The Ecological Correlates of Armed Conflict: A Geospatial and Spatial-Statistical Approach to Conflict Modeling

Abstract

Few studies of violent conflict focus explicitly on the environmental and ecological relationships that occur in locations that experience violence. Rather, traditional studies of focus on the political, economic, and social factors that affect conflict propensity, holding the state as the unit of analysis. This limits the ability of these studies to identify the geographic and biophysical factors that affect where violent conflict is physically located. I employ a spatially explicit approach to studying the environmental and ecological correlates of conflict. Using a geographic information system (GIS) platform, I overlay a grid of 100 x 100 km cells across the entire terrestrial surface of the earth. I then overlay a series of raster data sets across the terrestrial surface and identify the biophysical and geographic characteristics of each cell. I use logistic regression to explore which ecological characteristics are most closely correlated with the occurrence of violent conflict. Among the characteristics I explore are land cover/land use, vegetative productivity (with controls in place for rain-use efficiency), human population density, infrastructure expansion, forest cover/type, climate and precipitation. I control for political and economic variables including democracy and economic prosperity. I distinguish between conflicts over governance and conflicts over territory, and find that each type of conflict is associated with a distinct ecological, social, and political x profile. Armed conflicts fought over territory (succession, sovereignty, or independence) occur in cells that are marked by decreasing biomass productivity, high densities of crop and pasture lands, areas with higher incidence of distinct agricultural regimes and farming systems, open/fragmented land cover, and temperate climates. In contrast, armed conflicts fought over governance (control of the state, or right to rule) occur in areas with increasing populations and expanding infrastructure, in areas without much farming or grazing, in previously unconverted forest land, and in tropical climates. Identifying the environmental and ecological conditions that affect local conflict propensity holds important implications for organizations working in conflict-prone states, offering a richer understanding of the precipitants of local violence, and offering a greater return on investment through better-informed planning. Further, this approach bridges the gap between biodiversity conservation and human/social/economic development by demonstrating the interconnection of human and natural systems.