[the] thermal discharge together with all other significant impacts on the species affected, will assure the protection and propagation of a balanced indigenous community of shellfish, fish and wildlife in and on the body of water into which the discharge is to be made.” 40 C.F.R. §125.73(a). The standard for granting a § 316(a) variance is stringent; Congress intended that such variances be granted only rarely.
EPA Must Require Closed-Cycle Cooling to Adequately Address Entrainment and Impingement
Though both the U.S. Fish and Wildlife Service (in a letter dated December 22, 2010 to EPA, part of AR#042) and Sierra Club’s consultant identified weaknesses in the ichthyoplankton data report prepared by Kleinschmidt between 2008 to 2010, it is evident that impingement and entrainment at this facility is significant. Page 74 of the Fact Sheet states that the operation of the MTGS cooling water intake system (CWIS) has the capacity to result in an annual impingement mortality of around 1,700 adult and juvenile fish and an annual entrainment mortality of more than 16 million larvae. CRC and CLF support Sierra Club’s analysis that EPA’s conclusion in the draft permit is based on a flaw in the cooling tower configuration proposed by the applicant, and CRC and CLF concur with the Sierra Club that the best technology available to reduce the adverse environmental impacts of Mt. Tom’s cooling water withdrawals is conversion to a closed-cycle cooling system.
Section D of the Draft Permit requires the permittee to install, operate, and maintain wedgewire screens and reduce through-screen velocity to reduce impingement of juvenile and adult fish. The draft permit includes no deadline for installation. While CRC and CLF advocate for closed-cycle cooling instead of the wedgewire screens, any technology required under this permit should have a deadline for installation.
Outfall 001, noncontact cooling water, and thermal variance
CRC and CLF’s principal concern with this draft permit is the manner in which the thermal variance application was evaluated. This draft permit represents the first time the facility will be granted a § 316(a) variance to allow the facility to violate state water quality standards for temperature. Therefore, CRC and CLF reviewed the thermal modeling with scrutiny and conclude that there are a number of problems with the model. As CRC and CLF detail below, it appears that the review of the thermal variance has not followed standard practice in terms of data transparency, was not done using critical conditions (using a 7Q10 flow), and appears to not have undergone sufficient internal review by EPA staff.
Does the model represent “real world” conditions?
The 1974 thermal discharge analysis completed by MA F&G (AR#1) shows in Figure 6 that on August 26-27, 1974, the thermal plume extended across the full width of the Connecticut River (Attachment A). The ambient river temperature on that day was 77°F and the average river flow was 2,376 cubic feet per second (cfs). The Fact sheet in Section 7.3.1 states that this study was not considered to be a reasonable “worst case” impact to the receiving water because maximum discharge temperature and the delta T were well below permitted values.
On August 14, 2010, EPA conducted its own temperature measurements from noon to 4:00 PM (AR #34 and #35). Converting the temperatures on these graphs from Celsius to Fahrenheit, it can be assumed that the ambient river temperature on that day was 78-79°F (though the text to Attachment B says that the company reported the intake temperature to be approximately 80°F). Based on data from the USGS Montague gage for that time frame, and given the lag time downstream, flows during the survey were approximately 2,500 cfs. If it is conservatively assumed that the color categories on the graphs listed fourth through ninth (lime green to red) represent the thermal plume, the shallow plume likewise extended across the full width of the Connecticut River on that day. The deep plume also may cross the river, unless there is a different source of warm water coming from the South Hadley side. EPA considered these results to not represent a “worse case” discharge because the discharge and delta T were lower than the maximum allowed. EPA has said that these data were used to “ground truth” the CORMIX model (page 45 of the Fact Sheet), although it is not clear how this was done.
EPA asked FirstLight’s consultants to model our key scenarios. The low flow scenarios used a river flow of 3,000 cfs. River temperature during the summer low flow scenario was to be modeled at 83°F. In EPA’s information request letter to FirstLight dated February 15, 2011 (AR#39), EPA asked FirstLight to model the thermal discharge under spring conditions that assumed an air temperature of 65°F and summer low flow with air temperature of 95°F. Other than the model scenarios shown in table form in Attachments C, D, E, and F of the Fact Sheet, it is unclear what model inputs Kleinschmidt used for their model.
The color depictions produced by Kleinschmidt show four sets of plumes and none cross to the other side of the Connecticut River. In order to understand whether the model is reasonable, when actual temperature studies have shown otherwise, commenters would need to be able to see the inputs to the model.
The EPA Fact Sheet at page 47 states that, “These CORMIX models have been calibrated using Connecticut River water temperatures and MTS thermal plume profiles recorded in the field.” It is unclear how this could be the case. The CORMIX website’s frequently asked questions page indicates that the answer to question #11: “How do I calibrate parameters in the CORMIX models?” is “A. CORMIX does not have any user-adjustable parameters. However, it is suggested that modelers run a sensitivity analysis with representing a range of discharge (velocity, density) and/or environmental conditions (depth, velocity, density stratification) likely to occur at your site.” See http://www.cormix.info/faq.php#11. It is important that EPA provide information as to which temperatures and profiles recorded in the field were used to “calibrate” the model and whether a sensitivity analysis was run. EPA confirmed by email to CRC that no inputs or modeling methodology, other than that given in a report by Kleinschmidt dated May 2011 (AR#42), were provided by the permittee. Therefore, CRC and CLF don’t know what values for water density, river depth, air temperature, river bed morphology, or outfall geometry, for example, were evaluated in the model. Other thermal mixing zone studies using CORMIX models considered by federal agencies have shown the inputs and outputs of the model. See, i.e., “CORMIX THERMAL MIXING ZONE ANALYSIS AND DILUTION STUDY, Chesapeake Bay at Calvert Cliffs Nuclear Power Plant, Maryland (May, 2007), available at http://pbadupws.nrc.gov/docs/ML0906/ML090680538.pdf, and “Toledo Edison Company Bay Shore Station Thermal Mixing Zone Study, January 2003, available at http://www.epa.ohio.gov/portals/35/permits/FirstEnergyBayshore_ThermalMixingZoneStudy.pdf. Documentation of model inputs, outputs and a sensitivity analysis should have been provided to EPA to ensure quality control and adequate evaluation of model results.
In EPA’s §308 letter to FirstLight dated February 15, 2011 (AR#39), EPA asked for a series of maps that “shall represent an overhead view of the river, showing bank-to-bank or lateral isotherms at the surface, at mid-depth, and near the bottom of the river.” Only the surface level maps were ever produced by Kleinschmidt despite EPA’s original letter asking for three depths. EPA’s own study shown in Attachment B indicated that the deep plume (2.2 meters depth) had a different shape than the shallow plume (defined at 0.4 meters). This is important for understanding effects for fish movement and to benthic organisms. What depth the CORMIX model was run at is unknown, but it would be useful to see how the three depths would have looked.
Without knowing more detail about the model inputs, and when compared to actual field results from 1974 and 2010, CRC and CLF do not have a high degree of confidence that this model comes close enough to accurately portraying the thermal plumes. CRC and CLF think more scrutiny should be given to how Kleinschmidt ran the model, and also there should be a comparison of one or two model runs against actual field measurements that would truly “ground truth” conditions in the river.
EPA key findings
On page 63 of the Fact Sheet, EPA states that a key finding that supports the CWA §316(a) variance is that, “The summer two pump thermal plume is mostly associated with the west bank of the CT River downstream from the MTS facility for a maximum predicted distance of about 200 meters.” It is unclear how EPA makes this conclusion. The plume indicating temperatures 5 degrees or greater than ambient conditions is much longer than 200 m in Attachments F and G; Table 5 of the Fact Sheet lists the entire plume as being 1,285 m long.
River depths used
The morphology of the river bottom would affect the CORMIX model outputs, yet CRC and CLF do not know what Kleinschmidt used for assuming river depths in the area downstream of the thermal discharge. In EPA’s §308 letter to FirstLight dated February 15, 2011 (AR#39), EPA asked for a series of maps that “represent a longitudinal profile, showing temperatures at all depths along the longitudinal line of maximum temperatures, with isotherms.” The profiles shown in Attachment G of the Fact Sheet do not show the bottom of the river. At spring flows, the profile assumes a water depth close to shore of 6 meters deep (roughly 18 feet deep) and at summer flows, the profile close to shore assumes a depth of 2 meters deep (roughly 6 feet). In contrast, Figure 8 of the 1974 study clearly shows the shape of the river bottom, the location of the longitudinal transect, and then the isotherm bands have quite a different shape provided in the CORMIX model. (CRC and CLF also note that Page 7 of the 1974 study states that gill nets could not be set in the plume because of shallow depths.)
CRC and CLF include below a recreation map of the Connecticut River that shows rough depths that admittedly vary depending on river flow. Whether spring flows result in depths three times greater than during summer flows is unknown, but the depths in this map seems to call into question the CORMIX model outputs.
The profiles are also misleadingly presented because the scale for the spring plumes are very different than the summer plumes, making it look as though the orange and yellow zones are of comparable length. This is not the case. Water that is 5 degrees or hotter than ambient conditions extends beyond 500 m.
From “Recreational Guide to the Connecticut River, Hatfield MA to Holyoke MA. KM Digital Productions (www.ctrivermaps.com).
Low flow scenario wrongly set
EPA typically uses the 7Q10 (7 day low flow that occurs once every 10 years) to calculate discharge limits. The Fact Sheet Attachment A calculates the 7Q10 at MTGS as 1,772 cfs. However, EPA requested that FirstLight run the CORMIX model using a low summer flow of 3,000 cfs, almost twice the 7Q10. It appears to us to be arbitrary that EPA has decided that although heat is a pollutant, heat is assessed on a different basis than other pollutants. It is also not clear whether EPA considered climate change and a possible increase in variability of flow over the next several decades. Looking at a relatively dry and hot summer as in 2012, with the peaking conditions from hydropower projects upstream from MTGS, river flow levels do drop below 3,000 cfs. See below for graphs of July 1 to September 30, 2012 for the USGS at Montague City (upstream) and Holyoke at I-391 (downstream of the Holyoke dam). As EPA noted in the Fact Sheet at page 44, any discharge determined to qualify for a variance under CWA § 316(a) is automatically deemed to satisfy the state’s surface water quality standards. EPA should calculate a variance-based limit in the same way as other limits are set to meet state water quality standards. Since neither EPA nor the permittee have demonstrated that the variance assures protection of a balanced indigenous population under 7Q10 conditions, the variance-based limit cannot be issued based on the assumptions that were used in the modeling.
Fish assemblage and balanced indigenous population (BIP)
As page 53 of the Fact Sheet points out, no recent studies have been performed by the permittee to characterize the balanced indigenous population of fish in the vicinity of MTS. EPA may have wished to avoid endangered species consultation with National Marine Fisheries Service (NMFS) to do a site-specific study, instead choosing to review a number of reports. The only report that actually looked at fish both in the thermal plume and upstream of it was the 1974 Thermal Plume Study completed by MA F&G (AR#1). Two of the studies are the recent MTS impingement and entrainment study, and the third was completed in 1995 by Northeast Utilities. In 2008, EPA contracted (EPA Contract EP-C-06-033) with Midwest Biodiversity Institute and Kleinschmidt Associates to conduct a fish assemblage and habitat assessment of the entire Connecticut River from Lake Francis to the estuary. Reports on the upper river above Turner’s Falls and raw data for the entire river have been available to EPA Region 1 since 2010, so it is unclear to us why EPA did not use their own fisheries data in order to evaluate a BIP, particularly given that Kleinschmidt Associates was both a consultant to MTGS and EPA during the period of the fish assemblage work and EPA’s development of this draft permit.
EPA concluded on page 58 of the Fact Sheet that the overall Holyoke Pool fish community reflects a balanced indigenous population. EPA chose the Holyoke Pool, a 22-mile long stretch of river that extends from just downstream of the Sunderland Bridge to the Holyoke dam as its target area for this conclusion. A significant portion of the Holyoke Pool is actually upstream of MTGS. The only study cited that compared the downstream vicinity of Mount Tom to an upstream area represents data from 40 years ago. EPA relies upon fish avoidance of the plume in its findings on pages 63 and 64 of the Fact Sheet to justify the § 316(a) variance determination. Given the outstanding questions about the thermal model and the lack of relevant fish, mussel, and macroinvertebrate data for the area of river impacted by the thermal plume, CRC and CLF do not support EPA’s decision to grant a § 316(a) variance.
The Fact Sheet on page 62 states that EPA has determined that converting MTGS’s cooling system to closed-cycle cooling would represent BAT for controlling thermal discharges at the Facility. CRC and CLF support the analysis and conclusions contained in Sierra Club’s Comment Letter, which indicates that closed cycle cooling is comparable in cost to, or even cheaper than, wedge wire screens and would offer better reduction of entrainment and impingement than the solution proposed in the Draft Permit. CRC and CLF also concur with Sierra Club’s argument that MTGS’s analysis did not consider the cumulative impacts of Mount Tom’s thermal discharge in light of the effects of climate change. Given the economic and environmental benefits of the alternative identified in Sierra Club’s analysis, the Draft Permit’s § 316(a) variance is not justified and closed-cycle cooling must be required. In the event that MTGS ceases operation before the requirement to comply with this new permit is triggered, an adequate decision by EPA on the issue of closed cycle cooling and MTGS’s § 316(a) variance application will remain crucially important given that such determinations are relevant to the drafting of NPDES permits for other facilities with thermal impacts.
Comments Regarding Other Outfalls
Permit limits for Outfall 002 (wastewater treatment plant effluent plus drain water from the lime silo area from the FGD system and Bottom Ash Basin A).
• CRC and CLF support the increased frequency of pH monitoring from twice per month to once per week, particularly because of the additional flows to this outfall.
• The permittee should be asked to develop zero discharge strategies for the lime silo area to outfall 002.
• The TSS limit allows an average monthly TSS concentration of 30 mg/L and a maximum daily concentration of 100 mg/L. According to the Fact Sheet at page 25, the average has been 2.63 mg/L and the maximum 21.8 mg/L. Despite EPA’s rationale that the existing limits have been retained to be consistent with other power plants, CRC and CLF recommend the limits be lowered to something closer to an average of 10 mg/L and a maximum of 30 mg/L based on the mandate of the Clean Water Act to eliminate discharges of pollutants. Based on the plant’s performance, the permittee should have no trouble meeting these lower permit limits. This is particularly important as MTGS ramps down operation in anticipation of closure, given the reasonable possibility of between a five and ten-fold increase in actual concentration of TSS with reduced staffing or different facility ownership or operation.
• CRC and CLF support the addition of a Priority Pollutant Scan, but recommend that it be conducted 2-4 times per year initially, instead of once per year. This is particularly important if Bottom Ash Basin A is discharging into outfall 002; a PP scan needs to be done several times when this waste stream is part of the discharge to better characterize what is being adequately treated and what is not.
• If any of the new flows are intermittent, the toxicity testing should be geared towards times that these systems are draining into outfall 002. The Draft Merrimack Station NPDES permit requires toxicity testing 4 times per year (http://www.epa.gov/region1/npdes/permits/draft/2014/draftnh0001465permit.pdf). CRC and CLF recommend that the permittee be required to conduct toxicity testing 4 times a year for at least one year at the start of the new permit.
Outfall 003. This outfall receives stormwater from yard drains. There are no changes to the discharges and no changes to the permit limits. The Fact Sheet on page 35 indicates that, though the permittee considers the outfall to be active, the last time there was discharge that could be tested prior to the 1997 re-application was 1991. EPA should sunset this outfall and eliminate the discharge.
Outfall 004. This outfall was plugged in 2010 and the draft permit does not allow discharge from outfall 004. Stormwater that previously discharged to this outfall now infiltrates into the ground, according to page 35 of the Fact Sheet. A letter from Tighe & Bond to EPA dated November 17, 2008 suggested that a portion of stormwater from 004 would be redirected to outfalls 002 and 007, but it is not clear whether or not this happened. It is also unclear whether or not any mercury contained in the fine particles from the flue gas desulfurization (FGD) system scrubbed out of the smoke stack is now being land applied and volatilized into the air, negating any benefits from reducing mercury emissions. EPA should provide further information or analysis of the ground infiltration of stormwater previously discharged through this outfall.
Outfall 005. This outfall conveys intermittent discharge from traveling screen wash water, service water tank overflow, and fire pump water. CRC and CLF support the change in language regarding pH limits at this outfall.
Outfall 006. This outfall receives stormwater from reflecting pool overflow. The reflecting pool is bounded by the access road and the railroad line, and does not receive industrial discharges. EPA is proposing to eliminate the outfall in this permit. As noted in the Fact Sheet on page 36, there is still a hydraulic connection between the reflecting pool and the Connecticut River. Though not stated in the Fact Sheet, CRC and CLF assume that plugging this outfall would result in infrequent flooding along the access road. CRC and CLF recommend that EPA evaluate the effects of plugging this outfall. CRC and CLF do not concur that stormwater from outfalls 003, 007, and 008 would be representative of stormwater from 006, should a release occur.
Outfall 007. This outfall receives stormwater from highway runoff near the northern end of the site, and during the monitoring period did not discharge. The Draft Permit proposes to add booster fan drain water and baghouse roof drain water from the FGD system to this outfall. Comparing Figure 2 to Figure 3, the latter of which does not show outfall 007 to be upstream of Outfall 1, it is unclear how the new wastes are conveyed to the outfall 007 pipe. EPA should consider whether evaporation of this drain water, leading to zero discharge, could be utilized.
Outfall 008. This outfall receives discharges from Bottom Ash Basin A. EPA states in the Fact Sheet on page 39 that the permittee has requested permission to transfer bottom ash water through a bypass line from Basin A to the special wastewater basin (SWB) that goes to the wastewater treatment plant and ultimately outfall 002. The Fact Sheet states that, “EPA has not only agreed to allow the permittee to send wastewater from Basin A to the SWB if the WWTP is operating sufficiently below capacity, but the Draft Permit proposes to require the permittee to transfer wastewater from Bottom Ash Basin A to the SWB and WWTP to the extent practicable.” However, the Draft Permit in footnote “f” of Part I.A.6 says only, “The water from this basin may be discharged to the Special Wastewater Basin followed by the WWTP for treatment if necessary for enhanced treatment of this discharge.” This is not worded in a way that requires transfers to the WWTP, it only says the permittee may do it. If discharge and treatment through the WWTP truly offers a better level of treatment, the wording of footnote “f” should be made stronger.
EPA is looking for input on ways to handle bottom ash without any wastewater discharges for a plant of Mount Tom’s size and intermittent operation. CRC and CLF acknowledge that EPA’s draft rulemaking recommends a zero discharge BAT for bottom ash, but only for plants that are greater than MTGS’s size. While CRC and CLF cannot opine as to the scaling of particular technology to a plant the size of MTGS, a number of promising technologies are mentioned in a recent article from Power Engineering (http://www.power-eng.com/articles/print/volume-115/issue-2/features/ash-handling-options-for-coal-fired-power-plants.html). CRC and CLF suggest that EPA either hire an independent consultant to look into this question or ask the permittee to review the various available technologies and their costs. If the plant cannot find a way to convert to zero discharge, at a minimum Whole Effluent Toxicity (WET) testing must be required for Outfall 008. When the discharge is routed to Outfall 002, the same parameters must be tested at 002 in addition to WET testing.
EPA has based the draft permit limits for outfall 008 on a treatability study by Tighe & Bond described on page 38 of the Fact Sheet (AR#110) from April 2002. This study showed tests done on sedimentation column testing and ash basin sediment, comparing results from domestic and Chinese coal used at the facility. There was variability in the results depending on the source of coal. At a public meeting held on September 25, 2007 regarding the Mt. Tom’s emission reduction plans, which led to the installation of the Turbosorp system, FirstLight officials stated that they no longer use coal from China, and that they use a more expensive form from Colombia with less sulfur and higher mercury. If the coal used currently is still from Colombia, the testing should be re-done with an eye to revising the draft permit limits for outfall 008 if necessary.
Outfall 009. This outfall received discharges from Bottom Ash Basin B. Bottom ash Basin B has been capped in place and the new permit discontinues the discharge of Outfall 009.
Outfalls 010 and 011. These outfalls formerly discharged flyash transport water from two flyash basins, but the ponds have not been used for many years and the flyash has been removed from the site. These outfalls have been discontinued and removed from the draft permit.
* * *
CRC and CLF appreciate the opportunity to provide these comments. Please do not hesitate to contact CRC or CLF should you have any questions.
Andrea F. Donlon
Connecticut River Watershed Council
Caitlin Peale Sloan
Conservation Law Foundation
Attachment A: Figure 6 of 1974 MA Fish and Game report.
cc: Behany Card, DEP
Kathleen Keohane, DEP
Chris Duerring, DEP
Cindy Delpapa, Massachusetts Division of Ecological Restoration
Joshua Berman, Sierra Club