Impact of profile observations on the German Weather Service's NWP system

Citation
A. Cress et W. Wergen, Impact of profile observations on the German Weather Service's NWP system, METEOROL Z, 10(2), 2001, pp. 91-101
Citations number
21
Categorie Soggetti
Earth Sciences
Journal title
METEOROLOGISCHE ZEITSCHRIFT
ISSN journal
09412948 → ACNP
Volume
10
Issue
2
Year of publication
2001
Pages
91 - 101
Database
ISI
SICI code
0941-2948(2001)10:2<91:IOPOOT>2.0.ZU;2-E
Abstract
In preparation for a study on the potential impact of a space-borne Doppler wind lidar on the quality of NWP products, a series of assimilations and f orecasts were conducted to estimate the potential benefit of conventional w ind and temperature profile measurements over North America to numerical we ather forecasts for the Northern Hemisphere and specifically, Europe. A com parison of the forecast quality of a control run, using all available obser vations, to experiments omitting wind and temperature data from specific in struments (radiosondes, pilot stations and aircraft) makes it possible to e stimate the importance of the omitted data, and clarify whether winds deriv ed from the geostrophic relation are sufficient or whether observed wind pr ofiles result in a more realistic definition of the initial state for numer ical weather prediction systems in the extratropic regions. Very little impact on forecast quality was noted when wind or temperature o bservations from radiosondes and pilots were excluded from the assimilation process. However, a clear deterioration in forecast quality was observed w hen additionally all available wind or temperature measurements from aircra ft were also withheld. Comparisons of the relative utility of wind and temp erature observations over North America show that assimilations and forecas ts derive more benefit from wind data than from temperature data. The great est deterioration could be observed if both wind and temperature observatio ns were omitted from the assimilation cycle. By tracing the differences bet ween the control forecasts and the experimental forecasts to their initial difference, the regions around Hudson Bay, Novia Scotia, Buffin Bay and Nor thern Canada could be identified as sensitive areas, i.e. those where a mis sing observation could have a substantial effect on the forecast for the No rthern Hemisphere and Europe. Comparisons of the relative utility of radiosonde wind and temperature obse rvations over Canada and Alaska to numerical weather forecast quality, in c ontrast to the sonde and aircraft network over the United States, reveal th e importance of the conventional radiosonde network in the higher northern latitudes.