Coupling of hydrological and machine learning models for reservoir storage and inflow forecasting

Abstract: Forecasting near real-time water availability in reservoirs helps in effective water resources management, flood control, hydropower generation, irrigation, and drought mitigation. The conceptual framework is built around a deep tech exercise encompassing artificial intelligence, deep-learning, water physics, hydrological science, and geoinformatics, alongside an innovative interdisciplinary approach. The predictive modelling exercise yields meaningful results when it is complemented with human understanding of the physical processes in the river basin. This is achieved by combining machine learning and hydrological modelling. Three deep learning architectures have been used – fully connected deep neural network, long short-term memory network, and WaveNets. They take inputs such as storage, inflow, outflow, and precipitation leading up to the reservoir and use these inputs to predict the inflow and storage level 30, 60 and 90 days ahead. The visualization and dissemination of the results of this exercise and other critical hydro-meteorological datasets are done through an interactive dashboard to help stakeholders make informed decisions.

Date/Time
Date(s) - 21/03/2022
5:00 pm - 6:00 pm