Separating direct and indirect effects of global change: a population dynamic modeling approach using readily available field data

摘要

Two sources of complexity make predicting plant community response to global change particularly challenging. First, realistic global change scenarios involve multiple drivers of environmental change that can interact with one another to produce non-additive effects. Second, in addition to these direct effects, global change drivers can indirectly affect plants by modifying species interactions. In order to tackle both of these challenges, we propose a novel population modeling approach, requiring only measurements of abundance and climate over time. To demonstrate the applicability of this approach, we model population dynamics of eight abundant plant species in a multifactorial global change experiment in alpine tundra where we manipulated nitrogen, precipitation, and temperature over 7years. We test whether indirect and interactive effects are important to population dynamics and whether explicitly incorporating species interactions can change predictions when models are forecast under future climate change scenarios. For three of the eight species, population dynamics were best explained by direct effect models, for one species neither direct nor indirect effects were important, and for the other four species indirect effects mattered. Overall, global change had negative effects on species population growth, although species responded to different global change drivers, and single-factor effects were slightly more common than interactive direct effects. When the fitted population dynamic models were extrapolated under changing climatic conditions to the end of the century, forecasts of community dynamics and diversity loss were largely similar using direct effect models that do not explicitly incorporate species interactions or best-fit models; however, inclusion of species interactions was important in refining the predictions for two of the species. The modeling approach proposed here is a powerful way of analyzing readily available datasets which should be added to our toolbox to tease apart complex drivers of global change.