Radiosonde observations of partially trapped lee waves over Tasmania, Australia

Citation
Mj. Reeder et al., Radiosonde observations of partially trapped lee waves over Tasmania, Australia, J GEO RES-A, 104(D14), 1999, pp. 16719-16727
Citations number
13
Categorie Soggetti
Earth Sciences
Volume
104
Issue
D14
Year of publication
1999
Pages
16719 - 16727
Database
ISI
SICI code
Abstract
A well-defined family of lee waves over southern Tasmania, Australia, is an alyzed using a routine radiosonde sounding released from Hobart at 1100 UTC June 18, 1991. Pronounced fluctuations in the ascent rate of the radiosond e are analysed using a new technique called the horizontal projection metho d. The dominant horizontal wavelength is found to be 8249 m. An advanced ve ry high resolution radiometer (AVHRR) infrared satellite image at 1554 UTC is used to check this calculation and gives a horizontal wavelength of 8.9 +/- 1.4 km. The horizontal projection method takes account of the vertical variation of background flow and the trajectory of the radiosonde. When the Scorer parameter is a strong function of height, as is the case with the 1 100 UTC sounding, it is inappropriate to assume that the vertical wavenumbe r is constant with height. In particular, the vertical wavelength at low le vels is about 4.5 km, the waves become evanescent between the heights of 2. 5 km and 9 km, and above that the vertical wavelength is typically 10-15 km . As a consistency check, linear wave theory is used to calculate the reson ant wave modes supported by the background profiles of wind and potential t emperature. These are the wave modes that are most likely responsible for t he lee-wave clouds seen in the AVHRR satellite image. The horizontal wavele ngth for the resonant wave mode is 8188 m, in agreement with the estimates based on the AVHRR satellite image and from the horizontal projection metho d. The paper highlights the dangers of interpreting radiosonde soundings as vertical "snapshots" of the atmosphere and assuming that the vertical wave number can be characterized by a single value. It is shown that treating th e radiosonde sounding as a vertical profile and analysing wave-like structu res in the ascent rate using standard spectral methods gives an incorrect v ertical wavelength of 1906 +/- 238 m. The paper concludes that operational radiosonde soundings can be used to make accurate, quantitative measurement s of orographically generated waves, provided that the vertical variation i n the background flow and the motion of the radiosonde through the wave fie ld are taken into account.