CO budget and variability over the U.S. - PowerPoint Presentation

Slide1

CO budget and variability over the U.S.

using the WRF-Chem regional model

Anne Boynard, Gabriele Pfister, David Edwards

National Center for Atmospheric Research (NCAR), Boulder, Colorado, USA

NAQC – 9 March 2011Slide2

Motivation

Tropospheric CO is a

key species in tropospheric chemistry (tracer of pollution and precursor of O3) Air pollution monitoring is based on surface networks but little spatial coverage and no vertical information Satellite observations : good spatial coverage and some vertical sensitivity but little information at the

surface

Aircraft observations : vertical extension but

little spatial coverage

Regional chemistry-transport model

Can we distinguish the different factors that are driving the variations of pollutants at the scale of interest to AQ?=> Essential to understand how the surface and tropospheric variability is driven by 3 processes : emissions-chemistry-transportSlide3

Approach

Chemical boundary conditionsMOZART-42Meteorological boundary conditionsNCEP/GFS

Anthropogenic: US EPA NEI 2005

Biogenic: MEGAN

Wildfire: Fire INventory from NCAR

1

Emissions

1

[Wiedinmyer et al., 2006, 2010]

Regional CTM WRF-Chem

CO tracers

Anthropogenic

Chemical

Fire

Inflow

2

[Emmons et al., 2010]

Allows to separate out the different CO source contributions

Period simulation:

10 June – 10 July 2008

(2 weeks spin up)

Horizontal resolution:

24km x 24 km over the U.S.

Surface

observations (EPA)

Satellite

data (MOPITT)

Aircraft

data (ARCTAS campaign)

Model EvaluationSlide4

Model performance: comparison with surface data

Magnitude and variability well reproduced

On average good agreement: R=70% Slightly low bias: 28 ppbvSlide5

Rural site (Washington state)

Urban site (California state)

Model performance: Case studies

Surface CO

Surface CO tracers

Surface CO

Surface CO tracers

Increase due to anthropogenic and fire emissions underestimated in the model

CO inflow is dominant

First peak period: fire probably underestimated

Second peak period: mismatch probably due to an underestimate of fire emissions and a timing and magnitude problem in anthropogenic emissions

Decrease in relative contribution from transported pollution

Good agreement but some discrepancies…

Increase in the model but not as much as in the

obsSlide6

Model performance: comparison with satellite data

MOPITT (V4) Total CO Column

WRF-Chem AK Total CO Column

Globally, similar patterns observed by both WRF-

Chem

and MOPITT

On average, good agreement : R=83% & bias of 1±8%

Fire emissions underestimated by the model (California)

Boundary conditions overestimated by the model (South and West of U.S.)

Average over the period 24 June - 10 July

2008 (1e16 molecules cm

-2

)Slide7

Model performance:

comparison

with aircraft data

Aircraft CO

WRF CO

DC

-8 Flight, 26 June

2008 (1-minute merged data)

Altitude

CO

WRF-

chem

CO Fire

DC-8 Acetonitrile

WRF-

chem

DC

-

8

Underestimate by a factor of 3-4

Acknowledgments:

ARCTAS

science team

(Glen

Diskin

for CO data

and

Armin

Wisthaler

for

CH3CN data)

Good agreement but fire emissions underestimated by the model

ARCTAS mission

:

NASA’s Arctic Research of the Composition of the Troposphere from Aircraft and

Satellites

mission (Spring and Summer 2008)

Fire tracerSlide8

Surface CO tracer contributions over the U.S.

18±14%

14±8%

2±5%

63±19%

Average over the period 24 June - 10 July 2008 (

ppbv

)

Anthropogenic

Chemical

Fire

Inflow

Total CO

Over the Eastern U.S.: high CO concentrations due to anthropogenic emissions and CO produced chemically

In California: high CO concentrations due to anthropogenic and fire emissions

CO is coming from the West and the North

Note the different color scale for CO inflow

!

500

70

150

0Slide9

Can satellite observations be used for AQ monitoring?

Surface finest scale variability not captured in the FT but average behavior captured

Variability in CO inflow at the surface ≈ FT

At higher altitude, variability in inflow dominates the variability in anthropogenic CO

=> A sounder will observe most of the variability in boundary conditions

CO (

ppbv

)

CO Inflow (

ppbv

)

Thermal IR are sensitive in the lower FT (2-3km)

How much of the surface CO variability is reflected in the FT?

Anthropogenic CO (

ppbv

)

Is CO brought by long distance transport or produced locally?Slide10

Summary

Model performance

: good agreement with surface, aircraft and satellite data CO source contributions: Anthropogenic and CO produced chemically dominant over the Eastern coast CO inflow dominant over the Western and Northern U.S. AQ monitoring from satellite : Finest scale variability seen at the surface is not reflected in the FT but the average behavior is captured

Real need of sensitivity down towards the surface Multispectral retrieval has a real sensitivity down towards the surface as recently demonstrated by MOPITT V5 [Worden et al., 2010]

Plan to use multispectral techniques for future geostationary AQ observations (e.g

GEO-CAPE*)

for CO and O3*GEO-CAPE

: Geostationary Coastal and Air Pollution EventsSlide11

Thank you

for your attention!

