Human Mortality, Pollution Sources, and the Case for Tougher Clean Air Standards
Particulate Air Pollution: Abolish Proposal for Spatial Averaging
The proposal recommends a new monitoring initiative, spatial averaging. This new scheme is likely to create "sacrifice zones" where polluted air in yet undefined spatial averaging zones could be "cleaned up" simply by averaging pollution levels from new monitors placed in adjacent communities with cleaner air. If not modified, spatial averaging will very likely undermine the otherwise significant health protections that the new rule is designed to achieve. Worse, the proposal would subvert current policy whereby a violation at a single monitoring location is enough to trigger air quality improvements for the entire region. The agency has not provided an adequate scientific or public health rationale for reversing the present conservative, risk averse, enforcement procedures.
Two analyses illustrate how spatial averaging, as proposed by the agency, could serve primarily to hide existing pollution hot spots, instead of cleaning them up. EWG analyzed data from AIRS including the state and local air monitoring stations (SLAMS) and the national air monitoring stations (NAMS) for the 490 counties across the country with valid PM monitoring data for a three year period between 1993 and 1995. We then examined the impact of spatial averaging at the county level in two different ways. County boundaries were chosen because EPA's proposal suggests that county boundaries might delineate spatial averaging zones under the new rule. Regardless of the specifics of the spatial averaging zone boundary, the results of the following analyses hold true. It is the content of the proposal, not the nature of the boundary that creates the potential for sacrifice zones.
The first analysis searched for counties with more than one PM monitor, where the three-year average for at least one monitoring site exceeded the proposed annual PM2.5 standard, but the average from all the monitors was below 15 µg/m3. Under current rules, if one monitor exceeds the standard, the entire region must clean up its air. Under the EPA spatial averaging proposal, if one monitor exceeds the standard, these results can be averaged with lower pollution levels in a clean part of the county to bring the whole county or spatial averaging zone into "attainment".
We identified 35 counties that contain these potential "sacrifice zones" (Appendix 1). If the county boundary defined the spatial averaging zone, residents in these polluted areas would continue breathing unhealthy air because some other part of the county has cleaner air, and no remedy would be initiated.
EPA has suggested that the proposed 24 hour standard would serve as a backstop, and protect these counties against potential "sacrifice zones." Yet none of these counties would trigger an enforcement action under the proposed 24 hour PM2.5 standard of 50µg/m3, calculated at the 98th percentile.
EWG does not oppose spatial averaging as a concept. But both monitoring and spatial averaging must done in a way that is scientifically and statistically valid. The goal of any averaging technique must be to demarcate areas with unsafe levels of pollution and clean them up, not merely to make them appear clean through statistical artifact.
The second analysis examines the potential effect of additional monitors within counties as spatial averaging zones. According to the EPA proposal, new monitors within spatial averaging zones are encouraged. These monitors must be placed near populated areas, but they are not required to be placed in "hot spots" where the pollution is the worst, nor are they required to be placed systematically in such a way that provides a representative picture of pollution within the spatial averaging zone. Without major revisions, this proposal will create a strong incentive to place new monitors in cleaner locations to lower "average" pollution levels in the spatial zone -- again creating a lower number but doing nothing to clean the air. The purpose of more monitoring must be to better define polluted areas, not to create the statistical illusion of clean air.
To demonstrate how additional monitoring might be used to avoid pollution reduction via averaging, we analyzed the AIRS data for counties with just one monitoring site, where particulate levels at that site exceed the proposed annual PM2.5 standard. Our analysis revealed 40 counties with a total population of 5.5 million, each with just one monitoring site, where particulate levels currently exceed the proposed annual PM2.5 standard by less than 2.5µg/m3. In any of these 40 counties, compliance with the annual standard could be achieved easily by adding just one monitor at a less polluted location, rather than actually reducing pollution levels at the polluted site (Appendix 2). While this would technically comply with the standard in writing, residents at the more polluted monitoring site would be given no relief from air pollution that is above the health standard allowed by EPA.
An even more worrisome element of the proposal would allow independent parties to construct "special purpose monitors", with the promise that data showing poor air quality will not be used for regulatory purposes. Under the EPA proposal, if the data from these special monitors bring an area into a violation of the PM2.5 health standard, there is a three-year moratorium on the use of such data. If the data bring an area into compliance, however, there is no similar explicit moratorium on use of the information. While the draft rule does not specifically state that these data will be used, the absence of a prohibition on its use creates the strong supposition that only data that would moderate regulatory burdens from these special monitors will be used in the regulatory process. As drafted, this loophole provides major polluters with a risk-free incentive to set up monitors in clean areas of spatial monitoring zones, as it simultaneously eviscerates any independent efforts to monitor air where it is the most polluted. Any potential for such a double standard must be eliminated from the final rule.
Even if the same rule applies to both high and low measurements, allowing random siting of monitors by either private or government entities does not serve the purpose of protecting public health. Unless monitors are either sited uniformly and systematically, or sited at "worst-case" locations, the monitoring program will be little more than random guesswork, and clearly would not represent "good science" by any account. This problem needs to be addressed in the rule.
We strongly oppose the used of statistical techniques to hide pollution and to avoid cleaning up unsafe air breathed by millions of Americans. Instead, EWG recommends tough health standards that are backed up by a scientifically valid system of air monitoring. In most major U.S. cities many more monitoring sites are needed to achieve this goal. If it can not be met, then the worst cases need to be addressed, not statistically averaged.
Spatial averaging techniques must not be used in any metropolitan region that does not have a representative particulate monitoring network in place. In addition, we oppose any plan that allows areas to achieves compliance with the new health standard by:
- moving existing monitors to cleaner locations,
- adding monitors only at cleaner locations, and
- moving the pollution source (for example, a bus transfer station) and thus increasing pollution in cleaner areas.
Until such time as a representative monitoring system is in place, EWG strongly recommends that the EPA maintain the current rules for monitoring and enforcement where exceeding the standard in one location triggers a violation.