Livestock Settlement Dynamics in Drylands: Model application in the Monte desert (Mendoza, Argentina)

Abstract

Human settlements in arid environments are becoming widespread due to population growth, and without planning, they may alter vegetation and ecosystem processes, compromising sustainability. We hypothesize that in an arid region of the central Monte desert (Mendoza, Argentina), surface and groundwater availability are the primary factors controlling livestock settlements establishment and success as productive units, which affect patterns of degradation in the landscape. To evaluate this hypothesis we simulated settlement dynamics using a Monte Carlo based model of Settlement Dynamics in Drylands (SeDD), which calculates probabilities on a gridded region based on six environmental factors: groundwater depth, vegetation type, proximity to rivers, paved road, old river beds, and existing settlements. A parameter sweep, including millions of simulations, was run to identify the most relevant factors controlling settlements. Results indicate that distances to rivers and the presence of old river beds are critical to explain the current distribution of settlements, while vegetation, paved roads, and water table depth were not as relevant to explain settlement distribution. Far from surface water sources, most settlements were established at random, suggesting that pressures to settle in unfavorable places control settlement dynamics in those isolated areas. The simulated vegetation, which considers degradation around livestock settlements, generally matched the spatial distribution of remotely sensed vegetation classes, although with a higher cover of extreme vegetation classes. The model could be a useful tool to evaluate effects of land use changes, such as water provision or changes on river flows, on settlement distribution and vegetation degradation in arid environments.