The U.S. EPA revised the primary National Ambient Air Quality Standard (NAAQS) for sulfur dioxide (SO2) on June 2, 2010 by implementing a new 1-hour standard of 75 parts per billion (ppb) (calculated as the three-year average of the 99th percentile of the annual distribution of daily maximum 1-hour average concentrations). Based on an evaluation of the risks posed by acute exposure to SO2, EPA also revoked the two previous primary SO2 standards of 140 ppb (24-hour standard), and 30 ppb (annual standard) because they were not expected to provide any additional public health protection beyond the newly promulgated 1-hour standard.
In accordance with Section 107(d) of the Clean Air Act (CAA), States were required to submit attainment designation proposals for the new 1-hour SO2 NAAQS by June 3, 2011 (within one year of the promulgation date) so that EPA can promulgate the final attainment designations for all areas of the country by June 3, 2012 (within two years of the promulgation date). The CAA specifies three possible designations available to the States:
Nonattainment - which includes any area that does not meet or that contributes to ambient air quality in a nearby area that does not meet the primary or secondary NAAQS,
Attainment - which includes any classifiable area that meets the primary or secondary NAAQS,
Unclassifiable - which includes an area that cannot be classified on the basis of available information as meeting or not meeting the primary or secondary NAAQS
For criteria pollutants other than SO2, the typical approach for assessing the attainment status of a given area is to rely solely on air quality monitoring and to only propose 'nonattainment' for areas with monitoring data showing a violation of the NAAQS.
In the preamble to the final rule establishing the 1-hour SO2 NAAQS (75 FR 35550, June 22, 2010), EPA identified four unique challenges to implementing the SO2 NAAQS that require the development of a different attainment designation approach:
The unique source-specific impacts of SO2 emissions (i.e., ambient concentrations of SO2 are more significantly influenced by stationary industrial sources than other criteria pollutants like NO2 or ozone)
The challenges of monitoring short-term SO2 levels for comparison to the NAAQS in many situations
The superior utility that modeling offers for assessing SO2 concentrations
The need to develop a designation approach that ensures areas attain and maintain the new 1-hour SO2 NAAQS in a manner that is expeditious and accounts for potentially substantial SO2 emissions reductions from forthcoming national and regional rules
Recognizing these challenges, and in light of EPA's historical approach to the 3-hour SO2 NAAQS implementation process, EPA proposed a hybrid analytic approach for the 1-hr SO2 NAAQS designations that combines the use of monitoring and modeling to assess attainment with the NAAQS.
Under this approach, final attainment designations are expected to be based on 2008 - 2010 air quality monitoring data from the current ambient monitoring network using the following criteria for evaluating the attainment status of an area:
Nonattainment - An area where monitoring data or an appropriate modeling analysis indicates a violation
Attainment - An area that has no monitored violations and that has an appropriate modeling analysis, if needed, and any other relevant information demonstrating no violations
Unclassifiable - An area that has no monitored violations and lacks modeling analysis or other appropriate information sufficient to support an alternate designation
EPA Guidance for Attainment Designations Building upon the information provided in the final rule preamble regarding the new approach for the 1-hour SO2 NAAQS designation process, EPA issued a guidance memo to the Regional Air Division Directors on March 24, 2011 to provide information on the schedule and process for 1-hour SO2 NAAQS attainment designations. The memo describes the five factors EPA intends to evaluate in determining nonattainment boundaries and regarding the process for conducting an 'appropriate modeling analysis' to support nonattainment boundary recommendations and attainment demonstrations.
Attainment Designation Timeline
EPA will review state designation proposals and notify the States of any intended modifications to their recommendations within 120 days of issuing the final attainment designations (refer to the SO2 NAAQS implementation timeline in Figure 1).
EPA will also publish the State recommendations and any intended modifications for a 30-day public comment period in February 2012. If States believe the proposed modifications by EPA are inappropriate, they will have the opportunity to submit additional information on or after April 3, 2012.
EPA will issue final attainment designations for the 1-hour SO2 NAAQS based upon all information received from the States no later than June 3, 2012.
Once the attainment designations are finalized, the States must submit 'infrastructure'or 'maintenance' State Implementation Plans (SIPs) by June 3, 2013 (within three years of the promulgation date) pursuant to Section 110(a)(1)-(2) of the CAA to address any additional Federally enforceable control measures that are necessary to ensure attainment and maintenance of the NAAQS in attainment and unclassifiable areas. EPA's approval of these infrastructure SIPs is expected to be based on the adequacy of a State's modeling and any required monitoring to demonstrate attainment no later than August 2017. EPA also has the authority to redesignate areas based on the review of these SIPs including a redesignation to nonattainment if the SIPs are inadequate. Inadequate SIPs may also require EPA to impose a Federal Implementation Plan (FIP) whereby EPA takes over the SO2 NAAQS implementation process in a State.
For nonattainment areas, the States must submit attainment SIPs by February 2014 pursuant to Sections 191-192 of the CAA to identify the control measures that will be implemented to achieve compliance with the NAAQS by August 2017.
EPA must approve or disapprove of attainment SIPs by February 2015.
Modeling Requirements Given the relatively limited time between promulgating a SO2 NAAQS (with an entirely new form and averaging period) and the deadline for submitting initial attainment designations, EPA did not expect States to provide refined dispersion modeling information along with their initial designation recommendations. However, EPA fully expects modeling to be completed and submitted as part of the infrastructure SIP development process.
To support these future modeling efforts, the March 2011 guidance memo contains a comprehensive attachment that describes how refined dispersion modeling should be conducted for the purposes of assigning nonattainment boundaries and demonstrating attainment in unclassifiable areas.
The modeling guidance builds on previous memos available through EPA's Support Center for Regulatory Atmospheric Modeling (SCRAM) website related to Prevention of Significant Deterioration (PSD) modeling for the 1-hr SO2 and NO2 NAAQS (refer to August 23, 2010 and March 1, 2011 SCRAM memos). The recommendations generally conform to the guidance in the 40 CFR Part 51 Appendix W Guideline on Air Quality Models. The current version of AERMOD (v11103, issued April 13, 2011) is the preferred regulatory model for 1-hr SO2 NAAQS designation modeling. AERMOD v11103 includes model options specifically tailored to 1-hr SO2 NAAQS demonstrations including output options that:
produce results in the probabilistic form of the standard
allow users to identify sources culpable for causing or contributing to the modeled violations
consider time-varying background concentrations
For attainment demonstrations in the absence of monitoring data, the modeling should center on the dominant source(s) in the area expected to have the highest contributions to ambient SO2 concentrations. EPA outlines a step-by-step process for selecting SO2 sources to be modeled based on all of the stationary industrial sources in the area. These modeling procedures differ from the procedures typically applied in PSD modeling to screen out small and distant sources not expected to influence the modeling results. The recommended procedures have the potential to capture many sources in the modeling including facilities with SO2 emissions well below the 100/250 tpy PSD major source thresholds. Contrasting an April 2007 guidance document addressing attainment designation modeling for the ozone and PM2.5 NAAQS, EPA recommends that dispersion modeling for 1-hr SO2 NAAQS attainment demonstrations should be based on the maximum allowable emissions or federally enforceable permit limits and not actual emissions for the years in which monitoring data is being used to assess the attainment status of an area.
States are concerned about the resources that will be required for the infrastructure SIP development process. They are particularly concerned about EPA's current requirement to model short-term NSPS or SIP allowable emission rates that were developed to maintain compliance with the previous 3-hour primary SO2 NAAQS (which is much less stringent than the 1-hour SO2 NAAQS). EPA's recommendation to use the maximum 1-hour allowable SO2 emission rate for all sources in the modeling domain and to assume these maximum emission rates from individual sources are all occurring simultaneously can result in very large areas of modeled violations in areas with numerous grandfathered coal-fired power plants and other coal-fired industrial combustion units.
In the Midwest, many of the grandfathered coal power plants are only subject to state-wide SIP regulations that were derived in the late 1970s based on dispersion modeling studies or other analyses associated with the much higher 3-hour SO2 NAAQS. Therefore, when these high allowable SO2 emission rates are modeled for predicting impacts in 1-hour SO2 NAAQS designation modeling, modeled violations may be prevalent in large areas of the country while not necessarily indicative of actual violations. Essentially, areas adjacent to nonattainment counties may be unnecessarily brought into the nonattainment area, and unclassifiable areas with no monitoring data may be classified as nonattainment based on overly conservative dispersion modeling procedures.
Nonattainment Boundary Determinations For nonattainment boundary determinations, EPA recommends county boundaries as the analytical starting point. A five-factor test can be used to demonstrate violations are not occurring in nearby portions of the State that are outside the county with the violating monitor and that the excluded portions of the State do not contain emission sources that contribute to the monitored violations. The five factors include:
air quality data
EPA provided guidance for conducting refined source oriented air dispersion modeling for SO2 sources expected to cause or contribute to the monitored violations in the nonattainment county. The modeling domain for these nonattainment boundary determinations should center on the violating monitor and extend out 50 km in all directions to capture SO2 sources expected to contribute to the monitored concentrations. Similar to the recommended modeling procedures for demonstrating attainment in areas proposed for designation as attainment or unclassifiable, EPA recommends modeling maximum 1-hour allowable emission rates for the nonattainment boundary determination.
Summary of States' Attainment Recommendations Lacking sufficient time to conduct refined air dispersion modeling as recommended by EPA, no States with monitors indicating violations conducted dispersion modeling to evaluate compliance with the 1-hour SO2 NAAQS in attainment or unclassifiable areas. Only two States (Florida and Wisconsin) provided modeling results to support propose nonattainment area boundaries in their initial designation recommendations. Most States proposed unclassifiable designations for all counties either without a monitor or with a monitor showing no violations for the three year period from 2008 - 2010. Most States' nonattainment boundary designations were based on a qualitative analysis of the five factors recommended for consideration by EPA.
Unlike most States, and contrary to EPA's recommendations, Texas proposed attainment for all counties containing a monitor without a violation, and Ohio and Montana recommended attainment for all counties without any SO2 emissions sources expected to cause or contribute to a 1-hour SO2 NAAQS violation. Despite monitoring data for 2008 - 2010 showing noncompliance with the 1-hour SO2 NAAQS for a monitor, Florida recommended that EPA designate the county with the violating monitor as unclassifiable based on the expectation that 2009 - 2011 data will show compliance with the 1-hour SO2 NAAQS. Once the 2009 to 2011 monitoring data is available, other States may follow Florida's lead and recommend that areas with violations based on the 2008 - 2010 data be designated as unclassifiable if the 2009 - 2011 data shows compliance with the NAAQS. For any States that recommended partial county non- attainment designations based on the five-factor analysis, the portions of the counties not falling within the nonattainment boundary were recommended as unclassifiable.
The purpose of the five-factor test conducted by most States to support the nonattainment boundary proposals was to identify the largest SO2 emissions sources expected to contribute to the monitored violation based on the meteorology during the violation. In most cases, the proposed nonattainment boundaries are partial counties near the violating monitor and the suspect SO2 emissions source(s). Some States did, however, use the county boundaries as the presumptive nonattainment boundary with only limited analysis of whether the boundary should be larger or smaller to cover the actual area with ambient 1-hour SO2 concentrations expected to exceed the NAAQS.
Michigan conducted a relatively detailed analysis for the violating monitor near Detroit in order to refine the proposed nonattainment boundary using 1) statistical evaluations of the monitor data, and 2) a nonparametric wind regression (NWR) analysis of the meteorology, and SO2 emissions from nearby sources during episodes with the highest monitored SO2 concentrations. By comparing the monitored 1-hour concentrations from the violating monitor to a nearby compliant monitor (just 7 miles away), Michigan determined that the highest concentrations from the violating monitor and the compliant monitor were not well correlated, suggesting these monitors were not influenced by the same SO2 sources. In addition to pollution roses evaluating wind directions occurring during the highest monitored concentrations at the violating monitor, Michigan conducted a NWR analysis to further evaluate the source(s) expected to contribute to the monitored violation. This analysis showed that a group of industrial sources located in close proximity to the monitor was likely responsible for the monitored violation, and as such, Michigan was able to propose a small geographic area as nonattainment including the monitor and the culpable sources.
Similarly, Missouri conducted back trajectories using the HYSPLIT dispersion model for episodes contributing to a monitored NAAQS violation to determine the likely location of offending sources. The proposed partial county nonattainment boundaries were drawn to include the violating monitor and the primary offending sources.
Both Wisconsin and Florida conducted refined dispersion modeling using AERMOD, based on maximum hourly allowable emission rates from nearby industrial sources to support their nonattainment boundary proposals. In both of these limited cases, the modeling results in the form of the 1-hour SO2 NAAQS agreed reasonably well with the monitored design value, and the spatial extent of the modeled violations supported a nonattainment boundary that was a small portion of the entire county that may have been otherwise used as the presumptive boundary.
Future SO2 NAAQS Implementation Milestones Since many States have not completed modeling in support of the 1-hour SO2 NAAQS implementation process and EPA's review of the recommendations will not be completed for another eight months, much of the impact to industrial sources of this very stringent standard has not been fully characterized. At this stage, it is unclear to what extent EPA will propose designation boundaries that differ from what the States are recommending. In addition, given the compressed timeframe, industrial sources within proposed nonattainment boundary designations may not have had the opportunity to provide input to the States regarding the likelihood that their SO2 emissions are causing or contributing to a monitored violation.
Even if the States' proposed nonattainment boundaries are largely retained in EPA's final designations, one remaining hurdle in the SO2 NAAQS implementation process for States and facilities with significant SO2 emissions will be the state-wide refined air dispersion modeling required for infrastructure SIP development. Many of EPA's recommendations are expected to produce overly conservative modeled impacts that may not agree with monitored concentration in counties with monitors that do not show a violation of the NAAQS. Furthermore, rural counties that are currently unclassifiable because they do not contain a monitor to indicate actual ambient air quality may have large industrial SO2 sources that produce modeled impacts above the 1-hour SO2 NAAQS.
If States cannot resolve modeled violations through refinements to the modeled source parameters, meteorology, or model switch settings, they must develop rules in the infrastructure SIP to impose enforceable SO2 emissions reductions that will produce compliant modeled concentrations. The SO2 emission limitations in these new SIP regulations could pose a significant compliance burden on the affected sources and these limits may be established without collecting any ambient monitoring data showing a NAAQS violation to validate the results of the modeling.
Facilities should stay informed about any modeling State air agencies that are planning for their sources in the coming months and should also consider conducting their own source-oriented modeling to understand what the models may forecast for nearby SO2 impacts. Identifying any modeled violations using EPA's recommended procedures early in the 1-hr SO2 NAAQS implementation process will allow time for affected sources to identify modeling refinements to improve the accuracy of the predicted ambient impacts or set up ambient monitoring stations to demonstrate their facility does not cause or contribute to 1-hr SO2 NAAQS violations.