  SEQ CHAPTER \h \r 1 

TECHNICAL SUPPORT DOCUMENT

for

The Extreme One-Hour Ozone Attainment Plan Modeling

 for the 

San Joaquin Valley Nonattainment Area

EPA Region 9

September 2008



TECHNICAL SUPPORT DOCUMENT

The Extreme One-Hour Ozone Attainment Plan Modeling

for the San Joaquin Valley Nonattainment Area

I.  INTRODUCTION

 

This document provides a discussion of the modeling analysis provided in
the San Joaquin Valley Air Pollution Control District’s Extreme Ozone
Attainment Demonstration Plan, October 8, 2004.  It discusses the
adequacy of the analysis, and how the analysis meets EPA’s modeling
requirements and performance goals. The ozone air quality modeling in
the plan provides the basis for the attainment demonstration.  To be
approved, the State submission must show that plan demonstrates
attainment of the 1-hour ozone standard as expeditiously as practicable,
but not later than 2010.

II. EVALUATION OF THE MODELING DEMONSTRATION 

	As required by the Clean Air Act, California has used photochemical
grid modeling in its demonstration that the control strategy for the San
Joaquin Valley area will achieve attainment of the 1-hour ozone standard
by 2010.   

	The modeling analysis performed by CARB that is intended to simulate
the air quality levels that result from the emissions that occur in the
San Joaquin Valley, as a basis for determining the level of routine
emissions that can be allowed in the area and provide for  attainment.

	The modeling is based on a multi-stage analysis.  First, appropriate
episode days that represent days conducive to ozone formation in the San
Joaquin Valley are selected. Second, a modeling platform, including
photochemical model, domain, number of vertical layers, and initial and
boundary conditions most appropriate for the San Joaquin Valley area,
and other factors are selected.  Third, the air quality on the selected
episode is simulated, using the appropriate day-specific emissions. 
Finally, the performance of the selected modeling platform on each of
the modeled episode days is analyzed and compared to EPA performance
criteria.  When the modeling analysis is determined to be acceptable, it
then forms the basis for the attainment demonstration.  

The state’s analysis is provided in the plan in Chapter 5: Future
Ozone Air Quality and in Appendix D: Modeling and Attainment
Demonstration.  These documents provide the basis for the modeling that
was reviewed by EPA and compared to EPA guidance for each of these
factors. 

A. Episode Characterization 

	EPA’s Guideline sets forth a recommended procedure for selecting
ozone exceedance episodes appropriate for conducting a modeling
demonstration.  

	EPA guidance recommends that “a minimum of 3 primary days should be
simulated.”  EPA guidance also recommends that ”in addition to
considering the magnitude of the highest observed daily maximum
concentration in making this choice, data availability and quality,
model performance, availability of regional modeling analysis,
pervasiveness, frequency with which observed meteorological conditions
coincide with exceedances, and duration of observations greater than
0.12 ppm may be considered.”

	In the selection of the appropriate episodes, the state focused
primarily on episodes which occurred during the time period of the year
2000 Central California Ozone Study (CCOS), because of the enhanced data
availability and quality. The CCOS study had extensive measurements of
ozone, hydrocarbon, oxides of nitrogen, and other chemical species. The
data obtained during this study provided a wealth of detailed, high
sensitivity information about the SJV area’s atmospheric chemistry
with data on ozone, carbon monoxide, NO, and NO2.  More specific
information regarding the CCOS field study is available in Appendix D of
the Plan.

 

Five candidate episodes for the SIP modeling analysis were considered
for this modeling exercise. Three episodes that were measured during the
2000 Central California Ozone Study (CCOS) were considered: June 17-18,
July 31-August 2, and September 17-21. In addition, two episodes that
occurred outside the CCOS time period were considered: July 11-12, 1999
and August 6-16, 2002.  

	

Each of the three primary CCOS episodes was initially modeled by one the
participating CCOS agencies or their contractors. Of the three CCOS
episodes, the July 31- August 2, 2000 episode was considered to be the
most representative of the transport and formation of ozone in the San
Joaquin Valley.  The September 17-21 episode was also considered, but
the model performance did not meet the performance criteria, so it was
not pursued.

 	While two episodes that occurred outside the CCOS time period were
considered: July 11-12, 1999 and August 6-16, 2002, the CCOS episodes
had superior data availability and quality.    	

EPA believes that the extended episode from July 30-August 2, 2000 is an
acceptable episode for development of the 1-hour ozone standard
attainment plan.  The episode encompasses 4 exceedance days for the
study area (San Joaquin Valley, Sacramento and the San Francisco Bay
Area) and 2 exceedance days in the San Joaquin Valley.  It also has the
advantage of being during the CCOS study’s intensive data collection
period.  

B. Photochemical Modeling

CARB used the most current Comprehensive Air Quality Model with
Extensions (CAMx) photochemical grid model (which is based on well
established treatments of advection, diffusion, deposition, and
chemistry similar to the Urban Airshed photochemical grid Model (UAM),
for the attainment demonstration.  EPA believes that this model is an
appropriate model for use to develop the SJV SIP attainment
demonstration. 

C.  Modeling Domain

The modeling domain for the SJV extends from the Los Angeles County in
the south to the California/Oregon border in the north, and from the
Pacific Ocean into Nevada in the east and is shown in Figure D-6 (p.
D-41) of the plan.

EPA guidance states that:

	

It is recommended that the domain’s downward boundary be sufficiently
far from the CMSA/MSA that is the principal focus of the modeling study
to ensure that emissions from the CMSA/MSA occurring on the primary day
for each selected episode remain within the domain until 8:00 pm on that
day.  The extent of upwind boundaries will depend on the proximity of
large upwind source areas and the adequacy of techniques used to
characterize incoming precursor concentrations.  Large upwind emission
source areas should be included in the modeling domain to the extent
practicable.  Also, if large uncertainty is anticipated for the domain
boundary conditions, the upwind boundaries should be located at a
distance sufficient to minimize boundary effects on the model
predictions in the center of the domain.

EPA believes that the domain selected for the episode is sufficiently
large to characterize incoming precursors and minimize boundary effects
for the San Joaquin Valley.

D.  Horizontal Grid Cell Size

	The horizontal Grid cell size used for the photochemical modeling
exercise in the SIP was 4km by 4 km.   

	EPA recommends that the size of the horizontal grid cells should not be
greater than 5km by 5 km, and grid cell sizes smaller than 2km by 2 km
are not recommended because of potential model formation inconsistencies
for those grid sizes.  The grid cells should be small enough to reflect
emission gradients and densities in urban areas, particularly those
resulting from large point sources and major terrain or water features
that may affect air flow.

	EPA believes that the 4km by 4 km grid cell size is appropriate for
this modeling application.

E. Vertical Structure 

The vertical structure for the San Joaquin Valley modeling exercise
consists of 16 layers, and is shown in Table 1: Vertical Structure,
shown below.

Table 1

Vertical Structure of the Air Quality Modeling Domain

for the July/August 2000 episode based on the MM5

and CALMET meteorological models

Layer Number	MM5	CALMET

	Thickness

(meters)	Height

meters above ground level)	Thickness (meters)	Height (meters above
ground level)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

	24

26

56

66

74

133

161

182

205

232

265

356

488

666

926

1140	24

50

106

172

246

379

540

722

927

1159

1424

1780

2268

2934

3860

5000

	20

40

40

100

100

100

200

200

200

500

500

500

500

500

500

1000	20

60

100

200

300

400

600

800

1000

1500

2000

2500

3000

3500

4000

5000





EPA guidance recommends that a minimum of five vertical layers be used
in the modeling study, with at least three layers above the morning
mixing height.  EPA encourages greater detail in the grid cell size
particularly in modeling domains containing complex terrain or
land/water interfaces. 

EPA believes that the vertical structure meets EPA criteria and is of
sufficient resolution for the San Joaquin Valley domain.

F.  Initial and Boundary Conditions

The initial and boundary concentrations were based on a combination of
USEPA default values, South Coast Air Quality Management (SCAQMD)
District default values, and measurements taken during the CCOS study. 
EPA Guidance states that: 

To develop initial and boundary conditions, it is recommended that one
or more monitoring stations be sited upwind of the central urban area
along the prevailing wind trajectories that give rise to the exceedance.


The sampling and analysis program should provide data to calculate
hourly values for ozone, NO, NO2 and speculated hydrocarbons.

At the inflow boundaries, air quality data at the surface and aloft
should be used whenever available to specify the boundary conditions.

The CCOS study provided a number of surface sites that provided data to
develop initial and boundary conditions for ozone, NO, NO2 and speciated
hydrocarbons.

The top boundary concentrations developed for the domain were based on
the clean profile defined by EPA guidance, with the exception of the
ozone concentration of 70 ppb, which was based on ozonesonde
measurements collected at Granite Bay and Parlier during the CCOS study.



Table 2

Initial and Boundary Conditions for the CCOS Domain

Compound	Initial Conditions	Boundary Conditions

	

Non-SJV	

SJV	Top	Pacific (West and  part of southern)	Overland  Boundary

Reactive Hydrocarbons	60.7 ppbC	121.4 ppbC	22 pbbc	22ppbc	60.7 ppbC

Carbon Monoxide	200 ppm	200 ppm	350 ppm	350 ppm	200 ppm

Oxides of Nitrogen	2.0 ppb	2.0 ppb	2.0 ppb	 	2.0 ppb

Ozone	40 ppb	40 ppb	70 ppb	40 ppb	40 ppb

  

EPA believes that the EPA guidance criteria was met in choosing initial
and boundary conditions supported by measurements from the  CCOS field
study data, use of EPA default values, and SCAQMD default values.  EPA
also believes that, because of the relatively high level of the ozone
level used for the top boundary concentration of the domain, future
applications should determine the sensitivity of the model performance
to the ozone top boundary concentration value.

Model Performance Evaluation Data

EPA recommends that the data base used in the attainment demonstration
modeling meet the requirements for the enhanced ozone monitoring system
promulgated by EPA.

EPA believes that the CCOS study, described above, meets the objectives
of the EPA enhanced ozone monitoring system. 

H.  Base year Model Performance Evaluation

Model performance was evaluated for nine sub-regions of the modeling
domain, including the Fresno and Bakersfield areas in the San Joaquin
Valley, and well as for the Sacramento Delta and San Francisco Bay
Areas. 

EPA guidance recommends the following mathematical formulations be
applied as measures for model performance evaluation for the 1-hour
ozone standard:

Unpaired highest prediction accuracy – The measure quantifies the
difference between the highest observed value and highest predicted
value over all hours and monitoring stations.

Normalized bias test – This test measures the models ability to
replicate observed patterns during the times of day when available
monitoring and modeled data re most likely to represent similar spatial
scales.

Gross error of all pairs above 60 ppb - In conjunction with bias
measurements, this metric provides an overall assessment of base case
performance and can be used as a reference to other modeling
applications.  Gross error can be interpreted as precision.

EPA guidance recommends that the statistical performance be compared
with the following ranges:

Unpaired highest prediction accuracy: +/- 15-20% (0.8 -1.2),

Normalized bias test: +/5-15%, and 

Gross error of all pairs above 60 ppb: 30-35%

The model performance for each of these performance criteria is shown
for Fresno and Bakersfield, below, for the CAMx/CB4 and CAMx/SAPRC99f.  


Model performance with CAMx /CB4 is shown for Fresno in Table 3:  Fresno
Model Performance, below, and Bakersfield in Table 4: Bakersfield Model
Performance.   EPA performance criteria are met for Unpaired Peak ratio,
Normalized Bias and Gross Error for each of the episode days.



Table 3

Fresno Model Performance – CAMx /CB4

Day	Observed peak	Simulated Peak	Unpaired Peak Ratio	Normalized Bias %
Gross Error %

EPA Guideline

Performance Goals

	0.8-1.2	+/- 15%	35

July 29, 2000

July 30, 2000

August 1, 2000

August 2, 2000	129

118

118

131	130

128

124

127	1.01

1.09

1.05

0.97	-08

+03

-09

-10	18

17

20

22



Table 4

Bakersfield Model Performance – CAMX/CB4

Day	Observed peak	Simulated Peak	Unpaired Peak Ratio	Normalized Bias %
Gross Error %

EPA Guideline

Performance Goals

	0.8-1.2	+/- 15%	35

July 29,200

July 30,2000

August 1, 2000

August 2, 2000	128

115

116

151	123

119

114

129	0.96

1.04

0.98

0.85	-12

-12

-15

-13	20

18

19

21



Model performance with CAMx/SAPRC99f is shown for Fresno in Table 5:
Fresno Model Performance – CAMx/SAPRC99f, below, and Bakersfield in
Table 6: Bakersfield Model Performance – CAMx /SAPRC99f. EPA
performance criteria are met for Unpaired Peak ratio, Normalized Bias
and Gross Error for each of the episode days.  The CAMX/SAPRC99f
simulated peaks are higher on most days than the CAMx/CB4, so therefore,
the normalized bias is smaller.

Table 5

Fresno Model Performance – CAMx/SAPRC99f

Day	Observed peak	Simulated Peak	Unpaired Peak Ratio	Normalized Bias %
Gross Error %

EPA Guideline

Performance Goals

	0.8-1.2	+/- 15%	35

July 29, 2000

July 30, 2000

August 1, 2000

August 2, 2000	129

118

118

131	144 

138 

132 

137 	1.12

1.17

1.12

1.04	-03

+06

-05

-02	17

18

21

22



Table 6

Bakersfield Model Performance – CAMx/SAPRC99f

Day	Observed peak	Simulated Peak	Unpaired Peak Ratio	Normalized Bias %
Gross Error %

EPA Guideline

Performance Goals

	0.8-1.2	+/- 15%	35

July 29, 2000

July 30, 2000

August 1, 2000

August 2, 2000	128

115

116

151	149 

132 

120 

140	1.16

1.15

1.04

0.92	-02

-07

-10

-07	25

20

19

20



EPA believes that the model performance meets the performance criteria
for the 1-hour ozone standard, and is acceptable. 

I. Base Case Sensitivity to Natural Sources.

	Several simulations were conducted to determine the sensitivity of the
modeling analysis to variations in natural source emission estimations. 
To assess the sensitivity of the model to biogenic emissions, two
emission scenarios were evaluated; one scenario zeroing out the biogenic
emissions, and one scenario increasing the biogenic emission estimations
by 25%.  The sensitivity of the 1-hour ozone values to these changes is
shown for the modeling results in Bakersfield in Table 7: Sensitivity to
Natural Sources in Bakersfield, below.   

	An additional simulation was conducted to determine the sensitivity of
the model to wildfire emissions.  The results of the modeling scenario
with wildfire emissions removed from the emission inventory are shown
below.   The ozone values for this simulation are up to 7% less in the
Bakersfield area than ozone levels in the base case, indicating a fairly
substantial wildfire impact on ozone during the July, August 2000
episode.   

Table 7

Sensitivity to Natural Sources in Bakersfield

Day	Base case	No Fires

	+ 25% Biogenics

	No Biogenics



	ppb	ppb	% Change	Ppb	% Change	Ppb	% Change

July 31

August 1

August 2	129

119

135	0

-2

-9	0

-2

-7	2

1

1	2

1

1	-8

-8

-10	-6

-7

-7



III.  CONCLUSION 

EPA believes that each of the components in the modeling analysis is
acceptable.  The appropriate episode days that represent day conducive
to ozone formation in the San Joaquin Valley are selected.  The modeling
platform, including photochemical model, domain, number of vertical
layers, and initial and boundary conditions most appropriate for the San
Joaquin Valley area, and other factors are appropriate.  Finally, the
performance of the selected modeling platform on each of the modeled
episode days meets the performance criteria specified in EPA guidance. 
EPA believes that the modeling analysis in the plan is acceptable, and
that it, therefore, forms an appropriate basis for the attainment
demonstration.

References

Guideline for Regulatory Application of the Urban Airshed Model,
EPA-450/4-91-013 (July 1991); “Guidance on Use of Modeled Results to
Demonstrate Attainment of the Ozone NAAQS,” EPA-454/B-95-007 (June
1996);

Guidance for the 1-hour Ozone Nonattainment Areas that Rely on
Weight-of-Evidence for Attainment Demonstrations, Mid-Course Review
Guidance (March 28, 2002);

Guidance for Improving Weight-of-Evidence Through Identification of
Additional Emission Reduction Not Modeled (Nov 99).

“Guideline for Regulatory Application of the Urban Airshed Model,”
EPA-450/4-91-013 (July 1991).

	  Guideline for Regulatory Application of the Urban Airshed Model,
EPA-450/4-91-013 (July 1991); “Guidance on Use of Modeled Results to
Demonstrate Attainment of the Ozone NAAQS,” EPA-454/B-95-007 (June
1996);

	Guidance for the 1-hour Ozone Nonattainment Areas that Rely on
Weight-of-Evidence for Attainment Demonstrations, Mid-Course Review
Guidance (March 28, 2002);

	Guidance for Improving Weight-of-Evidence
Throug⁨摉湥楴楦慣楴湯漠⁦摁楤楴湯污䔠業獳潩⁮敒畤
瑣潩⁮潎⁴潍敤敬⁤丨癯浥敢⁲㤱㤹⸩഍ 錠畇摩汥湩
⁥潦⁲敒畧慬潴祲䄠灰楬慣楴湯漠⁦桴⁥牕慢⁮楁獲敨
⁤潍敤ⱬₔ偅ⵁ㔴⼰ⴴㄹ〭㌱⠠畊祬ㄠ㤹⤱഍ 䤠⹤഍
̍഍ഄ̍഍ഄ഍഍ግ䅐䕇†കഀ഍倓䝁⁅ᐠᔱ

഍഍഍ግ䅐䕇†ㄔᔲ഍ግ䅐䕇   11 

