                                       
                       Technical Support Document (TSD)

 Providing Responses to Comments Regarding the Environmental Protection Agency's Proposed Approval of the Attainment Demonstration for the Allegheny County PM-2.5 Nonattainment Area, under the 2012 National Ambient Air Quality Standard.
                                       

                           TSD Prepared October 2020
                           Timothy A. Leon Guerrero
                      Air Quality Analysis Branch, 3AD40
                U.S. Environmental Protection Agency, Region 3
                               1650 Arch Street
                       Philadelphia, Pennsylvania 19103
                                       
                                       
                                       
                 ____________________________________________
                        Reviewed by Alice Chow, Chief,
                      Air Quality Analysis Branch (3AD40)
                                       
                                       
                               _________________
                                  Date Signed



Purpose of the Technical Support Document

This Technical Support Document (TSD) describes the Environmental Protection Agency's responses to the comments received for the final attainment demonstration prepared by the Allegheny County Health Department and submitted by Pennsylvania to EPA Region 3 in early February of 2020 for the Allegheny County, PA PM-2.5 nonattainment area. EPA proposed to approve elements of a state implementation plan (SIP) revision submitted by the Pennsylvania Department of Environmental Protection (PADEP) on behalf of the Allegheny County Health Department (ACHD) to address Clean Air Act (CAA or "the Act") requirements for the 2012 annual fine particulate matter (PM-2.5) national ambient air quality standards (NAAQS or "standards") in the Allegheny County Moderate PM2.5 nonattainment area ("Allegheny County area"). The SIP revision was submitted on September 30, 2019 and published in the Federal Register (85 FR 35852) on June 12, 2020
During the public comment period, EPA received multiple comments from the Clean Air Council pertaining to specific parts of Allegheny County's PM-2.5 SIP. This TSD was developed to provide additional details for EPA's responses and to allow for the inclusion of figures and tables that are generally not capable of being included in the official Federal Register notice.
Specific comments from the Clean Air Council that this TSD will provide responses to include meteorological issues including average temperature, precipitation and vertical temperature inversion strength with respect to Allegheny County's base year (2011) selection and the assertion that Allegheny County's modeling analysis under estimates projected 2021 PM-2.5 concentrations because the base year meteorology is not conductive to PM-2.5 formation (specifically, fewer days with vertical temperature inversions).
This TSD includes a reexamination of the meteorological data from the Pittsburgh International Airport (and Allegheny County Airport) and Allegheny County's daily calculated inversion/ventilation rates included in its annual summary of inversion strength. It will also include a summary of Allegheny County's projected modeled design values, model generated tend lines and the most recent (2017-19) PM-2.5 design values for all monitors in Allegheny County to assess how well the model is predicting current county design values and to some extent the chances of all monitors complying with the annual and PM-2.5 NAAQS. Additional responses to specific comments are contained within EPA's formal response to public comments.

EPA's Analysis of Meteorological Trends in the Allegheny County, PA PM-2.5 Nonattainment Area

EPA downloaded daily average temperature and precipitation values for the Pittsburgh International Airport and the Allegheny County Airport using the Pennsylvania State Climatological website. Thirty-year averages using daily average temperature and daily precipitation information were processed using R, an open-source language and environment for statistical computing and graphics. This was done to double-check the information submitted in support of Clean Air Council's comment regarding the unrepresentativeness of the 2011 base year utilized in Pennsylvania's PM-2.5 SIP modeling demonstration.
In addition to the daily temperature and precipitation information, Clean Air Council also included information on inversion strength compiled by the Allegheny County Health Department as part of its daily air quality forecasting activities. Allegheny County's SIP recognizes the contributions local inversions have in the county's major river valleys, especially near the Liberty PM-2.5 monitor in southern Allegheny County. Clean Air Council believes Allegheny County's annual inversion day counts indicate that 2011 is unrepresentative due to the low number of inversion days when compared to long-term averages.

Pittsburgh International Airport's Average Daily Temperature and Precipitation (1990-2019):  EPA downloaded and processed daily average temperatures and precipitation values from the Pittsburgh International Airport (PIT) located in Findlay and Moon Townships, Allegheny County. PIT has both an Automated Surface Observing System (ASOS) unit, which are automated sensor suites designed to serve meteorological and aviation observing needs, and twice daily balloon-launched radiosondes ("radar wind-sonde"), whose position is tracked after release to give wind speed and direction information along with pressure, temperature and dew point (for relative humidity) measurements as the instrument ascents vertically in the atmosphere. This website information was used to develop a climatological data set that is representative of Allegheny County, PA.
Table 1 shows the 30-year monthly average temperatures for PIT over the 1990-2019 time period. The baseline year (2011) for Allegheny County's PM-2.5 SIP modeling demonstration is highlighted in pink while 2009, 2010, 2012, and 2013 are highlighted in gray. These years are highlighted because they are also included in the model attainment test calculation indirectly since the baseline design values for modeling purposes are "centered" around the 3-year base-year (2011) design value. The model attainment test takes into account all years in which the base year (2011) is included in the calculation of a (3-year) design value. Thus 2011 is included in design value calculations for 2009-11, 2010-12, 2011-13. Averages and standard deviation values for all months along with the annual averages are included in this table and are also highlighted in gray.
Base year (2011) average temperatures at PIT are generally within 1 standard deviation of the 30-year average for most months. The average annual temperature for 2011 is also within 1 standard deviation of 30-year annual average. This implies that the 2011 base year Allegheny County selected is not outside what would be expected for normal temperature distributions across the 30-year period (1990-2019). To be considered "unrepresentative", EPA would expect the average temperatures to be well outside the range indicated by the 30-year average temperature, +/- the standard deviation. Generally speaking, a true "outlier" would be several standard deviations from the calculated mean. While this might be the case for several months in the base year (January, July and November for example), the vast majority of PIT's 2011 monthly temperature averages fall well within the range they would be expected; +/- 1 standard deviation.
Table 2 is constructed similarly to Table 1 expect it shows monthly and annual precipitation totals for PIT. Base year 2011 precipitation totals at PIT were above the 30-year mean but again within 1 standard deviation of the constructed 30-year data (1990-2019).  EPA acknowledges the 2011 base year has periods of unusually high precipitation, especially in the February through May time period, but there are also months in the base year that were unusually dry (June and July). While precipitation could be expected to remove some PM-2.5 concentrations, EPA submitted an analysis during Allegheny County's public comment period that shows local PM-2.5 concentrations may be weakly correlated with humidity levels. Higher humidity levels may be present during the base year 2011 due to periods with higher than average precipitation in Allegheny County and could paradoxically lead to higher PM-2.5 concentrations.


Table 1.  Pittsburgh International Airport 1990-2019 Average Daily Temperatures
                                       


Table 2.  Pittsburgh International Airport 1990-2019 Monthly Precipitation Totals
                                       

Clean Air Council also questioned the validity of using PIT for determining inversion strengths in all of the major river valleys in Allegheny County, especially in the Mon Valley where the Liberty PM-2.5 monitor is located. To test this assertion, EPA reconstructed the 30-year average temperatures from the Allegheny County Airport (AGC) located near West Mifflin, PA and several miles north and northwest of the US Steel Irving and Clairton Coke Works. The average monthly and annual temperatures along with the standard deviations for each month and the annual average are shown in Table 3. There does not appear to be any appreciable difference in both site's temperatures over the 30-year period. EPA recognizes both airports are located in the higher terrain of the Allegheny Plateau physiographic province; airports would not normally be placed within the steep river valleys of southwest Pennsylvania for flight safety reasons. None the less, the similarity of the 30-year average temperatures at PIT and AGC would seem to support that the daily inversion strength calculations Allegheny County make, which are in part based on the surface temperature measurements, are applicable over a wide area, or at least over areas that have similar terrain features.

    Table 3.  Allegheny County Airport 1990-2019 Average Daily Temperatures
                                       


Allegheny County Health Department reported Inversion Strength Values:  The Allegheny County Health Department releases daily reports on pollutant dispersion conditions across the county. These reports quantify vertical temperature inversion strength based upon morning upper air rawinsonde measurements taken daily at Pittsburgh International Airport (PIT) in Findlay and Moon Townships and predicted wind speeds to calculate a ventilation rate. Similar ventilation index calculations were made by the Pittsburgh National Weather Forecasting office in the past. Allegheny County prepares these air dispersion conditions and outlook reports to assist in predicting periods of poor air quality over the county. Local (vertical temperature) inversions can lead to prolonged episodes of poor air quality. The 1948 Donora, Pennsylvania Smog Event is a classic example of how vertical temperature inversions can trap emissions in confined river valleys which can lead to extremely poor air quality that impacts respiratory sensitive individuals. These extreme pollution trapping events, while no longer as serious due to emission control programs enacted under the 1990 Clean Air Act amendments, still occur in southwest Pennsylvania and are well correlated with episodes of poor air quality. A basic description of these types of events was included in Problem Statement and Meteorology sections of Allegheny County's PM-2.5 SIP document.
Allegheny County has summarized its analysis of inversion strength over the county in annual reports available on its website. The latest report summarizes information from 2008 through 2019. A 2018 report was referenced by the Clean Air Council in its comments to Allegheny County's PM-2.5 SIP. Table 4 is taken from Allegheny County's Allegheny County Surface Temperature Inversion Analysis dated March 9, 2020 (Rev.4/8/'20). Liberty's mean quarterly and 98[th]% 24-hour PM-2.5 concentrations are added to this table for comparison to the inversion summaries.
At first glance, it does appear that the inversion (day) count and PM-2.5 concentrations are roughly correlated; more inversion days appear to coincide with higher PM-2.5 concentrations at Liberty. Inversion strength (the difference in vertical temperatures) and inversion depth appear less correlated though lower inversion depths should lead to higher concentrations since any trapped emissions will have a lower volume of air in which to mix.
EPA recognizes Clean Air Council's claim that Allegheny County's use of 2011 as a base year would coincide with a lower PM-2.5 potential due to the unusually low number of inversion days during that year. The chosen 2011 base year's number of inversion days is outside the standard deviation for the 2008-19 period providing some statistical validity to Clean Air Council's comment. Inversion strength and depth, however, are firmly within 1 standard deviation of the means meaning these values are not that unusual. EPA also notes that Liberty's PM-2.5 concentrations are falling over this period (EPA has already established Liberty's declining PM-2.5 concentrations are statistically significant).

Table 4.  Summary of Allegheny County's Yearly Inversion Strength and Liberty PM-2.5 Values (ug/m[3])
                                       

Comparison of 2021 Model Projected and 2019 Monitor Design Values

This section examines the base year (2011) and projected year (2021) PM-2.5 design value concentrations produced by the Allegheny County modeling analysis. Modeled values for all monitors in Allegheny County, except for Liberty, were derived using the CAMx 1.33 km results and passing them through EPA's Model Attainment Test Software (MATS version 2.6.1). Future PM-2.5 design values, similar to ozone, are calculated using Relative Response Factors (RRF). This allows the model to be used in a relative sense; the RRFs are applied to the base year weighted monitor design value concentrations to produce the projected design values. Allegheny County used a combination of the 1.33 km CAMx results representing the chemically reactor PM-2.5 species and a local area analysis using AERMOD for the non-reactive PM-2.5 species for the Liberty monitor.
EPA also pulled Allegheny County's most recent design values for comparison with the projected 2021 PM-2.5 design values. A modeled PM-2.5 design value trend line was constructed by using the change in each Allegheny County monitor's base year and projected year design value (divided by the time in between or 10 years). This model projected trend line was then used to calculate a projected 2019 design value, which then could be compared to the monitor's actual PM-2.5 design value for 2019 to gauge how well the model is predicting.
Table 5 shows modeled and monitored annual PM-2.5 values for each monitor in Allegheny County. Each monitor's modeled 2011 base year design value, actual 2009-11 monitored design value (if valid), model projected 2021 design value, modeled trend line (difference between the projected and base year divided by 10 years, so it's ug/m[3] change per year), actual monitor 2017-19 design value and the projected model 2019 design value. Each monitor's monitor ID and what model or combination of models was used to construct the base year and projected year annual PM-2.5 concentrations is included in the table. The Liberty monitor is highlighted given it is the highest monitor in the Allegheny County, PA PM-2.5 nonattainment area.

Table 5.  Allegheny County Monitor Modeled & Monitored Annual PM-2.5 Concentrations
                                       

Allegheny County's modeled 2011 base year annual design values are generally lower than the actual annual monitor design values for the 2009-11 time period. This may be due to more favorable meteorological set up (fewer inversion days) as noted by the Clean Air Council or faults in the model's base year inventory. Liberty's CAMx/AERMOD developed modeled trend line doesn't appear to be all that different from the CAMx only trend lines for the other Allegheny County monitors (-0.24 ug/m[3] per year). While not shown, Liberty's CAMx only trend line was -0.19 ug/m[3] per year or about 25% lower. Project 2019 design values using these trend lines show the model is almost uniformly over predicting current design values in Allegheny County; actual 2017-19 annual PM-2.5 design value concentrations are below what the model projected for 2019 using the linear trend lines. While the Liberty monitor is currently still above the annual PM-2.5 NAAQS, the model trend line indicates additional reductions (up to 0.5 ug/m[3] in the next 2 years) could bring this monitor into attainment by 2021.
Table 6 is identically constructed but shows modeled and monitored 24-hour PM-2.5 values for each monitor in Allegheny County. Each monitor's modeled 2011 base year design value, actual 2009-11 monitored design value (if valid), model projected 2021 design value, modeled trend line (difference between the projected and base year divided by 10 years, so it's ug/m[3] change per year), actual monitor 2017-19 design value and the projected model 2019 design value. Each monitor's monitor ID and what model or combination of models was used to construct the base year and projected year 24-hr PM-2.5 concentrations is also included in the table. The Liberty monitor is highlighted given it is by far the highest monitor in the Allegheny County, PA PM-2.5 nonattainment area.

Table 5.  Allegheny County Monitor Modeled & Monitored 24-hour PM-2.5 Concentrations
                                       

Allegheny County's modeled 2011 base year 24-hour design values are slightly lower than the actual 24-hour monitor design values for the 2009-11 time period. This may be due to a more favorable meteorological set up (fewer inversion days) as noted by the Clean Air Council or faults in the model's base year inventory. Liberty's CAMx/AERMOD developed modeled trend line doesn't appear to be all that different from the CAMx only trend lines for the other Allegheny County monitors (-0.66 ug/m[3] per year). While not shown, Liberty's CAMx only modeled trend line was -0.28 ug/m[3] per year, which is about 60% lower than the CAMx /AERMOD modeled trend line. Projected 2019 design values using trend lines from Table 6 show the model is almost uniformly over predicting current design values in Allegheny County. The Liberty monitor is currently just at the 24-hour PM-2.5 NAAQS (35 ug/m[3]). The (CAMx/AERMOD) model trend line indicates additional reductions (up to 1.3 ug/m[3] over the next 2 years) should provide additional reductions such that Liberty's 24-hour PM-2.5 design value will be slightly below the 24-hour PM-2.5 NAAQS by the projected 2021 attainment date.

Allegheny County Recent (2017-19) Monitored PM-2.5 Design Values

EPA has downloaded the most recent (2017-19) PM-2.5 design values for each monitor in Allegheny County. Figure 1 displays all PM-2.5 monitors currently operating in Allegheny County, PA. In addition to the PM-2.5 monitors already discussed in this TSD, the Allegheny County Health Department has recently deployed a near-road PM-2.5 monitor. This monitor is located near the Hosanna House Event Center and is the Parkway East Near-Road monitor on Figure 1.

      Figure 1.  Locations of all PM-2.5 Monitors in Allegheny County, PA
                                       

The most current PM-2.5 design values are shown in Figure 3 (annual PM-2.5 NAAQS) and Figure 4 (24-hr PM-2.5 NAAQS). All PM-2.5 monitor design values in Allegheny County currently meet the 24-hr PM-2.5 NAAQS (35 ug/m[3]) and only the Liberty monitor currently exceeds the annual PM-2.5 NAAQS (12.0 ug/m[3]). PM-2.5 design values at the Liberty monitor tend to be higher than other monitors in Allegheny County and this monitor has historically had the highest PM-2.5 concentrations in the county (and has been the subject of previous PM-2.5 nonattainment area designations in Allegheny County).


  Figure 3.  Allegheny County 2017-19 Annual PM-2.5 Design Values (ug/m[3])
                                       
                                       

  Figure 4.  Allegheny County 2017-19 24-Hour PM-2.5 Design Values (ug/m[3])
                                       
                                       
