link to homepage

Institute of Energy and Climate Research

Navigation and service

Brewer-Dobson Circulation

The Brewer-Dobson circulation (BDC) is an important factor for Earth’s climate as it determines the lifetimes of key greenhouse gases as well as the amount of water vapor, ozone [1] and aerosol, for example, in the lower stratosphere. Variability and trends in the BDC modify the amount and distributions of these radiatively active substances with implications for radiative forcing and surface temperatures. The BDC also impacts surface climate and extreme weather events through downward coupling , involving variability of the stratospheric polar vortex. Strength and variability of the BDC is often diagnosed by the age of the air masses, i.e. the time that has passed since an air mass had entered the stratosphere [2].
Despite the importance of the BDC for global and regional climate, understanding its decadal variability and long-term trend is challenging. BDC and its variability therefore represents a focus of current research. Climate models predict that increasing greenhouse gas levels will cause an acceleration of the stratospheric circulation . However, these results have been challenged by observations showing an increase of stratospheric age of air which has been interpreted as a slowing down of the circulation . In particular, the decadal mean age trend between 2002 and 2012 from MIPAS satellite observations shows a complex hemispheric pattern, with increasing age in the northern hemisphere and decreasing age in the southern hemisphere [2]. This pattern could not be simulated by climate models hitherto, leaving the atmospheric community with the puzzling situation, of deciding whether the stratospheric BDC is changing or not. Due its central significance, a reliable model representation of the BDC and of the water vapor distribution in the lower stratosphere is indispensable for reliable climate projections. Consequently, the inconsistency between climate models and observations constitutes a major uncertainty for current climate simulations.
IEK-7 research aiming at understanding inter-annual to decadal variability and long-term trends of the BDC and stratospheric water vapor from a process-oriented perspective, as described in the following, will help to reduce this uncertainty of current climate projections.

Variability and trends of the Brewer-Dobson circulation and age of air (H): From a suite of recent studies based on meteorological reanalysis data, it was shown that the discrepancy between models and observations regarding mean age trends may be explained by the fact that decadal variability is misrepresented in free-running models. Long-term trends from different reanalysis datasets (ERA-Interim, JRA–55, MERRA) are consistent with climate model results showing a robust strengthening of the BDC [3] and a related decrease in the mean age of air throughout large parts of the stratosphere [4]. However, the inter-annual variability of mean age simulated with the Lagrangian transport model CLaMS driven by ERA-Interim reanalysis meteorology agrees well with the variability observed by the MIPAS satellite instrument, revealing a strong impact of the quasi-biennial oscillation on mean age variability throughout the global lower stratosphere. Moreover, even mean age trends simulated by CLaMS are consistent with existing observations. These reanalysis-based simulations show a weak long-term ageing trend in the northern hemisphere above an altitude of about 24 km, exactly within the region of the balloon measurements of Engel and co-workers. They also show a distinct hemispheric pattern of the decadal (2002 to 2012) trend with increasing age in the northern and decreasing age in the southern hemisphere [2].

By separating the effects of the residual mean mass circulation and of eddy mixing on mean age, using a continuity equation-based approach, it was further shown that mixing plays a crucial role in shaping the decadal mean age trends and, in particular, the hemispheric asymmetry of the trend between 2002 and 2012 [4]. Consequently, the inability of current climate models to simulate observed mean age and related BDC trends is likely related to a misrepresentation of decadal variability in the models and the related effects of mixing. A similar analysis provided evidence that the climatological hemispheric asymmetry of mean age, with an older lower stratosphere in the northern than in the southern hemisphere, is related to stronger eddy mixing in the northern hemisphere [5]. For an even more detailed transport analysis in the future, a novel diagnostic has been implemented in the CLaMS model allowing the calculation of time-dependent 3D age spectra throughout the stratosphere in a transient simulation [6]. The age spectrum enables the full transit time distribution of the air to be considered, thereby providing new insight into transport pathways, and has been implemented in the flight-planning methodology and analysis of recent aircraft measurement campaigns (e.g. for the StratoClim and WISE campaigns in 2017, both coordinated by IEK-7).

Impact of BDC variability on stratospheric water vapor (H): The variability of water vapor transport deep into the stratosphere is closely related to the variability in the strength of the BDC modulating the stratospheric water vapor entry values [7]. The BDC itself is influenced by sudden stratospheric warmings (SSW). In a case study of a remarkable major SSW during boreal winter 2008/09, we investigated how transport and mixing triggered by this event affect the composition of the entire stratosphere in the northern hemisphere [8]. It was found that the major SSW event, at the same time, accelerated polar descent and tropical ascent of the Brewer-Dobson circulation, thereby influencing stratospheric entry values of water vapor. A systematic investigation of all major SSWs in the last 35 years [9] shows that this effect is strongly modulated by the phase of the quasi-biennial oscillation (QBO). Enhanced dehydration due to the major SSWs combined with a higher frequency of major SSWs after the year 2000 may also have contributed to the decrease in lower stratospheric water vapor observed after 2000.
The El Nino−Southern Oscillation (ENSO) is another important driver of water vapor variability. Not only ENSO-related temperature anomalies are confined to the tropical Pacific (180 - 300°E), but also anomalous wave propagation and breaking [10]. Thus, during El Nino a more zonally symmetric wave forcing drives a deep branch of the Brewer-Dobson circulation. During La Nina this forcing increases at lower levels (around 390 K) over the tropical Pacific, likely influencing the shallow branch of the BD circulation. In agreement with previous studies, wet (dry) and young (old) tape-recorder anomalies propagate upwards in the months following El Nino (La Nina).

Water vapor time series in the tropics (10°S–10°N) at 400 KWater vapor time series in the tropics (10°S–10°N) at 400 K from CLaMS simulation (black), HALOE (green) and MLS (red) satellite observations. QBO easterly phases are highlighted by grey shading, SSWs during QBO easterly phase by red and during QBO westerly phase by blue lines. Figure adapted from [9].


  1. Abalos, M. et al. 2013, Atmos. Chem. Phys. 13: 10787-10794, doi:10.5194/acp-13-10787-2013.
  2. Plöger, F. et al. 2015, J. Geophys. Res.-A. 120: 716-733, doi:10.1002/2014JD022468.
  3. Abalos, M. et al. 2015, J. Geophys. Res.-A. 120: 7534-7554, doi:10.1002/2015JD023182.
  4. Plöger, F. et al. 2015, Geophys. Res. Lett. 42: 2047-2054, doi:10.1002/2014GL062927.
  5. Konopka, P. et al. 2015, J. Geophys. Res.-A. 120: 2053-2066, doi:10.1002/2014JD022429.
  6. Ploeger, F. and Birner, T. 2016, Atmos. Chem. Phys. 16: (15), 10195-10213, doi:10.5194/acp-16-10195-2016.
  7. Fueglistaler, S. et al. 2014, J. Geophys. Res.-A. 119: 1962-1972, doi:10.1002/2013JD020772.
  8. Tao, M. et al. 2015, Atmos. Chem. Phys. 15: 8695-8715, doi:10.5194/acp-15-8695-2015.
  9. Tao, M. et al. 2015, Geophys. Res. Lett. 42: 4599-4607, doi:10.1002/2015GL064443.
  10. Konopka, P. et al. 2016, J. Geophys. Res.-A. 121: (19), 11486-11501, doi:10.1002/2015JD024698.