Evan M. Shipp
2088 Lester Avenue
Fresno, CA 93720
Jessica Hafer
San Joaquin Valley APCD
1990 E. Gettysburg Avenue
Fresno, CA 93726
April 7, 2008
Subject: Comments on the Proposed 2008 PM2.5 Plan
Dear Jessie:
As in the past, thank you for the opportunity to submit comments on the proposed air quality plan. As you know, I have been involved in writing plans for some time, and I know how much work is involved. You and your colleagues should be commended for producing a plan that is based on innovative science. There are still some outstanding technical issues from my comments on the first and second draft of this plan that could affect the amount of reduction needed to achieve the NAAQS. Attached are my comments.
Sincerely,
______
Evan M. Shipp
Comments on 2008 PM2.5 Plan – Proposed (March 13, 2008)
Evan M. Shipp (April 97, 2008)
I have reviewed the Proposed San Joaquin Valley 2008 PM2.5 plan and have the following comments. First, I would like to thank the technical staff of the ARB and SJVAPCD. They have done a good job in providing documentation of most of the technical milestones in the plan.
In the recent past, I have commented extensively on the technical issues involved in the modeled attainment demonstration. For completenesss and reference those comments are attached. Outstanding issues that were not addressed adequately in the District’s response to comments in the new proposed plan include:
· Missing data when meteorological conditions were strong enough to cause exceedances of the NAAQS
· Failure to include intermediate year predictions of PM2.5 concentrations
· Failure to recognize alternative methods of analysis under the weight of evidence provisions of the EPA modeling guidelines
· Assessment of 2007 PM2.5 Data that shows numerous exceedances of the NAAQS have occurred and should be analyzed to protect near term public health
· Assessment of meteorological representativeness of the 2004-2006 period that sets the baseline design value to determine whether worst weather conditions could produce more exceedances of the NAAQS in the future
· Inclusion of EPA model guidance required photochemical model evaluations
· District contradiction of using meteorological indicators and then stating that they are invalid
· Poor documentation of the relative reduction factors, and lack of documentation on how the photochemical was used to derive these
These general comments are detailed below:
1. Missing Reference Method Data During Periods When Meteorology and Non-Reference Monitors Show High PM2.5 Concentrations
As stated in previous comments on the first and second draft of this plan, missing data during periods of weather stagnation that are conducive to high PM2.5 levels should be considered when calculating the design value. This should be done under the weight of evidence provisions of the PM2.5 modeling guidelines. Failure to address this issue will result in analyses that underestimate the amount and timing of reductions needed to achieve the NAAQS.
As detailed in my comments on the second draft of the PM2.5 plan, (attached) BAM and meteorology show that there is good evidence from weather indicators that concentrations are much higher than characterized in the plan. Missing data during strong meteorological events that provide PM2.5 concentrations have not been addressed in the plan. The District dismisses the correlation of PM2.5 to these weather parameters (stability index) as being poor; however, they use these indicators of meteorological potential to produce PM2.5 forecast. In addition, as referenced in the PM10 plan, the indicators were developed and presented by me in a peer-reviewed paper for AWMA (Shipp, 1995). The correlation of these stability parameters to PM2.5 and winter PM10 is about an r of 0.7, which is considered good for meteorology to air contaminant correlation. In addition, the SJVAPCD has provided numerous CRPAQS analyses that use these meteorological indicators. Therefore, the District should apply these to assess the meteorological potential of the days where the data is missing and make adjustments to the design values. Again, this would be allowed under the weight of evidence provisions of the EPA guidelines. The District owes it to the citizens of the San Joaquin Valley to do a thorough job assessing the baseline concentrations to start the emissions reductions, and follow the letter and spirit of the law to protect public health, by analyzing the days with missing data that would be unhealthy.
2. Inclusion of Intermediate Year PM2.5 Predictions
To demonstrate reasonable further progress the plan needs to show that progress toward attainment will be seen in PM2.5 concentrations in the intermediate years before attainment. The plan only shows the attainment year concentrations. Since the plan indicates that most emissions reductions occur close to the attainment year, the modeling needs to show steady progress toward attainment.
3. Inclusion of 2007 PM2.5 Data
The District dismisses the comment made regarding the 2007 PM2.5 season data by saying that the data is not available yet. Because the real-time data indicates that numerous exceedances of the NAAQS occurred in 2007, every effort to include the data in the plan should be done. As an experienced air quality analyst, I have been able to have ARB expedite filter data when necessary. If the District cannot produce reference method data, they need to address fully why BAMS data is not included. I have shown that the correlation of BAMS to FRM’s is very good and the analysis should be done under the weight of evidence provisions of the EPA modeling guidelines. The bottom line is that the 2007 data shows that the SJV has levels well above the NAAQS, and those emissions levels have not been brought down sufficiently to protect public health. The design values need to be adjusted up by inclusion of the 2007 data.
4. Assessing the Representativeness of the 2004-2006 Period
In accordance with EPA guidelines, and as outlined in my comments on the 2nd draft of this plan, the meteorology for the time period that the design value was set needs to be assessed. My analysis for the 2nd Draft of the plan indicates that meteorological potential for PM2.5, during the assessment years 2004-2006, was lower than in past years, and a return to worst case conditions, such as occurred in 1999-2002 would bring PM2.5 concentrations to well over the 24 hour and annual NAAQS. The response to comments that the District assessed 100 years of data without providing any documentation on that analysis is completely inadequate. Again, the District owes it to the citizen’s of the SJV to do a thorough job on insuring that emissions reductions on worst-case meteorological days will still bring the valley into attainment of the 1997 NAAQS.
5. Inclusion of EPA Required Evaluation of the Photochemical Model
Since the proposed plan does not include a full model photochemical model evaluation, it does not comply with EPA modeling guidelines and regulations. The model evaluation is meant to assess the uncertainty of the photochemical modeling and can result in designation of alternate controls strategies depending on that uncertainty. A full list of the photochemical model evaluation metrics is included in the plan, but there is no mention of the specific model evaluation for the plan. Uncertainty assessment and discussion is an integral part of specifying the amount and chemical species to control.
6. Documentation of the Relative Reduction Factors in the Plan and Demonstrating Reasonable Further Progress
The process used to calculate the RRF’s and show attainment of the NAAQS is undocumented. It is unclear whether the RRF’s calculated from the photochemical model were the result of numerous runs at various percentage cuts in the inventory or just the 50% cuts as has been used in the past. Because photochemistry is non-linear, the RRF’s should be based on multiple cuts in the inventory. In addition, a run containing the attainment inventory should be documented. According to the ARB, the District was given RRF’s based on a future and base year model runs. This is not sufficient to demonstrate reasonable further progress toward attainment. A majority of the emissions reductions in the plan occurs at the end attainment year, and therefore the modeling system using the base and end year inventory is not sufficient
The plan does not contain detailed documentation of the RRF’s that were used to scale the design values and show attainment. The District answered this concern in the proposed plan by stating that the files were too big to download and distribute. It is the District’s and/or ARB’s responsibility to condense this information into an easily understandable format showing the output of the photochemical model before and after control for each chemical species and how that affects the design value. If multiple cuts are run, the District should include nitrate, sulfate, and VOC isopleths.
An evaluation of the result of control of individual NOx, VOC, EC, and SO2 needs to be performed. In addition, the District should show alternative species strategies and give more detail on the way the model was used to show that a NOx only strategy is viable. As in the comment above on analyzing intermediate years, the plan needs to provide documentation on speciated RRF’s and predicted species concentrations for intermediate years.
In addition to the more general comments above, I am providing specific comments (labeled S-number) on the text in the plan:
S-1. Page 3-2 Paragraph 1
The paragraph should state that PM2.5 emissions could be retained in the SJV air basin for up to one month. This was seen in the 2000-2001 CRPAQS episodes, and it is relevant to the assessment of the representativeness of episodes. Long-term episodes are more difficult to control than short term ones. In addition, assessment of representative meteorology should include how long episodes last.
S-2. Page 3-2 Last Paragraph
The description of subsidence does not account for the humidity increase at the surface as water vapor is trapped by the dry warm air above it. During major winter PM2.5 events, the descending air does not usually penetrate to the surface and a subsidence inversion is formed.
S-3. Page 3-19 first paragraph in 3.3.3
The District determined that the annual standard is more stringent than the 24-hour NAAQS. This needs to be documented.
S-4. Page 3-19 second paragraph in 3.3.3
The statement in this paragraph supports the idea that the District does not have enough data to determine that the area is in compliance with the PM2.5 annual and 24 hour NAAQS. The plan states that the District is in compliance with the 24 – hour NAAQS.
S-5. Page 3-20 section 3.3.4
The paragraph states that the District will perform the SMAT analyses. ARB did perform these analyses.
S-6. Page 3-22 Section 3.4 first paragraph
As stated above, without rigorous analysis and EPA approvals the text should not state that some sites already comply with the NAAQS.
S-7. Page 3-23 Top
As stated in previous comments, the 2004-2006 period had missing data and better dispersive meteorology and therefore is non-representative. Method 2 projections may be misleading.
S-8. Page 3-23 Regarding Lack of SMAT Documentation
The District and ARB should provide detailed documentation on how the SMAT calculations were derived from the photochemical model.
S-9. Page 3-27 Prior Accuracy of the PM10 Methodology
The District is only in attainment of the PM10 NAAQS due to flagging of data that is over the NAAQS by the Natural Event Action Plans. These exceedances were not predicted by the PM10 plan.
S-10. Page 3-30 Air Monitoring and Trends
As stated above, the 2007 FRM data should be analyzed in the plan. My comments on the 2nd draft show that the 2007 data would change design values. ARB needs to expedite the processing of the 2007 FRM data and include it in the plan. It addition, the weight of evidence should include BAMS data which could be adjusted by the method portrayed in my comments on the 2nd Draft of this plan.
S-11. Page 3-31 Weight of Evidence
The total attainment demonstration approaches using CMB and PMF are not independent checks because they still depend on design value calculations that are flawed.
S-12. Page 9-3 Last paragraph
As stated above, PM10 is dependent on NEAP’s and non-representative years of 2004-2006.
S-13. Page A-1
Because of the exceedances in 2007, the table should show 2005-2007 design values.
S-14. Page A-2
Quarter completion problems should be investigated. Is missing data due to aerosol loading during poor meteorological conditions?
S-15. Table A-3 and A-8
Missing data in 2004 at the highest sites during the highest quarters is shown in table A-3 and illustrates the point regarding the loss of high concentration data. Also, table A-8, shows the loss of data from 1999-2003 during the worst meteorological years.
S-16. Page A-14 through A-24 Trends Charts
As stated in previous comments, the trend charts need to reflect fluctuations in meteorology by using adjustment methods and/or analyzing meteorological parameters that correlate to PM2.5.
S-17. Page F-61 Regional Modeling
Here the protocol states that regional model evaluation will be included. As stated above, this evaluation is not included in the plan.
S-18. Figure H-4
Without a meteorological evaluation, this chart is not definitive. The 2001 meteorology was much worst than it was in 2006. This may be the difference in the meteorological trend and not necessarily emissions.
S-19. Bottom of page H-8
Again, as stated above, these comparisons with 1999-2001 and 2004-2006 need a meteorological evaluation.
S-20. Figure H-8
The upward trend in 1998-2001 is caused by strong meteorology that has not returned in the 2004-2006 period the plan is based on.
S-21. Figure H-14 Speciated Trends
The species trends look pretty flat, and may not account for the entire PM2.5 trend. This should be explained.
S-22. Page H-18
Again, the emissions trend looks pretty flat and would not account for the reduction in PM2.5 seen between 2001-2005.
S-23. Page H-23
The PMF and CMB model corroborate, but this is only a piece of the attainment process. More corroboration is needed for other characteristics of the attainment demonstration. These might include other photochemical modeling, meteorological adjusted trends, and alternate design value calculations.