Comment 7:EPA did not provide the public with the criteria for making its BART determination, which appears inconsistent with the BART Guidelines and the intent of the Regional Haze Rule.
Response 7:As noted previously, EPA was not conducting its own BART analysis. We were reviewing the adequacy of NDEP's BART analysis. NDEP correctly noted that RGGS is not the size of a facility for which application of the BART Guidelines is mandatory.
After receiving significant comments on our initial proposed rule (76 FR 36450), EPA independently and thoroughly reviewed NDEP's NOXBART determination and concluded that NDEP provided the public with information regarding the criteria it was applying in making its BART determination. See “Revised NDEP BART Determination Review of NV Energy's Reid Gardner Generating Station Units 1, 2 and 3” revised October 22, 2009. NDEP adequately informed the public about the basis for its NOXBART determination for RGGS, stating: “NDEP concluded, based on a review of the economic analysis, that the $/ton of NOXremoved increased significantly for the LNB with OFA and SNCR, and ROFA with SCR technologies without correspondingly significant improvements in visibility.”Id.page 6. We are approving NDEP's determination that NOXBART for RGGS is an emissions rate that is achievable by installing and operating LNB with OFA and SNCR because NDEP reasonably weighed the small incremental visibility improvement that would result from installation of SCR against its higher cost. NDEP adequately disclosed the factors it considered in its BART determination.
Comment 8:EPA fails to explain what level of incremental cost or visibility improvement would justify SCR. EPA should disclose the dollar limit and rationale for what constitutes “cost effectiveness,” and how its method is consistently applied across other facilities and states.
Response 8:EPA's approval of NDEP's BART determination is based on finding that the State adequately considered the appropriate factors for BART and provided a reasonable explanation for selecting a NOXemissions rate that can be achieved with SNCR. NDEP explained that requiring SCR technology would result in a small incremental visibility improvement over SNCR when weighed against the incremental cost-effectiveness of SCR. As stated in our proposed approval, our modeling analysis was performed “in a manner that more closely adheres with current EPA regulatory guidance on CALPUFF modeling.”See77 FR 21903 (April 12, 2012). Our analysis found that the average and incremental visibility improvement would be significantly lower than the visibility improvement relied upon by NDEP. In addition, EPA's revised cost analysis also indicated lower cost per ton of pollutant removed for SCR. In our analysis, we evaluated the cost-effectiveness of both technologies (SCR and SNCR with LNB and OFA) based on using the Control Cost Manual (CCM) for including appropriate costs.
Our modeling shows that there would be a very small improvement in visibility at the GCNP from using SCR at RGGS. Based on this analysis we have determined that we can approve NDEP's determination that RGGS is required to comply with a NOXemissions rate that can be achieved with SNCR as BART. Although the values that EPA considered for cost-effectiveness and visibility improvement differ from NDEP's analysis, we conclude NDEP's analysis reasonably weighed the small visibility improvement against the cost to eliminate SCR.
One comment faults EPA, stating: “EPA further fails to explain what level of incremental cost or visibility improvement would justify the incremental cost.”SeeConsortium Letter at page 6. EPA's BART guidelines did not establish bright-line thresholds for cost-effectiveness or visibility improvement, choosing to allow the states to exercise discretion to choose such values when appropriate. EPA stated:
We agree with the suggestion that the use of a comparison threshold, as is done for determining if BART-eligible sources should be subject to a BART determination, is an appropriate way to evaluate visibility improvement. However, we believe the States have flexibility in setting absolute thresholds, target levels of improvement, or de minimis levels since the deciview improvement must be weighed among the five factors, and States are free to determine the weight and significance to be assigned to each factor. For example, a 0.3, 0.5 or even 1.0 deciview improvement may merit stronger weighting in one case versus another, so one `bright line' may not be appropriate.See70 FR 39129 (July 6, 2005).
The same rationale should apply to cost-effectiveness. A bright line for cost-effectiveness may not be appropriate for every case and is dependent on case specific factors relating to economics and technology. In this case-by-case determination, the small amount of visibility improvement did not justify the cost of SCR.
Comment 9:EPA should explain the amount of incremental visibility improvement from SNCR to SCR that would justify the incremental cost increase of SCR, since no threshold is established in rulemaking or guidance.
Response 9:EPA is not setting generally applicable thresholds for incremental cost-effectiveness or visibility improvement for the reasons discussed above. EPA's BART Guidelines established presumptive emissions limits for SO2and NOXat electric generating units at facilities generating more than 750 MW. But EPA did not extend those presumptive emissions limits to electric generating units at smaller facilities, such as RGGS.
EPA did not establish presumptive cost-effectiveness or visibility improvement values. EPA left weighing the factors to the state providing the state considered the five factors and exercised its discretion reasonably. Here, EPA proposed to find that NDEP reasonably eliminated SCR when it weighed the cost-effectiveness against the small incremental visibility improvement associated with requiring SCR rather than SNCR.
BART is a case-by-case analysis that is initially evaluated by the states. Provided the state exercises its discretion reasonably and meets the requirements of the CAA and regulations, EPA may approve it. EPA's approval is not a ministerial act. In this rulemaking, EPA has carefully reviewed the basis for NDEP's determination. There is no reason, and none is provided in the comment, to support the assertion that EPA should establish thresholds for cost-effectiveness or visibility improvement, or challenge EPA's authority to approve a BART determination without them.
Comment 10:EPA's use of incremental visibility improvement to find that the cost of SCR is unjustified contradicts its finding that SCR is cost-effective (77 FR 21901).
Response 10:The commenter mischaracterizes EPA's proposed approval. The commenter is correct that we did not find the average and incremental cost-effectiveness of SCR to be cost prohibitive. Nevertheless, our evaluation supported NDEP's determination that the small amount of visibility improvement at GCNP did not justify the cost of SCR.
The comment states that EPA has invented a “sixth factor” by “concatenating incremental visibility and incremental cost.”SeeConsortium Letter, page 7. EPA has not invented an additional factor in the BART analysis but has approved a reasonable conclusion reached by NDEP when it weighed these two factors. NDEP's weighing two factors in the analysis does not create a sixth factor. The comment does not explain how weighing two factors in the five-factor analysis constitutes stringing together and joining those factors into a sixth factor.
National Parks Conservation Association and Sierra Club wrote to EPA on June 29, 2012, concerning several regional haze actions. We are treating this letter as a late comment on our proposed action and including it in our docket as such. This letter indicates that NPCA and Sierra Club understand that our approval is based on finding that NDEP reasonably weighed visibility improvement and cost-effectiveness rather than inventing an additional BART factor. The letter provides:
In many cases, EPA has summarily concluded that the incremental costs of concededly superior controls are not warranted by the visibility benefits determinations, which are routinely at odds with the Agency's own analysis demonstrating that installing the most effective controls will yield needed visibility improvements.SeeLetter dated June 29, 2012, page 1.
EPA's analyses are also based on weighing the five BART factors. The relative weight of the cost-effectiveness and visibility improvement varies depending on the facility at issue. For the three 100 megawatt units at RGGS, EPA concludes that notwithstanding differing conclusions about both cost and visibility improvement, NDEP reasonably determined that a small visibility improvement at GCNP does not justify the cost of SCR. Our approval of NDEP's NOXBART determination on this basis is consistent with our actions on other regional haze SIPs.See, e.g.,77 FR 24385 (Apr. 24, 2012) (Final Approval of Maine SIP).
D. Cost Analysis
Comment 11:The incremental cost difference between SCR and SNCR is less than EPA estimated because the cost of SCR is overestimated and the cost of SNCR is underestimated, making SCR look relatively more expensive than is the case.
Response 11:The comment does not provide any basis for EPA to revise its proposed approval of NDEP's NOXBART determination. Our proposal stated:
Based on our revised cost estimates, we do not consider these [EPA's] average and incremental cost effectiveness values for SCR and LNB and OFA as cost prohibitive. Our analysis of this factor indicates that costs of compliance (average and incremental) are not sufficiently large to warrant eliminating SCR from consideration. The incremental cost effectiveness values for Units 1 and 2 are around $4,500/ton. Although EPA does not consider this incremental cost prohibitive, we note that the State has certain discretion in weighing this cost. Because RGGS is not a facility over 750 MW and therefore not subject to EPA's presumptive BART limits, the State may exercise its discretion more broadly in this particular determination.See77 FR 21901 (April 12, 2012).
Even if the average and incremental cost-effectiveness between SCR and SNCR were somewhat different, NDEP's BART determination would still be approvable based on its reasonable weighing of the cost and visibility improvement factors.
Comment 12:EPA incorrectly estimated the cost-effectiveness of SCR (i.e., dollars per ton of emissions removed on an annual basis) by assuming that SCR can achieve an annual average emission no lower than 0.083 to 0.098 lb/MMBtu, despite substantial evidence that SCR can achieve 0.05 lb/MMBtu or lower on an annual basis.
Response 12:EPA disagrees with this comment. Regarding the accuracy of the cost effectiveness calculations of SCR, the commenter is correct that we estimated cost-effectiveness of SCR based on annual average emission rates ranging from 0.083 to 0.098 lb/MMBtu. However, we indicated in our proposal that we did not find SCR to be cost prohibitive at these emission rates. As a result, although we did consider more stringent SCR emission rates, such as 0.06 lb/MMBtu, when evaluating visibility improvement, we did not also revise our cost estimates to reflect the more stringent SCR emission rates, since we had already indicated that did notfind SCR to be cost prohibitive at the less stringent SCR emission rates. It would not have been in any way determinative to our decision to find that SCR was “even more” cost-effective or that the incremental cost-effectiveness value between SCR and SNCR was “even more” incrementally cost-effective.
Regarding the emission rate achievable by SCR, the BART Guidelines state that: “[i]n assessing the capability of the control alternative, latitude exists to consider special circumstances pertinent to the specific source under review, or regarding the prior application of the control alternative” (70 FR 39166, July 6, 2005).12
In other words, the BART emission limits are not required to represent the maximum level of control ever achieved by a given technology. Limits set as BACT under the PSD program, or emission rates achieved from the operation of individual facilities under an emission trading program (e.g., Clean Air Interstate Rule), may provide important information, but should not be construed to automatically represent the most appropriate BART limit for all facilities.
12Although NDEP's BART analysis for RGGS need not conform to the BART guidelines because the capacity of RGGS is smaller than 750 MW, the BART guidelines do provide useful guidance in setting appropriate BART limits.
The coal composition is also an important component of estimating the NOXemissions rate that a facility can achieve. RGGS is capable of purchasing coal on the spot market so there is likely to be variability in the NOXemissions rate that would be achievable with SCR or SNCR. As previously discussed in the response to Comment 2, RGGS receives its coal by rail line and has access to different ranked coals with varying nitrogen content, which influence the NOXconcentration in the exhaust going to either SNCR or SCR controls. EPA's policy is to set an emission limit that would reasonably accommodate the various coal sources under these circumstances.
EPA disagrees with this comment, but even if we accepted the premise that RGGS is capable of continuously meeting an emission limit of 0.05 lb/MMBtu, the comment does not provide any basis for EPA to change our approval of NDEP's SIP or our FIP. Assuming the cost of achieving 0.05 lb/MMBtu was equal to the cost of achieving 0.083 to 0.098 lb/MMBtu, using a NOXemissions rate of 0.05 lb/MMBtu for SCR would likely result in lower average and incremental cost per ton values. Thus, we would calculate SCR to be more cost-effective (i.e., lower dollars per ton) on an average and incremental basis. As stated above, EPA did not determine the average or incremental cost of SCR to be prohibitive. Rather, EPA's approval of NDEP's determination that NOXBART for RGGS for Units 1 and 2 is an emissions limit of 0.20 lb/MMBtu that can be achieved by installing and operating LNB with OFA and SNCR is based on our determination that NDEP reasonably weighed the visibility improvement against the other factors in rejecting SCR. EPA does not believe this analysis would be significantly altered by slightly lower incremental cost numbers.
Comment 13:EPA did not correct all the errors in the State's cost calculations for SCR (e.g., lack of multiple unit discounts, high reagent costs, incorrect capital recovery factor), which would have further reduced the cost and improved the cost effectiveness of SCR, thereby reducing the incremental cost difference with SNCR.
Response 13:EPA partially agrees with this comment. EPA's revised cost-effectiveness values are consistent with EPA's regulations and the parameters set forth in the CCM. EPA explained in promulgating the BART Guidelines that “[s]tates have flexibility in how they calculate costs.”See70 FR at 39127 (July 6, 2005). A state may deviate from the Control Cost Manual provided its analysis is reasonable. EPA independently evaluated NDEP's cost-effectiveness calculation, stating in our proposal:
We received several public comments that NDEP's cost calculations were overestimated and based on methodology inconsistent with EPA's Control Cost Manual (CCM). [footnote omitted]. We agree that NDEP included inappropriate costs and our analysis excludes those costs that are not allowed by the CCM.See77 FR 21901 (April 12, 2012).
Our proposal noted that we did not revise the cost-effectiveness calculation to adjust for all of the discrepancies with the CCM because based on our initial adjustments we found that SCR was not cost-prohibitive. It would not have been in any way determinative to our decision to find that SCR was “even more” cost-effective or that the incremental cost-effectiveness value between SCR and SNCR was “even more” incrementally cost-effective.
As discussed above, EPA is approving NDEP's determination that NOXBART is an emissions limit achievable with SNCR rather than SCR. The basis for our approval is that when NDEP weighed the small visibility improvement of moving from an emissions limit achievable with SNCR to one based on SCR against the incremental cost-effectiveness of SCR, NDEP determined that NOXBART for RGGS for Units 1 and 2 is an emissions limit of 0.20 lb/MMBtu that can be achieved by installing and operating LNB with OFA and SNCR. NDEP has discretion in determining how to weigh the factors in reaching a BART decision under the CAA and regional haze regulations. NDEP's rationale for its decision, although based on different values than EPA calculated and modeled, was reasonable. Therefore, EPA is approving NDEP's determination.
The comment implies that correcting each of the costs listed as incorrect and substituting a SCR emissions limit of 0.05 lb/MMBtu rather than 0.06 lb/MMBtu for SCR would yield a very low incremental cost difference between SCR and SNCR. However, that implication is not supported by the comment. The comment does not calculate an alternative average or incremental cost-effectiveness differential between SCR and SNCR. Therefore, EPA is approving NDEP's conclusion that the incremental cost-effectiveness is not justified when weighed against the small visibility improvement.
Comment 14:EPA did not consider the adverse non-air quality impacts of SNCR due to ammonia injection, which would increase the cost of SNCR and reduce the incremental cost difference with SCR.
Response 14:As noted previously, EPA was reviewing the State's BART determination to evaluate whether NDEP reasonably applied the requirements of the CAA and the regional haze regulations. EPA anticipates that ammonia emissions will be quite low because these units are equipped with baghouses and wet scrubbers that each can be expected to remove most ammonia slip associated with SNCR or SCR. To the extent the commenter is concerned that considering costs due to ammonia injection would lower the incremental cost-effectiveness value between SCR and SNCR, EPA reiterates that our proposed approval of NDEP's RGGS NOXBART determination is not based on agreeing with NDEP that SCR is not cost-effective. EPA's proposed approval states that SCR is cost-effective. Nonetheless, the BART determination is a multiple-factor analysis. NDEP has discretion to determine how to weigh the factors. Our independent analysis of the two critical factors demonstrated that the NDEP reasonably weighed the cost of SCR controls against the small visibility improvement to determine that SNCR is NOXBART for RGGS.
Comment 15:In determining the average and incremental cost-effectiveness, EPA should have used actual emissions for the baseline value of each unit rather than each unit's annualized maximum permitted heat input multiplied by each unit's maximum permitted NOXlimit, which is closer to the potential to emit (PTE).
Response 15:EPA disagrees with this comment. Again, we note that EPA was not performing its own BART analysis, but was reviewing the adequacy of NDEP's BART analysis. The commenter is correct in noting that, in our review of NDEP's evaluation of the cost of compliance, we did not modify the estimate of baseline annual emissions that NDEP used in its cost calculations. We agree that NDEP's baseline more closely represents the sources' PTE, and results in higher baseline annual emissions than the methodology proposed by the commenter, which would rely almost entirely on past actual annual emissions. Because the regional haze regulations and BART Guidelines are not prescriptive regarding the calculation of baseline emissions, stating that “the baseline emissions rate should represent a realistic depiction of anticipated annual emissions for the source”13
, the commenter's proposed methodology is a potentially acceptable way to calculate baseline annual emissions. NDEP used a methodology that resulted in a higher estimate of baseline annual emissions, and we consider the methodology used by NDEP to be within the discretion afforded to states.
1370 FR 39167, July 6, 2005.
E. Cost of Compliance
Comment 16:Use of EPA'sAir Pollution Control Cost Manual(“CCM”) is not required since RGGS is less than a 750 megawatt facility.
Response 16:EPA agrees that the states are not required to use the CCM for electric generating units smaller than 750 MW but that it is generally a good guide concerning costs to include and exclude. EPA performed an independent average and incremental cost-effectiveness calculation using the CCM to evaluate whether NDEP had reasonably weighed small visibility improvements against the incremental cost-effectiveness of requiring SCR rather than SNCR. EPA's analysis resulted in different cost-effectiveness and visibility improvement values. Although the values for these factors differed from NDEP's values, our analysis supported approving NDEP's NOXBART determination to establish an emissions limit of 0.20 lb/MMBtu achievable from installing and operating SNCR.
Comment 17:EPA'sAir Pollution Control Cost Manualis out of date, and thus substantially underestimates current market costs of control technologies including SCR, which misrepresents the cost-effectiveness of chosen technologies.
Response 17:EPA disagrees with the comment. The CCM is a valuable resource to guide the states in evaluating costs that should be included or excluded. The states have discretion to rely on specific capital and annual cost information that is updated or specific to the facility under consideration.
F. Visibility Analysis
Comment 18:EPA underestimated the visibility improvement that would result from SCR by assuming an emissions limit of 0.06 lb/MMBtu (about 84 percent efficiency) instead of 0.05 lbs/MMBtu (about 90 percent efficiency) or lower, which was achieved at 21 coal-fired EGUs in 2011, 11 of which are dry-bottom, wall-fired units like RGGS.
Response 18:EPA disagrees with this comment. As noted previously, the purpose of EPA's independent analyses assessing anticipated visibility improvements and cost-effectiveness of SCR were to evaluate the reasonableness of NDEP's determination based on weighing small incremental visibility improvement against the incremental cost-effectiveness of SCR. The modeling that NDEP relied on assumed that SCR would reduce NOXbetween 78 percent and 82 percent. Although NDEP's assumptions for SCR performance were within the range of emission rates achieved nationwide, EPA determined that for the purposes of visibility modeling and calculating cost-effectiveness of SCR, assuming an 85 percent reduction efficiency to meet an emissions limit of 0.06 lb/MMBtu was reasonable for RGGS. As noted by the commenter, other coal-fired facilities do achieve emission rates of 0.05 lb/MMBtu or lower, and some BART determinations have established a NOXemission limit of 0.05 lb/MMBtu for SCR. However, as noted in Response 12, emissions information reported to EPA's Clean Air Markets program show that among coal-fired boilers operating with SCR nationwide, there is significant variability in actual NOXemission rates achieved, ranging from below 0.05 to greater than 0.10 lb/MMBtu.
EPA's assumption that RGGS could meet an emission limit of 0.06 lb/MMBtu is reasonable and within the expected performance range of SCR. The commenter does not provide a basis,e.g.,modeling that compares visibility benefits expected from a NOXlimit of 0.05 versus 0.06 lb/MMBtu, to change our approval of NDEP's determination that NOXBART for RGGS is an emissions limit of 0.20 lb/MMBtu that can be achieved by installing and operating LNB with OFA and SNCR for the three units at RGGS. EPA anticipates that even if we modeled SCR to achieve 0.05 lb/MMBtu instead of 0.06 lb/MMBtu, the visibility benefits of SCR would still be smaller than the benefits modeled by NDEP. For example, if the post-SCR impact at GCNP is scaled by 0.05/0.06, it decreases from 0.20 dv to 0.17 dv. Relative to the 0.59 dv base case impact, the benefit of SCR would correspondingly increase from 0.38 dv to 0.42 dv, roughly 10 percent higher. However, as discussed in the Technical Support Document (“TSD”) for our proposed rule, EPA's estimates of visibility impacts are more than 50 percent lower than those relied on by NDEP due to differences in modeling procedures. The net effect of using 0.05 lb/MMBtu as the NOXemissions factor would not change the fact that EPA's estimate of SCR's benefit would remain substantially smaller than that estimated by NDEP. As noted in previous responses, NDEP determined that the visibility benefits of SCR based on its modeling do not justify the cost. Thus, additional modeling of SCR at a lower emission rate is not likely to change NDEP's consideration of the visibility factor, or our determination that NDEP's process for weighing the factors is reasonable.
Comment 19:The small visibility improvement from SCR is the result of underestimating the base case emissions and the amount of NOXthat could be removed by SCR. The commenter provided an alternative, larger estimate of SCR benefits by scaling the EPA modeling results.
Response 19:EPA disagrees with this comment. EPA performed an independent modeling analysis to ensure NDEP's NOXBART determination was reasonable. Although estimates of the visibility improvement would be larger if EPA had used higher base case emissions, the scaling method used by the commenter does not accurately reflect the effect of a different base case, which would require new modeling. Even if the commenter's scaling method results were accurate, the estimated visibility improvement remains small. The scaled benefits of SCR provided by the commenter are 0.7 dv at GCNP, and 1.9 dv cumulatively over the five Class I areas; the comparable scaled figures for SNCR would be 0.4 dv and 1.1 dvcumulatively. Thus, using the commenter's method, the incremental visibility improvement of SCR over SNCR would be 0.3 and 0.8 cumulatively. This is only slightly larger than the EPA-estimated benefit increase of 0.2 dv at GCNP, and is the same as the EPA-estimated benefit increase of 0.8 dv cumulatively. EPA's decision to approve NDEP's BART determination would be unchanged. See also the response to comment 20.
Comment 20:A commenter states that EPA used NDEP's NOXbaseline emission rates and control scenario emission rates to determine modeled visibility impacts. Because NDEP's emission rates are based on an annual average instead of a maximum 24-hr average, the commenter alleges that EPA underestimated visibility impacts, and provides its own estimate of 24-hr average baseline and control scenario emission rates.
Response 20:We acknowledge that we used NDEP's baseline and control scenario emission rates, based on annual average emission factors, in the visibility modeling supporting our proposed approval. As noted in our proposal, NDEP modified the baseline emission rates and control scenario emission rates that Nevada Energy included in the BART analysis.14
NDEP did not, however, perform updated modeling to determine the visibility improvement associated with the revised baseline emission rates and revised control scenario emission rates. The absence of modeling results complicated our ability to evaluate the adequacy of NDEP's analysis. To evaluate the adequacy of NDEP's analysis, we performed our visibility modeling using NDEP's revised baseline and revised control scenario emission rates. Again, the purpose of our modeling was to evaluate the adequacy of NDEP's analysis which is not directly comparable to any modeling decisions we might make in our own BART determination as part of a FIP, such as at San Juan Generating Station.
1477 FR 21903.
Regarding the use of control scenario emission rates based upon annual average emission factors (in lb/MMBtu) instead of 24-hour average emission factors (lb/MMBtu), we disagree with the commenter that these emission rates do not provide acceptable estimates of visibility benefits. The methodology for calculating control scenario model emission rates described by the commenter involves applying the estimated control efficiencies of a particular technology to the baseline (pre-control) model emission rate. While this methodology has been used by EPA, it does not preclude the use of other methodologies for calculating control scenario emissions. In the case of control technology performance, engineering estimates of a particular technology's post-control level of performance will often be expressed in terms of lb/MMBtu, either on a 30-day or annual average basis. To the extent that the engineering estimate represents a more accurate depiction of future anticipated emissions at a particular facility, it may be appropriate to rely on the specified post-control level of performance rather than on a control efficiency applied to a pre-control emission rate. In fact, using model emission rates based on an annual average, instead of a 24-hour average, results in more stringent emission rates. As an example, the RGGS Unit 1 model emission rate calculated by the commenter for SCR and LNB with OFA is 99 lb/hr.15
By comparison, the RGGS Unit 1 model emission rate used by EPA for this same technology is 73 lb/hr.16
15See Table 13, National Park Service comment letter dated June 4, 2012, from Susan Johnson (NPS) to Thomas Webb (EPA).
16As used in Model Scenario c16 that is based on the more stringent level of SCR+LNB+OFA performance of 0.06 lb/MMBtu. See Technical Support Document, Appendix C, Docket Item No. EPA-R09-2010-0130-0077-11 and -15.
Regarding the use of baseline emission rates based upon the annual average maximum instead of the 24-hour average maximum, we agree with the commenter that the BART guidelines state: “Use the 24-hour average actual emission rate from the highest emitting day of the meteorological period modeled (for the pre-control scenario).”See70 FR 39170 (July 6, 2005). We note, however, that because the capacity of RGGS is less than 750 MW, NDEP is not required to adhere to the BART guidelines, and is therefore afforded some flexibility when evaluating the five statutory factors in its analysis of RGGS. We disagree th