Abstract
Six properties appear desirable for any set of ozone attainment criteria: (1) sufficient stringency to protect public health, (2) simplicity and understandability, (3) sensitivity to real changes in air quality, (4) stability against meteorological fluctuations, (5) use of as much data as possible, and (6) equivalence between the stringency the criteria appear to mandate, and what they actually mandate. We consider how the federal attainment criteria might be improved with respect to Properties 4 through 6 while being equally strong on 1 through 3. Whether the federal standard meets Property 1 has been the subject of debate, but our analysis would apply also to a modified standard. We show that there are subtleties in how improvements might be made. In particular, basing the attainment status on a statistic with low variance may not lead to a more stable criterion, and although it is easy to find a criterion that makes it hard for a district with bad air to reach attainment, or a district with good air to get out of attainment, it is hard to find a criterion that does both. This suggests using different criteria for districts that are out of attainment from districts that are in attainment. Initially the burden of proof would be on the district to prove that its air is of attainment quality. However, once the district has achieved attainment the burden of proof would shift; the district would remain in attainment unless there was strong evidence to the contrary. An evaluation of a set of criteria based on this idea showed improvements over the federal criteria with respect to the last three desirable properties-stability, use of data, and equivalent stringency-with some sacrifice in simplicity and sensitivity.