"Changes in stratospheric temperatures and their implications for changes in the Brewer-Dobson circulation, 1979-2005", by Young, P. J., K. H. Rosenlof, S. Solomon, S. C. Sherwood, Q. Fu, and J.-F. Lamarque has been accepted and can be found on the Journal of Climate in press page.
From analyzing at the "annual cycle" of temperature trends from the satellite-mounted MSU and SSU instruments, and IUK radiosonde data, we found evidence of a significant strengthening of the Northern Hemisphere (NH) branch of the Brewer-Dobson circulation (BDC) in December, extending throughout the depth of the stratosphere. Similarly, we found a significant strengthening of the Southern Hemisphere branch of the BDC during August, extending to the middle stratosphere, The results also suggested a weakening of the lower stratosphere NH branch of the BDC during March, although there is some suggestion that this is not indicative of a secular trend (see also Free, 2011).
Models predict an increase in the strength of the BDC with increased greenhouse gas concentrations (e.g. Butchart et al., 2010), so any significant change could be a fingerprint of climate change.
The results are consistent with the MSU temperature/BDC studies of Fu et al. (2010) and Lin et al. (2010), but extend the analysis to the middle and upper stratosphere through the use of the SSU data.
Application to Chemistry-climate models (CCMVal2)?
In the paper we defined a "BDC index", which tracked the strength in the BDC by subtracting tropical temperatures from extratropical temperatures (see Yulaeva et al. (1994) and Young et al. (2011) for more on the tropical/high-latitude "see-saw" in temperatures). As a part of the response to reviewers (but not in the paper) we looked at how trends in the our BDC index correlated with modeled tropical vertical velocity trends, which are a more direct measure of BDC strength. The above figure shows that trends in the BDC index correlate quite closely with 21st century tropical (20°S/N and 30°S/N) vertical velocity trends using data from three chemistry-climate models (CCMs) from the recent CCMVal-2 exercise. Although not perfect for all altitudes/models, it perhaps illustrates the utility of our index for monitoring changes in the strength of the BDC.