Mechanisms coupling stratosphere and troposphere in observations and
Mechanisms coupling stratosphere and troposphere in observations and modelling results B. Hassler, W. Steinbrecht, P. Winkler, Met. Obs. Hohenpeissenberg; M. Dameris, C. Schnadt, S. Matthes, DLR, Oberpfaffenhofen; C. Brühl, B. Steil, MPI Mainz; M. Giorgetta, MPI Hamburg Abstract total ozone fluctuations related to different influences Kopplung von Dynamik und Atmosphärischer Chemie in der Stratosphäre Global Maps Temperature, Wind, NCEP Reanalysis, (1958 -) (http: //wesley. wwb. noaa. gov/cdas_data. html) KODYACS Partners DLR-Oberpfaffenhofen, MPIs Mainz + Hamburg ECHAM-DLR (ECHAM 4. L 39(DLR)/CHEM): 1000 to 10 h. Pa = 0 to 30 km ECHAM-MPI (MA-ECHAM/CHEM): 1000 to 0. 1 h. Pa = 0 to 80 km State-of-the-art fully coupled chemistry-climate models; chemical species feed back into radiation scheme; transient runs (1960 to 1999), changing source gases (CFC, CO 2, CH 4, . . . ) , Solar Cycle, volcanic aerosol, nudged QBO Station data (Hohenpeißenberg) mod_data. public. html) Models Observations at single points: Ozonesondes 1967 -2002, Hohenpeißenberg, Canada, USA, Japan (http: //www. woudc. org) Lidars, (1987 -2002), Hohenpeißenberg, Haute Provence, Table Mountain, Hawaii, Lauder, (Tokio) (http: //www. ndsc. ncep. noaa. gov) no data local ozone fluctuations related to different influences Total Ozone, TOMS/SBUV, 11/1978 -12/2002 (http: //code 916. gsfc. nasa. gov/Data_services/merged/ Regression allows to separate individual influences. • determine j, f, q, e. . . by least squares • use only significant predictors (> 90%) regression Global Observations: Proxy time series (predictors) stand for individual influences Linear Regression fits a sum of individual influences (predictors) ΔO 3/ΔT = j*lin_trend + f*Sol_Cyc + q*QBO + e*ENSO + t*tropop. _height + a*aerosol + to recreate the observed ozone w*wind_60°N + rest (Eq 1) fluctuations. nature Ozone and temperature show substantial variations on many time-scales. We are interested in interannual, decadal, and trend-like variations. The purpose of this study is to • attribute and quantify ozone and temperature variations due to different influences (natural and anthropogenic). • compare long-term observations with state-of-the-art chemistry-climate model simulations. • identify coupling mechanisms between stratosphere and troposphere. Transient simulations with the 0 to 80 km altitude ECHAM-MPI and the 0 to 30 km altitude ECHAM-DLR models show very realistic influences of QBO, solar cycle, increasing (anthropogenic) source gases, etc. , both in the global distribution and as a function of altitude and season. Generally the 0 -80 km model gives better results. Tropospheric meteorological conditions (e. g. T at 400 h. Pa) have a large influence on lower stratospheric ozone and temperature. Although not shown here, solar cycle, QBO and polar vortex strength are found to have significant regional effects on the troposphere in both observations and models. Various natural (and anthropogenic) influences combine to create ozone fluctuations. no data no data Above: Global maps of contributions to total ozone fluctuations. Contribution size is defined as 2 of the corresponding time series term in Eq. 1. Left column: observations (1979 -2002). Middle : ECHAM-MPI transient run (1978 -1999). Right: ECHAM-DLR transient run (1978 -1999). Blue and green colours indicate negative correlation between predictor and total ozone, red and yellow colours indicate positive correlation. White regions indicate that the influence is not statistically significant at the 90% confidence level. Only data for December, January and February (DJF) are shown. • models reproduce observed size and geographical variation of ozone fluctuations very well • zonal symmetry at low latitudes, localized features at high latitudes (“Aleutian Anticyclone”) • differences between observations and both models for localized high latitude features • results for total ozone and 50 h. Pa temperature (not shown) are very similar • ECHAM-DLR shows weaker QBO signal in ozone, but stronger QBO signal in temperature (at 50 h. Pa, not shown) References: Austin, J. et al. (2003): Uncertainties and assessments of chemistry-climate models of the stratosphere. Atmos. Chem. Phys. , 3, 1 -27. Giorgetta, M. A. et al. (2002): Forcing of the quasi-biennial oscillation from a broad spectrum of atmospheric waves. Geophys. Res. Lett. , 29, 10. 1029/2002 GL 014756. Kistler, R. et al. (2001): The NCEP-NCAR 50 -Year Reanalysis: Monthly Means. Bull. Am. Met. Soc. , 82, 247 -267. Steinbrecht, W. et al. (2003): Global distribution of total ozone and lower stratospheric temperature variations. Atmos. Chem. Phys. , 3, 1421 -1438. Steinbrecht, W. et al. (2004): Enhanced upper stratospheric ozone: Sign of recovery or solar cycle effect? J. Geophys. Res. , 109, doi: 10. 1029/2003 JD 004284. Above: Size of local ozone fluctuations related to different influences as a function of season and altitude, for Hohenpeißenberg. Influence size is defined as 2 of the corresponding time series term in Eq. 1. Left column: observations. Middle: ECHAM-MPI transient run. Right: ECHAM-DLR transient run. Blue and green colours indicate negative correlation between predictor and total ozone, red and yellow colours indicate positive correlation. White regions indicate that the influence is not statistically significant at the 90% confidence level. Below 30 km observations are from the period 1967 to 2003, model data are from 1960 to 1999. Above 30 km only data after 1987 are available/used. • general agreement between station data and models, upper stratosphere missing in ECHAM-DLR • good agreement for T(400 h. Pa), trend, solar-cycle, QBO • most influences show seasonal variations • weak QBO in ozone for ECHAM-DLR (but strong QBO in temperature, not shown) • regression works less well (small R 2) than for global data: small scale fluctuations not accounted for, data sampling problems, silent layers (no variance - no regression), midlatitude stations in areas of poor regression Contact: Birgit Hassler, birgit. hassler@dwd. de; Wolfgang Steinbrecht, wolfgang. steinbrecht@dwd. de
- Slides: 1