1. Local Accumulation of PM2
Local Accumulation of PM2.5 from Prescribed Burning with Consequences for Attainment
K. Baumann*, S. Lee#,
M. Zheng*, A. Unal#
*) School of Earth and Atmospheric Sciences
#) School of Civil and Environmental Engineering
Contributors: K. Manomaiphiboon, J. Schauer, L. Naeher, D. Blake, M. Chang
M. Clements, P. Gustafson, J. Greenlee, A. Braswell, D. Chan, K. Redmond
Funded by DoD/EPA/State P2 Partnership, US Army IC, Fort Benning & IMA-SRO, Fort McPherson

2. Exceedances of Annual PM2.5 NAAQS in GA
Metro Atlanta Macon Columbus Augusta Coastal Sites
Metro area always > Annual NAAQS
Regionally decreasing trend…
…Correlates with increasing rainfall
Columbus 2001 < Annual NAAQS
if excl. 24h-NAAQS

3. Problem: ESA versus CAA
PM2.5 Eceedance at Columbus-OLC near Fort Benning for SE winds in Winter 2001/02
Manage 94,000 acr, Burn 1/3 per yr, ~ 520 acr/day
Military Installations in SE-US occupy habitat of endangered species (red-cockaded woodpecker), and are required to maintain ecosystem by prescribed burning, risking violations of the NAAQS.
Clean
Air
Act
Endangered
Species

4. PM2.5 Wind Roses: Seasonal Differences Across GA Indications for Regional and Local Transport
GRF
FAQS Period
Jul’00 - Sep’03
MAY - OCT
NOV – APR
JST Period
AUG’99

5. Prescribed Burning in Georgia > 1,000,000 acres annually, majority in winter half
Columbus
OLC site

6. Seasonal Differences in Diurnal Cycles of PM2.5
Summer
Winter
PM2.5 sources at OLC near Columbus drive nighttime averages in winter.
Midday minimum due to BL mixing in winter seems compensated by SOA in summer. Summer stagnation with high O3 also leads to high PM2.5 (e.g. 2000). Annual PM2.5 NAAQS (15 mg m-3) sensitive to: - SOA formed under regional stagnation in summer; - Primary PM2.5 from local sources at night in winter.

7. PM2.5 at Columbus-OLC in Oct-Dec 2001
Critical parameters driving [PM2.5] at OLC:
size of burn, distance and plume trajectory
atmospheric divergence (horizontal wind speed)
{vertical} boundary layer stability (T difference)
BL mixing depth at night (BLHnight)

8. PCA and Regression Results

9. Emission Factors in g/kg BiomassPM
ΔX: excess mass concentration
ΔC: excess carbon concentration
NMHC : non-methane hydrocarbons
PC : particulate carbon
Assume Biomass-C ~ 50%
[Ward and Hardy, 1991; Sinha et al, 2004;
Fine et al., 2002; Hays et al., 2002;
Schauer et al., 2001]
x1000
Gas
Organic carbon (OC) is a dominant species of particulate emission, followed by EC and K.
Enhanced emissions (2-3 orders > background) of Aromatics (benzene, toluene, xylenes), NH3 and Biogenics (isoprene, a-/b-pinene), important precursors for formation of secondary PM.
Emission Factors (EF) higher at smoldering than flaming; even more as C content increases.
Smoldering EF in g/kg biomass mostly higher relative to other studies.

10. Direct Emissions of VOCs
Flaming vs Smoldering
Average source profiles of 28 VOCs measured at 5 flaming and 5 smoldering PB locations at Fort Benning between February and May 2003

11. VOC Source Apportionment at OLC Receptor
Contributions from 7 identified source categories to the total 28 VOC measured at OLC.
Significant contributions from prescribed burning only at night and early morning.
Biogenic contributions highest at midday, but significant only towards warmer season.

12. High-Vol Sampling and GC/MS Analyses
Quantification of >100 Particle-phase Organic Compounds
POC
Retene
Pimaric acid
Abietic acid
Sandaracopimaric acid
Levoglucosan
Five consecutive 5-h samples taken at OLC between February 5th 1200 and 6th 1300.
Note: Less than 10 % of total OC identified via GC/MS !!

13. Influence from February 5th Burn: Source Apportionment
Nighttime avg 53% wood ?
Low P front moving through GA on 3rd and 4th, with cold dry air moving in behind it from NE, causing below normal T under clear skies. Prescribed burning of 937 acres on 2/ at ~28 km to east, smoldering until 2/6 am.

14. Findings
Climatological effect of precipitation-rich years beneficial for PM2.5 attainment.
Local impacts from high PB and wild fire intensities cause violation of 24h [PM2.5] and also the exceedance of annual NAAQS.
Annual PM2.5 NAAQS is sensitive to i) SOA formed under regional stagnation in summer; ii) Primary PM2.5 from local sources at night in winter; iii) regional transport within air sheds.
PCA and regression analysis yields strong correlations between temperatures, wind speeds, forecasted chance for precipitation and fire occurrence, with [PM2.5] at OLC.
Strongest direct impact from PB emissions encountered under clear skies nocturnal inversion and local cool air pooling (along the Upatoi Creek), with contributions of ~52 % to total OC, and up to 15 % to total VOC.
Substantial aromatic & biogenic HC emissions have implications for O3 and PM2.5 modeling.
Emission Factors (EF) higher at smoldering than flaming; even more as C content increases.
Indications for higher emissions of polar, more oxygenated POC under “real world” burning conditions.

15. Questions?
Thank you!