Web Based GLOF Susceptibility Model Using Geo informatics

Abstract

Glacial lake outburst floods (GLOFs) are becoming major concerns for the Northern Pakistan region. Several disasters have occurred due to outburst floods from glacial lakes and the frequency of occurrence is increasing with the looming crisis of climate change. A total of 498 lakes were identified through digitization on Google Earth Pro. This project identifies the lakes susceptible to GLOF occurrences in the Hunza River Basin using Analytic Hierarchy Process and Machine Learning techniques, and then a model is created which predicts the susceptibility of any lake, if the required parameters are provided. Simulations for the extent of flood in the case of an outburst for most susceptible lakes are also created. Moreover, this project shows a spatio-temporal change in glacial lakes of Hunza River Basin between the years 2016 and 2021 so that the effects of climate change can be analyzed, and the patterns can be further discussed by researchers. The machine learning model, flood extent simulations, and spatio-temporal change are displayed on a web application. In our study area of Hunza River Basin, 10 lakes are found to be highly susceptible. Based on the machine learning model, the web application classifies the GLOF susceptibility of any input lake into one of 3 categories: high, medium, or low. Moreover, the web application consists of a few easy-to-use buttons using which the spatiotemporal change can be observed. Simulations of the most susceptible lakes in our study area can be seen with some climatic data for the region as well.