University of Tehran Press Iranian Journal of Soil and Water Research 2008-479X 55 11 2025 01 20 Modeling soil loss due to gully erosion in the data-scarce regions Modeling soil loss due to gully erosion in the data-scarce regions 2173 2189 100562 10.22059/ijswr.2024.381049.669782 FA Bahram Choubin Department of Soil Conservation and Watershed Management Research, Isfahan Agricultural and Natural Resources Research and Education Center, AREEO, Isfahan, Iran Omid Rahmati Department of Soil Conservation and Watershed Management Research, Kurdistan Agricultural and Natural Resources Research and Education Center, AREEO, Sanandaj, Iran Seyed Masoud Soleimanpour Department of Soil Conservation and Watershed Management Research, Fars Agricultural and Natural Resources Research and Education Center, AREEO, Shiraz, Iran Samad Shadfar Soil Conservation and Watershed Management Research Institute, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran Ahmad Najafi Eigdir Department of Soil Conservation and Watershed Management Research, West Azarbaijan Agricultural and Natural Resources Research and Education Center, AREEO, Urmia, Iran Journal Article 2024 08 18 Gully erosion is recognized as a detrimental form of land degradation and soil loss worldwide. Considering the time-consuming and costly nature of field monitoring, this research aimed to develop models for estimating the volume of soil lost due to gully erosion in the Choopanlu watershed, located in West Azerbaijan province, Iran. The study commenced with field monitoring to identify gullies in the area. Following this, digital layers of factors influencing gully erosion were prepared to facilitate gully clustering and selection. These factors included topographical characteristics (elevation, slope, aspect, surface curvature, and relative slope position index), vegetation, land use, soil, lithology, and hydroclimate indicators (distance from stream, drainage density, topographic wetness index, annual precipitation, and frequency of heavy rainfall events). Subsequently, the volume of soil lost due to gully erosion during the three-year period (2021-2023) was measured as the dependent variable for the selected gullies through field observations. In this study, three machine learning models including random forest (RF), support vector machine (SVM), and artificial neural network (ANN) were employed using a cross-validation approach. Cochran's formula results indicated that among the 67 identified gullies in the field, a minimum sample size of 58 gullies was required. Following clustering, this number of selected gullies was chosen from the three identified clusters. The annual soil erosion caused by the selected gullies (i.e., 58 gullies) was estimated to be 172 tons in 2021, 196 tons in 2022, and 208 tons in 2023. According to the modeling results, it can be inferred that the RF model demonstrated the best performance, followed by the SVM model with moderate performance, and the ANN model exhibiting the poorest performance in modeling soil loss due to gully erosion.  Gully erosion is recognized as a detrimental form of land degradation and soil loss worldwide. Considering the time-consuming and costly nature of field monitoring, this research aimed to develop models for estimating the volume of soil lost due to gully erosion in the Choopanlu watershed, located in West Azerbaijan province, Iran. The study commenced with field monitoring to identify gullies in the area. Following this, digital layers of factors influencing gully erosion were prepared to facilitate gully clustering and selection. These factors included topographical characteristics (elevation, slope, aspect, surface curvature, and relative slope position index), vegetation, land use, soil, lithology, and hydroclimate indicators (distance from stream, drainage density, topographic wetness index, annual precipitation, and frequency of heavy rainfall events). Subsequently, the volume of soil lost due to gully erosion during the three-year period (2021-2023) was measured as the dependent variable for the selected gullies through field observations. In this study, three machine learning models including random forest (RF), support vector machine (SVM), and artificial neural network (ANN) were employed using a cross-validation approach. Cochran's formula results indicated that among the 67 identified gullies in the field, a minimum sample size of 58 gullies was required. Following clustering, this number of selected gullies was chosen from the three identified clusters. The annual soil erosion caused by the selected gullies (i.e., 58 gullies) was estimated to be 172 tons in 2021, 196 tons in 2022, and 208 tons in 2023. According to the modeling results, it can be inferred that the RF model demonstrated the best performance, followed by the SVM model with moderate performance, and the ANN model exhibiting the poorest performance in modeling soil loss due to gully erosion.  https://ijswr.ut.ac.ir/article_100562_02ab6edad4755681bbf6b8b5a9bc2bd9.pdf