The Effects Of Habitat Heterogeneity And Dispersal Limitation On Species Diversity In Subtropical And Tropical Forests

Understanding the mechanisms of maintenance of species diversity is one of the central subjects in ecology, but there still remain many unresolved questions. The debate about mechanisms of maintenance of species diversity arose from contradictory explanations about basic patterns and maintenance processes of species diversity, and also from the inability of detecting differences between theoretical and empirical community patterns by classical curve fitting methods. The objective of this research was to reanalyze and to compare the contribution of random replacement, habitat heterogeneity and dispersal limitation to species diversity pattern in a 24-ha a stem-mapped forest dynamics plot in Gutianshan (GTS) National Natural Reserve in East China, a 25 ha plot in Fushan in Taiwan (FS) and a 50 ha plot in Barro Colorado Island (BCI) in Panama, by a more accurate spatial statistic framework. In general, there were two different approaches for explaining the maintenance of species diversity. The habitat heterogeneity hypothesis assumed that species diversity in a particular area depends on the diversity of habitats in this area. Species with similar niches can only coexist in different habitat types. If the environmental variables are spatially structured, their structure will be reflected by species distributions through induced spatial dependence, called habitat-association. In contrast, neutral theory assumed that species are competitive equivalent and dispersal limited, and thus leaving no room for habitat heterogeneity. Dispersal limitation is one of the main reasons for generating competitive equivalent species, and can also directly affect species diversity by changing immigration and emigration of local community.Spatial autocorrelation was detected in most of those patterns in GTS plot, after I systematically checked the spatial pattern of each species and species diversity. This finding suggested that results in most empirical tests of the importance of habitat heterogeneity and dispersal limitation might be biased by using classical statistics, which has identical and independent distribution (i.i.d.) sample assumption. Spatial summary statistics also found that most of the species have aggregative pattern. This result strongly suggested that nonrandom processes were involved in producing the observed species diversity pattern.In order to separate the effects of habitat heterogeneity and dispersal limitation on species diversity more precisely, we applied a series of nested spatial process models with a progressive increase in complexity. Specifically, four models were assessed:(1) the homogeneous Poisson process for examining the effect of a pure random process; (2) the heterogeneous Poisson process for examining the effect of habitat heterogeneity; (3) the Poisson cluster process for examining dispersal limitation; and (4) the heterogeneous Thomas model for examining the joint effects of dispersal and habitat heterogeneity. These different spatial models were applied to generate species-area relationships (SARs) at different spatial scales and the results obtained from each of the above four models were compared across scales. Inferences on the effect of each of these models on SARs were then drawn from comparing the goodness-of-fit of the models. Those analyses demonstrated that although dispersal and habitat heterogeneity each could significantly contribute to the SARs, each alone was insufficient to explain the SARs. Their joint effects sufficiently explained the real SARs, suggesting that heterogeneous habitat and dispersal limitation are two predominant mechanisms for maintaining the spatial distributions of the species in these two forests. Although the importance of habitat heterogeneity and dispersal limitation in explaining species diversity has been separately stressed in forest or other communities, our results for the first time highlighted the significance of their joint effects in determining the SARs.Finally, we re-assessed the relative contribution of habitat association and dispersal limitation in the two forest communities by a new developed analysis framework based on spatial point process framework. Results showed that all of species in FS and BCI plot are significantly affected by habitat hetergeneity or dispersal limitaion. Most of species are affected signficantly by the joint effect of the two processes. However, the importances of the two processes are different among species. At community scale, dispersal limitation is account much more variance of species distribution than habitat heterogeneity does. The ratio of explained variance by dispersal limitation and habitat heterogeneity is round 7 to 3. At population scale, the relative strength of dispersal limitaion and habitat heterogeneity to species distribution is different among species. Dispersal limitation again palys a more inportant role than habitat heterogeneity among near 70% species in FS and BCI plot.