The Weather Prediction Center Day 8-10 Forecast Experiment: Assessment and Verification

The Day 8-10 Experiment conducted by the Hydrometeorology Testbed (HMT) at the Weather Prediction Center (WPC) assesses the skill and utility of experimental guidance to make daily temperature and precipitation forecasts in the day 8-10 period. Since January 2017, an automated blend of available operational guidance is created once per day for maximum and minimum temperatures and 24-h quantitative precipitation forecasts for days 8, 9, and 10. To compare with the automated blend, a forecaster-generated blend is created within experiment sessions twice per month that incorporate conceptual knowledge and current model trends. To increase the sample size of forecaster-generated blends, the blends have been produced every weekday since June 2017.

Verification of maximum and minimum temperatures and 24-h accumulated precipitation will be presented for days 8, 9, and 10. Temperature and precipitation forecasts from the automated blend, forecaster blend, and several of the individual NCEP operational inputs will be evaluated against the UnRestricted Mesoscale Analysis (URMA) and Stage IV dataset, respectively. Verification will be conducted using the Model Evaluation Tools (METv6.0) software package for both the Continental United States (CONUS) and Alaska. Preliminary CONUS results of daily maximum and minimum temperatures indicate a general cool forecast bias in the cool season (Oct–Mar) and a warm bias in the warm season (Apr–Sep). 24-h forecast maximum temperature mean error scores for all datasets were within ± 7°C for the day 8, 9, and 10 forecasts. The largest CONUS mean absolute errors scores for 24-h precipitation were incurred during the convective warm season months with values around 7 mm (0.25 in). During the cool season months, the data showed lower mean absolute errors but a consistent wet bias, over-forecasting of 24-h precipitation by approximately 1 mm (0.04 in). Probabilistic verification of a sample of day 8-10 forecasts of high-impact weather events will also be highlighted.