This project focused on assessing and identifying avenues to improve Reclamation’s current temperature modeling use of seasonal predictions of input meteorology, as well as on several related investigations. The effort developed a new overview for Reclamation and stakeholders of the methods and performance of the meteorologic forecasts currently being applied for water temperature simulation. To facilitate this effort, the current spreadsheet-based Local Three-Month Temperature Outlook (L3MTO) method was duplicated in a set of Python scripts, which were transitioned to Reclamation for potential use in operations at Shasta Lake and other California facilities. These scripts make hindcast analyses feasible and provide flexibility in examining alternative inputs and ease in conducting supporting analyses. The scripts were used to test or demonstrate several variations on the approach, including using Sub-seasonal to Seasonal (S2S) forecasts from other sources, and to assess the skill of the approach and its likely upper limits of performance. In particular, the work showed that the temperature model input forecasts had mean monthly skill in the first lead month of the forecast and indicated potential for improvement through further development of the input climate forecasts as well as using alternatives to the climate-conditioned deterministic analog selection method. The project also implemented a California-wide daily gridded meteorological analysis using a tool called the Gridded Meteorological Ensemble Tool (GMET), and a demonstration of a linked Structure for Unifying Multiple Modeling Alternatives-mizuRoute-River Basin Model (SUMMA-mizuRoute-RBM) model for the Shasta and Trinity Lake drainage areas, demonstrating the feasibility of a distributed strategy for stream temperature simulation and potentially prediction.