Fire-susceptibility mapping in the natural areas of Iran using new and ensemble data-mining models

Fires have increased in northeastern Iran as its semi-arid climate landscape is desiccated by human activities. To combat fire outbreaks any region, susceptibility must be mapped using accurate and efficient models. This research the forests rangelands of Golestan Province new data-mining Fire effective factors, including elevation, slope angle, annual mean rainfall, temperature, wind effect, topographic wetness index (TWI), plan curvature, distance to river, road, village were obtained from several sources. The relative importance each variable was determined a random-forest algorithm. Fire-susceptibility maps produced R 3.0.2 software GAM, MARS, SVM algorithms, ensemble three models: GAM-MARS-SVM. four fire-susceptibility validated area under curve. results show that village, elevation greatest predicting susceptibility. GAM-MARS-SVM model achieved highest precision mapping. map best detected high risk areas Province. can very useful for creating enhancing management strategies preventing fires, particularly higher-risk portions