Predicting the biomass of aerial and underground parts of palatability species caused by grazing using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Awareness of the extent and impact of vegetation from livestock grazing disorders can be a solution to develop rangeland ecosystem management strategies to achieve sustainability and continuous production in these ecosystems. In this study, the production of aerial and underground organs of grazable livestock species in the southeastern rangelands of Sabalan under the influence of different grazing intensities and distance from the village as the focus of the crisis was investigated. In addition, the development and evaluation of ANFIS model was presented in order to predict the production of aerial and underground organs of food species and compare the results with the regression model. For evaluation of regression and ANFIS models the Root Mean Square Error (RMSE) and correlation coefficient (R2) were used. The results showed that different grazing intensities, distance from village and interaction between them were significant effect on the production of aerial and underground organs of palatability species at (p≤0.01). Also, with increasing grazing intensity, the production of aerial and underground organs of these species decreased. The results of ANFIS section showed that in low grazing intensity and distance of about 400 meters, the highest amount of production of palatable species is observed. The lowest production of these species is predicted to be close to the village (200 meters). Moreover, the highest amount of underground biomass at farther distances (600 meter) and lowest amount of that was observed at 200 meter. In addition, ANFIS model with higher accuracy (R2 = 0.98 and R2 = 0.95) and lower error (RMSE = 0.9792 and RMSE = 1.168) than less accurate regression model (R2 = 0.92 and R2 0.77) which also had more errors (RMSE = 2.2835 and RMSE = 3.8954), predicted the production of aerial and underground organs, respectively.