The purpose of this study was to obtain the regional model of erosion according to the specific climatic, adaptive, and other conditions of the Toroq watershed located in the east north of Khorasan Razavi province. To conduct this research, first, the homogeneous units were prepared using slope maps, lithology, land use, and erosion forms in a Geographic Information System environment. Then, to optimize the number of homogeneous units, the cluster analysis method was used in Statistical Product and Service Solutions (SPSS) software. The diagnostic analysis confirmed the accuracy of cluster analysis inhomogeneous regions. Field operations were carried out in homogeneous units with the establishment of a rainfall simulator and also the application of 30-min rainfall intensity with a return period of 10 years. Also, the collected soil samples were analyzed in the laboratory. After performing statistical analyses in the SPSS environment, the variables affecting erosion were determined and prioritized. Then, through the use of multivariate linear regression and step-by-step and interpolation methods, the equations for estimating the amount of erosion were determined. Finally, the multivariate linear model of plot erosion was prepared using the step-by-step method using two variables of plot slope and land use. The model was selected for estimating erosion after examining different validation methods based on less RE and less RMSE, higher R, low significance coefficient (Sig < 0.05), and also fewer inputs.

这项研究的目的是根据位于霍拉桑·拉扎维省东北北部的托罗克分水岭的具体气候，适应性和其他条件，获得区域侵蚀模型。为了进行这项研究，首先，在地理信息系统环境中，使用坡度图，岩性，土地利用和侵蚀形式制备均质单元。然后，为了优化同类单元的数量，在统计产品和服务解决方案（SPSS）软件中使用了聚类分析方法。诊断分析证实了聚类分析不均匀区域的准确性。现场操作是在均质单元中进行的，建立了降雨模拟器，并应用了30分钟的降雨强度，回归期为10年。也，收集的土壤样品在实验室中进行了分析。在SPSS环境中进行统计分析后，确定影响腐蚀的变量并确定优先级。然后，通过使用多元线性回归以及逐步和插值方法，确定了估算侵蚀量的方程。最后，利用样地坡度和土地利用两个变量，采用分步法建立了样地侵蚀的多元线性模型。在基于较少的RE和较少的RMSE，较高的R，较低的显着系数（Sig <0.05）和较少的输入检查了不同的验证方法之后，选择该模型来估计侵蚀。然后，通过使用多元线性回归以及逐步和插值方法，确定了估算侵蚀量的方程。最后，利用样地坡度和土地利用两个变量，采用分步法建立了样地侵蚀的多元线性模型。在基于较少的RE和较少的RMSE，较高的R，较低的显着系数（Sig <0.05）和较少的输入检查了不同的验证方法之后，选择该模型来估计侵蚀。然后，通过使用多元线性回归以及逐步和插值方法，确定了估算侵蚀量的方程。最后，利用样地坡度和土地利用两个变量，采用分步法建立了样地侵蚀的多元线性模型。在基于较少的RE和较少的RMSE，较高的R，较低的显着系数（Sig <0.05）和较少的输入检查了不同的验证方法之后，选择该模型来估计侵蚀。