Predicting soil properties of rangeland under the effect of grazing intensity and distance from plant using adaptive neuro-fuzzy inference system (ANFIS) and response surface methodology (RSM) approaches (Case study: Goshchi rangeland of Urmia)

Background and objectives: One of the basic demands in rangeland studies for its proper management is to investigate and understand how the spatial distribution of vegetation and effective factors in this regard. Knowing the behavior of plant species in relation to ecological factors, especially soil factors, is an effective guide in the improvement programs and proper use of pastures. The present research was conducted to evaluate the soil properties including: organic carbon, nitrogen, phosphorus, potassium, electrical conductivity and acidity in ghoshchi rangeland of Urmia as representative rangeland of the Azerbaijani vegetation climate in West Azerbaijan province. Methodology: The prediction of these properties was done in two exclosure and grazed sites using adaptive neuro-fuzzy inference system (ANFIS) and response surface methodology (RSM) approaches in order to evaluate the capability of these new methods in estimating the soil properties and comparing the results. The obtained data were used to train adaptive neuro-fuzzy inference system (ANFIS) in MATLAB software platform and response surface methodology (RSM) in Design Expert software platform. Results: Correlation coefficient (R2) between predicted data by ANFIS model and RSM method and measured data for organic carbon, nitrogen, phosphorus, potassium, electrical conductivity and acidity were obtained equal to 0.75, 0.93, 0.44, 0.95, 0.68 and 0.95 respectively, Along with 0.84, 0.93, 0.24, 0.96, 0.54 and 0.98. The results of present research showed that the factor of grazing conditions (exclosure and grazed) and distance from the plant (near plant and between plants) had a significant effect on acidity, nitrogen, potassium and organic carbon, but on electrical conductivity and phosphorus, they had no meaning. Thus, the highest amount of organic carbon, nitrogen, and phosphorus was observed in the conditions of exclosure and at the near the plant, which shows the sensitivity of these factors to the grazing factor. The results of RSM method showed that the two factors of grazing conditions and distance from the plant had no significant effect on phosphorus and electrical conductivity while had a strong significant effect on other factors. R2 is lower than expected in both methods of ANFIS and RSM for exactly these two factors. Conclusion: The results of both ANFIS and RSM method showed that both of them can accurately predict the soil properties under the investigated conditions. This indicates that the ANFIS and RSM method have a more accurate prediction of the factors under the significant effect of the input variables and they are not accurate enough regarding the non-significant outputs.