Application of M5 algorithm of decision tree in simulation and investigation of effective factors of erosion in rangelands and forests

Interrill erosion is the process of soil erosion that occurs on small, un-vegetated areas between ridges or furrows caused by raindrops falling on sloped land. The impact of raindrops can cause the soil to detach and be carried away by runoff. Interrill erosion can be a significant contributor to overall soil erosion and is considered a problem in agricultural areas, construction sites, and other areas with disturbed soil. The aim of this paper was to identify the factors affecting interrill erosion using the M5 algorithm of decision tree in four different regions. The M5 algorithm is considered to be a robust and powerful method for time series forecasting and has been widely used in a variety of applications. To study interrill erosion, 200 soil samples were collected from two rangelands and two forests in Ramadi, Iraq. The soil samples underwent analysis to determine various chemical and physical properties, and the amount of interrill erosion was calculated using the Kamphorst rainfall simulator. The results showed that in the studied areas, the properties of clay, silt, sand, geometric standard deviation and geometric mean particle diameter had the greatest role in interrill erosion. The highest amount of interrill erosion occurred in the disturbed rangeland with a value of 7 tons/hectare and the lowest amount in the protected forest with a value of 3 tons/hectare.