After promulgation of the New Source Performance Standards (NSPS) and Emissions Guidelines (EG) for Large and Small Municipal Waste Combustors (MWCs), the Environmental Protection Agency (EPA) entered a new regulatory arena – regulating the remaining risks to public health and the environment after Maximum Available Control Technology (MACT) is applied. The residual risk from MWCs is expected to be negligible; however, the public, and some state and local regulators are now looking for ways to assure continuation of the exemplary emissions performance being measured at many of these retrofit sources. Hence, the question now becomes: how low can an achievable emissions limitation be? Confidence should not be placed in a source’s ability to continually meet the low emissions limitations embodied in the MWC EGs and NSPSs. Contrary to assertions in the Response to Comments for the Small MWC regulations [1], the Environmental Protection Agency could not have properly considered and incorporated measurement uncertainty into its dioxin guidelines; no one knew the uncertainty of total dioxin measurements above 28 ng/dsm3 corrected to 7 percent O2 until 2001 when the work supporting this paper was performed. When the 13 ng/dsm3 corrected to 7 percent O2 NSPS for MWCs was developed, the data needed to determine measurement uncertainty of most Section 129 pollutants had not even been collected. Further, asserting that the data used to derive the NSPS emissions limitations include measurement error, and therefore, any data-derived emissions limitations inherently consider that error, is only true if the measurement error is much smaller (say less than 10 percent) than the short and long term variations in emissions performance. Beginning with a set of three total dioxin measurements that averaged 4 ng/dsm3 corrected to 7 percent O2, the emissions limitation meeting the 95 percent statistical confidence level criterion underlying many NSPS, is almost 15 ng/dsm3 corrected to 7 percent O2. If the statistical criterion is changed to inclusion of “almost all” the expected results when these facilities continue to emit as they did during the original data acquisition, the emissions limitation becomes almost 18 ng/dsm3 corrected to 7 percent O2. Consequently, sources must not agree to standards that do not properly consider measurement method precision if they want to avoid exceedances when everything is working properly.

This content is only available via PDF.
You do not currently have access to this content.