GIS-based multi-criteria decision analysis for mapping flood-prone areas in Dehradun city, India

Dehradun city, capital of Uttarakhand state in India, is severely affected with increasing incidents of flooding, causing tremendous losses of life, infrastructure, and eroding years of development. This research aims to map flood hazard zones by conducting flood hazard assessment based on flood hazard index (FHI), which is constructed using 11 indicators, viz. elevation, slope, rainfall, LULC (land-use land-cover) runoff potential, and distance from drainage, drainage density, soil, geomorphology, geology, water table depth, and TWI (Total Wetness Index). The indicators are ranked and reclassified and then weighted using AHP (analytic hierarchy process) methodology. These are combined using weighted linear combination method in GIS (geographic information systems) with necessary inputs of weights of layers. Landsat8 OLI image of spatial resolution 30m, DEM 2015 of CARTOSAT-I of spatial resolution 2.5 m, and Survey of India (SOI) toposheets, scale 1:50,000, for Dehradun city are used for creating base layer and deriving indicators. The annual rainfall data of different rain gauge stations are interpolated in GIS to obtain rainfall map. As a result of final analysis, FHI is classified into 4 flood hazard zones spatially—very high risk with 29.51% area, high risk with 29.83% area, medium risk with 35.50 % area, and low risk with 5.16 % area. Thus, almost 59% of city areas are under high to very high risk. It is observed that the areas along Bindal and Rispana rivers are primarily at high risk. The results are validated by using ROC-AUC (area under the ROC curve) method using 151 high flood point samples spatially distributed well in the study area. As a further scope of research, number of socio-economic indicators may be added and flood risk assessment may be carried on. Also there is a scope of comparative assessment of other MCDA (multi-criteria decision analysis) methodologies to improve zonation accuracy in results.