Ensemble streamflow forecasting over a cascade reservoir catchment with integrated hydrometeorological modeling and machine learning

Abstract. A popular way to forecast streamflow is to use bias-corrected meteorological forecast to drive a calibrated hydrological model, but these hydrometeorological approaches have deficiency over small catchments due to uncertainty in meteorological forecasts and errors from hydrological models, especially over catchments that are regulated by dams and reservoirs. For a cascade reservoir catchment, the discharge of the upstream reservoir contributes to an important part of the streamflow over the downstream areas, which makes it tremendously hard to explore the added value of meteorological forecasts. Here, we integrate the meteorological forecast, land surface hydrological model simulation and machine learning to forecast hourly streamflow over the Yantan catchment, where the streamflow is influenced both by the upstream reservoir water release and the rainfall-runoff processes within the catchment. Evaluation of the hourly streamflow hindcasts during the rainy seasons of 2013–2017 shows that the hydrometeorological ensemble forecast approach reduces probabilistic forecast error by 10 % and deterministic forecast error by 6 % as compared with the traditional ensemble streamflow prediction (ESP) approach during the first 7 days. The deterministic forecast error can be further reduced by 6 % in the first 72 hours when combining the hydrometeorological forecast with the long short-term memory (LSTM) deep learning method. However, the forecast skill for LSTM using only historical observations drops sharply after the first 24 hours. This study implies the potential of improving flood forecast over a cascade reservoir catchment by integrating meteorological forecast, hydrological modeling and machine learning.