Earthquake risk mapping in the Himalayas by integrated analytical hierarchy process, entropy with neural network

Earthquakes are natural disasters that threaten human lives and infrastructure, especially in seismo-tectonically active regions. Therefore, mapping and assessment of earthquake risks are indispensable for disaster preparedness and mitigation. In this study, a novel approach has been adopted by integrating the subjective and objective multi-criteria decision-making (MCDM) models, i.e. analytical hierarchy process (AHP), entropy, and artificial neural network (ANN), to estimate the earthquake risk in the Himalayan tectonic region. Integration of AHP and Entropy has been applied to assess the vulnerability and the coping capacity, whereas ANN has been used to estimate the earthquake probability. The hazard map is generated using earthquake intensity and probability thematic layering information. Subsequently, the earthquake risk was evaluated by combining the hazard, vulnerability, and coping capacity maps. The results indicate that more than 31% of the area may be under high to very high risk, whereas about 27% of the population and 31% of the buildings may be at high to very high risk of earthquake hazards. The receiver operating characteristic (ROC) curve indicates good results, with the area under the curve of approximately 0.83. The results presented in this study may be helpful for the government agencies involved in disaster mitigation to mitigate and prepare strategies for earthquake hazards in the Himalayan region.