Estofex starts objective forecast verification
In the process of "learning how to forecast thunderstorms" ESTOFEX wants to monitor the quality of its forecasts. This is of course done by comparing the weather that was forecast with that what occurred in reality. Since the spring of 2006, lightning location data and (mostly unverified) severe weather reports from the European Severe Weather Database have been plotted on the maps of old forecasts. This gave the possibility to subjectively see if the threat level and thunderstorm areas had more or less been placed correctly.
Now we take a next step, by analyzing all forecasts issued since April 30th 2006. The results can be used to become aware of systematic forecast errors. This data is made available to our visitors as well and can be found in a new section of the ESTOFEX website.
1 Comments:
Don't be too hard on yourselves re: the False Alarm Rate. I would regard a forecast a success if storms break out within the yellow line area; storms will never cover every 40km grid cell within that zone.
More important is to get the nothing forecasted/nothing happened zones right - which you are! If *that* value was low, then I'd say you were producing lots of false positives.
Post a Comment
<< Home