Reducing Air Pollution In Los Angeles

1
Data Needs for Evaluation of Radical and NOy
Budgets in SCOS97-NARSTO Air Quality Model
Simulations
Gail S. Tonnesen University of California,
Riverside Bourns College of Engineering Center
for Environmental Research and Technology
February 14, 2001, SCOS97-NARSTO DataWorkshop
2
Acknowledgments

• Funding for related projects
• U.S. EPA
• American Chemistry Council
• Datasets
• Draft prerelease datasets provided by ARB

3
Trace Gas Governing Equations

• j1,N Coupled PDEs
• ?Cj ??t ? ?v.?Cj D?2Cj P(C) ? L(C)Cj Ej ?
  Dj
• Operator Splitting
• ?Cj ??t ? v.?Cj
• ?Cj ??t D?2Cj Ej ? Dj
• dCj ?dt P(C) ? L(C)Cj
• Gear solver is the gold standard for stiff ODEs

4
Model Evaluation

• Verification, Validation or Evaluation?
• Oreskes et al., 1994.
• Comparisons with ambient data.
• Validation of component processes.
• Indicators for testing O3 sensitivity.
• Sensitivity and uncertainty analysis.

5
Family Definitions NOx NO NO2 (NO3
2 N2O5 HONO HNO4) NOz HNO3 RNO3
NO3 PAN NOy NOx NOz total
oxidized nitrogen. HC VOC (or ROG)
CH4 CO Ox O3 O NO2 NOz 2 NO3
3 N2O5 HNO4 HOx OH HO2 RO2
6
Fundamental Photochemistry

• Tropospheric gas phase chemistry is driven by the
  OH radical
• Radical Initiation
• Radical Propagation
• Radical Termination
• NOx termination

7
PSS Equilibrium
NO2 h? ? NO O O O2 ?
O3 O3 NO ? O2 NO2
NO2 O3 ? NO3 O2 NO3 h? ? NO2 O
P(Ox) RO2 NO ? RO NO2 HO2
NO ? OH NO2
8
Radical Initiation
O3 h? ? O(1D) O(1D) H2O ? 2 OH
HCHO h? ? 2 HO2 CO HO2 NO ? OH
NO2 HONO h? ? OH NO PAN ? RO3 NO2
9
Radical Propagation
OH CH4 O2 ? CH3 O2 H2O CH3O2 NO ?
NO2 CH3O CH3O O2 ? HO2 HCHO HO2
NO ? NO2 OH 2x( NO2 h? O2 ? O3 NO
) Net Reaction CH4 4 O2 ? 2 O3 HCHO H2O
10
Radical and NOx termination
OH NO2 ? HNO3 HO2 HO2 ? H2O2 HO2 RO2 ?
ROOH RO2 NO? RNO3 RO3 NO2? PAN N2O5 H2O ?
2 HNO3
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Model Evaluation

• Local Diagnostics
• Instantaneous reaction rates at a given site.
• Examples P(OH), P(Ox), P(Ox)/P(NOz)
• Cannot get production rates from time-series!
• Cumulative Trajectory Diagnostics
• cumulative history of reaction rates and other
  loss processes in an air parcel integrated over
  hours or days.
• Examples H2O2, HNO3, O3, O3/NOz

15
Data Needs for Local Diagnostics

• Radical Initiation
• J-values HCHO, O3, H2O, HONO, H2O2, PAN
• OH Chain Length
• ? kOH HCi /(? kOH HCi kOH NO2 )
• kHO2 NO /(kHO2 NO kHO2 (RO2 2 HO2 ) )
• Radical Termination
• NO2 OH, HO2 RO2, NO RO2, O3
• NOx Termination, P(NOz)
• NO2 OH, NO RO2, NO2 RCO3, NO3, N2O5
  H2O
• Pg(Ox)
• NO, HO2, RO2.

16
Data Needs for Cumulative Diagnostics

• Radical Initiation Termination (approximate)
• (2 peroxides NOz )
• OH Chain Length (approximate)
• Ox / (2 peroxides NOz )
• 2 peroxides/NOz
• NOx Termination, P(NOz)
• HNO3, speciated RNO3, NO3-, PAN
• P(O3), P(Ox)
• O3, O3 NO2 NOz

17
Model Domain and Parameters

• 1997 Southern California Ozone Study (SCOS97).
  Aug 3 to 5, 1997
• CMAQ and CAMx
• MM5 16 layers
• CB4 chemical mechanism
• Gear CMAQ, CMC CAMx
• Bott Advection Scheme
• No Aerosols
• Includes process analysis diagnostic outputs.

18
Uncertainties In CMAQ vs CAMx Comparison

• Timing in CAMx - are emissions calculated as PST
  or PDT?
• Vertical mixing - CAMx has less vertical
  dispersion in early morning?
• Emissions - CMAQ may be missing large point
  sources.
• Problem with isoprene in CAMx

19
Peak Model Ozone on Aug 5 (3rd day)
Difficult to analyze effects accumulated over 3
days, so...
20
Start Evaluation with spinup (1st day)
Comparison of O3 at 1500 PDT
21
Comparison of O3 aloft before start of 2d day
Errata all units are ppbV
22
Pg(Ox) 700-800 PDT
23
Pg(Ox) 800-900 PDT
24
Pg(Ox) 900-1000 PDT
25
Pg(Ox) 1000-1100 PDT
26
Pg(Ox) 1100-1200 PDT
27
Cumulative Pg(Ox) 700-1900 PDT
28
CO conc. at 900 PDT in LA inversion breaks up
2 hours later in CAMxis timing of emissions
wrong?
29
Cumulative P(OH) 700-1900 PDT, Aug 3.
30
H2O at 1200 PDT
31
contribution of O1D to OH initiation,
cumulative for Aug 3.
32
HO2 initiation, cumulative for Aug 3.
33
RO2 radical initiation, cumulative for Aug 3.
34
Reactions of NO3 O3 with isoprene, cumulative
for Aug 3.
35
Reactions of OH with isoprene, cumulative for Aug
3.
36
Total new radical initiation, Layer 1, cumulative
for Aug 3.
37
Total OH Production, Layer 1, cumulative for Aug
3.
38
HNO3 mixing ratio, 2400 PDT, Aug 5.
39
HNO3 produced by OHNO2, Layer 1, cumulative for
Aug 5.
40
HNO3 produced by OHNO2, Later 3, cumulative for
Aug 5.
41
HNO3 produced by N2O5H2O, cumulative for Aug 5.
42
E-W Slice through LA, cumulative for Aug 5.
43
Fraction HNO3 of total NOz, cumulative for Aug 5.
44
Net Production of PAN, cumulative for Aug 5.
45
Production of organic nitrates, cumulative for
Aug 5.
46
Total Production of NOz, cumulative for Aug 5.
47
Ox production efficiency per NOx, cumulative for
Aug 5. (Note regions of gray within red are
areas in which P(NOz) is negative).
48
Indicators to Evaluate O3 Sensitivity

• Indicators based on HNO3 or NOz may fail in CAMx
  simulations due to large contribution of N2O5H2O
  to P(HNO3).
• Alternative Use indicators based on radical
  propagation efficiency, O3 is VOC sensitive for
• HO2NO gt 93
• OHHC lt 80

49
Indicator of O3 sensitivity HO2NO
(cumulative for Aug 5).
50
Indicator of O3 sensitivity OHHC (cumulative
for Aug 5). (Note colormap is inverted)
51
Conclusions

• Minor problems with emissions, vertical
  dispersion and time zone need to be corrected
  before full evaluation.
• More serious issue w.r.t. N2O5 chemistry.
• Uncertainty in fate of NOx is a critical issue
  for O3 sensitivity and weekend effects.
• Validation of HOx budgets is equally important.

52
Recommendations

• Should adopt an up-to-date mechanism
• SAPRC99, CB4-99, RACM2.
• Use NOy data to better characterize N2O5
  chemistry and NOx fate.
• Use sensitivity studies to evaluate effects of
  uncertainty in N2O5 chemistry.