Despite decades of debate, many fundamental questions regarding the ecological drivers of community assembly remain unanswered. The extent to which communities are shaped by deterministic versus stochastic forces is a well-known example, with abundant evidence for both types of forces occurring in natural systems. However, there is a lack of consensus on how to quantitatively assess these forces. To address the challenge, we propose a general mathematical framework to provide quantitative assessment of ecological stochasticity under different situations in which deterministic factors drive the communities more similar or dissimilar than null expectation. We developed a new normalized index (NST), tested it with simulated communities considering abiotic filtering, biotic interactions, environmental noise, and spatial scales, and applied it to a field study on a groundwater microbial community.
Results/Conclusions
Compared to previous approaches, the new index (NST) showed obviously higher accuracy and precision, for which the coefficients were over 0.9 in most simulated scenarios. However, all approaches showed limited performance at large spatial scale or under very high environmental noise. We then applied this index to estimate stochasticity in the succession of a groundwater microbial community in response to vegetable oil (EVO) injection. Phylogenetic diversity was analyzed based on 16S rRNA gene sequence, and functional gene diversity was analyzed by GeoChip, a comprehensive high-throughput microarray of microbial functional genes. A total of 11,517 bacterial OTUs and 12,868 functional genes were detected. NST based on functional gene beta diversity revealed reasonable succession of ecological stochasticity, i.e. community assembly processes shifted from deterministic to highly stochastic post-EVO input, and that, as EVO was consumed, the groundwater community gradually returned to be more deterministic similar to its pre-EVO state. NST determined by bacterial beta diversity showed a similar pattern. Null model algorithms and community similarity metrics had strong effects on quantitatively estimating ecological stochasticity, from which preferred algorithm and metrics are suggested based on reasonableness of the results.