Non-stationary flood frequency analysis and attribution of streamflow series: A case study of Periyar river, India

Abstract

The occurrence of extreme hydrological flood events has become frequent recently, especially in India. Flood analysis methods need to capture variations in trends and identify key factors causing this variation. Non-stationary flood frequency analysis is one of the few methods which can account for these variations. In this study, the non-stationarity of Periyar river discharge was investigated through the flood frequency analysis of one stationary and 12 non-stationary models, produced using Generalized Extreme Value distribution with annual precipitation, urban extent and time as a linear function of its location and scale parameter. The analysis explained that both, climate change and anthropogenic activities are accountable for the non-stationarity of the Periyar river discharge. Also, the 50-year return level of the stationary model was found to be equal to the 39-year return level of the optimal non-stationary model, thus, the assumption of stationarity may lead to the unsafe design of structures.