Share This Paper. Figures from this paper. Citations Publications citing this paper. References Publications referenced by this paper. Emission of trace gases and aerosols from biomass burning Meinrat O. Andreae , Pierre Merlet. Earth's early atmosphere. James F Kasting.
Earth and cosmos Robert S. Evolution of the Atmosphere.
A production of amino acids under possible primitive earth conditions. Atmospheric homeostasis by and for the biosphere: the gaia hypothesis James E.
Middle Atmosphere Group
Lovelock , Lynn Margulis. It should be noted that future ozone data sets depend on the representative concentration pathway RCP only in the troposphere. The simulations RCP2. The total column ozone for the three latitude bands for the years — is presented in Figure 4. It shows clearly the chlorine related decrease of total ozone since about and a recovery after about A solar cycle dependence is clearly visible for the tropical and northern high latitudes.
At high southern latitudes the total ozone depends only weakly on solar variability. For the piControl simulation, constant monthly mean ozone fields averaged over the years — are used. In the stratosphere, background aerosols are ignored, and only sulfate aerosols resulting from volcanic eruptions are accounted for. This data set contains monthly mean zonal averages of the aerosol extinction, the single scattering albedo, and the asymmetry factor as a function of time, pressure, and wavelength i.
The data are given at 40 pressure levels and interpolated to the actual hybrid model layers during the simulations.
For piControl and the years after , zero stratospheric aerosols are assumed. The vertically integrated and globally averaged time series of monthly mean aerosol optical depth is presented in Figure 2. Differences between the two configurations are small except near to the upper lid and for the zonal wind in the tropics.
Compared to ERA40 data Figure 5 top row , simulated winds, temperature, and stream function show very similar structures. Tropopause temperatures are slightly underestimated in particular at high southern latitudes. There are large differences in variability between the LR and MR configurations in the tropics bottom two rows of Figure 6. This is caused by the QBO being internally generated in the version with high vertical resolution MR, see section 3. Furthermore, it should be noted that the observed standard deviation, in particular of ERA40 temperatures above 10 hPa, is strongly overestimated due to artificial jumps in the absolute temperatures introduced by the changes in the observation system.
Standard deviations at polar latitudes are smaller than in ERA Again, differences between MR and LR in mean quantities are small except for the equatorial region. The westerly stratospheric jet is slightly stronger in MR than in LR. Comparison to ERA40 shows agreement in many features. However, the QBO seems to also cause a slightly increased variance of the NH stratospheric polar night jet. Compared to ERA40, the simulated interannual DJF variability, in particular of the polar night jet, seems well captured by the model see also the following subsection.
This indicates different seasonalities of the simulated and real QBOs. Krismer et al. Seasonal variability of the quasi biennial oscillation, submitted to Journal of Advances in Modeling Earth Systems , , hereinafter referred to as Krismer et al. Charlton and Polvani [ ] and Charlton et al. They discarded such events that were not preceded by westerlies over at least 20 consecutive days and the final warmings that were identified when the jet did not return to westerlies for at least 10 consecutive days.
In this study we follow the approach of Charlton and Polvani [ ] but use a minimum of 10 days instead of 20 days of westerlies to separate two events. In order to allow for proper statistics we have counted major warmings in the full historical ensembles. In LR and MR, major warmings occur with very similar frequency of 7. The seasonal distribution from reanalysis data shows a convex distribution peaking in January, while the distribution simulated in the LR configuration is rather flat, and MR even shows a minimum in January Figure 8.
The simulated frequency was 5. However, this analysis was based on only 29 years, simulated with an uncoupled model version at T42 horizontal resolution with 39 vertical layers. One prerequisite is a relatively high vertical resolution of, in general, less than 1 km. In the AMIP simulations i. Also in the coupled historical simulations the amplitude is too strong. While at preindustrial times, the coupled model simulates a QBO of realistic 28 month period, the historical simulations show an increase of the length of the QBO period by about 3 months until To determine the origin of this trend in the QBO period is beyond the focus of this study.
It is, however, consistent with the simulated strengthening in equatorial stratospheric upwelling see following section that has also been identified by Kawatani et al. With the satellite era, global temperature records have become available. An analysis of data from satellites, radiosondes, and lidars by Randel et al. Numerical simulations have indicated that the cooling is related to both the increase in CO 2 and the depletion of ozone [ Langematz et al. Time series obtained with the MR model not shown looks very similar except for the higher variability from year to year and among ensemble members which is related to the QBO.
The future evolution depends strongly on the scenario. In the lower stratosphere at 47 hPa the temperature continues to decrease in all scenarios. At the end of the 21st century, however the trend levels off for those scenarios in which the CO 2 mixing ratio stays constant RCP4. Figure 2.
Middle Atmosphere Dynamics
At 1 and 9 hPa RCP2. In particular close to the stratopause, the radiative effect of ozone seems to play an important role as indicated by the slow temperature reduction after the decades of strong ozone depletion at the end of the 20th century. Temperature differences between the extreme low and high emission scenarios in reach about 13 K at 1 hPa.
As in all the following discussions of trends we will concentrate on the LR simulation because the MR responses are in general very similar except for a more complicated structure in the equatorial region. The historical trend comparison of the years — with the preindustrial state as well as all scenarios comparison of the years — with the years — show the expected distinct picture of tropospheric warming and middle atmospheric cooling.
The maximum of the cooling, however, is located at different altitudes.
In RCP2. In RCP4.