Time Series Modeling of Annual Maximum Daily Streamflow for Flood Forecasting

Abstract

Abstract Flood is a devastating phenomenon responsible for the loss of human lives, destruction of roads, buildings, electric systems and damage to hydraulic structures, leading to a great economic loss to the country. This study aims to identify a suitable time series model.Forecasting plays a major role in environmentally sustainable flood management. For the present study, daily streamflow records for four gauging stations were sourced from the Public Works Department. The Annual Maximum Daily Stream Flow (AMDSF) were determined from the daily data and analyzed through an Auto-Regressive Moving Average (ARMA) time series model. NCSS 9 software was used to determine the most suitable ARMA model for all four stream gauging stations. To further, the model is calibrated and validated using observed and forecasted data. The preferred ARMA model is utilized to forecast the river basin flood and is also helpful in framing proper environmental planning.