National Weather Service Baltimore/Washington's Model Verification Internal Web Display

With the ever growing repository of model data available to the world of meteorology today, forecasters find themselves struggling to identify the model(s) or model blends to use for a particular element without an easy to interpret verification database. Model blends, such as the Super Blend or the National Blend of Models may outperform certain pieces of guidance over the long run for certain elements of the forecast, but continue to struggle during stretches of time, during notable events, or with certain forecast elements, especially when terrain is involved.

The National Weather Service Baltimore/Washington Weather Forecast Office has developed a local verification page that displays model verification statistics and biases in a visual and easy to digest manner. The model verification statistics are computed using the BOIVerify application that is widely used at weather forecast offices around the United States. After computation the data is organized and displayed in a format that a forecaster at first glance can determine the model(s) or blend(s) that is (are) performing best for a particular element on any given day.