SUMMARY OF RESULTS BASED ON THE
NORPEX-98, WSR99 AND WSR00 FIELD PROGRAMS
1) Of the 63 targeted events, 70% of the cases
showed a clear improvement in the quality of 24-96 hour lead time forecasts,
measured in terms of error reduction in surface pressure, tropospheric
wind, and precipitation forecasts. Only 13% of all cases showed a degradation,
while in the rest of the cases the adaptively taken data had neutral impact.
2) The average error reduction within the verification regions that were preselected based upon large expected societal impact/threat (spanning the continental US and Alaska) is on the order of 10-25%. This can be compared to the 10% error reduction due to enhancements in data quality/quantity in the global observing network over the NH extratropics over the past 25 yrs (including new satellites etc). When comparing the numbers above one should keep in mind that the targeted data has a relatively large impact over a region only (though the regions are selected based on their importance) while overall imporvements to the global observing system enhance the forecasts everywhere.
3) The 10-25% rms error reduction transletes to a
gain in forecast accuracy that is equivalent to using a forecast (based
on non-adaptive observations only) that is issued 12-24 hours later.