Mr. Michael Koerber
Technical Director
Lake Michigan Air Directors Consortium
2350 E. Devon Avenue
Suite 242
Des Plaines, Illinois 60018

Re: Additional Control Strategies for Modeling

Dear Merrylin and Mike:

You have invited members of the Implementation Strategies and Issues and Regional and Urban Modeling Workgroups to submit proposals for additional control strategy runs for modeling. These runs are to supplement the Round 3 modeling runs that were approved by the ISI Workgroup and the OTAG Policy Group at the meetings held in St. Louis on January 15-16, 1997.

The Midwest Ozone Group supports the compromise package of Round 3 modeling runs designed to begin assessing the effects of the incremental geographic application of various control strategies. Our support for these runs is, of course, premised on the carefully crafted set of caveats which accompanied the presentation of these modeling runs. These caveats make it expressly clear that the Round 3 runs do not represent final OTAG control zones and that more modeling should be done subsequent to the performance of Round 3.

We strongly support the conclusion that the Round 3 modeling runs should be viewed as work preliminary to OTAG's final recommendations. These runs are a necessary iterative step building on the increasing body of evidence as to the limited geographic extent of ozone transport. We firmly believe that no OTAG final recommendations can be made without additional post-Round 3 modeling runs, which not only assess even more spatially and directionally focussed geography and control levels, but also are subject to the application of the ISI Workgroups Set 2 criteria and cost analysis.

We understand that Steve Gerritson's cost impact mini workgroup is examining the cost impact issue. We expect that cost analysis will focus on runs that are more likely to be acceptable by all or most Policy Group members.

The issue that must now be addressed with respect to modeling runs is what set of control strategies should be examined to provide policy makers with a realistic basis upon which to formulate OTAG's final recommendations to EPA. Few, if any of the modeling runs examined thus far have sought to address this issue. Specifically, none of these runs have been developed with enough sensitivity to allow the policy group to determine whether emissions from any source or group of similar sources in a state has a significant enough impact on the nonattainment areas involved to justify the imposition of additional controls that go beyond those requirements now established under the Clean Air Act.

We are concerned that the overly broad zone approach which is a basis for the Round 3 incremental geographic modeling runs is not appropriate for policy purposes, since the criteria used to establish these zones assumes a "significance" level of 2-6 parts per billion over potentially a couple of hours and a criteria for grouping states that is, at best, overly restrictive. Such a stringent test places too much reliance on the model given the known concerns with emissions errors and model biases. Moreover, the resulting zones are so large and arbitrarily selected that the model results will obscure rather than clarify the issue of transport into non-attainment areas.

In addition to concerns about the use of a broad zonal approach in formulating policy runs, the Midwest Ozone Group strongly believes that the control levels that are being examined in Round 3 are not appropriate to form a basis for OTAG's final recommendations to EPA. The control levels assumed in Round 3 were selected for their relationship with each other to test geographic variability. The control strategies in the policy stage of the OTAG process must be more refined and must address the fundamental findings that have grown out of all modeling to date, namely:

• The object of the policy level control strategies must be the attainment of the current ozone NAAQS in the 3 serious and severe problem areas of Lake Michigan, Atlanta and the northeast. The boundaries of these 3 areas should be the boundaries formally designated by EPA as being in serious or severe nonattainment.
  
• Because the greatest ozone reductions occur in the same area where the emission reductions occur, the greatest level of emission reduction should be imposed at or near the 3 problem areas and diminish with distance from those areas. The data thus far generated indicates that on the highest ozone days, most reductions of the highest ambient ozone concentrations are a function of local controls, rather than controls on emissions remote from the nonattainment areas.
  
• The imposition of policy level control strategies should be designed to assure that no source or region is subject to the imposition of controls that go beyond existing requirements unless the imposition of those controls would result in a "significant" improvement in ground level ozone in those areas and on those days which give rise to the serious and severe nonattainment status of the 3 primary problem areas.
  
• Policy level control strategies should treat all NOx sources in problem areas comparably.
  
• VOC emission reductions are particularly important within the 3 problem areas themselves.
  
• Where policy level strategies are to apply to Level 0 controls, those controls should be consistently at the levels mandated by the Clean Air Act.

Accordingly, the Midwest Ozone Group recommends that additional modeling be conducted to develop the type of data that would be necessary to allow policy relevant decisions to be made by OTAG in the formulation of its final recommendations to EPA. We recommend that these additional modeling runs, which are described in the following tables, be a combination of refinements to the Round 3 incremental geographical modeling runs that are already underway and the performance of a new matrix of modeling runs, based on roll-out concepts. Only the roll-out concept provides enough sensitivity to allow determinations to be made as to whether the imposition of controls on particular sources will achieve ozone reductions in the 3 problem areas at the times and in the places that could be expected to influence the attainment status of those areas.

* These runs should utilize revised geography from that assumed in the Round 3 incremental modeling runs already initiated. This geography should move western Pennsylvania into and retain only the northern (OTR) portion of Virginia in Region 3. Central and southern Virginia should be moved into Region 4, thereby defining Region 3 as the Northeast Ozone Transport Region which was established by the Clean Air Act. In addition, Indiana should be moved into Region 2.