies and data collection, data analysis and data interpretation. Detailed guidelines could in turn form the basis of improved training curriculum and capacity building in its application to real world situations. Decisions made when working through the flow chart could be documented as part of the pre-feasibility assessment process commonly undertaken by international donors prior to agreeing to fund a particular project. 5-13 A decision support system for mangrove restoration Step 1: Calculate hydroperiod. To assess the suitability of hydroperiod at the site for mangrove restoration activities, topographic surveys should be carried out at each potential restoration site and hydroperiod calculated. Using results of hydroperiod calculations, three options exist: 1.1 The first option is where the hydroperiod is less than 35%. These sites in the Ngao river are occupied by mid-high intertidal species as Rhizophora spp. and were not included in the case studies. The recommendation action at these sites is to conduct further research to assess the use of mid-high intertidal species as part of restoration efforts at these sites (Decision 1) 1.2 The second option is where hydroperiod is in the range 35-55%, the range currently occupied by both the “Aegiceras forest” and the “Sonneratia-Avicennia forest” in the Ngao river as detailed in Chapter 2 of the thesis. The recommendation for these sites is to conduct additional assessment of site exposure through subsequent steps in the flow chart (Go to step 2). 1.3 The third option is where hydroperiod is greater than 55%. The recommendation for these sites is to conduct further research to assess suitability for restoration (e.g through hydrologic restoration or hybrid engineering initiatives) (Decision 2) Step 2: Measure distance of the site from the river as a proxy for degree of site exposure To assess the level of site exposure, simple measurements of distance from the shoreline or alternative measurements of relative site exposure (Thampanya et al., 2002). Chapter 2 of the thesis concluded that distance from the Ngao river and difference in physical effects of wave action in sites close to the Ngao river (exposed sites) as compared to those located away from the Ngao river (protected sites), was likely to be the major factor affecting the distribution of communities dominated by S.alba and A.alba trees (more tolerant to exposure) from those dominated by A.corniculatum trees (less tolerant to exposure). Using results of distance measurements, two options are available: 2.1. For sites 0-10m from the shoreline, Chapter 2 of the thesis shows that in both forest types these sites are typically occupied by a mixed S.alba and A.alba community. Results of Chapter 3 of the thesis also showed greater densities of A.alba propagules and seedlings and S.alba 5-14 seedlings at these sites, especially at higher hydroperiods. Chapter 4 of the thesis showed that more transplanted S.alba seedlings survived and growth was higher in sites away from the river although the preference for sheltered sites was not as strong as in A.corniculatum seedlings. The recommendation for these sites is to conduct additional assessments of the degree of natural recruitment at the site before making a decision on restoration option (Go to step 3) 2.2 For sites >10m from the shoreline, Chapter 2 of the thesis found that in both forest types these sites are typically occupied by communities dominated by A.corniculatum suggesting a reduced tolerance of this species to exposed conditions. Results of Chapter 3 also showed the highest densities of A.corniculatum propagules occurred at these sites. Chapter 4 showed that transplanted A.corniculatum seedlings grew best in sites away from the Ngao river. The recommendation for these sites is to conduct additional assessment of degree of shading at the site (Go to Step 4). Step 3: Assess levels of natural recruitment at the site To assess levels of recruitment at the site, simple measurements of recruitment should be measured using techniques similar to those described in Chapter 3 of the thesis. For sites located in exposed sites (0-10m from the river), decisions on restoration options for the site will depend on whether natural recruitment to the site is sufficient to allow for natural recovery of mangroves to occur. In most exposed sites, results of Chapter 3 of the thesis showed that in both forest types natural recruitment of S.alba and A.alba seedlings is relatively low. Results of Chapter 3 of the thesis also showed that recruitment to the site was dependent on proximity to mature trees with a close relationship between presence of A.alba and S.alba mature trees in the canopy and density of propagules and seedlings. Using results of recruitment measurements, two options are available: 3.1 For sites with high levels of natural recruitment, no mangrove restoration activities are recommended as based the site is likely to recover naturally (Decision 3) 3.2 For sites with low levels of natural recruitment, the recommendation for these sites is to carry out planting of nursery raised S.alba and A.alba seedlings to increase the rate of recovery of these sites (Decision 4). 5-15 Step 4: Assess level of light availability at the site To assess levels of light availability at the site, simple measurements should be made of the presence and absence of light gaps. Alternatively simple measurements of light availability can be made using light sensors. For sites located in sheltered sites, decisions on restoration options for the site will depend on degree of light availability and specifically whether the site is shaded. Chapter 3 of the thesis noted that seedlings of A.corniculatum did not persist in the denser “Aegiceras forest” compared to the more open “Sonneratia-Avicennia forest” possibly related to differences in light availability in the former. Chapter 4 of the thesis showed higher survival and growth of seedlings of both species in light gaps, a difference which was more pronounced in the “denser “Aegiceras forest” compared to the more open “Sonneratia-Avicennia forest”. Using results of light gap assessments, two options are available: 4.1 For sites with high degree of light availability and presence of light gaps, the recommendation for the sites is to conduct additional assessment of natural recruitment at the site (Go to Step 5). 4.2 For sites with low degree of light availability and no light gaps, the recommendation for these sites is not to conduct any restoration at these sites as based on the results of Chapter 4 of the thesis, as seedling survival under the forest canopy is likely to be low in these sites (Decision 5). Step 5: Assess levels of natural recruitment at the site For sites with sufficient light availability, decisions on restoration options for the site will depend on whether natural recruitment to the site is sufficient to allow for natural recovery of mangroves to occur. For sites with high levels of natural recruitment, no mangrove restoration activities are recommended as the site is likely to recover naturally (Decision 6). For sites with low levels of natural recruitment, the recommendation for these sites is to carry out planting of nursery raised A.corniculatum seedlings to increase the rate of recovery of these sites (Decision 7). 5-16 Figure 5-1: Flow chart summarising the recommended steps incorporating the evidence base developed under the project to a hypothetical mangrove restoration project Using a hypothetical case study where hydroperiod was 52%, the site was exposed, and there was low natural recruitment, the decision on the most suitable restoration option would be Decision 7: Planting of A.corniculatum seedlings. Incorporated together with the improved evidence base on lower intertidal forests, developed through the three case studies, the decision support system could serve as a practical tool for the integration of scientific knowledge about mangrove ecosystems into restoration planning and ultimately improve the results of restoration and assist in the recovery of these essential ecosystems. 5-17 5.4 References BALL, M. 1980. Patterns of secondary succession in a mangrove forest of Southern Florida. Oecologia, 44, 226-235. BALL, M. C. 1988. Salinity Tolerance in the Mangroves Aegiceras corniculatum and Avicennia marina I. Water Use in Relation to Growth, Carbon Partitioning, and Salt Balance. Aust. J. Plant Physiol., 15, , 447-64. DUKE, N. C. 1996. Mangrove reforestation in Panama: an evaluation of planting in areas deforested by a large oil spill. In: FIELD, C. D. (ed.) Restoration of Mangrove ecosystems. Okinawa, Japan: International Society for Mangrove Ecosystems (ISME). ELLISON, A. M. 2000. Mangrove Restoration: Do We Know Enough? Restoration Ecology, 8, 219-229. EWEL, K., TWILLEY, R. & ONG, J. I. N. 1998. Different kinds of mangrove forests provide different goods and services. Global Ecology & Biogeography Letters, 7, 83-94. LEWIS, R. R., HODGSON, A. B. & MAUSETH, G. S. 2005. Project facilitates the natural reseeding of mangrove forests (Florida). Ecological Restoration, 23, 276-277. MCKEE, K. L. & FAULKNER, P. L. 2000. Restoration of Biogeochemical Function in Mangrove Forests. Restoration Ecology, 8, 247-259. MCKEE, K. L., MENDELSSOHN, I. A. & HESTER, M. W. 1988. Reexamination of Pore Water Sulfide Concentrations and Redox Potentials near the Aerial Roots of RhizophoraMangle and Avicennia-Germinans. American Journal of Botany, 75, 1352-1359. SMITH, T. J. 1992. Forest structure. In: ROBERTSON, A. I. & ALONGI, D. M. (eds.) Tropical Mangrove Ecosytems. Washington, DC. : American Geophysical Union SOUSA, W. P., KENNEDY, P. G., MITCHELL, B. J. & ORDÓÑEZ L, B. M. 2007. SUPPLY-SIDE ECOLOGY IN MANGROVES: DO PROPAGULE DISPERSAL AND SEEDLING ESTABLISHMENT EXPLAIN FOREST STRUCTURE? Ecological Monographs, 77, 53-76. THAMPANYA, U., VERMAAT, J. E. & DUARTE, C. M. 2002. Colonization success of common Thai mangrove species as a function of shelter from water movement. Marine Ecology Progress Series, 237, 111-120. 5-18