Flood Frequency Analysis Using ERA5-Land Based Precipitation for Kosi-Mahasetu Station in North Bihar, India

Abstract

Floods are the biggest concerns in developing countries. Lower catchment areas of Kosi River basin inside North Bihar recurrently affected by floods due to heavy downpour. Flood frequency information is therefore vital but unavailability of historical data at public domain makes difficult to investigate flood events. This study has utilized ERA5-Land based rainfall data (1981-2020) over Kosi-Mahasetu CWC rain gauge in Supaul district for evaluating the rainfall frequency and development of different distribution curves. The mean annual rainfall at Kosi-Mahasetu station was 518.12 mm (SD =115.66, CV =0.12). The coefficient of skewness and Kurtosis was 0.184 and 0.607. The distribution functions such as Normal, Log 10 Normal and Gumble were applied over historical rainfall data to calculate rainfall intensity for 5, 10, 25, 50, and 100 years of return period. Based on Kolmogorov-Smirnov (K-S) test, it is observed that Log 10 normal distribution showed significant results. During 2007 flood event, the region received highest rainfall as ~779 mm during monsoon season with a return period of 41 years. Similarly, 1999, 2019, 1996, 2008, 1995, and 1993 flood events received ~748, ~730, ~717, ~677, ~673, ~624, ~616 mm rainfall with a return period of 20.5, 13.67, 10.25, 8.2, 6.83, 5.86, and 5.12 years, respectively. The study also exhibited that a total ~24% land area of North Bihar inundates recurrently. This study found helpful to predict and prepare for the extreme hydrological events by considering the return period of major flood events.