Verification over the western United States of surface temperature forecasts from the National Digital Forecast Database

Experimental gridded forecasts of surface temperature issued by National Weather Service offices in the western United States during the 2003-2004 winter season are verified against surface observations and gridded analyses. The 5 km horizontal resolution gridded forecasts issued at 0000 UTC for forecast lead times at 12 h intervals from 12-168 h were obtained from the National Digital Forecast Database (NDFD). Forecast skill is determined relative to observations at over 3000 locations archived by MesoWest. Forecast skill is also determined relative to Rapid Update Cycle (RUC) analyses at 20 km resolution that are interpolated to the 5 km NDFD grid as well as objective analyses obtained from the Advanced Regional Prediction System Data Assimilation System (ADAS) that rely upon the MesoWest observations and RUC analyses.

For the West as a whole, the experimental temperature forecasts issued at 0000 UTC during the 2003-2004 winter season exhibit skill at lead times of 12, 24 and 48 h on the basis of several measures-oriented and distributions-oriented verification approaches. Sub-grid scale temperature variations and observational and analysis errors undoubtedly contribute some uncertainty regarding these results. However, these skill measures have a generally small sensitivity to the choice of verifying data set or to whether the entire grid or only grid locations near observations are used. Morning temperature biases that are nearly independent of forecast lead time may result from the methodology used to derive the hourly temperature forecasts from daily maximum and minimum temperature forecast grids. Evaluation of the forecast grids, using a variety of skill measures based on clearly defined verification goals, should be encouraged at the National Weather Service offices as well as nationally.