1. A modelling study on trends and variability of the tropospheric chemical composition over the last 40 years S.Rast(1), M.G.Schultz(2) (1) Max Planck Institute for Meteorology, Bundesstr. 53, D-20146 Hamburg, Germany (2) ICG-II, Forschungszentrum Jülich, D-52425 Jülich, Germany

2. The RETRO project Primary objective: understand, detect and assess long- term changes and interannual variability of the tropospheric chemical composition over the last 40 years (1960-2000).

3. The Global Circulation Chemistry Transport Model ECHAM5-MOZ Based on global circulation model ECHAM5 Tropospheric chemistry comprises 63 transported species and 168 reactions (chemistry of MOZART2.4) ERA40 analysis is used for driving the atmosphere dynamics Spectral resolution: T42L31 (about 2.8ºx2.8º, upper limit 10 hPa) Tracers undergo: advection, vertical diffusion, dry deposition, chemical reactions, wetdeposition, emissions

4. RETRO Emissions I - Anthropogenic CO Emissions Anthropogenic emissions Tg(CO)/yr: 320 (1960), 477 (2000), 49% increase Wildfires Tg(CO)/yr: 330 ± 86

5. RETRO Emission II - Anthropogenic NOx Emissions Anthropogenic NOx emissions (Tg(NO 2 )/yr): 37.4 (1960), 90.2 (2000), 141% increase Wildfires (Tg(NO2)/yr): 14.6 (1960), 12.7(2000)

6. RETRO Emissions III - Other Anthropogenic Emissions Tg(species)/yralcoholsethanebutane 19618.74.85.4 200012.46.98.9 Anthropogenic emissions increase CO and NOx emissions are quantitatively the most important anthropogenic emissions

7. RETRO Emissions - Other Anthropogenic Emissions Anthropogenic emissions increase CO and NOx emissions are quantitatively the most important anthropogenic emissions Species19602000 NOx (Tg(NO 2 )/year) 37.490.2 CO320477 Butane5.48.9 Alcohols8.712.4

8. Natural Emissions from Vegetation (MEGAN) I f(T,  ) IsopreneTerpene f(T) Large interannual variability in the tropics (Latin America) High values of emissions in warm years (1998)

9. Lightning Emissions Lightning emissions do not show a tendency for the GCM-chemistry models MOZECH and LMDz Strong increase in lightning emissions for CTM TM4

10. ECHAM5 – water budget ERA40

11. Lightning Emissions - Correlation Correlation is generally low: R=0.1 (MOZECH-LMDZ) R=0.2 (MOZECH-TM4)

12. Mass of troposphere Mass of troposphere using the meteorological tropopause is slightly increasing by +0.7% (global warming leads to higher tropopause) Mass of troposphere using the chemical tropopause (150ppbv O 3 ) is slightly decreasing by -0.5% (ozone concentration is in general increasing) Chemical tropopause is below the meteorological tropopause in contrast to literature (Wu et al.)

13. Mass weighted OH concentration in troposphere OH is mainly located in lower and middle troposphere  no difference between chemical and meteorological tropopause Mean concentration (12.9±0.1)  10 5 molec/cm 3 (Spivakovsky et al. (2000): 11.6  10 5 molec/cm 3, Poisson et al. (2000): 12.4  10 5 molec/cm 3 ) No longterm tendency

14. Lifetime of Methane Lifetime of methane using model methane concentration and meteorological tropopause Total lifetime includes a constant soil sink of 30Tg/yr and stratospheric sink of 40Tg/yr. Trend in total lifetime:  0 Stevenson found a total lifetime of (8.67±1.32) yr (2006).

15. The Ozone budget Ozone budget is calculated by P+L+S+D=Δ P (  0): Ozone production rate L (  0): Ozone loss rate S (  0): Inferred stratospheric influx D (  0): Dry deposition rate Δ: Rate of change in ozone burden Δ  (1.2 to 1.6) Tg/yr  can be neglected S PL D

16. Ozone burden Ozone burden: about 10% difference between chemical and meteorological tropopause Correlation R=0.987 Warm years have extra hight ozone burden (1998), increase about 20% over 41 years Δ=1.6Tg/yr Δ=1.2Tg/yr Ozone Ozone from stratosphere Δ=0.9Tg/yr Δ=0.5Tg/yr chem. tropopause met. tropopause

17. Time series analysis Seasonal trend decomposition based on Loess time series analysis: Decompose time series in a trend, a seasonal component and a remainder Trend: long term trend over several years Seasonal component: Seasonal variation and its changes over the years „white noise“ remainder (R.B.Cleveland et al., 1990)

18. Time series decomposition for ozone production and loss Monthly O 3 production Tg/month Seasonal component Trend component Remainder (Negative) monthly O 3 loss Tg/month Seasonal component Trend component Remainder

19. Ozone production and loss Trend in ozone production: (26.3±6.4)Tg/yr 2 Trend in ozone loss: -(24.6±6.4)Tg/yr 2 Warm years (1998) show higher chemical activity No significant trend in seasonality

20. Time series of dry deposition of ozone Monthly O 3 dry deposition Tg/month Seasonal component Trend component Remainder Ozone dry deposition trend: (3.13±0.73)Tg/yr 2 Correlation with ozone burden: R=0.96

21. Inferred Ozone Influx from Stratosphere There is net ozone production in the upper troposphere between the chemical and meteorological tropopause

22. Summary of variability patterns 2000

23. Comparison with other models Short comparison with values given by Stevenson (2006). Mean values of 25 models which are representative for the year 2000 P/Tg/yrL/Tg/yrD/Tg/yrBurden/TgS/Tg/yr 5110±606-(4668 ±727)-(1003 ±200)344 ±39552 ±168 6114-5880-907400676 Stev. MOZ

24. Conclusion Natural emissions are rather constant (climate variability more important than climate change) Global mass weighted OH concentration has no trend Methane lifetime has a trend with respect to OH. Must be due to tempereature change and/or changes in OH distribution Ozone burden increase about 20% over 41 years, decoupled from OH?