Identifying User-Defined Weather Events of Significance through Data Mining

Data mining is used by organizations to sift through vast amounts of raw data and turn it into useful information. By using software to look for patterns in large batches of data, businesses can learn more about their data and develop more effective use of what otherwise might not be seen. As the information load for the forecaster has increased, creating evaluation criteria, then filtering the forecast models with software that searches for any pattern match using that criteria, is potentially beneficial.

Based on forecaster requests, NOAA’s Global Systems Division (GSD) created a software tool to assist in pre-screening gridded forecasts. It is integrated into the AWIPS Graphical Forecast Editor (GFE) and called the experimental Forecaster Event Awareness Tool (FEAT). This tool alerts the operational forecaster to weather events of significance. Rules defining significant weather are used to systematically screen the forecast grids to identify events of potential importance. In this tool, when multiple model forecasts are in good agreement on the occurrence of a significant weather event, then a higher weight is applied to the results. The size of the area of coverage is also an important factor. Forecasters are interested and actively engaged with this tool to identify events that concern them by creating rules of varying complexity.

This presentation will showcase a few examples of general data mining scenarios. For example, FEAT can search for unusually high gradients in fields of selected parameters, typically corresponding to areas of rapidly changing weather. Weather Forecast Offices can set specific custom rules for automated Red Flag warnings to reflect their local conditions. The magnitude of differences between the latest model forecasts and the official forecast can also be monitored, alerting forecasters if a preset criterion is met. Such alerts relieve forecasters of some routine monitoring work, leaving more of their time and attention for more critical forecast decisions and better provide decision-support services.