Soil texture and plant degradation predictive model (STPDPM) in national parks using artificial neural network (ANN)

Abstract

Soil and plants are interconnected; so destruction in the soil causes degradation in plants. In this study, predictive model of soil and plants degradation was developed using artificial neural network. For sampling of soil, parallel transects by systematic random method were carried out. Soil profiles were drilled in four depths of 5–0, 10–5, 15–10 and 15–20 cm, and the soil texture was examined by hydrometer method. According to the Margalef and Simpson indices, the diversity and richness of plant species were calculated. Totally, the gathered data from 600 vegetation sample plots and 480 soil profiles, physical properties of soil and human and ecological factors were introduced into the artificial neural network model. Among the proposed models, the highest value of R in soil texture and biodiversity index is clay = 0.6960, sand = 0.5657, silt = 0.5913 and Margalef index = 0.5406. Based on the sensitivity analysis results, distance from the road, slope and direction of the slope in soil model and distance from the road, soil moisture content and direction of the slope in plant model were identified as the most effective variables in predicting soil and plant cover degradation.