Get reliable rangeland science

Anthropogenic activities, species invasions, and ecological factors are rapidly altering rangeland ecosystems, challenging the sustainability of plant species habitats. To address this, reliable prediction models are needed to forecast and map species distribution under varying ecological conditions. This study compares three machine learning methods —Multilayer Perceptron (MLP), Radial Basis Function (RBF), and Support Vector Machine (SVM) —in predicting the distribution of Festuca ovina in Eshkevarat Protec ted Area eastern Guilan province in Iran. This 30,347-ha mountainous protected area is located in the hills of the Alborz Mountains, ranging from 200 to 3600 m in elevation. We analyzed F. ovina distribution in 305 randomly selected plant sample plots, recording 10 ecological variables in each plot. Three machine learning models were developed to predict the likelihood of F. ovina distribution. Results showed that the RBF mode l had more misclassifications (11 samples) compared to MLP and SVM models (10 samp les), suggesting that MLP and SVM were more accurate for distribution modeling. Additionally, MLP demonstrated a higher R2 value (0.87) compared to SVM (0.85), indicating that MLP was the most precise model for predicting F. ovina distribution. Thus, we developed the F. ovina Distribution Model (FODM) using the MLP model. Sensitivity analyses revealed that soil texture, soil depth, electrical conductivity (EC), pH, and vegetation density significantly influenced F. ovina distribution, with sensitivity coefficients of 0.48, 0.47, 0.45, 0.41, and 0.41, respectively. Based on the finalized FODM, we designed an Environmental Decision Support System (EDSS) tool to assist rangeland managers in mapping F. ovina distribution. The practical application of the EDSS tool demonstrated its effectiveness in using the FODM for decision-making and land management. This tool is a valuable resource for rangeland managers, enabling them to make informed decisions regarding F. ovina restoration and effectively utilize the predictive capabilities of the FODM in real-world applications.