COMDREEF · Community disassembly rules and the erosion of ecosystem functions in fragmented landscapes

Community disassembly rules and the erosion of ecosystem functions in fragmented landscapes

This Project aimed to provide a better understanding of the causes and consequences of species responses to the loss and fragmentation of the Atlantic Forest of Brazil. To address this goal, this Project made use of an extensive dataset of more than 25,000 birds, small mammals and amphibians captured in over 90 forest patches of different sizes and connectivity located in six landscapes varying from 10 to 100% forest cover in the state of São Paulo. This dataset was analysed to determine: (i) which assembly rules are structuring communities, and do these rules change with habitat loss and fragmentation or among taxa; (ii) how community disassembly caused by habitat loss and fragmentation affects functional diversity; (iii) which species are more important for ecosystem functioning; (iv) the existence of extinction and functional diversity thresholds; and finally, (v) to determine the minimum area requirements necessary to sustain a viable and functional tropical forest ecosystem. These questions were tested with the use of analytical frameworks developed in several areas within biology, drawing together advances in phylogenetics, population biology, community ecology, landscape ecology, and multivariate statistics. In this Project, I found that the community composition of Atlantic Forest birds is strongly influenced by habitat loss at the patch and landscape scale, while species richness is largely unaffected. Modelling results revealed that the compositional changes observed in the Atlantic Forest was only matched by models ruled by species turnover akin to what would be observed along natural gradients. Such deterministic changes in community composition reveal that communities are not structured by neutral forces, and that assembly rules do not change with habitat loss or fragmentation (Banks-Leite et al. 2012 Ecology). Observed deterministic changes driven by habitat loss are caused by a directional loss of species with a particular set of traits. Within the bird community, for instance, species that have narrow geographic ranges, low relative abundance, and are specialised in specific habitats are more likely to go extinct (Hatfield, Orme, Tobias and Banks-Leite in prep). Considering birds, mammals and amphibians altogether, the single most important trait is endemism to the Atlantic Forest, as endemics are more sensitive to habitat loss than species that can occur in other biomes, showing that extinction in this biome will lead to global extinction (Banks-Leite et al. in prep). The extinction of Atlantic Forest bird species is also likely to lead to a reduction in the number of ecosystem functions performed. In particular, there is less bird-mediated pollination and weaker control of insect populations in degraded and fragmented areas of Atlantic Forest (de Coster, Metzger & Banks-Leite in prep). However, this Project also found that species with very different sensitivities to habitat loss present similar traits, which suggests a certain level of functional redundancy among bird species. Furthermore, although the loss, or reduced abundance, of major trophic groups is obviously related to reduced provision of ecosystems functions, we also found out that further sub-divisions of trophic guilds based on species foraging mode and morphometrics are not necessarily related to resource partitioning. In other words, species with different morphological measures and foraging modes broadly overlap in their diet, and are therefore unlikely to provide different ecosystem functions (Thomas & Banks-Leite in prep). The fact that there is a high degree in functional trait redundancy among bird species, and that species with different traits occupy similar niches suggest that all species are equally important for ecosystem functioning. These results also suggest the existence of an ecological buffer against functional meltdown, at least up to a certain level of habitat degradation. In this project, I also observed the existence of extinction thresholds for birds, mammals and amphibians of the Atlantic Forest, and through the use of phylogenetic analyses I have also observed the existence of a threshold in phylogenetic diversity (as a proxy for functional diversity). Community composition and phylogenetic beta-diversity does not vary much between continuously forested areas (100% forest cover) and fragmented landscapes with up to 30% forest cover. However, below this value community composition and phylogenetic beta-diversity change dramatically. These results are caused by an interplay between the loss of sensitive endemics and a gain of widespread generalists. Changes in phylogenetic beta-diversity could indicate that the functions being provided in areas below the 30% forest cover threshold are very different to the functions provided in more pristine areas. The minimum area needed to sustain a viable and functional ecosystem is 30% of forest cover at the landscape scale. However, most of the Atlantic Forest is below this amount of forest cover, which means much of this biome has to be restored and that payment for ecosystem services could be an advantageous strategy to protect biodiversity. Taking the average PES payment across the biome and costs of restoration, it would cost just over 0.0006% of Brazil’s GDP per year to target restoration at 30% cover. Such effort would lead to an increase ecosystem functions by nearly 14%. Results obtained from this project have been published in Ecology (2012,93:2560-2569) and as two MSc theses. Additional analyses were published in Conservation Biology (2012, 26:1100-1111), Journal of Applied Ecology (2012, 49:1325-1333), PLoS ONE (8(3): e58093), Ecological Indicators (2013, 31:82-88). Results have been presented at SCB 2011 (Auckland, New Zealand), ATBC 2012 (Bonito, Brazil), ATBC 2013 (San Jose, Costa Rica) and Intecol 2013 (London), and as invited talks at Imperial College, University of São Paulo, Oxford University, University of Helsinki, and the Natural History Museum. The most important result obtained during this project is the recognition that at least 30% of Atlantic Forest is needed to preserve biodiversity and ecosystem functions and how much it will cost to restore the Atlantic Forest to reach this target. These results are not only novel to science but also give a crucial tool to conservation managers and practitioners in preserving one of the most endangered Biodiversity Hotpots of the world. Cristina Banks-Leite Imperial College London Grand Challenges in Ecosystems and Environment Silwood Park Campus Munro Building N2.2 Ascot, SL5 7PY, United Kingdom Email: c.banks@imperial.ac.uk

Data: CORDIS, © European Union