7.2A How strong are the Southern Hemisphere storm tracks?

Wednesday, 10 June 2009: 2:10 PM
Pinnacle A (Stoweflake Resort and Confernce Center)
Yanjuan Guo, SUNY, Stony Brook, NY; and X. Xia and E. K. M. Chang

The true amplitude of the Southern Hemisphere storm tracks is still poorly known, as the NCEP-NCAR and ECMWF ERA40 reanalyses differ significantly in this measure. The difference is mainly due to the lack of conventional observations in the Southern Hemisphere as well as the different ways the two reanalyses assimilate satellite radiance observations.

In this work, the true amplitude is investigated using the recently available COSMIC (Constellation Observing System for Meteorology, Ionosphere, and Climate) radio occultation dataset. The COSMIC dataset has the advantage of having global coverage (about 2000 profiles per day) with similar biases and errors in both hemispheres. Furthermore, it is not assimilated in the reanalyses, thus it can be treated as an independent data source to validate the reanalyses. Daily geopotential height on a 2.5 x 2.5 deg lon-lat grids from Aug 2006 to Apr 2008 is obtained by analyzing COSMIC profiles using the Cressman analysis method, and its 24-hour difference filtered variance is computed to represent the strength of the baroclinc waves. In addition, eddy momentum flux can be computed by deriving wind components using the geostropic approximation, and it turns out the errors in the flux resulted from this approximation is not large in the mid-latitudes. The ratio of the amplitude of the variance averaged over the Southern Hemisphere to that in the Northern Hemisphere is regarded as a value close to the true inter-hemisphere ratio. Comparing the ratio based on the COSMIC dataset with that based on the reanalyses, and assuming that the reanalyses are reliable in the Northern Hemisphere, the bias of the reanalyses in the Southern Hemisphere can be inferred. In this way, it is found that the amplitude of the Southern Hemisphere storm tracks in the NCEP-NCAR reanalysis is significantly biased low by about 30% at 300 hPa. The amplitude in the ECMWF ERA40 is much closer to that inferred from COSMIC observations.

This work gives the first quantitative estimate of the true amplitude of the Southern Hemisphere storm tracks made independent of numerical model outputs. In addition to exploring the biases in the reanalyses, these results can also be applied to assess the performance of GCM simulations. For example, we are examining how realistic the IPCC AR4 models simulate the southern hemisphere storm track amplitude by comparing their inter-hemisphere ratios in eddy heat and momentum fluxes with the ratio derived based on COSMIC data.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner