Although there is much need to establish the links between different dimensions of biodiversity (such as species richness, functional diversity, genetic diversity, variability across sites) and ecosystem services that are important to human societies, this knowledge is still quite sparse. Understanding these links is relevant, among other aims, to establish whether goals of biodiversity conservation and sustainability and delivery of ecosystem services are compatible. There are many empirical and theoretical reasons to anticipate that local communities composed by species that are closer to their expected optima (driven by abiotic and biotic factors) should be more productive (within and across species) than local communities composed by species away from their optima. If that is the case, then a rationale can be made that the targets of biodiversity conservation would coincide with their services. However, this prediction may not hold across all types of environments. As such, identifying, understanding and mapping the links and potential trade-offs between complex biodiversity patterns and their associated productivity are important endeavours. Here we propose a statistical framework that models communalities and differences in patterns of species co-occurrences in space based on observed and unobserved (latent) factors and mechanisms. The model links variation among species attributes (traits, phylogenies, physiology) and environmental characteristics to predict species distributions and their patterns of co-occurrence. We then connect model predictions with species productivity at the level of local communities, establishing the conditions and where biodiversity patterns match or trade-off with local productivity.
Results/Conclusions
Commercial and sport fisheries are important industries in the Great Lakes region where sixty-five million pounds of fish per year are harvested from the lakes. Moreover, Great Lakes angler spending has a large economic impact for Ontario, contributing with more than 30 000 full-time jobs. We applied the proposed framework to model biodiversity patterns of fish species across 700 lakes in the province of Ontario, Canada; and linked model predictions to fish biomass in the studies lakes. Our results indicate that predictions are environmentally dependent - fish biodiversity patterns are strongly linked to fish biomass in southern lakes (relatively warmer and higher energy systems) in contrast to northern lakes (relatively colder and higher energy systems). The framework allows establishes regional spatial maps based on predictability that can facilitate policy and conservation plans.