Anthropogenic climate change has already started to affect the distribution of species. Species are not only valuable in their own right, but also because they are responsible for the capture, conversion and flow of energy and nutrients through ecosystems. It has proven challenging to study the impacts of climate change on ecosystem processes, first because effects on single species cannot be extrapolated to the complex network of species interactions, second because it is difficult to manipulate entire ecosystems, and third because it is not clear how results from one location can be used to predict responses across entire regions when species show biogeographic turnover in composition and traits. Our approach to these issues is two?fold. First, we will manipulate a small, spatially discrete food web (the microbial?faunal food web inhabiting water?filled bromeliads) to determine the role of species interactions in determining ecosystem responses. Second, we take advantage of the fact that our focal food web occurs over a broad biogeographic gradient to examine the generality of food web responses. We concentrate on precipitation because it is understudied (compared to temperature) and has potentially profound impact for ecosystems, and specifically on Neotropical ecosystems, which are expected to lose more species than their temperate counterparts. The general aims of this project are: (1) to understand the interaction between biogeographic changes and climate change, and (2) to disseminate a robust, multi?regional theory of how climate affects ecosystems. Our project comprises three hierarchical tasks: (i) To determine if the responses of populations and multipartite interactions (from bacteria to metazoa) and ecosystem function (carbon and nitrogen dynamics, decomposition, carbon dioxide and methane emission) to altered precipitations differ between countries; (ii) To use a biogeographic analogue experiment inspired from geneticists' twin studies to determine whether this variance between countries is driven by biogeographic changes in species composition or differences in local conditions; (iii) To disentangle the direct effects of precipitation change mediated by organism physiology from the indirect effects mediated by interactions between species. To answer these questions, we will experimentally change precipitation entering bromeliad ecosystems from baseline levels in 3 field sites covering the range of faunal diversity in general in the Americas: French Guiana, the centre of bromeliad radiation and a hotspot for bromeliad faunal diversity, Costa Rica which has a moderate species pool, and Puerto Rico, a Caribbean site with a depauperate species pool. If we understand the mechanisms underlying biogeographic effects, we can consider how our results can be extrapolated to unstudied portions of the biogeographic gradient. We will experimentally increase or decrease precipitation entering bromeliads to study effects on the bromeliad ecosystem. Manipulations will involve either a 40% decrease in rainfall by deflecting rain with cone?shaped shelters, a 40% increase by concentrating rain with inverted (funnel?shaped) shelters, or no change (shelter sides vertical). Our major findings will be disseminated to scientists, students, stakeholders, and public schools. Ecologists have a limited timeframe in which studies on consequences of climate change will be useful to society, so need to seek shortcuts by which results from particular fieldsites can be extrapolated to other regions with differing species pool. This project will provide a fresh approach on how to predict the ecosystem consequences of climate change.