Precipitation and temperature inputs are the primary factors affecting net primary plant production across rangeland ecoregions. Skillful seasonal climate forecasts have the potential to directly aid in informing management for multiple agricultural and natural resource applications across the western United States. A key limitation in using these forecasts has been the availability of high-resolution climate forecasts across rangeland ecoregions given the high spatiotemporal heterogeneity in precipitation and temperature across these regions. This study examines the skill, or the correlation between of a climate forecast to actual data. Seasonal climate forecasts at lead times of up to 7 months lead times are statistically downscaled from models participating in the North American Multi-Model Ensemble (NMME). The skill of seasonal climate forecasts has typically been assessed at regional scales which may obscure forecast utility for site-specific management applications. Skill assessments are also typically only conducted on individual models, or the ensemble mean of all models which may overlook value-added skill by subsampling models. Here, we evaluated the skill of statistically downscaled monthly precipitation and temperature forecasts derived from seven NMME models for multiple rangeland ecoregions across the western United States. We found that multimodel aggregate forecasts often had synergistic skill when compared to both individual model forecasts as well as the ensemble mean. We also found that forecast skill was dependent on location, the month in which the forecast was made, and the forecast length. We suggest that a site and time-specific optimization of multimodel ensemble skill would increase the potential utility of these forecasts to inform rangeland management decision-making.