modeling the contribution of wildfire emissions to ambient...

Modeling the Contribution of

Wildfire Emissions to

Ambient PM2.5 & Black Carbon

Serena Chung1, Rodrigo Gonzalez-Abraham 2

Farren Herron-Thorpe3,

Brian Lamb1, Sim Larkin4, Tara Strand 5

1Laboratory for Atmospheric Research, Washington State University

2Molina Center for Strategic Studies in Energy and Environment 3Washington Department of Ecology

4U.S. Forest Service Pacific Northwest Research Station 5Scion Research

NW-AIRQUEST Annual Meeting

June 19, 2014

Introduction

• Funded by the Joint Fire Sciences Program

• Objectives:

– Quantify the contributions for fires to ambient

• BC

• PM2.5

– Evaluate inter-annaul variability

• Approach:

– Modeling: WRF-BlueSky-CMAQ Modeling

• 36-km CONUS, Jun-Jul-Aug, 1996-2005

– Observations: IMPROVE Network

2

Modeling Framework

3

SMOKE

CMAQ

MEGAN

WRF

BlueSky “pthour” method w/ WRAP plume-

rise

Large-Scale Reanalysis

Meteotrology

3D Hourly Meteorology

Historical Fire Database

Hourly Fire Emissions &

With Plume Rise

Biogenic Emissions

NEI 2002 Anthropogenic

Emissions

3D Hourly Speciated Emissions

3D Hourly BC, OC, PM2.5, etc Concentrations

BlueSky Framework Options

4

Fuels

Total

Consumption

Time

Rate

Emissions

Plume

Rise

FCCS

NFDRS

Hardy

Manual CONSUME

FEPS

EPM

FOFEM

Manual

FEPS

EPM

WRAP

FOFEM

Manual

FEPS

FOFEM

CONSUME

Manual WRAP

FEPS

Manual

Dispersion

CALPUFF

HYSPLIT

Area Burned & Modeled PM2.5 Emissions

5

Monthly

PM2.5 Emission (Gg)

from BlueSky

(16% and 77% of PM2.5

emissions were assumed to be

BC and organic aerosol,

respectively.)

Monthly

Acres-Days Burned x 10-6

(http://capita.wustl.edu/fsan/FedFireHist.htm)

Fire Emissions: August of 1997-2005

6

August-Total Fire BC Emissions

Averaged over 1997-2005

August-Total Fire BC Emissions

Maximum during 1997-2005

Modeled BC Concentrations: August of 1997-2005

7

August-Mean BC Concentrations

from Fires

Averaged over 1997-2005

August-Mean BC Concentrations

from Fires

Maximum during 1997-2005

8

1

2

3 4

5 6 Modeled vs Observed BC Concentrations

August of 1997-2005

9

1

2

3 4

5 6 Modeled vs Observed PM2.5 Concentrations

August of 1997-2005

Modeled vs Observed BC Concentrations

August of 1997-2005

10

1

2

3 4

5 6

Modeled vs Observed PM2.5 Concentrations

August of 1997-2005

11

1

2

3 4

5 6

August 2000 Results

12

Fire BC Emission

(total for the month)

Fire Contribution to

BC Concentration

(averaged over the month)

Modelel vs Observed BC & PM2.5 for August 2000

13

Fire Contribution to

BC Concentration

(averaged over the month)

Modelel vs Observed BC & PM2.5 for August 2000

14

Modeled vs Observed BC Concentrations:

August of 1997-2005

15

1

2

3 4

5 6

Modeled vs Observed PM2.5 Concentrations:

August of 1997-2005

16

1

2

3 4

5 6

Summary

17

• For the August 2000 high fire season in central Idaho and western Montana

– On average, model predictions for BC agree well with IMPROVE data

– The model tends to underpredict PM2.5

– The model sometimes over predicts BC and PM2.5 ”close” to the fires, but

under predicts further downwind.

• Possibly because too much smoldering emissions are trapped in the first model in the BlueSky’s “pthour” method.

– Modeled BC to PM2.5 ratios tend to be higher than observed

• 16% and 77% of fire PM2.5 emissions were assumed to be BC and organic aerosol, respectively

• For other periods, the model tends to underpredict predict both PM2.5 and BC

• Insufficient secondary organic aerosol (SOA) formation could explain some of the underprediction

Modeled vs Observed BC Concentrations

August of 1997-2005

18

1

2

3 4

5 6