Over the western U.S, 27% of the electricity demand is met with hydropower on an average year. Water resources availability in the western U.S. are projected to undergo changes in seasonality in the future due to the decline in snowpack. Future projection of hydropower generation which are based on regression relationship between observed monthly flow and hydropower generation combining with simulated flow have the advantage of reflecting the complexity in hydropower operations. However they do not capture the non-stationary generation due to the shift in natural flow seasonality and associated changes in reservoir storage which affect the generation. Physically – based hydropower models based on large scale regulated flow simulations can capture the change in hydropower generation however they would need to be calibrated in order to provide further insight onto power system resource adequacy studies and understand the impact on overall power system operations. In this study we simulate historical and future monthly hydropower generation using a process-based model. We use an integrated water modeling framework consisting of hydrologic simulations based on the Variable Infiltration Capacity Model (VIC), CMIP5, and large-scale river routing - water management model (MOSART-WM) over the Western Electricity Coordinating Council (WECC) region. Reservoir operations characteristics, in addition to generation, are calibrated to reproduce historical observations. Estimates are provided for run of the river reservoirs and storage reservoirs over the Western U.S. We provide a new projection of potential hydropower generation and evaluate the effect of non-stationarity with respect to previous regression-based estimates.