%0 Journal Article %T Estimation of Rainfall Erosivity Map in Areas with Limited Number of Rainfall Station (Case study: Semnan Province) %J Iranian Journal of Soil and Water Research %I University of Tehran %Z 2008-479X %A Amini, Elham %A Zolfaghari, Ali %A Kaboli, Hasan %A Rahimi, Mohammad %D 2022 %\ 11/22/2022 %V 53 %N 9 %P 2027-2044 %! Estimation of Rainfall Erosivity Map in Areas with Limited Number of Rainfall Station (Case study: Semnan Province) %K Auxiliary variables %K Network precipitation data %K Random forest model %R 10.22059/ijswr.2022.343710.669279 %X Water erosion is one of the most important challenges of agriculture and watershed management in the world and it has been considered by many researchers. To estimate water erosion, many experimental models have been proposed, of which the Revised Universal Soil Loss Equation (RUSLE) is one of the most widely used models for estimation of soil erosion. Rainfall erosivity (R) is one of the factors in this model. Direct calculation of R required meteorological gauge stations which are available at a limited number of synoptic stations. In this study, the attempt was to estimate rainfall erosivity using available data such as annual rainfall. Semnan province, with an area of 96816 km2, has a limited number of synoptic and rain gauge stations, makes it difficult to estimate rain erosivity in this province. In this study the auxiliary variables including digital elevation model (DEM), normalized vegetation index (NDVI), land surface temperature (LST) and global precipitation network data "Open Land Map Precipitation" (LMP) were used for spatial prediction of annual rainfall. The rainfall map of the study area was prepared using auxiliary data and using random forest (RF) model. Also in synoptic stations, the amount of erosivity was determined based on the EI30 index and average annual rainfall. Finally, the relation between rainfall and erosivity and annual rainfall was determined using nonlinear regression. Root mean square error (RMSE) and correlation coefficient (r) of RF model for prediction of annual rainfall were 16.9 mm and 0.98, respectively. The results of the rainfall map in the study area showed that the rainfall varied between 70-420 mm year-1.  Rainfall classification maps showed that near the half of the study area has annual rainfall less than 100 mm, 30% of the province has annual rainfall of between 100 and 150 mm and only about 17% of the province has annual rainfall more than 150 mm year-1. The maximum and minimum values of erosivity were 380 and 39 MJha-1mm h-1year-1 in the northern and southern part of the study area, respectively. Our results indicated using new method of data mining, it is possible to spatial prediction of rainfall and erosivity, especially in areas with small number of synoptic stations.   %U https://ijswr.ut.ac.ir/article_91206_6ed42d048b5906860dc52c210c8fdecd.pdf