Accurate snowfall estimation in mountainous regions is critical for hydrological forecasting, avalanche hazard assessment, and infrastructure development planning. For the mountains of Western Canada, the High-Resolution Deterministic Prediction System (HRDPS) provides short-term precipitation forecasts at 2.5 km grid spacing and captures most precipitation-generation mechanisms in complex terrain. However, systematic biases in HRDPS can compromise the accuracy of snowpack simulations that rely on its precipitation outputs. The Canadian Precipitation Analysis (CaPA) offers a potential solution by combining HRDPS precipitation estimates with surface observations through Optimal Interpolation (OI). Despite this improvement, previous research has demonstrated that the operational CaPA configuration underestimates winter precipitation in mountainous terrain due to sparse station coverage and inherent limitations of the OI methodology. This study evaluates a revised CaPA configuration for the mountains of Western Canada that incorporates three key enhancements: (i) adjusted wind quality control parameters to assimilate more stations during winter, (ii) integration of precipitation observations from additional networks in mountainous regions, and (iii) a refined OI methodology that accounts for topographic effects. These modifications yielded systematic improvements in precipitation estimates compared to both the operational CaPA and HRDPS products. The precipitation datasets were subsequently used to force snowpack simulations with the SVS2 (Soil, Vegetation and Snow) land surface scheme. Evaluation against observed snow water equivalent and snow depth revealed notable improvements across Western Canada's mountains, particularly in the Coastal Mountains. However, challenges persist in adequately representing topographic effects within the OI framework. These findings inform the development of the next operational CaPA version.