Long-term dynamics of a semiarid grass steppe under stochastic climate and different grazing regimes: A simulation analysis

We built a grid-based spatial explicit stochastic model that simulates grazing events and basic processes like seedling establishment, growth or mortality of the dominant species in the grass steppes of Patagonia. After evaluating the model with field data, we performed simulation experiments aimed to explore the interaction of precipitation and grazing regimes on vegetation dynamics. Grazing generated a reduction in tussock density which results in a decline in aboveground net primary production (ANPP). Both response variables presented a non-linear behavior including high temporal variability and delay effects, which may prolong for decades. There was a clear threshold in the response of the variables to stock density, though changes become evident only when a highly selective grazing scenario was used. Under high stock density conditions, precipitation use efficiency (PUE) was 82% lower than the values for non-grazed runs. The inter-annual variability of precipitation was more important than the grazing regime in explaining differences in tussock density. Simulation results highlight important issues regarding rangeland management: grazing regime might be as important as stocking density as a degradation agent, temporal lags might obscure degradation processes for decades, the definition of monitoring variables need to consider their response time constants.