2020 vision: Modelling the near future tropospheric composition

1
2020 visionModelling the near future
tropospheric composition

• David Stevenson
• Institute of Atmospheric and Environmental
  ScienceSchool of GeoSciencesThe University of
  Edinburgh
• Thanks to
• Ruth Doherty (Univ. Edinburgh)
• Dick Derwent (rdscientific)
• Mike Sanderson, Colin Johnson, Bill Collins (Met
  Office)
• Frank Dentener (JRC Ispra), Markus Amann (IIASA)

2
Talk Structure

• Chemistry-climate model STOCHEM-UM
• Several transient runs 1990 ? 2030
• the 1990s
• how modellers use observations
• comparisons with ozone-sonde data
• the 2020s
• what is needed to predict the future?
• a believable model (hence the first bit)
• a computationally efficient model
• future emissions
• climate change
• other things?
• What are the results telling us?

3
STOCHEM

• Global Lagrangian 3-D chemistry-climate model
• Meteorology HadAM3 prescribed SSTs
• GCM grid 3.75 x 2.5 x 19 levels
• CTM 50,000 air parcels, 1 hour timestep
• CTM output 5 x 5 x 9 levels
• Detailed tropospheric chemistry
• CH4-CO-NOx-hydrocarbons
• detailed oxidant photochemistry
• Interactive lightning NOx, C5H8 from veg.
• 1 year/day on 36 processors (Cray T3E)

4
Model experiments

• Several transient runs 1990 ? 2030
• Driving meteorology
• Fixed SSTs (mean of 1978-1996)
• SSTs from a climate change scenario (is92a)
• shows 1K surface warming 1990s-2020s
• Shorter run with observed SSTs 1990-2002
• New IIASA global emissions scenarios
• Business as usual (BAU)
• Maximum reductions feasible (MRF)
• Stratospheric O3 is a fixed climatology
• Vegetation (land-use) also a fixed climatology

IIASA International Institute for Applied
Systems Analysis (Austria)
5
IIASA Emissions scenarios
Global totals there are significant regional
variations
Courtesy of Markus Amann (IIASA) Frank Dentener
(JRC)
6
Model experiments
BAU, observed SSTs 1990-2002
BAU, fixed SSTs 1990-2030
MRF, fixed SSTs 1990-2030
BAU, is92a SSTs 1990-2030
2030
1990
7
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Ozone-sonde data from Logan et al. (1999 - JGR)
8
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Ozone-sonde data from Logan et al. (1999 - JGR)
9
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Modelgt obs
Modellt obs
Ozone-sonde data from Logan et al. (1999 - JGR)
10
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Ozone-sonde data from Logan et al. (1999 - JGR)
11
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Ozone-sonde data from Logan et al. (1999 - JGR)
12
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Model overestimatesby gt1 std dev
Model underestimatesby gt1 std dev
Ozone-sonde data from Logan et al. (1999 - JGR)
13
Hohenpeissenberg Ozone-sonde model vs
observations (monthly data for the 1990s)
Model overestimatesby gt1 std dev
Identify where andwhen the model iswrong
Model underestimatesby gt1 std dev
Ozone-sonde data from Logan et al. (1999 - JGR)
14
Ny Alesund (79N, 12E), Spitzbergen
Ozone-sonde data from Logan et al. (1999 - JGR)
15
Resolute (75N, 95W), Canada
Ozone-sonde data from Logan et al. (1999 - JGR)
16
Sapporo (43N, 141E), Japan
Ozone-sonde data from Logan et al. (1999 - JGR)
17
Wallops Island (38N, 76W), Eastern USA
Ozone-sonde data from Logan et al. (1999 - JGR)
18
Ascension (8S, 14W), Mid-Atlantic
Ozone-sonde data from Thompson et al. (2003 - JGR)
19

• The model has some skill at simulating
  tropospheric ozone, but is far from perfect.
• Careful comparisons with other gases (NOx, NOy,
  etc.) also needed, but there is much less data.
• For climate-chemistry model validation, lengthy
  climatologies, including vertical profiles are
  most useful.
• If you want modellers to uses the data, provide
  it in easy-to-use formats (were lazy!)
• MOZAIC (operational aircraft data) and satellite
  data are examples of the sort of datasets needed.
• If you trust the model, it may be useful for
  future predictions

20
Model experiments
BAU, observed SSTs 1990-2002
BAU, fixed SSTs 1990-2030
MRF, fixed SSTs 1990-2030
BAU, is92a SSTs 1990-2030
2030
1990
21
(No Transcript)
22
BAU 2020s
23
A large fraction is due to ship NOx
Change in surface O3, BAU 2020s-1990s
BAU
24
MRF 2020s
25
Change in surface O3, MRF 2020s-1990s
MRF
BAU
26
BAUclimate change 2020s
27
Change in surface O3, BAUcc 2020s-1990s
MRF
BAU
BAUcc
28
?O3 from climate change
Warmertemperatures higher humidities increase
O3 destruction over the oceans
But also a role from increases in isoprene
emissions from vegetation?
29
Zonal mean O3 ?O3 (2020s-1990s)
1990s
BAU ?O3
MRF ?O3
BAUcc ?O3
30
Zonal mean OH ?OH(2020s-1990s)
1990s
BAU ?OH
MRF ?OH
BAUcc ?OH
31
CH4, ?CH4 OH trajectories 1990-2030
Current CH4 trend looks like MRF coincidence?
All scenarios show increasing OH
32
Radiative forcings from ?O3 and ?CH4 (2020s-1990s)
BAU MRF BAUcc
?O3 0.04 -0.07 0.01
?CH4 0.14 0 ?
0.18 -0.07 ?
33
Conclusions

• Model development and validation is ongoing, is
  guided by observations
• Anthropogenic emissions will be the main
  determinant of future tropospheric O3
• Ship NOx looks important
• Climate change will introduce feedbacks that
  modify air quality
• We can estimate the radiative forcing
  implications of air quality control measures
• NB Many processes still missing