REAL-TIME FORECASTING OF THE WESTERN-AUSTRALIAN SUMMERTIME TROUGH - EVALUATION OF A NEW REGIONAL MODEL

Citation
Lm. Leslie et Tcl. Skinner, REAL-TIME FORECASTING OF THE WESTERN-AUSTRALIAN SUMMERTIME TROUGH - EVALUATION OF A NEW REGIONAL MODEL, Weather and forecasting, 9(3), 1994, pp. 371-383
Citations number
NO
Categorie Soggetti
Metereology & Atmospheric Sciences
Journal title
ISSN journal
08828156
Volume
9
Issue
3
Year of publication
1994
Pages
371 - 383
Database
ISI
SICI code
0882-8156(1994)9:3<371:RFOTWS>2.0.ZU;2-R
Abstract
The real-time prediction of the location, strength, and structure of t he summertime heat trough is a major forecasting problem over Western Australia. Maximum temperatures, wind strength and direction along the west coast, low-level coastal cloud, and thunderstorm activity are vu lnerable to forecast errors in the heat trough. This study has three m ain parts. First, prediction errors of the operational Australian regi on numerical weather prediction (NWP) model were quantified over the p eriod December 1991 to February 1992. Second, a newly developed region al NWP model, which will be the next operational regional model, was c ompared with the current operational model. The new model has more eff icient numerics than the present operational model, allowing higher-re solution forecasts and a more sophisticated representation of physical processes. The third part was a set of sensitivity experiments to ass ess the relative importance of the differences. The dominant errors in the current operational model are a large westward bias in the trough location, a wide spread of errors in the intensity of the low in the northern section of the heat trough, a sizable range of coastal pressu re gradient errors, and a northward bias in the latitude of the subtro pical ridge axis between longitudes 110 degrees and 120 degrees E. It was demonstrated that these errors are reduced significantly in the ne w model, especially the subtropical ridge error, which has been virtua lly eliminated. The sensitivity studies revealed the importance of eac h of the differences between the models, and that the relative impact varies from case to case.