OOS 31-4 - Decomposing uncertainty in species abundances in tropical and temperate forests

Wednesday, August 8, 2012: 2:30 PM
B110, Oregon Convention Center
Zhanqing Hao and Dingliang Xing, Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang, China
Background/Question/Methods

Niche-based selection and neutral dispersal are the two fundamental ecological processes in community assembly. While a number of recent studies have shown that plant traits and dispersal limitation play important roles in local community assembly, how they contribute to determining species abundances is not fully investigated. In this study, using a recently developed MaxEnt model together with eight plant traits at species level and species distributions in six worldwide large-scale stem-mapping forest dynamics plots, we tested the importance of trait-based filtering, dispersal limitation from metacommunity and stochastic drift in determining species relative abundances (SRA) at local scale (20×20, 50×50, and 100×100 m). We used uniform distribution as non-informative prior and plot scale SRA as neutral prior in the model to quantify contributions of different processes.

Results/Conclusions

Traits and neutral prior together explained 68%~98% variations of the local SRA at 20×20 m scale, and the amount increased with spatial scale. Permutation tests showed that trait constraints were significant in five out of the six studied forests. Decomposing of the variation showed that the contribution of pure traits varied from 0 to 6% in different forests at different scales while neutral prior and local stochastic processes accounted for 50 – 87% and 9 – 28% variations respectively. There was also an unresolved association between trait constraints and neutral priors in four of the six forests, which explained 7 – 35% variations of local SRA. The effects of trait constraints differed when different priors were used. The proportion of local communities where trait constraints have a significant effect is larger in two forests and smaller in another two forests when neutral prior was used.
Our results show strong evidence for local scale trait-based filtering in five forests while neutral prior and stochastic drift explain more variations in most cases. This suggesting that when neutral processes play important roles in local community assembly, niche-based selection also play a role in shaping local SRA. Our results also highlight the importance of interactions between local environmental filtering and dispersal limitation from regional species pool in community assembly.