Predicting depositional areas of landslide susceptibility comparing four datasets extracted from landslide area: a case of study after rainfall-induced landslides by Ida Hurricane in 2009 on Ilopango Lake, El Salvador.

Hurricane Ida and low-pressure system 96E crossed Central American countries in 2009. However, in El Salvador, the torrential rainfalls caused many flooding and landslides. As a result, over 200 causalities and the destruction of several villages, and bridges occurred along the mountain slopes. The remote analysis allowed us to prepare an inventory of landslides that occurred after the Hurricane in a basin located in the northern part of Ilopango Caldera. Five groups of data sets were created using selected pixels of each landslide area in order to evaluate the capacity to predict the lowest and the entire landslide area. Multivariate Adaptive Regression Splines (MARS) were employed to model the spatial distribution of the following five data sets: i) the highest cell (data set MAX), ii) the highest 10% of cells (data set SUP), iii) the lowest cell (data set MIN), iv) the lowest 10% of cells (data set INF), and v) the entire landslide area (data set BODY). To calibrate and validate the models were selected randomly in groups of 75% and 25% of the mapped landslides, respectively. In order to evaluate the robustness of the results, ten calibration and validation samples were extracted for each instability data set. The analysis revealed that the most important predictors were Slope Length Factor, Normalized Difference Vegetation Index (NDVI), Terrain Ruggedness Index, Lithology (pyroclastic rocks), Topographic Position Index, and Aspects NE and NW. The receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC), calculated for each of the five instability data sets, indicated that calibrating the models with the lowest landslide pixels (MIN data sets) allows to obtain the most accurate prediction of the validation the depositional area and the entire landslide bodies (BODY and INF data set), achieving AUC values ranging between 0.88 and 0.84.