Comparison of three photochemical mechanisms (CB4, CB05, SAPRC99) for the Eta-CMAQ air quality forecast model for

O3 during the 2004 ICARTT study

Shaocai Yu$,&, Golam Sarwar+, Rohit Mathur+, Daiwen Kang$,&, Daniel Tong$,&,

&Atmospheric Modeling Division, ARL, NOAA , RTP, NC 27711. $On assignment from Science and Technology Corporation,

+ NERL, U.S. EPA, RTP, NC 27711

Introduction (Motivation)Introduction (Motivation)

Photochemical mechanisms Critical module of air quality models 30% or more uncertainties30% or more uncertainties due to errors in reaction rate constants and yields due to errors in reaction rate constants and yields

(Russell and Dennis, 2000) (Russell and Dennis, 2000)

Three widely used photochemical mechanisms: CB4: Carbon Bond Mechanism-version IV with lumped structure approach (Gery

et al., 1989) CB05: updated version of CB4 with more inorganic reactions and organic species

(Yarwood et al., 2005) SAPRC99: different scheme (lumped molecule approach) for condensing the

organic chemistry and more detailed organic chemistry (Carter, 2000)See Table

Useful and interest: study how these three photochemical mechanisms affect the CMAQ simulation of

O3 by comparing to observations

CB05 and SAPRC: more species and reactions; better characterization of radical recycling; but need more time

Objectives

Evaluate influence of CB4, CB05 and SAPRC99 on

the spatial and temporal variations of O3 with

AIRNOW Obs over the eastern US

Examine impact of CB4, CB05 and SAPRC99 on

Eta-CMAQ simulations for O3, its gas precursors

with 2004 ICARTT field data

Model Description (Configuration)Model Description (Configuration)

Eta-CMAQ model:

Eta forecast model provides meteorological fields for CMAQ

CB4, CB05, SAPRC99: photochemical processes

Emissions processed using Emissions processed using SMOKE processing systemSMOKE processing system

12 km12 km horizontal grid resolution horizontal grid resolution

22 Vertical layers between surface and 100 22 Vertical layers between surface and 100 mbmb

This result: July 14 to August 18, 2004

Observations

EPA AIRNOW network:

Hourly O3 at 614 sites in E US.

Ground Data at Four AIRMAP sites

Model domain and surface sites (AIRNOW, AIRMAP)

AIRMAP sites

Observations

2004 ICARTT Data

Vertical profiles (O3, CO, NO, NO2, HNO3, SO2) from aircraft (P-3 and DC-8).

Surface-level data over the ocean on Ron Brown ship

Tracks of (a) P-3, (b) DC-8

P-3

DC-8

•P-3: Northeast; •DC-8: Eastern US

Results: Max 8-hr O3 at AIRNOW sites

Very close

58.8

21.5

5.3

-5.7

73.3

32.6

15.1

3.9

77.3

37.1

20.7

10.6

-10.0

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

<40 40-60 60-75 >75

O3 concentrations (ppb)

NM

B (%

)

CB4

CB05

SAPRC

Results: O3 Vertical profiles (7/14-8/15)

Models reproduced vertical structure of Obs•P3: SAPRC99>CB05>CB4>Obs•DC-8: similar to P3 although slightly close to Obs

(1) P-3

Daily Layer Means

Hei

ght (

m)

Results: CO and HNO3 Vertical profiles (7/14-8/15)

CO:Consistent Underpredictions.partly due to inadequate representation of biomass burning

effects from outside the domain

Obs>CB05~CB4>SAPRC

Daily Layer Means

HNO3:Very good

Slight underprediction

(1) P-3

Conc. (ppb)

Conc. (ppb)

Hei

ght (

m)

Results: NO2 and NO Vertical profiles (7/14-8/15)

NO2:P3: good at high altitudes

Underestimate at lower altitudesOrganic nitrate can react back to NO2 in CB05 and SAPRC

Daily Layer Means

NO:Under predictions of NO

Aircraft and lightning NO emissions are not in inventory

Results: NOy, SO2, H2O2Vertical profiles (7/14-8/15)

NOy, Consistent overestimation

SO2: Overestimation at low altitudes but good at high altitudes

Daily Layer Means

H2O2:CB4: significant overestimation

Its H2O2 formation rate is 62% higher than CB05 (Luecken et al., 2008)

CB05 and SAPRC: close to ObsCB05: slightly higher than ObsSAPRC: slightly lower than Obs

CB05 produces more new HO2, enhancing H2O2

Results: Time-series evaluation on ship

Gas species (NMB ,%)

Con

ce

ntr

atio

n

020406080

100120140160

O3 (Obs)O3 (CB4)O3 (CB05)O3 (SAPRC)

ppbv

0100200300400500600700 CO (Obs)

CO (CB4)CO (CB05)CO (SAPRC)

ppbv

010203040506070 NOy (Obs)

NOy (CB4)NOy (CB05)NOy (SAPRC)

ppbv

0

10

20

30

40 NO2 (Obs)NO2 (CB4)NO2 (CB05)NO2 (SAPRC)

ppbv

0

50

100

150 O3+NO2 (Obs)O3+NO2 (CB4)O3+NO2 (CB05)O3+NO2 (SAPRC)

ppbv

0

4

8

12

16 NO (Obs)NO (CB4)NO (CB05)NO (SAPRC)

ppbv

012345 PAN (Obs)

PAN (CB4)PAN (CB05)PAN (SAPRC)

ppbv

02468

10 SO2 (Obs)SO2 (CB4)SO2 (CB05)SO2 (SAPRC)

ppbv

0

0.4

0.8

1.2

1.6 Isoprene (Obs)Isoprene (CB4)Isoprene (CB05)Isoprene (SAPRC)

ppbv

7/14 7/24 8/3 8/137/19 7/29 8/8Time (UTC, 2004)

CB4 CB05 SAPRC

O3 24.4 36.5 42.3

O3+NO2 21.5 32.2 37.9

CO -4.3 -7.4 -13.3

NOy 43.3 37.6 16.4

NO2 -6.2 -13.9 -10.2

NO -30.6 -46.9 -46.4

PAN 28.6 24.6 -11.3

SO2 68.4 82.0 85.6

ISOP -61.3 -58.7 -60.7

Slightly better:CB4: O3, CO, NO2, NO, SO2

CB05: Isop

SAPRC: NOy, PAN

O3

NOy

NO2

NO

PAN

SO2

Isoprene

CO

O3+NO2

Results: Time-series data on ship

O3 production efficiency (N)

O3-NOz slope (Olszyna et al., 1994):Upper limit of N

SAPRC>CB05>CB4 but all are lower than ObsConsistent with O3 concentrations

Results2. Time-series evaluation at AIRMAP sites

Castle Springs (CS)

O3

NOy

NO

CO

SO2

High O3 period (721-7/23)

Low O3 period (724-7/27)

020406080

100120 O3 (Obs)

O3 (CB4)O3 (CB05)O3 (SAPRC)

ppbv

0100200300400500 CO (Obs)

CO (CB4)CO (CB05)CO (SAPRC)

ppbv

0

0.5

1

1.5NO (Obs)NO (CB4)NO (CB05)NO (SAPRC)

ppbv

0

2

4

6

8 NO2 (Obs)NO2 (CB4)NO2 (CB05)NO2 (SAPRC)

ppbv

02468

101214 NOy (Obs)

NOy (CB4)NOy (CB05)NOy (SAPRC)

ppbv

0

5

10

15

20 SO2 (Obs)SO2 (CB4)SO2 (CB05)SO2 (SAPRC)

ppbv

7/14 7/24 8/3 8/13 8/187/19 7/29 8/8time (UTC, 2004)

NO2

Results2. O3 production efficiency at AIRMAP sites (O3-NOz slope)

0

20

40

60

80

100

120

O3 (Obs)O3 (CB4)O3 (CB05)O3 (SAPRC)

0 2 4 6 8 10 12

CS

O3 (Obs)O3 (CB4)O3 (CB05)O3 (SAPRC)

0 2 4 6 8 10 12

MWO

0 2 4 6 8 10 12

O3 (Obs)O3 (CB4)O3 (CB05)O3 (SAPRC)

TF

NOz (ppb)

O3 (

pp

b)

Results2. O3 production efficiency at AIRMAP sites (O3-NOz slope)

N

O3-NOz slope:Upper limit of N

SAPRC>CB05>CB4 but all are lower than ObsConsistent with O3

concentrationsSAPRC is close to Obs

Contacts:

Brian K. Eder

email: eder@hpcc.epa.gov

www.arl.noaa.gov/

www.epa.gov/asmdnerl

ConclusionsBased on 35-day (7/14-8/18/2004) simulations over eastern U.S:

At AIRNow sites for max 8-hr O3, SAPRC has the highest O3, followed by CB05 and CB4

Consistent with the results of N

For high concentrations>75ppb, CB05 is better than CB4 and SAPRC

For concentration ranges<75 ppbv, CB4 is better (less overestimation) than CB05 and SAPRC

Vertical profiles from P-3 and DC-8: Models captured vertical structures of gaseous species (O3,CO,NOy,HNO3,H2O2 etc.) Overestimated O3: SAPRC>CB05>CB4

Consistent overestimation: NOy,

Consistent underestimation: CO, NO, NO2

H2O2: CB4 significantly overestimated but CB05 and SAPRC are close to Obs

• CB05: slightly higher than Obs; SAPRC: slightly lower than Obs)

Contacts:

Brian K. Eder

email: eder@hpcc.epa.gov

www.arl.noaa.gov/

www.epa.gov/asmdnerl

Conclusions (Continued) Results on the basis of ship time-series data:

Slightly better: CB4: O3, CO, NO2, NO, SO2

CB05: Isoprene SAPRC: NOy, PAN

Modeled upper limit of N by O3-NOz slopes on basis of ship and AIRMAP data: SAPRC (5.3 to 10.2)>CB05(4.5 to 7.5)>CB4(4.0 to 6.1)

Are consistently lower than observations (8.5 to 11.8)Background O3 conc. is ~10 ppb higher than Obs

Disclaimer

The research presented here was performed under the Memorandum of Understanding

between the U.S. Environmental Protection Agency (EPA) and the U.S. Department of

Commerce's National Oceanic and Atmospheric Administration (NOAA) and under

agreement number DW13921548. This work constitutes a contribution to the NOAA Air

Quality Program. Although it has been reviewed by EPA and NOAA and approved for

publication, it does not necessarily reflect their policies or views.

CBMIV (Operational) vs. CB05 (experimental) performance

Max 8-hr O3, CONUS, June 15-Aug. 31, 2008

• Lower bias and error with CB05 at moderate-high O3 mixing ratios• Higher error in regional statistics due to over-prediction at low mixing ratio range

- Could it be for reasons other than chemical mechanism?

Results

7/19/047/18/047/17/04

July 16-22, 2004: Evidence of effects of long range transport (Alaskan fire)

(1) MODIS (satellite) observations for AOD

(2) TOMS (satellite) observations for absorbing aerosol index

Significant underpredictions of PM2.5 by the model during July 16 to 26 are mainly due to inadequate representation of biomass burning (carbonaceous aerosol) effects from outside the domain (Alaskan fire)