Soil compaction assessment at different grazing intensities using Adaptive Neuro-Fuzzy Inference System (Case study: Sabalan south eastern rangelands, Ardabil province)

Grazing intensity should follow the management roles to prevent the soil compaction. Soil compaction decreases plant root penetration so that at 2 MPa the root extension and above ground biomass will stop. In this research, digital penetrometer was used to assessment the soil compaction level in Sabalan region in Ardabil province caused by livestock. Grazing intensity and distance from village were evaluated as independent variables. The results show that the grazing intensity and distance from village have significant effect on soil compaction at 1 percent of probability level. Grazing intensity at closer distance (200 m) increases the soil compaction and reduce the soil penetration. Modeling of different grazing intensities was made by ANFIS approach at MATLAB software. For assessment of models’ operation, root mean square error (RMSE) and correlation coefficient (R2) were used and the best model was determined. The results of best ANFIS model for prediction of soil compaction (RMSE=7.223 and R2=0.967) was compared with the result of regression model (RMSE=11.518 and R2=0.918). The results show that the ANFIS model had more R2 and less RMSE consequently more accuracy than the regression model.