Predicting the plant diversity under the effect of different grazing intensities at Sabalan southeast rangelands

One of the most important issue at natural resources management is plant diversity conservation and detecting the different effects of usage methods on the its quality and quantity. This research was done with the goal of investigating the disaster center (village) at evaluation of rangeland damage with the accent of plant cover combination and diversity in Sabalan mountain southeast rangelands. Plant cover sampling was performed at three villages named Alvares, Aldashin and Asbmarz with the grazing intensity low, med and high using 1 m2 plots in the route of 600 m transect. After the detection of plant combination from the analysis of variance and compare means, the diversity and unique indexes were extracted. Then, development and evaluation of regression and ANFIS models to predict of plant diversity was done and comparing the results was performed. RMSE and R2 indexes were used for evaluation of regression and ANFIS models. The results showed that with the increment of grazing intensity, the plant diversity decreased and the max plant diversity was happened at distance of 400 m plus min plant diversity was occurred at distance of 200 m near the village. Moreover, the prediction part results showed that ANFIS model predicted the plant diversity with higher accuracy (R2=0.91) and lower error (RMSE=0.601) relative to the regression model with lower accuracy (R2=0.83) and higher error (RMSE=1.230). Consequently, the ANFIS approach can predict the relations parameters to rangelands, accurately.