DANGEROUS GEOGRAPHY: SPATIAL DISTRIBUTION OF LIVESTOCK RAIDING IN NORTHWESTERN KENYA

Abstract

In a previous study, Ember and associates (2012) found that livestock-related violence involving the Turkana was higher in dry months, drier years, and when months were drier than expected between the years of 1998–2009. This article has data on livestock-related violence from media reports, together with localized and georeferenced spatial and rainfall maps, to explore the question of whether and how strongly landscape variations (e.g., in topography, rainfall variability, and hydrology) predict the place and intensity of livestock raids. Our findings broadly replicate a longstanding land-use dilemma well documented in ethnographic accounts. During the wet season, Turkana herders camp in relatively safe areas at the center of the district. With the onset of the dry season, they must move away to access pasture in blocks along the boundaries with hostile ethnic groups. Much of the raiding by the Turkana on others occurs along the borders where the rainfall pattern is relatively reliable. By contrast, most of the raiding against the Turkana occurs while the herds are on transitional moves, splitting from, and coalescing at, the margins of expansive plains, en route to patches of dry-season ranges. While most of the results are consistent with ethnographic reports, we do find some spatial patterns that were not previously apparent, particularly regarding water. We suggest that policy interventions at mitigating livestock raids need to take account of links between spatial conditions and the onset of localized raiding incidents.