Coupled daily streamflow and water temperature modelling in large river basins
Realistic estimates of daily streamflow and water temperature are required for effective management of water resources (e.g. for electricity and drinking water production) and freshwater ecosystems. Although hydrological and process-based water temperature modelling approaches have been successfully applied to small catchments and short time periods, much less work has been done at large spatial and temporal scales. We present a physically based modelling framework for daily river discharge and water temperature simulations applicable to large river systems on a global scale. Model performance was tested globally at 1/2 × 1/2° spatial resolution and a daily time step for the period 1971–2000. We made specific evaluations on large river basins situated in different hydro-climatic zones and characterized by different anthropogenic impacts. Effects of anthropogenic heat discharges on simulated water temperatures were incorporated by using global gridded thermoelectric water use datasets and representing thermal discharges as point sources into the heat advection equation. This resulted in a significant increase in the quality of the water temperature simulations for thermally polluted basins (Rhine, Meuse, Danube and Mississippi). Due to large reservoirs in the Columbia which affect streamflow and thermal regimes, a reservoir routing model was used. This resulted in a significant improvement in the performance of the river discharge and water temperature modelling. Overall, realistic estimates were obtained at daily time step for both river discharge (median normalized BIAS = 0.3; normalized RMSE = 1.2; r = 0.76) and water temperature (median BIAS = −0.3 °C; RMSE = 2.8 °C; r = 0.91) for the entire validation period, with similar performance during warm, dry periods. Simulated water temperatures are sensitive to headwater temperature, depending on resolution and flow velocity. A high sensitivity of water temperature to river discharge (thermal capacity) was found during warm, dry conditions. The modelling approach has potential to be used for risk analyses and studying impacts of climate change and other anthropogenic effects (e.g. thermal pollution, dams and reservoir regulation) on large rivers.