HOW WELL DO PM AND METEOROLOGICAL HOW WELL DO PM AND METEOROLOGICAL MEASUREMENT SYSTEMS QUANTIFY MASS MEASUREMENT SYSTEMS QUANTIFY MASS

CONCENTRATIONS, PARTICLE SIZES, CONCENTRATIONS, PARTICLE SIZES, CHEMICAL COMPONENTS AND CHEMICAL COMPONENTS AND METEOROLOGICAL FEATURES?METEOROLOGICAL FEATURES?

Dr. John G. WatsonDr. John G. Watson(johnw@dri.edu)(johnw@dri.edu)

Desert Research InstituteDesert Research Institute

Good comparability for most PM2.5Good comparability for most PM2.5MeasurementSampler Code Model Manufacturera Size FRMb FEMc Principle

AN100 RAAS 100 Andersen PM2.5 yes - GravimetricAN300 RAAS 300 Andersen PM2.5 yes - GravimetricRP2K R&P 2000 Rupprecht & Patashnick PM2.5 yes - GravimetricRP225 R&P 2025 Rupprecht & Patashnick PM2.5 yes - GravimetricAN400 RAAS 400 Andersen PM2.5 no no GravimetricM1ST SASS Speciation Sampler Met One PM2.5 no no GravimetricDICHOTF SA-246B Andersen PM2.5 no no GravimetricSFS Sequential Filter Sampler DRI PM2.5 no no GravimetricMINIVOL25 MiniVol Portable Airmetrics PM2.5 no no GravimetricMOUDI Model 100 MSP PM2.5 no no GravimetricBAM25 BAM-1020 Met One PM2.5 no no Beta AttenuationTEOM25 TEOM 1400a Rupprecht & Patashnick PM2.5 no no Inertial MassDUSTTRAK DustTrak 8520 TSI PM2.5 no no Light ScatteringGREENTEK GT640A GreenTek PM2.5 no no Light ScatteringRAD25 M903 (nephelometer) Radiance Research PM2.5 - - Light ScatteringRAD M903 (nephelometer) Radiance Research TSP - - Light ScatteringHIVOL10V GMW-1200 Andersen PM10 yes - GravimetricMINIVOL10 MiniVol Portable Airmetrics PM10 no no GravimetricBAM10 BAM-1020 Met One PM10 no yes Beta AttenuationTEOM10 TEOM 1400a Rupprecht & Patashnick PM10 no yes Inertial Mass

Seasonal Variations of PMSeasonal Variations of PM2.52.5

Winter

FallSummer

Spring

Longitude Longitude

Good relationships between Good relationships between bscatbscat and and PM2.5PM2.5

Winter

0

50

100

150

200

250

300

350

400

ACP

ALT1

ANG

IBA

CBO

DGBR

ES BTI

CARP CH

LCL

OCO

PED

IED

WFE

DL FEL

FELF

FREM

FRES FS

FHE

LMKC

WLV

R1 M14

MRMOLD

OLW PA

CPI

XLSE

LMA

SFA

SNFH

SOH

SWC

TEH2

VCS

Sites

Bsp

(Mm

-1)

0

10

20

30

40

50

60

70

80

90

100

PM

2.5 (µg

/m3 )

s1s2Bsp

PM2.5

Summer

0

50

100

150

200

ACP

ALT1

ANG

IBA

CBO

DGBR

ES BTI

CARP CH

LCL

OCO

PED

IED

WFE

DL FEL

FELF

FREM

FRES FS

FHE

LMKC

WLV

R1 M14

MRMOLD

OLWPA

CPI

XLSE

LMA

SFA

SNFH

SOH

SWC

TEH2

VCS

Sites

Bsp

(Mm

-1)

0

10

20

30

40

50

PM

2.5 (µg

/m3 )

Bsp

PM2.5

Bsp

PM2.5

Bscat/PM2.5 relationship better by site Bscat/PM2.5 relationship better by site type and in winter (low coarse, more type and in winter (low coarse, more

datadataWinter, 24-hr average

0

20

40

60

80

100

120

140

160

180

200

0 20 40 60 80 100 120 140 160 180 200

Measured PM2.5 (µg/m3)

Pre

dict

ed P

M2.

5 (µg

/m3 )

Boundary sites (r=0.63)Source sites-woodburning (r=0.96)Source sites-oilfields (r=0.90)Source sites-motor vehicles (r=0.46)Source sites-dairy (r=0.66)Desert sites (r=0.45)Intrabasin gradient (r=0.90)Interbasin transport sites (r=0.95)Community exposure sites (r=0.92)

Nitrate and FRM PM2.5 wellNitrate and FRM PM2.5 well--quantified quantified during winter. Much lost in summerduring winter. Much lost in summer

Nitrate and FRM PM2.5 wellNitrate and FRM PM2.5 well--quantified quantified during winter. Much lost in summerduring winter. Much lost in summer

Variation of Nitrate (NOVariation of Nitrate (NO33--))

• Low NO3- found in

summer (<3.5 µg/m3) (Note: different scales)

• Nitrate centered at urban areas.

Summer

Winter

• High NO3- found in

winter.

• Elevated NO3- found in

rural areas between urban centers.

Continuous nitrate biased low, but Continuous nitrate biased low, but consistent with filter nitrateconsistent with filter nitrate

Daily average

y = 0.64x - 0.17r = 0.96, N = 168

y = 0.63xr = 0.96, N = 168

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Non-volatile + volatile NO3- concentration by RAAS (µg/m3)

NO

3- con

cent

ratio

n by

R&

P84

00N

(µg/

m3 )

Continuous nitrate sufficient to detect Continuous nitrate sufficient to detect vertical mixing phenomenonvertical mixing phenomenon

2001-2004

0

2

4

6

8

10

12

14

16

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Hour (PST)

NO

3- (µg/

m3 )

0

0.5

1

1.5

2

2.5

3

BC

(µg/

m3 )

NO3 (Winter 01) NO3 (Winter 02)

NO3 (Winter 03) NO3 (Winter 04)

BC (Winter 04)

Continuous sulfate not as well related Continuous sulfate not as well related to filter measurementsto filter measurements

Daily average

y = 0.63x + 0.52r = 0.64, N = 96

y = 0.94xr = 0.53, N = 96

0

1

2

3

4

0 1 2 3 4

SO42- concentration by FRM (µg/m3)

SO

42- c

once

ntra

tion

by R

P84

00S

(µg/

m3 )

Total carbon comparable (except for Total carbon comparable (except for RP5400), EC varies by detection method.RP5400), EC varies by detection method.

12/1/03-8/31/04

0

2

4

6

8

10

12

14

16

18

20

Dec-03

Jan-0

4Feb

-04Mar-

04Apr-

04May

-04Ju

n-04

Jul-0

4Aug

-04Sep

-04Oct-

04Nov

-04

PM

2.5 m

ass

conc

entra

tion

(µg/

m3 )

0

5

10

15

20

25

BC

/ P

M2.

5 (%

)

PM2.5 mass IMPROVE ECMAAP BC (670 nm) 2-Aethalometer BC (880 nm)7-Aethalomter BC (880 nm) Photoacoustic BC (1027 nm)Sunset BC (660 nm) Sunset EC R&P 5400 EC

Variation of Elemental Carbon (EC)Variation of Elemental Carbon (EC)

• Higher EC around urban centers.

• Elevated EC found near the Fresno Supersite and Bakersfield. Rural sites show limited summer-winter contrast.

Summer

WinterE

C

EC

0102030405060708090

100

% o

f rep

orte

d d

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Laboratory

within 25%25% to 50%50% to 75%>75%

InterlaboratoryInterlaboratory comparisons not too good for organic compounds comparisons not too good for organic compounds ((SupersiteSupersite meeting, meeting, SchantzSchantz et al, 1Cet al, 1C--1)1)

Distribution of Distribution of LevoglucosanLevoglucosan, a Wood, a Wood--Burning Burning Marker, Compared to OC DistributionMarker, Compared to OC Distribution

Annual Avg Winter Avg*FEL 6 26CHL 7 32YOSE 9 38EDW 12 52OCW 14 58HELM 19 81PIXL 19 82ANGI 23 98COP 32 138BAC 49 209BTI 50 215SNF 57 244SJ4 58 247S13 63 269LVR 68 291FEDL 75 323M14 101 433FRS 121 521SDP 128 551FSF 202 868

Levoglucosan Concentrations (ng/m3)

* Predicted concentration based on mass concentration measurements

Annual OC Distribution

Meteorological data quality summaryMeteorological data quality summary• Surface network OK, CIMIS, NWS, and PG&E

nonstandard heights. ARB and Bay area best. RAWS are more for local rather than regional analyses. Some CIMIS wind vanes mis-alinged by up to 30 degrees.

• Surface spatial coverage is very good.• Sonics on tower out of alignment, but agree with

mechanicals when 230 degrees added. Mounted upside down. 97 m sonic not good. Good comparability with mechanical.

• Sodars had poor return at lowest levels at Angiola. Agreed well at 100 m.

• Rawinsondes have slow response time for T and RH. Not too good for short distances, e.g. surface vs. valleywide transition.

Meteorological data quality summaryMeteorological data quality summary• RASS not good < 100 m, but appear to sufficiently

capture changes in the valleywide layer. Reasaonable comparisons with audit rawinsondes, considering limitations of both methods.

• Hourly averages appear to adequately represent vector directions and speeds for shorter time periods ( 5 min).

• RH are sufficient to estimate locations and frequency of fogs (RH>95%).

• Profilers can detect light winds aloft.

CIMIS comparison with ARB and NOAACIMIS comparison with ARB and NOAA

Sonic vs. mechanical wind direction, after Sonic vs. mechanical wind direction, after correctionscorrections

SodarSodar OK at 100 mOK at 100 m

CRPAQS measurement evaluation CRPAQS measurement evaluation studiesstudies

• Measurement evaluations did a good job of determining how well measurement systems did or didn’t work.

• With a few exceptions, measurements appear adequate to determine general features, test conceptual models, and to serve as input and to evaluate air quality modeling.

• There are individual times and locations where data are compromised, but this is evident from similar measurements at the same or nearby sites. The network is redundant.

• Given that many of the measurement methods were new and innovative, a broader perspective is needed on their overall accuracy and precision relevant to central