Title: Assessing the impact of meteorological parameters for forecasting floods in the northern districts of Bihar using machine learning

Authors: Vikas Mittal; T.V. Vijay Kumar; Aayush Goel

Addresses: School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi, 110067, India ' School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi, 110067, India ' Department of Electronics and Communication Engineering, Bharati Vidyapeeth's College of Engineering, New Delhi, 110063, India

Abstract: India is the second largest flood affected country in the world. Every year floods have a deleterious effect on people, agriculture and infrastructure. Due to its high population density and poor infrastructure, the damage caused by floods in India is exacerbated forcing millions of people to migrate from one place to other. Therefore, there is a need to devise flood mitigation strategies that would forecast future floods in real time. In this paper, machine learning techniques have been used for forecasting floods in the northern districts of Bihar. Experimental results showed that, in addition to traditional meteorological parameters rainfall and temperature, certain parameters like vapour pressure, cloud cover, wet day frequency, crop evapo-transpiration and surface evapo-transpiration had a severe impact on the performance of a flood forecasting model.

Keywords: natural hazards; floods; forecasting; artificial intelligence; machine learning; supervised learning; classification.

DOI: 10.1504/IJW.2021.126818

International Journal of Water, 2021 Vol.14 No.4, pp.219 - 239

Received: 01 Jul 2021
Accepted: 09 Nov 2021
Published online: 08 Nov 2022 *