Water Temperature Ensemble Forecasts: Implementation Using the CEQUEAU Model on Two Contrasted River Systems

In some hydrological systems, mitigation strategies are applied based on short-range water temperature forecasts to reduce stress caused to aquatic organisms. While various uncertainty sources are known to affect thermal modeling, their impact on water temperature forecasts remain poorly understood. The objective of this paper is to characterize uncertainty induced to water temperature forecasts by meteorological inputs in two hydrological contexts. Daily ensemble water temperature forecasts were produced using the CEQUEAU model for the Nechako (regulated) and Southwest Miramichi (natural) Rivers for 1-5-day horizons. The results demonstrate that a larger uncertainty is propagated to the thermal forecast in the unregulated river (0.92-3.14 degrees C) than on the regulated river (0.73-2.29 degrees C). Better performances were observed on the Nechako with a mean continuous ranked probability score (MCRPS) <0.85 degrees C for all horizons compared to the Southwest Miramichi (MCRPS approximate to 1 degrees C). While informing the end-user on future thermal conditions, the ensemble forecasts provide an assessment of the associated uncertainty and offer an additional tool to river managers for decision-making.