Curve Estimation Modeling between Area and Volume of Landslides in Tajan River Basin, North of Iran

-

Determining landslide size could be a difficult and expensive task. In this research, size parameters of 142 landslides recognized in Tajan River Basin, northern Iran, have been assessed. The dataset was prepared through the extensive field surveys and using the satellite imagery available via Google Earth. Dependence between landslides area (A-m2), volume (V-m3), and depth (D-m) was appointed by the Pearson correlation coefficient (r) at 0.05 and 0.01 levels. Then, the relation between the area and volume variables has been investigated using 10 curve estimation (CE) models. Coefficient of determination (R2), F statistic, and RMSE were calculated to compare the models with each other. Results showed that the power law fit the data better than other CE models. Although, the quadratic and cubic relationships have represented high R2 and low RMSE, they have resulted negative estimated volumes, and also their F statistic is less than its value in power law. To achieve a better result, the estimated volumes were compared with the observed ones using paired test. Results indicated that the estimated volumes were in conformity with the observed ones and there was no statistically significant difference between them (R2=0.801, sig=0.633). Although, the estimated depths were significantly different from the observed ones. The mean depth was estimated 5.5 m which was close to mean of the actual depths (5.53 m).