The National Weather Service produces Public Forecasts for use on the local, regional, and national level. To determine and improve our level of customer service, these forecasts must be verified for accuracy. Migration to the AWIPS platform left the Weather Forecast Offices (WFOs) with little means to facilitate this, thus a verification program was written.

The Coded Cities Forecast (CCF), various Model Output Statistics (MOS), and observed data are used to statistically measure forecaster performance. Statistics calculated include the traditional percent improvement over MOS, Brier Score, Bias, and Mean Absolute Error. Additionally, Probability of Precipitation (PoP) reliability is measured, as are 5 and 10 degree forecast errors and their percentage of occurrence.

Real time output is generated 2 times per day and is viewable via the AWIPS text workstation. This output includes a basic set of statistics and output for the previous 10 forecasts for both the forecasters and MOS. User defined output can be generated on demand for any given time period, with two scoring options available. The standard version computes all statistical output for the specified forecaster. The abbreviated version calculates a subset of the standard output for each forecaster, stratified by statistical data type and station. A graphical representation of the statistical output can also be generated using the enclosed Quattro Pro file.