Climate change is expected to bring about major change in freshwater availability, the productive capacity of soils, and eventually also in patterns of human settlement. However, considerable uncertainties exist with regard to the extent and geographical distribution of these changes. Predicting scenarios for how climate related environmental change may influence human societies and political systems necessarily involves an even higher degree of uncertainty. The direst predictions about the impacts of global warming warn about greatly increased risks of violent conflict over increasingly scarce resources such as freshwater and arable land. We argue that our best guess about the future has to be based on our knowledge about the relationship between environment and violent conflict in areas that already experience forms of environmental change that we think will increase with climate change. Previous rigorous studies in the field have mostly focused on national level aggregates. This article represents a new approach to assess the impact of environment on domestic armed conflict by using geo-referenced (GIS) data and small geographical, rather than political, units of analysis. It addresses some of the most important factors assumed to be strongly influenced by global warming: land degradation, freshwater scarcity, and population density and change. The preliminary results indicate that the relationships between local level demographic/environmental factors and conflict are not uniform. High levels of population growth in already densely populated areas do not appear to increase the risk of conflict, while land degradation appears to have some effect. Also, water scarcity in densely populated areas appears to increase the risk of conflict generally, but decrease the risk of territorial conflict.