DOWNSTREAM DEVELOPMENT OF BAROCLINIC WAVES AS INFERRED FROM REGRESSION-ANALYSIS

Authors
Citation
Ekm. Chang, DOWNSTREAM DEVELOPMENT OF BAROCLINIC WAVES AS INFERRED FROM REGRESSION-ANALYSIS, Journal of the atmospheric sciences, 50(13), 1993, pp. 2038-2053
Citations number
34
Categorie Soggetti
Metereology & Atmospheric Sciences
ISSN journal
00224928
Volume
50
Issue
13
Year of publication
1993
Pages
2038 - 2053
Database
ISI
SICI code
0022-4928(1993)50:13<2038:DDOBWA>2.0.ZU;2-E
Abstract
The structure and evolution of transient disturbances in the Northern Hemisphere winter season are examined using one-point regression maps and longitude-height sections derived from the European Centre for Med ium-Range Weather Forecasts (ECMWF) operational analyses for seven win ter seasons. With the use of unfiltered time series of normalized 300- mb meridional wind perturbations at a grid point in the Pacific storm track as the reference time series, regression statistics for perturba tions in the horizontal wind, geopotential height, temperature, and ve rtical velocity are derived. The resulting perturbation fields exhibit characteristics of mid-latitude baroclinic waves, such as a westward tilt with height in the velocity and height fields and eastward tilt i n the temperature field, with typical wavelengths of 4000 km and perio ds of around 4 days. The main difference between the results of this w ork and previous similar analyses is in the propagation characteristic s of the baroclinic wave trains. The wave trains found here exhibit ch aracteristics of downstream development, with successive perturbations developing toward the downstream side of existing perturbations. An a nalysis of the eddy kinetic energy budget of the wave train indicates that downstream radiation of ageostrophic geopotential fluxes by exist ing perturbations triggers the development of new eddies downwind, wit h conversion becoming important only during the later part of the life cycle of a downstream developed wave. The Appendix suggests that the difference between these results and those of previous works is mainly due to time filtering obscuring the delicate signal of downstream dev elopment.