How the SO2 NAAQS Came Crashing Down On Us

The sulfur dioxide (SO2) landscape looked very different in 2008 than it does today.  At that time, the annual, 24-hour, and 3-hour National Ambient Air Quality Standards (NAAQS) for SO2 had been in place for a long time.  Except for one (1) county in the entire United States, there were no areas that were classified as nonattainment with any of the SO2 NAAQS levels.  The need to demonstrate direct compliance with the NAAQS levels for SO2 was only triggered by projects that were major modifications or new sources that were subject to Prevention of Significant Deterioration (PSD) permitting requirements.  As a result, air dispersion modeling for SO2 was reserved for those few projects that were significant for SO2 emissions, and the SO2 NAAQS levels were comfortably high such that they impacted permitting decisions for only a small fraction of those PSD significant projects.  As a result, that special group of dispersion modeling scientists, along with their unique attributes and occasional idiosyncrasies, weren’t in a position of high visibility in the air permitting arena.  The primary issue during a PSD permitting evaluation was the requirement to install a control technology or establish an emission limit to be consistent with Best Available Control Technology (BACT) requirements.  And so it had been for a long time. However, U.S. EPA had been performing their SO2 NAAQS review in the background, and things changed abruptly in July 2010 when U.S. EPA finalized a rulemaking establishing the new 1-hour SO2 NAAQS and those quirky dispersion modelers suddenly got much more recognition.

The final SO2 NAAQS rule published in July 2010 was notable not for the language in the rule itself, but the implementation approach outlined in the preamble to the rule.  As with the proposed 1-hour NAAQS rule, the final NAAQS as established was extremely stringent to nobody’s surprise.  The way that U.S. EPA proposed to implement the standard was quite surprising, however.  In the proposed rule, U.S. EPA indicated a plan to establish an enhanced network of ambient SO2 monitors to establish which areas were or were not attaining the new 1-hour NAAQS.  The use of ambient monitoring data to establish NAAQS attainment is as old as the Clean Air Act (CAA) itself.  However, U.S. EPA’s intended implementation approach laid out in the preamble to the final rule looked very different.  The final rule outlined a “hybrid” approach to evaluating NAAQS attainment that would utilize existing SO2 monitoring data in combination with ambient concentrations predicted by air dispersion modeling (specifically with the AERMOD air dispersion model).  Air dispersion modeling was given significant weight in the preamble, to the point that an attainment designation in an area could only be established with both ambient monitoring and air dispersion modeling data.

U.S. EPA’s suggested approach would have required that dispersion modeling be performed and results submitted by states to U.S. EPA as part of a “Maintenance SIP” submittal in June 2013.  The approach as outlined and the use of dispersion modeling represented a major departure from the standard NAAQS implementation process and a significant new burden on a wide range of industries.  The AERMOD dispersion model is inherently conservative in both its inputs and underlying assumptions, making it very difficult to demonstrate modeled compliance with the 1-hour NAAQS, particularly for multiple facilities located close together that exhibit overlapping impacts.  Since the Maintenance SIP submittals would be required to include emission rates that “work” with modeling for the 1-hour NAAQS, they would have required more stringent new instantaneous emission limits for the vast majority of SO2 emitting industrial facilities.  In many cases, the new limits would have prompted the installation of add-on control devices or changes to the operational strategies of facilities (e.g., types of fuels fired, etc.) and resulting in an unanticipated round of New Source Review (NSR) permitting for affected facilities.  To justify this approach, U.S. EPA opted to consider the public comments on the proposed NAAQS rule that focused on the timing requirements and logistical difficulties around establishing an enhanced ambient monitoring network.

While it is clear that ambient monitors are the only appropriate way to establish NAAQS compliance, U.S. EPA’s surprise dispersion modeling-based implementation approach had to be taken at face value by industry, prompting both regulatory agencies and industry to take action. Legal challenges to the approach were filed.  States began evaluating their resources and capabilities to conduct such complicated and widespread dispersion modeling (many states began requesting specific modeling information from facilities).  Many facilities began to conduct preliminary exploratory dispersion modeling to evaluate modeled concentrations.  The more the states reviewed their approaches to modeling, and the more exploratory dispersion modeling that was conducted, the more clear it became that conducting dispersion modeling across a wide range of sources and determining which sources would be required to make emission reductions and to what extent would be a virtually impossible exercise in most areas of the country.  U.S. EPA attempted to put some definition around the dispersion modeling process in September 2011.  They were flooded with comments from the public and from states urging them to reconsider the dispersion modeling implementation approach.  Some of the comments included:  (1) dispersion modeling to assess NAAQS attainment is inconsistent with the provisions of the Clean Air Act; (2) the conservative nature of the dispersion model makes it an inappropriate tool to evaluate NAAQS compliance; and (3) the inputs as specified by U.S. EPA, including the modeling of potential-to-emit rather than actual emissions, were exacerbating the primary issues with the dispersion modeling approach.

In response to the public comments, U.S. EPA put a temporary hold on the dispersion modeling process for NAAQS implementation purposes, and scheduled three public stakeholder meetings to solicit input on how best to continue with the 1-hour NAAQS implementation process.  The meetings included separate venues for environmental groups, state regulatory agency representatives, and industry and consultant representatives.  ALL4 attended the industry stakeholder group meeting and noted the following observations:

• Environmental groups continue to put pressure on U.S. EPA that every area of the country needs to be evaluated against the 1-hour NAAQS.  Environmental groups are attempting to apply pressure by conducting their own dispersion modeling for specific facilities using publicly available information, and reporting the unfavorable (and conservative, potentially inaccurate) results.  If the implementation process were in the hands of environmental groups, all facilities would be performing SO2 dispersion modeling right now.
• U.S. EPA agrees on the need to at least qualitatively evaluate all areas of the country since, unlike NOX and ozone that are either oriented around mobile sources or are regional in nature, SO2 can be a more source-oriented pollutant.  Whether the differences between ozone and SO2 means that SO2 should be evaluated in every corner of the country is debatable.
• Under the assumption that all areas need to be evaluated, U.S. EPA is open to a threshold approach for narrowing the number of areas that would need to be specifically considered.  Establishing an annual emissions threshold on an individual facility basis was discussed (i.e., if a facility emits greater than 1,000 tons per year it would be required to evaluate the 1-hour NAAQS).  Establishing population-weighted emission thresholds was also discussed.  This approach would evaluate areas that meet criteria for a factor that accounts for total emissions and total population.
• The conversation turned from selecting which areas to evaluate to the methodology that would be used to assess NAAQS compliance in those areas.  The overwhelming comment:  ambient monitors are the only appropriate way to assess NAAQS compliance.  Dispersion models are not the correct tool for the evaluation for various reasons.
• In a marked shift from the final 1-hour SO2 NAAQS rule, U.S. EPA expressed openness to an implementation process that is based entirely on data from ambient monitors.  While promising, the conversation then turned to the logistics around establishing an expanded monitoring network.  Who will pay for the monitors and who will be responsible for running them?
• Federal and State funding for new ambient monitors is scarce, so U.S. EPA directly asked those industry groups and companies in attendance about their potential willingness to assist in funding an increased monitoring network.  The responses were mixed, ranging from a willingness to further explore funding to requests that U.S. EPA determine their national health priorities and fund them directly.

While no specific decisions were made on the implementation process, the meeting was helpful to understand U.S. EPA’s current thinking on the next steps.  U.S. EPA is now tasked with establishing new rulemaking or guidance to specify the next steps in the NAAQS implementation process.  The stakeholder group attendees strongly urged that U.S. EPA use the formal rulemaking process to allow for public comments on the next steps of the implementation process.  Our hope is that the final rulemaking will include the following components based on what we believe is an appropriate approach along with the industry stakeholder feedback to U.S. EPA:

• NAAQS designations will be based exclusively on ambient monitoring data.
• New ambient monitors will be placed in areas using both emissions information and population data.  Only those areas with a combination of high SO2 emissions and relatively high population will be evaluated.  This approach is the most reasonable because it identifies those areas where the general public could be exposed to elevated SO2 concentrations.
• The specific location of ambient monitors within those areas being evaluated will account for population weighting such that the monitors will be located in population centers.  Dispersion modeling will not be utilized for siting ambient monitors since it would result in facility fenceline monitoring locations that do not realistically represent areas of peak SO2 concentrations to which the general public would be exposed.

After collection of the new ambient monitoring data, those areas in which elevated concentrations are identified will not be reclassified as nonattainment areas; rather Maintenance Plans will be established to allow progress towards attainment with the 1-hour NAAQS (side note that areas in which ambient monitors are currently located will receive their designations from U.S. EPA in the near future).

While there is a lot of information to consider and a lot of uncertainty around how the final implementation process will look, our advice to facilities remains the same.  Despite the temporary hold on what would have been a burdensome dispersion modeling process, the 1-hour SO2 NAAQS remains a liability across industry (remember that modeling requirements could be triggered by factors beyond your control such as another facility performing major permitting nearby).  Consider the following:

• Understand the ambient concentration predicted for your facility by dispersion modeling.  This knowledge is critical in the event that dispersion modeling is re-inserted into the implementation process or if modeling is triggered by another mechanism such as a nearby facility conducting PSD permitting.
• Remain in tune with U.S. EPA’s ongoing preparation of a final rulemaking for the remainder of the 1-hour implementation process.  Given the amount of information and comments that need to be considered, it would be surprising if U.S. EPA finalized a proposed rule in less than a year.
• Consider submitting comments on your thoughts around the implementation process.  U.S. EPA prepared a White Paper outlining the discussion points for the stakeholder meeting.  Written comments and thoughts to U.S. EPA are due by June 29, 2012.
• Consider your facility’s and company’s role in the implementation process.  Initiating discussions with state agencies on new ambient monitors and the potential for funding those monitors could put you in a position to influence the process of locating ambient monitors in the future.

As the SO2 NAAQS implementation process continues to develop, we will be there to follow the twists and turns.  Stay tuned for additional updates.