Coach acquires knowledge from the WFO forecasters and stores it in the context of the current weather pattern. For instance, when a forecaster is formulating the forecast, he or she will identify the forecast problems of the day, expected weather events, confidence in each of several model forecasts, and any relevant subjective guidance or heuristics. Approximately 24 hours later, the forecaster will enter followup information into the Coach system. Such information as observed weather events, subjective evaluation of model and local forecasts, and any unspecified heuristics is stored in a database linked to the initial forecast entry. By capturing knowledge in this manner, the job of automatically distributing relevant expert information becomes straightforward.

Expert forecasting knowledge is furnished in real time and displayed in the context of the current weather pattern. This knowledge can come from several sources, both internal and external to Coach. Data such as objective forecast verification scores and subjective "followup" information from the forecasters are obtained directly from the Coach system. However, vast knowledge sets relevant to operational forecast obviously exist from other sources. A long-term goal of this project is to acquire and process as much of these external knowledge sets as possible so they can be distributed succinctly in a pattern or phenomenon-based context. Results so far indicate this goal will be difficult to achieve on a WFO scale (approximately 300 km), and likely requires the assistance and cooperation of other agencies and research organizations involved in operational forecasting.

For more information see http://www.srh.noaa.gov/tulsa/coach.html.