Effects of local and landscape factors on paddy field arthropod diversity: from species to communities

Agricultural landscapes are dynamic mosaics where environmental conditions vary across spatial scales. Identifying the spatial scales of environmental variables relevant to biodiversity is therefore essential for safeguarding biodiversity and the ecosystem services they provide. Although species richness is a useful aggregate metric the captures the responses of biodiversity as a whole, few studies have examined how species-specific responses to environmental variables give rise to community-level patterns. Failure to account for the heterogeneity in species responses may undermine potential processes that structure biodiversity. In this study, we sampled paddy-field inhabiting arthropods across the first growing season and examined how local (organic farming, water depth, and crop height) and landscape factors (forest cover and connectivity) affect their abundance and distribution. We built a spatio-temporally explicit joint species distribution model (JSDM) to answer three main questions. 1) How do species differ in their response to local and landscapes factors. 2) Can trophic guild explain the variation in species responses to environmental covariates, and 3) How do species-specific responses scale-up to structure species richness at the community-level? Our results show that the mean importance of environmental covariates was low overall (mean R2 = 0.21), suggesting the prevalence of stochasticity in structuring species abundances. However, environmental importance varied substantially across species (highest R2 = 0.87 and lowest R2 = 0.01) and that trophic guild explained only approximately 15% of species responses, which indicates high idiosyncrasy in species responses both within and among guilds. At the community-level, we found that species richness responds negligibly to local factors while increasing with forest cover at the landscape scale. The importance of each environmental covariate was reflected by the variability of species responses. Lastly, we found no evidence of dispersal limitation in structuring species abundances. Taken together, our study suggests that many paddy field arthropods exhibit transient dynamics and that environmental effects are highly species-specific, even within the same trophic guild. Most importantly, our study highlights the need to understand biodiversity from the species level to understand the processes that structure them in dynamic mosaic landscapes.