Short-Term Stream Flow Forecasting With The Aid Of Global Climate Change Indices

To more efficiently use the limited amount of water under the impact of global climate change or to resourcefully provide adequate time for flood and drought warning, there is an acute need to seek an advanced modeling technique for improving streamflow forecasting on a short-term basis. This study aims to expand forecasting capacity by incorporating sea surface temperature (SST), spatiotemporal rainfall distribution from the Next Generation Radar (NEXRAD), meteorological data, and historical stream flow data to forecast discharges in a semi-arid watershed in South Texas, U.S.A. Comparative study was conducted by comparing the performance of traditional time-series model against the outputs of neural network (NN) and genetic programming (GP) models. The case study elicits microclimatological factors and the resultant stream flow rate in a river system given the influence of dynamic basin features, such as soil moisture, soil temperature, ambient relative humidity, air temperature, and precipitation. SST data were acquired from three locations including the Atlantic Ocean, the Gulf of Mexico, and the Pacific Ocean. Five numeric evaluators were defined and applied to evaluate all models involved. Research findings show that GP-based models have relatively better performance in most cases compared to neural network (NN) time-series models based on 16-week historical data. SST and meteorological data may significantly improve the GP-derived stream forecasting model that is highly recommended. The developed GP models can do pretty well forecasting for next 30-day discharges with r-square value of 0.84 and percentage error (PE) of 7%