The effect of pyro-convective fires on the global troposphere:

comparison of TOMCAT modelled fields with observations from ICARTT

Sarah Monks

Outline:

• Brief introduction to the relevance of the work done

• Discuss the effect of forest fires on the troposphere

• Present some of the results from comparing the global chemical transport model to observations

Introduction• Over recent years it has become accepted that emissions from one

region can be easily transported to another

• Intercontinental transport can affect regional air quality

• The International Consortium for Atmospheric Research on Transport and Transformation (ICARTT) 2004 took place over the Atlantic Gain a better understanding of:

- Regional air quality

- Intercontinental transport of Ozone and it’s precursors

• O3 is formed photochemically from NMHC and CO in the presence of NOX.

• Ozone (O3) is an important trace gas: It is a greenhouse gas Its distribution affects the oxidising capacity of the troposphere It is harmful to living organisms

Biomass Burning in Alaska, 2004

• Extensive forest fires in Alaska in 2004 during ICARTT

• Forest fires emit a wide variety of reactive trace gases into the atmosphere including:- CO, hydrocarbons and NOX (=NO + NO2)

• In the summer of 2004 the Alaskan forest fire emissions dominated the distributions of trace species over the Atlantic

• Pyro-convection lifts the trace species up to the free troposphere – effect global troposphere

Projects Aims

• Test whether a global chemical transport model (CTM) can reproduce ICARTT in situ observations of trace species

The effect of model resolution The effect of convection and vertical diffusion Comparison of Eulerian and Lagrangian models (see poster) Test TOMCAT’s treatment of mixing and chemistry (see poster)

• Quantify the impact of the forest fires

TOMCAT MODEL• Eulerian 3D global chemical transport model (CTM)

• Forced by meteorological analyses from ECMWF

• 31 vertical levels (1000 hPa – 10 hPa)

• Variable horizontal resolution: used 5.6° x 5.6° and 2.8° x 2.8°

• Chemistry scheme: 41 chemical species, 118 chemical reactions

• Subgrid-scale processes: convection and vertical diffusion are parameterised

Observations

• ICARTT Aircraft measurements: - CO, O3, C2H6, C3H8

DC8

BAe-146

Falcon

• MOPITT satellite instrument:- CO observations

Additional Alaskan Forest fire emissions

Initial full chemistry simulations: Emission region: 202.5 - 219.37E

and

60.95 – 66.5N Mixing up to ~250hPa to simulate

pyro-convection Emitted at a constant rate from 15th

July 2004 to end of August

Sensitivity simulation: Emission region extended to: 202.5 – 239E and

61 – 67N Mixing reduced to ~350 hPa Increased surface fluxes to represent increased

emission rate

Trace Species

Mass flux in initial runs (Tg)

Mass flux in sensitivity

run (Tg)

CO 30 46.8

NOX 0.5 0.78

C2H6 0.27 0.42

C3H8 0.08 0.12

Impact of Forest Fires on CO• Control Run:

No forest fire emissions

• With Alaskan forest fires:Large increase in CO – greater than US

anthropogenic emissions33% increase in CO plume over the

Atlantic

• MOPITT CO total column:Plume of CO over Canada and

Atlantic observed in the same location

Comparisons to ICARTT DataMore detailed comparisons to in situ observations from ICARTT

•Model can reproduce CO vertical profiles on days when the aircraft did not target fire plumes

•Model is unable to capture the magnitude forest fire enhancements on the 18th, 20th and 23rd July

Extensive vertical mixing on the 23rd resulting in a nearly uniform concentration throughout the troposphere

Ethane and Propane

• Model cannot reproduce magnitude of enhancements in forest fire plumes

• The peak in propane is underestimated by a factor of 2.6 whereas ethane is only underestimated by a factor of 1.2 Suggests the 0.08 Tg of

propane emitted from Alaska is not sufficient

Emission Region

20th July 2004 at 500 hPa(with Alaskan forest fires):

• There is a plume to the North of the BAe-146 flight region with concentrations of up to 230 ppbv of CO and1800 pptv of ethane

• In July the forest fires moved more to the east

• The emissions will have been subjected to different meteorological conditions

• This caused the plume sampled by the aircraft on the 18th and 20th to be displaced in the model

Modelled Pyro-Convection20th July 2004 at 350 hPa(with Alaskan forest fires):

• Concentrations over Alaska are much too high (greater than 200 ppbv)

• Mixing scheme which is

simulating pyro-convection is

too strong

• This will lift too much CO up

and therefore TOMCAT will

underestimate CO at lower

levels

Switching off ConvectionNorth America•Over the US convection and vertical diffusion

are needed to prevent a build up of pollution in

the boundary layer

•Demonstrates the importance of including a

convection scheme in a model

Atlantic and Europe

•Greatly improves the gradients

of some of the profiles over the

Atlantic and Europe

Changes to Alaskan Forest Fires Emissions

Total CO column:

• 44% increase in CO in the plume

over the Atlantic

• CO is still underestimated by

13% in this region

CO concentration on the

20th July at 500 hPa:

• Extending the emission region

moves the plume closer to the

flight track

• CO in the plume increases from

230 ppbv to 245 ppbv

Forest Fires Impact on the O3 Burden

18th July at 400 hPa:

-31% increase in ozone

-Concentrations of 60-78 ppbv

20th July at 500 hPa:

-Ozone increases by 47 %

-concentrations of up to 78 ppbv.

The tracer fields show the displaced plumes. The concentrations of these plumes are better matches to the observed O3 at these altitudes.

Contribution of forest fires to ozone burden:

• The largest impact on the ozone budget can be seen in the emission region,

where the contribution to ozone is greater than 15 ppbv

• Ozone is perturbed by 1ppbv throughout the whole Northern Hemisphere

Forest Fires Impact Regional Air Quality

Over Europe

• The fires contribute 1-2 ppbv

• The largest impact from the

fires in Europe can be seen

over the UK, with up to 4 ppbv

contributed from the fires

Summary

• The model can reproduce observed CO and O3 on days when the aircraft did

not target fire plumes

• The performance of the model was shown to be very sensitive to the chosen emission region

• Parameterised convection and vertical diffusion are needed in a CTM, however, they were overly strong on certain days

• Alaskan forest fires contributed a large quantity of trace species to the global troposphere which will effect tropospheric chemistry

• The forest fires affected regional air quality as far as Europe with up to a 4 ppbv contribution over the UK.