August 20, 1991 MEMORANDUM SUBJECT: Hadson Power - Buena Vista PSD Permit Class I Increment Issue FROM: Joseph A. Tikvart, Chief Source Receptor Analysis Branch (MD-14) TO: Alan J. Cimorelli, Lead Meteorologist Technical Assessment Section, Region III (3AT12) In response to your request, the Model Clearinghouse has reviewed your proposal for modeling the increment consumption in Shenandoah National Park attributable to multiple sources. The modeling is required as part of the Hadson Power Plant Prevention of Significant Deterioration (PSD) permit application. We have previously negotiated your proposal over the telephone and through telefax. Thus we agree that we can support it. The proposed technique is technically defensible, given the time frame available to work out a solution. As you point out there is no guidance specifically applicable to this type of situation; thus there are no conflicts in that regard. Also there are no historical precedents that would conflict with your proposal. While we agree with your proposal, we want to emphasize that the proposed modeling technique is unique to this situation and should not be regarded as precedent for other multiple source rural or urban modeling problems. On pages 2 and 3 of the August 19, 1991 revised attachment to your memorandum you have listed eight characteristics of this model application. We agree that these characteristics, when taken in combination, are unique, that the situation is not covered by current guidance, and that your proposed solution to the modeling problem is appropriate. It is clear that the case for straight line Gaussian models for multiple sources each having their own on-site meteorological data, and even including long-range transport, leaves a bit to be desired as a permanent solution. Over time we will need to be thinking about how we can establish credible techniques that develop and utilize the spatial and temporal variations in meteorological data. If you have any questions, please contact Dean Wilson or me. cc: Dan deRoeck (Statement of Regional Office Position Follows) BUENA VISTA PSD PERMIT CLASS I INCREMENT ISSUE REGION III PROPOSED SOLUTION DATE: JULY 15, 1991 REVISED: AUGUST 19,1991 Over the past couple years VA has had a dramatic increase in the number of PSD permit applications seeking to construct co-generations plants throughout the state. This major growth activity has resulted in a heightened interest in the consumption of Class I increment in the Shenandoah National Park (SNP). The Buena Vista (BV) project (Hadson Power) is one of the proposed VA co-gens. BV is to be located approximately 60 Km southwest of the SNP within the Appalachian mountains (see the attached map). The major pollutant of concern is SO2 with proposed allowable emissions of 355 tons/yr. The plant is to locate in complex terrain and therefore, as required in the "Guideline on Air Quality Models", the applicant collected one year on-site meteorological data. Moreover, agency policy requires that an intermediate terrain analysis be performed if it is possible for the design concentration to occur on terrain which is between stack height and plume height. Since this was the case for BV, an intermediate terrain analysis was performed which utilized the ISC and Complex I models. To perform the analysis, the applicant developed an ISC/Complex I hybrid model. In addition, the applicant was required to show that the hybrid model was equivalent to guideline techniques; an adequate demonstration was submitted by BV. A Class I increment analysis was required for two areas in Virginia: Shenandoah National Park (SNP) and the James River Face. These analyses were performed using the hybrid model and BV on-site meteorological data. The analysis showed compliance with the increments at the James River Face while predicting violations of the class I increment in both the central and northern parts of the SNP. These violations resulted from the impact of two increment consuming sources. The central area of violation was caused by a Coors facility which is located within a kilometer or two of the SNP and approximately 100 km from BV. The northern area of violation was caused by a Potomac Electric power station (PEPCO Station H) located in Maryland some 60 km northeast of the SNP and more than 200 km from BV. Both Coors and PEPCO did collect a year of meteorological data for purposes of their previously approved PSD permit. However, none of the three meteorological data bases (BV, Coors or PEPCO) were collected during the same year. Analyses performed by Hadson Power have shown that impacts from PEPCO and Coors are greatly reduced when the on-site meteorological data bases from these previously permitted sources are used. Given the unusual and non-guideline nature of this situation, VA, in a letter dated 6/24/91, has requested that we provide them with advice in resolving this issue. In a recent letter to VA, Hadson Power has requested that, for purposes of the Class I increment analysis, that they be permitted to use the Coors meteorological data base when modeling southern parts of the SNP and the PEPCO data for modeling northern areas. Hadson bases this request on the opinion that the Coors and PEPCO data are most representative of the major impacts. We are considering accepting the general approach as stated in the above paragraph. Specifically we would require the following: 1. Hadson would be required to develop two receptor grids covering the entire Class I area. One grid would be designed to determine maximum impact on the Class I increment in the area where Coors dominates the impacts. The second would be designed for the area where PEPCO dominates. 2. For the Coors area all sources contained in the increment consuming inventory will be modeled using the Coors data. 3. For the PEPCO area all sources contained in the increment consuming inventory will be modeled using PEPCO data. 4. This procedure applies only to the Class I area. To perform the required Class II area impact analysis BV meteorological data is to be used exclusively. The Class II area is centered on the BV plant and is synonymous with the area of significant impact. Approval of this approach would be strictly case-specific and should not be construed as general guidance. A solution to the general issue raised by this case will be handled once Mr.Ken McBee, Chief of Virginia's Modeling Section, has defined the general problem. As a result of discussions held at the May, 1991 National Modelers meeting, it was agreed that a solution to the general problem would be sought once the problem was clearly defined. Mr. McBee agreed to provide this definition. We believe that the above stated approach, applied to BV, is justified on the merits of this specific case. Present policy does not cover this situation which has the following characteristics: 1. Predictions of the applicant source's impact are needed at a distance which is greater than 50 km from the source. That is, the Class I receptor area of interest is displaced more than 50 km from the applicant source. 2. The analysis includes multiple increment consuming sources. 3. While most of the sources in the inventory are more than 50 km from the receptor area the Coors facility is within a couple kilometers. 4. The applicant source and the two most dominant sources in the inventory (Coors and PEPCO) all have on-site meteorological data. 5. None of the meteorological data sets are coincident. 6. Significantly different impacts are predicted from the various meteorological data bases. 7. Predicted increment violations occur in two areas of the SNP. Impacts in each area are dominated by either Coors or PEPCO. Impacts from the applicant source are very small in comparison to the dominant sources regardless of which meteorological data base is used. 8. The impacts in each of the two areas of violation would be best represented by the on-site meteorological data from Coors and PEPCO respectively.