Reduced-complexity model of stream temperature

Abstract

In this article, a process-based model of time-varying stream temperatures is developed. The integrated compartment-based model balances a need to account for the major physical components of the heat balance at specific sites while maintaining computational efficiency. The ability of the model to simulate instantaneous temperatures is demonstrated at six study locations in the sub-tropical, low-gradient region of Southern Louisiana, USA. The model yielded Nash–Sutcliffe efficiency (NSE) scores of 0.9 and above and an average root mean squared error (RMSE) of 1.07°C for hourly water temperature over a six-month period. The parameter sensitivity analysis demonstrates the variable significance of vaporization and bed sediment conduction across the different study locations. The results show that the sediment heat exchange may have a significant seasonal effect on the energy budgets of streams where flows are a large component of the heat balance in temperate climates. The model is well-suited to serve as a simple physically based thermal impact assessment tool for streams and potentially many other types of waterbodies as well. The model also complements the widely used meteorological- and statistical-based methodologies applied in the development of tributary and headwater temperature boundary conditions for large-scale river temperature studies.