Discrete Measurements of PM2.5 Mass and Composition in the Southeastern U.S.:

Regional and Seasonal Trends between 1999 and Today

K. Baumann, M.E. Chang, R.J. Weber, A.G. Russell, W.L. Chameides

School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta

Instruments, Data Quality, Sites

PM2.5 - O3 – Reactive Gases Relationships

Estimating Organic Mass

Acknowledgement: J.L. Waynick, TN-DEC, K.J. Olszyna, TVA

M. Bergin, D. DiPasquale, F. Ift, K. Patel, W. Younger, J. Zhao all Georgia Tech

Funding provided by U.S.-EPA through NCER Grant R826372 and GA-EPD

2

Particle Composition Monitor “PCM”

Special tests and procedures for eliminating positive water bias, OC artifacts and other details described in paper accepted to JGR “Atlanta Supersite” special section, coming out soon…

For each PCM sample 17 (!) components (incl. 5 field blanks) are being analyzed via IC & TOT.

The following species are being quantified and reported.

Channel 1:

NH3

Na+, K+, NH4+, Ca+2

Channel 2:

HF, HCl, HONO, HNO3, SO2,

HCOOH, CH3COOH, (COOH)2

F-, Cl-, NO3-, SO4

=,

HCOO-, CH3COO-, C2O4=

Channel 3:

EC, OC, “SVOC”

3

Assessing Accuracy of PCM Measurements

S-compounds and mass agree well, volatile species esp. NO3- more difficult to measure accurately

15

10

5

0

30 m

in P

ILS

(u

g m

-3)

12108642

discrete PCM (ug m-3

)

SO42-

r = 0.97

slope = 1.04 +-0.04i-cept = 0.12 +-0.21

70

60

50

40

30

20

10

0

30 m

in T

EO

M (

ug

m-3

)

403020100

discrete PCM (ug m-3

)

Will. TowerMass r = 0.92slope = 0.96 +-0.07i-cept = 2.52 +-1.23

3

2

1

0

30 m

in P

ILS

(u

g m

-3)

1.51.00.50.0

discrete PCM (ug m-3

)

NO3- r = 0.61

slope = 0.47 +-0.09i-cept = 0.18 +-0.05

excl 1st 2 PCM

30

20

10

0

30 m

in T

EC

O (

pp

bv)

1086420

discrete PCM (ppbv)

SO2 r = 0.99

slope = 1.38 +-0.033i-cept = 0.00 +-0.08

70

60

50

40

30

20

10

0

30 m

in T

EO

M (

ug

m-3

)

403020100

discrete PCM (ug m-3

)

LaPorteMass r = 0.95slope = 1.01 +-0.05i-cept = 0.98 +-0.81

excl 1st 3 PCM

6

5

4

3

2

1

0

30 m

in P

ILS

(u

g m

-3)

4321

discrete PCM (ug m-3

)

NH4+

r = 0.92

slope = 0.91 +-0.06i-cept = 0.26 +-0.11

4

Measurement Sites

36.5

36.0

35.5

35.0

34.5

34.0

33.5

33.0

32.5

-88.0 -87.5 -87.0 -86.5 -86.0 -85.5 -85.0 -84.5 -84.0 -83.5 -83.0 -82.5 -82.0

Atlanta

major point sources point sources w/ CO:NOx > 1

SCISSAP FAQS sites

50x50 km

Wansley

Johnsonville

Bowen

McDonough

Branch

Cumberland

Urquhart

Augusta

Macon

Columbus361

NashvilleDickson

SuperSite

Hendersonville

Gallatin

Arkwright

SchererGriffin

Southern

Center for the

Integrated

Study of

Secondary

Air

Pollutants

(US EPA)

Fall-line

Air

Quality

Study

(GA-EPD)

5

Special Setup at Hendersonville

Set up to capture the Nashville urban plume.

Utilizing an instrumented 48 m tower during

SOS’99, 16 June - 22 July 1999.

Measurements between 42 and 4 m agl :

Positive vertical gradients for

60-70 % of all daytime, and

70-80 % of all nighttime samples of

PM2.5 mass, SO4=, NO3

-, and NH4+!!

Direct emissions and/or secondary formation of

fine PM aloft

6

Seasonal Comparison of PM2.5 Composition: Suburban vs Rural

24.3 +-9.3

32%

3%

13%2%

19%

18%

4%

1%

8%

27.6 +-11.6

31%

2%

10%

21%

22%

5%

0%

8%

1%

SO4=

NO3-

NH4+

EC

OC

OOE

LOA

Others

UnID

Summer-99

Rural Dixon

Suburban

Hendersonville

•Differences in composition are insignificant regionally but not seasonally•Differences in mass related to different BL dynamics

15.9 +-6.9

20%

7%

12%

1%31%

6%

10%

1%

12%

Fall-99 Winter-99/00

20.1 +-15.2

20%

8%

11%

0%29%

16%

2%

2%

12%

10.6 +-3.6

21%

11%

12%

1%31%

5%

6%

1%

12%

12.2 +-4.7

18%

10%

11%

2%35%

5%

4%

1%

14%

7

Sequential measurements at GA sites captured different episodes, e.g. July 4th weekend at Augusta!

11.4 +-2.8

21.0%

5.8%

6.8%

2.1%

23.1%

9.2%

7.0%

23.3%

1.7%

24.3 +-9.4

30.5%

2.3%

9.4%

1.5%16.0%

6.4%

7.0%

1.9%

25.0%

Macon, GA

Late June

Augusta, GA

Early July

Columbus, GA

Late July

21.4 +-5.9

28.5%

1.5%

9.2%

1.8%26.3%

10.5%

6.8%

13.7%

1.7%

Regional Comparison of Average PM2.5 Composition in 2000

Williams Tower LaPorte

Both Houston, TexAQS sites from mid August to mid September

16.4 +-7.9

30.2%

3.1%

9.4%

5.3%18.4%

7.4%

7.2%

14.9%

4.1%

16.4 +-7.7

29.8%

3.1%

10.3%

4.4%19.8%

7.9%

6.1%

3.6%

14.9% SO4=

NO3-

NH4+

EC

OC

OOE

LOA

Others

UnID

8

PM2.5 Mass Balance and Maximum Hourly Ozone

0

10

20

30

40

50

60

Atl

anta

H.v

ille

Dix

on

H.v

ille

Dix

on

H.v

ille

Dix

on

Mac

on

Au

gu

sta

Co

lum

bu

s

LaP

ort

e

W.T

ow

er

Gri

ffin

Gri

ffin

PM

2.5

(m

g m

-3)

0

20

40

60

80

100

120

O3

-h m

ax

(pp

bv

)

SO4= NO3- NH4+ EC OC OOE LOA Others UnID O3-h max

Summer-99 Fall-99 Winter-99/00 Summer-00 Jul-01 Jan-02

Comparison of Seasonal and Regional Averages: PM2.5 & O3

•Seasonal [PM2.5]-mass /-SO4= / [O3] correlation: high in summer, low in fall & winter

•Houston TexAQS measurements governed by local emissions, episodes and meteorology•BL dynamics possibly causing sub-regional differences between H.ville & Dixon

•Unidentified [PM2.5] mass highest and most variable in summer (and fall)

9

Comparison of Seasonal and Regional Averages: Reactive Gases

Reactive Gases and Maximum Hourly Ozone

0

1

2

3

4

5

6

7

8

Atl

anta

H.v

ille

Dix

on

H.v

ille

Dix

on

H.v

ille

Dix

on

Mac

on

Au

gu

sta

Co

lum

bu

s

LaP

ort

e

W.T

ow

er

Gri

ffin

Gri

ffin

LO

A, S

O2

, HN

O3

, NH

3 (

pp

bv

)

0

20

40

60

80

100

120

O3

-h m

ax

(pp

bv

)

LOA SO2 HNO3 NH3 O3-h max

Summer-99 Fall-99 Winter-99/00 Summer-00 Jul-01 Jan-02

•Strong seasonal trend of acidic gases (LOA, HNO3) pointing to photochemical sources•Neutralizing NH3 less variable seasonally but regionally: highest in metro areas

•High abundance of precursor gases combined correlates with high [PM2.5]

10

Seasonal/Regional Aerosol Acidity Based on [SO4=/NO3

-/NH4+]

•Aerosol is closely neutralized / slightly alkaline in fall & winter but more acidic in summer•Acidity caused by insufficient NH3, or unaccounted for organic amines (with higher OM/OC)?

PM2.5 Charge Balance

-250

-200

-150

-100

-50

0

50

100

150

200

250

Atl

anta

H.v

ille

Dix

on

H.v

ille

Dix

on

H.v

ille

Dix

on

Mac

on

Au

gu

sta

Co

lum

bu

s

LaP

ort

e

W.T

ow

er

Gri

ffin

Gri

ffin

(ne

m-3

)

SO4= NO3- NH4+ Net Acidity

Summer-99 Fall-99 Winter-99/00 Summer-00 Jul-01 Jan-02

11

Major PM2.5 Mass Fractions and OM-to-OC ratios for Closure

0

20

40

60

80

100

Atl

anta

H.v

ille

Dix

on

H.v

ille

Dix

on

H.v

ille

Dix

on

Mac

on

Au

gu

sta

Co

lum

bu

s

LaP

ort

e

W.T

ow

er

Gri

ffin

Gri

ffin

PM

2.5

ma

ss

fra

cti

on

(%

)

0

2

4

6

8

10

OM

/OC

{inc

l SV

OC

} for c

los

ure

(NH4)2SO4 TOC.4 TOC.6 OM/OC OM/OC {svoc}

Summer-99 Fall-99 Winter-99/00 Summer-00 Jul-01 Jan-02

Estimating Organic Mass (OM) Using Mass Closure

•OM/OC highly variable within sites/periods; average 1.4 too low!•Factors need to be higher in summer: more oxygenated species

•Factors seem to increase from urban, suburban, to rural sites (aging?)•Incl. SVOC from XAD-backup filter leads to range: 1.5 (Atlanta) to 2.7 (Griffin)

12

Summary

• Positive vertical gradients of PM2.5 mass and sulfate point to atmospheric aerosol formation.

• Insignificant regional differences in PM2.5 composition, but noticeable seasonal differences,

esp. %-SO4=, likely due to higher SO2 emissions and photochemical activity in summer.

• Based on SO4=/NO3

-/NH3+ system, PM2.5 in SE-US is slightly alkaline in winter but more acidic

in summer ((NH4)HSO4 only or some species possibly not accounted for?).

• OM/OC= 1.4 seems mostly too low but is highly variable reflecting different air masses.

• Higher factors necessary in summer due to more oxygenated species from photochemistry.

• General trend for higher factors away from urban areas pointing to secondary processes.

• Different factors might have to be applied for OC from quartz front and XAD backup filters

due to different volatilities (true for Atlanta Supersite 08/99, see JGR paper)

Developing lab techniques for WSOC and ISOC

Speciation of OC via GC-MS: collaboration with Dr. Mei Zheng !!

Outlook

13

PCM

14

OM/OC Estimates With & Without “SVOC”

    OM/OC for closure OM/OC {svoc} for closure

  AVG STD AVG STD

Summer Atlanta 2.1 0.7 1.5 0.3

-99 H.ville 2.4 0.7  

  Dixon 2.9 1.8    

Fall-99 H.ville 1.7 0.5  

  Dixon 2.2 0.9  

Winter H.ville 1.6 0.3    -99/00 Dixon 1.6 0.4    

   

Summer Macon 2.5 0.6 1.7 0.1-00 Augusta 3.4 1.7 1.7 0.3

  Columbus 2.0 0.5 1.6 0.5  LaPorte 3.5 3.8 2.3 2.6  W.Tower 3.5 3.2 2.4 2.3

Jul-01 Griffin 2.9 1.2 2.7 1.0Jan-02 Griffin 2.2 1.0 1.9 0.7

Note, due to technical difficulties, only undenuded quartz filter samples were taken at H.ville & Dixon!