Modeling of peak discharges and frequency analysis of floods on the Jhelum river, North Western Himalayas

Abstract

The modeling of peak flood discharges and flood frequency analysis at various sites on a river is essential for planning, design, and management of hydraulic structures. The first and the foremost aim of this study is to choose the best-fit flood model among Log Pearson type 3 (LP3), Generalized Extreme Value (GEV), and Gumble (EV1) for each of the eight sites on the Jhelum River and for the same purpose goodness-of-fit tests like Anderson–Darling (A–D) and Kolmogorov–Smirnow (K–S) and distribution graphs (P–P plot and Probability difference graph) were used. The parameters of these models were determined by L-moments. The outcomes of the study reveal that the LP3 model is best-fit for Khanabal, Sangam, Awantipora, Padshahi Bagh, Ram Munshi Bagh, and Asham, and GEV is the best fit for Sopore, and Baramullah sites. Furthermore, peak discharges for 2-, 5-, 10-, 25-, 50-, 100-, 200-, and 500-year return periods were estimated and the analysis depicts that the discharge rate determined by distribution models at a return period of 5 years or more would surpass the safe carrying capacity (990.85 cumecs) of the Jhelum river. The study further shows that there exists a high positive correlation (R2 = 0.99) between observed and predicted peak discharges of LP3 and GEV models. Thus, indicating LP3 and GEV as best-fit models for modeling and flood frequency analysis of annual peak discharges on the Jhelum River.