Background on MACT Floor Statistics
Posted: November 7th, 2013Author: All4 Staff
Back in August of 2013, the U.S. Court of Appeals for the District of Columbia Circuit made a decision regarding a petition that the National Association of Clean Water Agencies (NACWA) and Sierra Club made to object to the approach that U.S. EPA used to develop maximum achievable control technology (MACT) standards for sewage sludge incinerators that are regulated under Section 129 of the Clean Air Act (CAA). The decision directs U.S. EPA to provide justification for their use of alternate statistical measures for determining the emission limits that sources must meet under Section 112 and Section 129 standards. This focus on the alternate statistical measures could have significant impact on several rules that U.S. EPA has promulgated over the past few years under Sections 112 and 129 of the CAA. This blog post provides some of the background information to help frame the importance of this court decision.
First, it is probably worthwhile to review a couple of basic components to the method used to develop MACT standards. The first item to clarify is around the applicability of MACT in general. Typically MACT is linked to the regulation of hazardous air pollutants (HAPs) under Section 112 of the CAA. But in reality, MACT is a generic term that is equivalent to “…the maximum degree of reduction in emissions…that the Administrator, taking into consideration the cost of achieving such emission reduction, and any non-air quality health and environmental impacts and energy requirements, determines is achievable for new or existing sources (units)…” Since both Section 112 and Section 129 contain the phrasing in the previous sentence, the MACT concept applies to both regulatory programs.
The second key component to review involves how MACT is determined and introduces the statistical concept that should be used to develop the MACT limits. Again, both Section 112 and 129 use the same language to frame the basis for a MACT emission limitation. Section 129(a)(2) of the CAA states “Emission standards for existing units…shall not be less stringent that the average emission limitation achieved by the best performing 12 percent of units in the category.” Section 112(d)(2) contains the same wording but goes just a little further and states that the emission standards “ may be more stringent than the average emission limitation achieved by the best performing 12 percent of units in the category.” Section 112 also includes a provision to address small populations of sources (i.e., fewer than 30 sources). In both Section 112 and 129 the term “average” is used and it is this specific language that is at the core of the DC Circuit Court ruling.
Finally the statistical tools used by U.S. EPA to develop the MACT emission limits are worth reviewing. U.S. EPA has used different approaches to characterize the concept of “average.” In plain English, the average of a population most commonly would be represented by the mean. The mean of a population is the sum of values of the population divided by the number of values. The median of a population is the value that has the same number of values above it as below it and can also be utilized to represent an average. An upper confidence limit (UCL) is the value that, with some assigned degree of certainty (e.g., 95, 99, or 99.9 percent) it can be expected that the mean (i.e., average) of a population will be below. The upper predictive/prediction limit (UPL) is the value that with some assigned degree of certainty it can be expected that the mean of a population that includes future values (i.e., future mean) will be below. There are subtle nuances to these four statistical terms regarding population size and whether the values are distributed normally or log normally. However, in order of lowest to highest, the mean or median value of a population will result in the lowest value while the UPL will result in the highest value. Thus emission limitations that are based on mean/medians will be more stringent and emission limitations that are based on the UPL will be less stringent.
As mentioned in the Mark’s blog post, the statistical tools play an important role in the DC Circuit Court’s ruling. It is the use of the UPL that is specifically at issue in the DC Circuit Court case and secondarily how statistical tools other than an “average” are, or are not, appropriate for developing MACT emission limits.