Influence of Dam Spill Flow on Downstream River Water Temperatures

River water temperature forecasting is very important for river management, including power plant operation. A substantial amount of research has been performed to forecast river water temperatures using different means. However, no research has utilized nor investigated the effect of adding the spill flow feature to forecast river water temperatures. This research will utilize weather data, water flow through the dam (turbine flow and spill flow), and spatially distant water temperatures to forecast river water temperature at a specified location of interest. Two models (Gradient Boosting Regression and Random Forrest Regression) were trained and tested with and without the spill flow feature to compare results. Also, a SHAP values study was conducted on Gradient Boosting Regression and Random Forrest Regression to rank features by importance. The study showed that the spill flow feature enhances the models' accuracy in predicting river water temperatures.