Report 35:  Structure of the Tropical Lower Stratosphere as Revealed by Three Reanalysis Data Sets

Pawson, Steven and Michael Fiorino
May 1996, 35 pp. 

---

While the skill of climate simulation models has advanced over the last decade, mainly through improvements in modeling, further progress will depend on the availability and the quality of comprehensive validation data sets covering long time periods. A new source of such validation data is atmospheric "reanalysis" where a fixed, state-of-the-art global atmospheric model/data assimilation system is run through archived and recovered observations to produce a consistent set of atmospheric analyses. Although reanalysis will be free of non-physical variability caused by changes in the models and/or the assimilation procedure, it is necessary to assess its quality. That is, how good is reanalysis and how close should our climate models agree with these data.

A region for stringent testing of the quality of reanalysis is the tropical lower stratosphere. This portion of the atmosphere is sparse in observations but displays the prominent quasi-biennial oscillation (QBO) and an annual cycle, neither of which is fully understood, but which are likely coupled dynamically. We first consider the performance of three reanalyses, from NCEP/NCAR, NASA and ECMWF, against rawinsonde data in depicting the QBO and then examine the structure of the tropical lower stratosphere in NCEP and ECMWF data sets in detail.

While the annual cycle and the QBO in wind and temperature are quite successfully represented, the mean meridional circulations in NCEP and ECMWF data sets contain unusual features which may be due to the assimilation process rather than being physically based. Further, the models capture the long-term temperature fluctuations associated with volcanic eruptions, even though the physical mechanisms are not included, thus implying that the model does not mask prominent stratospheric signals in the observational data. We conclude that reanalysis offers a unique opportunity to better understand the dynamics of QBO and can be applied to climate model validation. (pdf file)

UCRL-MI-123395

---