1. Tropospheric ozone: past, present (and future)
David Stevenson
Institute of Atmospheric and Environmental Sciences,
University of Edinburgh Thanks to all the ACCENT Photocomp & Royal Society modellers

2. Motivations
IPCC (2007): Tropospheric O3 is the third largest greenhouse gas contributor to radiative forcing of climate change: Wm-2 (CO2: 1.66 Wm-2; CH4: 0.48 Wm-2)

3. Radiative forcing from tropospheric O3
Forster et al. (2007)
IPCC-AR4
WG1 Chapter 2

4. Radiative forcing based on emissions not concentrations
Ozone is a secondary pollutant – emissions of: CH4 CO
NOx
NMVOC have driven up its concentration
Forster et al. (2007)
IPCC-AR4
WG1 Chapter 2

5. Motivations
IPCC (2007): Tropospheric O3 is the third largest greenhouse gas contributor to radiative forcing of climate change: Wm-2 (CO2: 1.66 Wm-2; CH4: 0.48 Wm-2)
Ground level O3 is also a serious air pollutant (it is a reactive oxidant), affecting human health, & damaging crops & natural vegetation.

6. The biosphere-atmosphere boundary
Ozone enters a plant via stomata;
attacks plant cells
Ozone crosses the fluid lining of the lungs, and stimulates a variety of responses at the cell level
Mudway and Kelly (2000)

7. Ozone reduces the lung function of healthy children
Lung function reduces at higher O3
Lung function improves at higher O3
New Jersey, Berry 1991
Ontario, Kinney 1996
New Jersey, Lioy 1985
California southern, Kinney 1996
San Bernadino, Higgins 1990
New Jersey, Kinney 1996
Indiana state, Lippmann 1983
Fairview Lake, Spektor 1988
Fairview Lake, Spektor 1991
pooled FE est.
pooled RE est.
-.015
-.01
-.005
.005
.01
Change for 10μg/m3 increase in O3
Ozone reduces the lung function of healthy children
Courtesy of Ross Anderson

8. High levels of ozone increase your chances
of dying
Meta-analysis based on 98 US cities
Bell et al. (2006, Environmental Health Perspectives)

9. Motivations
IPCC (2007): Tropospheric O3 is the third largest greenhouse gas contributor to radiative forcing of climate change: Wm-2 (CO2: 1.66 Wm-2; CH4: 0.48 Wm-2)
Ground level O3 is a serious air pollutant (it is a reactive oxidant), affecting human health and damaging crops and natural vegetation.
Or is it even more important for climate?

10. AspenFACE: Exposure of tree stands to elevated CO2 and O3
Control (360 ppm CO2) (x = 36 ppb O3)
+CO2 (545 ppm)
+O (x = 52 ppb)
+25% Biomass
-23% Biomass
LAI  Leaf duration  Leaf size  Ps  Water stress
LAI  Leaf duration  Leaf size  Ps  Pests  Water stress  Antioxidant gene expression
NPP estimates for aspen from King et al New Phytol. 165:
Components of Aspen Productivity (NPP)

11. Indirect and direct radiative forcings from tropospheric ozone
Symbols are direct forcings (IPCC, 2001)
Blue and red curves are indirect ozone forcing, due to ozone impacts on vegetation
(high ozone sensitivity)
(low ozone sensitivity)
Suggests that the indirect forcing may be similar in magnitude to the direct forcing.
Sitch et al. (Nature, 2007)

12. Radiative forcing from tropospheric O3
If we believe this O3 indirect effect, then
Tropospheric O3 approaches CO2 as the No.1 GHG!
Forster et al. (2007)
IPCC-AR4
WG1 Chapter 2

13. How has ozone changed since 1750?
No ice-core data (O3 is too reactive)
Very sparse/poor quality observations
I.e. some evidence that P-I surface O3 in Europe was <10 ppb

14. Observed trends in surface O3 since the 1970s at various relatively remote sites
Oltmans et al., 2006

15. NH mid-lats, mid-troposphere
Ozonesonde observations since the 1970s
Large interannual variability
Regionally different trends; regionally different AQ measures
Logan et al., 1999; O3 sonde data
NH mid-lats, mid-troposphere

16. Summary of observed ozone
A very few measurements before 1900 suggest surface O3 in Europe was <10 ppb before industrialisation
Before 1970s, a few observations show increases in surface O3
Since ~1970, surface/sonde monitoring networks have expanded
Most sites show increases in ozone, some show strong increases, but significant levels of variability (time and space)
Models needed to produce a global picture

17. Modelling tropospheric ozone
Dynamical core: GCM or NWP analyses
Stratosphere-troposphere exchange
Tracer transport
Convection
BL-free troposphere exchange
Chemical mechanisms
Reducing complex schemes
Photolysis rates (clouds/aerosols)
Surface exchange
Biosphere: emissions/deposition (stomatal uptake)

18. What can models tell us?
Give a global view of the spatial and temporal distribution of ozone and its precursors (more detail than observations alone)
Allow us to diagnose when and where ozone chemical production and destruction is taking place
If we have faith in the models, we can use them for hindcasts/forecasts, and sensitivity experiments (e.g., what happens to ozone if emissions and/or climate change?)

19. Ensemble mean of 26 models
Year 2000
Ensemble mean of 26 models
Annual Zonal Mean
Annual Tropospheric Column
ACCENT Photocomp: Stevenson et al., 2006, JGR

20. Comparison of ensemble mean model with O3 sonde measurements
ACCENT Photocomp: Stevenson et al., 2006, JGR
Individual models in
grey
UT 250 hPa
Model ±1SD
Observed ±1SD
J F M A M J J A S O N D MT
500
hPa LT
750
hPa
90-30°S
30°S-Eq
30°N-Eq
90-30°N

21. standard deviation (%) Annual Zonal Mean
Year 2000
Inter-model
standard deviation (%)
Annual Zonal Mean
Annual Tropospheric Column
Models show large variations in the crucial tropical UT region
ACCENT Photocomp: Stevenson et al., 2006, JGR

22. Seasonal variation of surface ozone
Ensemble mean of 26 ACCENT Photocomp models

23. Broadly in agreement with observations
Remote NH: Spring peak
Polluted NH: summer peak
Tropics: biomass burning (dry) season
Remote SH: winter peak
Broadly in agreement with observations

25. Multi-model ensemble mean ozone P, L, NCP
Ship NOx
 = 0.997
Surface
ACCENT Photocomp

26. Multi-model ensemble mean ozone P, L, NCP
 = 0.975
ACCENT Photocomp

27. Multi-model ensemble mean ozone P, L, NCP
 = 0.930
ACCENT Photocomp

28. Multi-model ensemble mean ozone P, L, NCP
 = 0.870
ACCENT Photocomp

29. Multi-model ensemble mean ozone P, L, NCP
Mid-trop net destruction
 = 0.792
ACCENT Photocomp

30. Multi-model ensemble mean ozone P, L, NCP
Mid-trop net destruction
 = 0.700
ACCENT Photocomp

31. Multi-model ensemble mean ozone P, L, NCP
Mid-trop net destruction
 = 0.600
ACCENT Photocomp

32. Multi-model ensemble mean ozone P, L, NCP
 = 0.505
ACCENT Photocomp

33. Multi-model ensemble mean ozone P, L, NCP
 = 0.422
ACCENT Photocomp

34. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.355
ACCENT Photocomp

35. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.300
ACCENT Photocomp

36. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.250
ACCENT Photocomp

37. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.200
ACCENT Photocomp

38. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.150
ACCENT Photocomp

39. Multi-model ensemble mean ozone P, L, NCP
Upper-trop net production lightning
 = 0.099
ACCENT Photocomp

40. Assumes no change in biomass burning or soil NOx between P-I and present

41. O3 radiative forcing since 1750
Model tuned to get low ‘observed’ PI ozone, by reducing soil and lightning NOx, and increasing biogenic VOC
‘Preindustrial’ conditions:
Typically anthropogenic emissions set to zero, biomass burning to ~10% present-day
W m-2
-0.1
+0.3
Lower panel all
ACCENT models:
Gauss et al., 2006, ACP
Forster et al. (2007)
IPCC-AR4
WG1 Chapter 2

42. Conclusions (past & present O3)
The direct radiative forcing from tropospheric ozone is W m-2 (range to W m-2)
An indirect effect of O3, via reduced growth of vegetation, may add a further 0.2 to 0.4 W m-2, suggesting O3 may approach CO2 in terms of radiative forcing
O3 affects human (as well as plant) health, and legislation exists in most countries to limit emissions of O3 precursors
Ozone precursor emissions from ships and aircraft are not currently regulated, and are growing fast
Strong legislation to reduce all emissions will bring important benefits for both air quality and climate, and may be an important short- to medium-term tool to reduce radiative forcing, especially if O3 does have a larger forcing than currently thought
Models appear to realistically simulate present-day ozone, but a more stringent test will be for them to reproduce longer term regional trends

44. Ozone in the future…
Will depend strongly on the trajectory of anthropogenic emissions, in particular NOx, but also CH4, CO and VOCs.
IPCC SRES probably too pessimistic; new projections from IIASA expect air quality legislation to significantly reduce NOx emissions by 2050
Climate change is likely to impact ozone

45. Under current legislation, NOx emissions should reduce in most places:
SRES B2
IIASA ‘current legislation’
Courtesy Markus Amann, IIASA

46. Ship and aircraft emissions
Both have essentially no legislation to regulate them
Likely to keep increasing if nothing is done
Ships a particularly large NOx source
In the 2050 scenario used here, optimistically assumed that ship emissions will be controlled…

47. NOx emissions
Ships x 0.5
W. Europe x 0.41
N. America x 0.23
China x 1.03
Reduce almost everywhere – global total down ~40%

48. Methane emissions
Courtesy Markus Amann, IIASA

49. Tropospheric O3 responds approximately linearly to anthropogenic CH4 emission changes across models
MOZART-2 [West et al., PNAS 2006; Fiore et al., in prep]
TM3 [Dentener et al., ACP, 2005]
GISS [Shindell et al., GRL, 2005]
GEOS-CHEM [Fiore et al., GRL, 2002]
IPCC TAR [Prather et al., 2001]
Courtesy of Arlene Fiore
Anthropogenic CH4 contributes ~50 Tg (~15%) to tropospheric O3 burden
~5 ppbv to surface O3