Understanding the impacts of global environmental change in complex natural systems is a key challenge in ecology. Paleoecological records have recently yielded evidence of the dynamic nature of community structure in response to climate change. Important pioneering work comes from studies in ancient communities, such as Cambrian fauna, but the degree to which generalities can be made about changes in community structure in response to climatic changes, remains to be resolved. Several thousand fossil rodent middens have been dated and analyzed from the deserts of North and South America spanning the last 50 kyr. They are commonly reported as local, discontinuous series of 10-50 middens, and potentially lend themselves to novel analyses of discrete and changing plant and invertebrate assemblages (soil food webs) through thousands to tens of thousands of years, depending on the site. Each midden contains discrete invertebrate deposits (soil food webs) that allow invertebrate species richness, composition, abundance, body sizes, and trophic level be estimated along environmental conditions, such as precipitation, primary productivity, and plant species richness. In collaboration with paleoecologists Claudio Latorre (Pontificia Universidad Católica de Chile) and Julio Betancourt (National Research Program, Water Mission Area, U.S. Geological Survey & Adjunct Professor, School of Natural Resources and the Environment, University of Arizona) we are starting to study ancient and modern food webs to improve our understanding of the influence of environmental change on food web structure.