The best model to predict functional changes in the grazing gradient of semi-arid ecosystem (Case Study: Gorgan plain)

Rangelands as dynamic ecosystems respond to the disturbances such as climatic fluctuations and grazing management patterns. Prediction of spatial trend of these changes could be contributed to management planning and rehabilitation of degraded landscape. In this study, changes in the function of ecosystem, affected by livestock grazing, were investigated along grazing gradient. To this end, in the winter rangelands of Gorgan, functional features of the rangeland ecosystem including: nutrient cycling, infiltration and stability were measured at the set of points along grazing gradient. Measured parameters were analyzed using statistical methods and regression models to determine changes in ecosystem function along grazing gradient. The results showed a significant trend in functional indices along the grazing gradient. Points close to the livestock camp had the minimum value of functional indices while those at the end of the grazing gradient were found to be maximized. Comparison of different regression models using Akaike Information Criterion revealed that, as per infiltration index four-parameter sigmoid model and also for stability and nutrient cycling indices, three-parameter sigmoid model had the lowest AIC value and were the best models to predict functional changes along grazing gradients. The results of this study may be promising as rangeland model input to identify critical areas and can be used to predict management effects on productivity and sustainability of rangeland ecosystems.