Prediction of watershed processes based on morphometric features using feature selection and neural network algorithms

Abstract

Alluvial fans of 4 watersheds in Iran were extracted semi-automatically using GIS and digital elevation model (DEM) analysis. The relationships between 25 morphometric features of these watersheds, the amount of erosion, and formation material were investigated using the self-organizing map (SOM) method. A feature-selection algorithm was used to select the most important parameters affecting erosion and formation material. The group method of data handling (GMDH) algorithm was employed to predict erosion and formation material based on morphometries. The results indicated that the semi-automatic method in GIS could detect alluvial fans. The SOM algorithm determined that the morphometric factors affecting the formation material were fan length, minimum height of fan, and minimum fan slope. The main factors affecting erosion were fan area and minimum fan height. The feature selection algorithm identified minimum fan height, maximum fan height, minimum fan slope, and fan length to be the morphometries most important for determining formation material, and basin area, fan area, maximum fan height and compactness coefficient (Cirb) were the most important characteristics for determining erosion rates. The GMDH algorithm predicted the fan formation materials and rates of erosion with high accuracy (R2 = 0.94, R2 = 0.87).