In a modern Met Service, issuing weather alerts and warnings is core business. The number of warnings has risen significantly as a result of the availability of more detailed models and improved skill in forecasts. At KNMI, forecasters have to monitor more then 25 warnings and alerts, each for a different weather parameter, region and forecast lead-time. Since some of these warnings are only issued a few days per year, it is very hard for a forecaster to stay alert and ahead of (severe) weather developments.

As part of an innovation programme, a pilot project started at KNMI to explore whether automatic determination of (severe) weather events could help the forecaster in putting him on the right track, at the right moment. Relevant information (a “trigger”) is sent to the forecaster by a number of trigger-generating modules. These modules determine severe weather events by combining warning thresholds and area definitions with real time observations.

Trigger generating modules determine upcoming severe weather events by combining warning threshold and area definition with model forecast (HIRLAM, ECMWF). Trigger-modules have been implemented for observations (amount of precipitation, snowfall, number of lightning) and forecasts upto 10 days ahead (avg. windspeed and gusts, amount of precipitation, stormsurge and heathwaves).

Trigger generation for deviations in actual products and observations can also be implemented, helping the forecaster to stay ahead. The trigger concept is also useable for monitoring automatic production. Here, a trigger is generated depending on the content of the product resulting in appropriate action (e.g. quality control) by the forecaster.

Trigger information is sent to a central webbased dashboard in an XML trigger file. The information in the trigger is displayed on the dashboard and the forecaster gets a quick overview of relevant events. For further support, intranet pages are provided on which the forecaster can find more detailed information as well as production support.

The challenge of the project is to present a variety of triggers in such a way that forecasters find the system supportive, i.e. to make a system that effectively filters the amount of information, in stead of increasing it. If this challenge is met, the way forecasters work will change. They will most certainly more easily find the needle in the haystack of meteorological information.