in section 9.3b recommendation #10 was, “Portions of the 1971 ground surveys by Ainsworth and Associates, Inc. of Greenfield, MA should be resurveyed to identify changes in bank position since the opening of the Northfield Mountain Pumped Storage Project.”
CRC Recommendation: There are several possible ways FirstLight could have analyzed bank movement over time. What they have produced in this report is inadequate and should be re-done.
The Peer Review memo points out several flaws with FirstLight’s use of the Energy Grade Line to parse the river into 4 segments and then use those sections in the extrapolation process. We have also found two instances, described below, in which other relicensing reports describe river fluctuations that influence the river outside of the river segment that FirstLight created:
- Section 2.4 and Figure 2-5 of the October 2016 Alden Report included in Appendix C of Study 3.1.3, notes a Northfield tailwater surface elevation fluctuation range of 5 feet at lower flows. The Report indicates that the RFP stated that, “the fluctuation is not an artifact of operations at Northfield but results from downstream control of the river.”
- Study Report 3.6.6 goes into much detail about the conditions under which the Governor Hunt Boat Launch located just downstream of the Vernon Dam are affected by operation of Turners Falls Dam and Northfield Mountain Pumped Storage. Though Study 3.6.6 concludes that the operation of Vernon Dam has the most control, the operation of the downstream facilities do affect river levels in this area below the Dam.
CRC recommendation: We re-iterate the Peer Review recommendation to look at operational effects in the entire reach, and we do not concur with page iii of Volume I of Study 3.1.2 Report which states that, “The results of the hydraulic and BSTEM models indicated that hydropower operations can only potentially impact erosion processes within the hydraulic reach where the project is located due to the varying hydraulic characteristics of the TFI.”
River bank Transects
On November 23, 2016, FirstLight filed an Answer to CRC’s November 21, 2016 motion to intervene and comments on FirstLight’s application for a temporary license amendment. In the Answer (page 4), FirstLight stated, “CRC’s crude analysis reflects a gross misunderstanding of how field surveying is actually performed. Even when using the same techniques, protocol, and equipment, differing results can occur—particularly when measurements are taken from opposite banks of the river when many locations cannot have permanent ground markers. When plotted, a difference of one foot—or even a few inches—in location from one year to the next can erroneously show dramatic movement in banks. Nonetheless, FirstLight’s survey techniques have improved significantly with advances in technology and the results not only show remarkable consistency, but also verify bank stability.”
In this statement, FirstLight now reveals that there are some transects that do not have permanent ground markers, and are surveyed from the bank across the river. They say that the survey techniques have improved, but that sometimes a difference of a few inches can show dramatic movement in banks. CRC and members of the Connecticut River Streambank Erosion Committee (CRSEC) have long requested that methods to assess banks be written down in a Quality Assurance Project Plan (QAPP). The QAPP submitted as part of the RSP to Study 3.1.1 did not have a protocol for bank transect surveys. Section 4.2.4 included two sentences on methods: “Transect surveys typically entailed surveying the complete cross-section starting at one riverbank, across the channel bed, and up the other riverbank. Permanent markers are typically placed on both banks denoting the start/end points of the cross-section survey to allow for direct comparison of past and future surveys.” We note that the Full River Reconnaissance (FRR) conducted in 2004 by New England Environmental showed diagrams of cross-sections with the location of bank pins indicated in each profile.
FirstLight’s November 23, 2016 Answer on page 4-5 indicates that some of the transect profiles in Appendix E to Study 3.1.2 were inadvertently flipped, but that the data are correct and FirstLight will file an errata report, as needed. To date, no errata report has been filed for this study.
The Field 2007 report Section 7.3a describes rates of erosion at long term cross-sections. It states that the highest peak discharge since 1990 at the Montague gauge was 1998, a year in which the greatest one-year change in bank position did not occur at any of the cross sections. It also said that the most significant period of bank recession for several cross sections occurred in the early 1990’s with average rates of recession ranging between 1.7 and 4.5 ft/year, but that no flood discharges were recorded during this period. These observations seem to be at odds with the results of Study 3.1.2 that reveal erosion only happens during high flow events.
- FirstLight submit SOPs for review to state and federal agencies for bank transect survey work.
- FirstLight submit an addendum that explains which transects have permanent markers and which do not.
- FirstLight explain what transect data was input into the BSTEM model – if FirstLight has determined that some transect data can erroneously show dramatic movement in banks, yet this was used to “calibrate” the BSTEM model, did FirstLight smooth out or modify the transect data to fix these errors?
- FirstLight should clarify if the flipped profiles were input into BSTEM on the wrong bank side, or if the presentation in Appendix E doesn’t reflect how the profiles were input into BSTEM.
- Appendix E transect profiles should be corrected and should be re-drawn showing no vertical exaggeration, and the typical operating range of the river elevations should be shown on each profile.
Hydraulic Modeling and Sheer Stresses
The Field 2007 report used two-dimensional numerical hydraulic modeling using bathymetric data. That report stated in Section 6.0 that, “While erosion does occur where high flow velocities and sheer stresses approach near the bank (Figure 17), significant amounts of erosion also occur where flow velocities near the bank are low (Figure 18 and Appendix 4).” Study 3.1.2 comes to the opposite conclusion, that erosion occurs only during high flow events. However, we aren’t sure what velocities were used in Study 3.1.2.
Section 4.2.3 of the Study 3.1.2 Report states that two hydraulic models were utilized for this effort – HEC-RAS model developed as part of the Study No. 3.2.2 Hydraulic Study and a River2D model created specifically for the Causation Study. Although we have reviewed Study 3.2.2, the Hydraulic Study, we have not seen any of the background data that would tell us what kind of river velocities are experienced at the banks during any flow range or operational parameter.
As for the River2D model, it appears from the description that the River2D model for this study may be different than the 2-D model developed for Study 3.3.9. This should be clarified, particularly with regard to comments that USFWS submitted regarding the roughness co-efficient in this study and other technical comments.
Sheer stresses at flows below 30,000 cfs apparently were not assessed under the RIVER2D model constructed for Study 3.1.2. Pumping and generating with 3 or 4 units apparently was not assessed either, see page 5-29. It is unclear what sheer stress numbers were used in BSTEM for lower flows. If no modeling was done for flows under 30,000 cfs, then it seems like a foregone conclusion that no impacts would be seen if they weren’t even assessed.
The pumping data in Study Report 3.1.2 Volume II Table 220.127.116.11-2 for inflow >30,000 cfs does not quite match what was listed in the September 2016 Alden Report for Study 3.1.3. Also, the pumping velocity for 1 and 2 pumps does not match what was listed in the Alden Report. Please see Table 7 in the September 2016 Alden Report and accompanying text for flow and pump use scenarios.
Study 3.1.2 Volume II Section 5.5.1 shows a few velocity maps, but it appears some areas with velocity vectors were overlain upland areas, so it is not clear what velocities are estimated for the river in contact with the bank.
See also the maps on the next two pages copied from Appendix B of Study Report 3.3.9 – Scenario 36 Map 3 (river flow 40,100 cfs with 4 units generating) compared to Scenario 12 Map 3 (river flow 4,900 cfs, 4 units generating). Though one would need to zoom in, a look at these maps hint that there may be areas where bank velocities are higher at the bank under the lower flow scenario (4,900 cfs) than the higher flow scenario. We could find no evidence in Study 3.1.2 of that dynamic being considered.
Red circles in the next two pages highlight areas that should be compared to look at the higher flow (40,000 cfs) vs. a lower flow (4,900 cfs) under maximum generation of 20,000 cfs. Note the higher velocities near the banks under the lower flow scenario. Velocities of water in contact with the bank are not known.
See also our comments on page 15 later in this letter on Study 3.1.3. Table 7 in the September 2016 Alden report indicates significant uptake of sediments to the upper reservoir during low flows over a given year. This seems to undermine the hypothesis in Study 3.1.2 that erosion only occurs during naturally high flows.
FirstLight should explain in detail how the RIVER2D model was used. If flows less than 30,000 cfs were indeed not modeled, this appears to be inconsistent with the RSP and FirstLight should explain the rationale.
FirstLight should provide detailed data on hydraulic modeling results and sheer stresses at each of the transects to allow for a complete review.
Cyclical process of erosion
Section 7.4 of the Field 2007 report described a cyclic process of erosion, started by the creation of a notch or undercut at the base of the bank by the individual removal of particles. The notch grows taller and steeper, and eventually there will be a topple or slide, and the mass of sediment can be washes away from the bank by water currents. When the report came out in 2007, this description made sense to CRC and the members of CRSEC, because it described a condition we saw happening out on the river.
This process appears to have been lost in the complexity of Study 3.1.2.
Study 3.1.1 Appendix K November 2013 photos were published online, and during a quick look through those photos, we immediately found two examples in which erosion at the toe of the slope appears to be creating other erosion above it.
The photo below is a cropped version of DSC_0764.jpg from Study 3.1.1 Appendix K photos from November 2013. The location of this photo is river left across from Bennett Brook.
|Undercutting (yellow) and slumping (red)|
The photo below is a cropped version of DSC_0997.jpg from Study 3.1.1 Appendix K photos from November 2013. The location of this photo is river right downstream of Kidds Island.
|Note undercutting at the water line and leaning trees with slumping bank above.|
Study 3.1.2 has ignored a key process of erosion. See the Peer Review memo for specific ways to remedy this so that the study can meet the goals stated in the RSP.
Boat Wave Analysis
Section 4.2.8 provides many pages of detail on the input and results of the boat wave analysis. CRC did not have enough time to review this section in any detail.
However, for the purposes of Study 3.6.4, Assessment of Day Use and Overnight Facilities Associated with Non-Motorized Boats, as well as Study 3.6.1 Recreation Use/User Contact Survey, we would like to make note of Table 18.104.22.168-1, which gives results for total numbers of boats (power boats only, or all types?) counted in a season. The data have been copied below. Also Tables 22.214.171.124-3 and -4 give average numbers each day of the week for all weather and for sunny days only.
Table 126.96.36.199-1: Total Measured Number of Boats
|Wave Logger ||Location ||Dates ||Number of Boats |
|WLOG-1 ||Schell Bridge||May 21 – Aug 28||2,133|
|WLOG-2 ||Rt. 10 Bridge||May 21 – Sep 14||2,650|
|WLOG-3 ||French King Bridge||May 21 – Sep 14||7,365|
|WLOG-4 ||French King Bridge||May 21 – Sep 14||7,263|
Progress Reports Lacking
We note that the FERC Determination on the Initial Study Report (ISR) dated January 22, 2015 required FirstLight to provide stakeholders updates after each study task was completed. Since that time, stakeholders received one or two updates in 2015 in addition to the Updated Study Report (USR) filed in September 2015. No progress reports have been provided to stakeholders since the September 2015 USR.
3.1.3 Northfield Mountain Project Sediment Management Plan
In May of 2010, the upper reservoir of Northfield Mountain was drained for routine dredging. Draining the reservoir caused sediments to become entrained in the project’s works, and the pumped storage project ended up being off line for five months while the company cleaned out the sediment. During the summer of 2010, the company began dumping excavated sediment into the Connecticut River until the EPA issued a cease and desist order. EPA Clean Water Act Administrative Order Docket No. 10-016 required that FirstLight, “shall submit a report identifying the measures that it will adopt to prevent discharges of sediments associated with draining the pumped-storage reservoir in the future and a schedule for their implementation.” FERC letters dated August 10, 2010 and January 20, 2011 requested, “a plan and/or procedures to avoid or minimize the entrainment of sediment into the project’s works during similar drawdowns in the future.”
Fulfilling the EPA and FERC requirements were essentially the purpose of this study, but the study was incorporated as a relicensing study and the RSP stated that the study purpose “is to better understand sediment transport and dynamics between the Connecituc River and Upper Reservoir. After a few years of monitoring SSC and conducting annual bathymetric surveys in the Upper Reservior, FirstLight will evaluate management meausures to avoid or minimize the entrainment of silt into the Project works and Connecticut River during future Upper Reservoir drawdowns.”
The study involved the following elements described below.
Upper reservoir bathymetry surveys and sediment accumulation
FirstLight has chosen to conduct annual bathymetry studies to understand the accumulation rate of sediment in the upper reservoir. The study concludes on page 4-1 that the accumulation rate of sediment in the upper reservoir, based on two different methods, is ~4,000 to ~8,000 cubic yards/year.
In Appendix C, Alden Research Laboratory used the bathymetric studies performed in 2010, 2011, 2012, and 2013 to estimate an average of 17,600 cubic yards of “sediment uptake to the Deposition Zone” (page 36 of September, 2016 Alden Report).
Alden used the higher accumulation rate to calibrate a FLOW-3D model. If FirstLight thought the accumulation rate was too high, it is not apparent from the study.
FirstLight should explain the order of magnitude discrepancy between its sediment accumulation rate and Alden’s.
Suspended Sediment Monitoring
Seasonal patterns and trends observed in relation to flow
The study concludes that suspended sediment concentrations (SSC) in the Connecticut River were relatively low and without an apparent trend when flows from Vernon Dam were below 12,000 cfs; increase between 12,000-35,000 cfs, and were significantly higher when flows exceeded 35,000 cfs.
CRC has examined the Figures and graphs in the appendices. What we have observed is that during spring high flows (above 35,000 cfs), SSC levels increase above 40 mg/L. In the spring, flows between 20,000-35,000 cfs experience low SSC levels (generally below 20 mg/L). However, in the summer and fall, SSC levels above 20 mg/L can be triggered by moderate high flow events in the range of 12,000- 30,000 cfs.
We do not concur with the three flow thresholds identified in the report, therefore, and think the report should note the possibility of seasonal thresholds.
Patterns and trends observed in relation to flow, Vernon operations, and Project operating conditions.
Figure 4.2.1-15 shows box plots indicating that Northfield Mountain tailrace samples analyzed during pumping had higher SSC than during generating. This fits with the hypothesis that sediment in the river is deposited in the Northfield Mountain upper reservoir. What wasn’t investigated in detail is whether pumping concentrations exceeded river concentrations. Are there situations where pump concentrations are higher than ambient river sediment? During April 2014, during high flows (40-70,000 cfs) – no. See graph copied below from Appendix D (p. 437 in pdf file).
When high spring flows stabilized and declined, there were hints that SSC levels when pumping (yellow and gray) were higher than the ambient river SSC levels (blue dots). See above and below (copied from Appendix D (pdf page 438). CRC doesn’t have an explanation for a mechanism that would cause higher SSC levels when pumping than is in ambient river levels, other than erosion caused during pumping, but it may be worth additional thought.
Below shows a scenario (graph copied from Appendix D, pdf page 430) in September 2013 when flows from Vernon were stable and elevated (~14,000 cfs) for about three days, and the SSC levels at the Northfield tailrace rose and fell with the Northfield pumping and generating cycles.
Other than Table 4.2.2-3, there is no analysis of SSC levels compared to the number of units pumping and generating. Better analysis is needed, showing 1-4 units pumping (only 2 units pumping is shown for a single date), and 1-4 units generating (1-3 units are shown, each on a single date) and concentrations when idle (only a single date is shown). The analysis should include multiple dates, separated by season.
Sediment Management Techniques Explored
Physical change to upper reservoir intake channel and operational changes
Alden Research Laboratory was contracted by FirstLight to create a 2-D sedimentation model for the upper reservoir and to evaluate sediment management alternatives. Their work was summarized in a report called “Engineering Studies of Sedimentation at the Northfield Mountain Project” and dated May, 2014 (“2014 Alden Report”). It was included as an appendix the December 2014 Sediment Management Plan report.
The 2014 Alden Report compared an existing conditions model to three sediment management strategies. One management strategy was to narrow the intake channel of the upper reservoir, which would increase the velocity of water exiting the upper reservoir. The two other strategies involved lowering the minimum upper reservoir level to 928 or 920 ft, to flush out some of the sediment build-up. All of these options led to a 4-5% reduction in sediment accumulation in the reservoir and intake area compared to current operating procedures.
A sensitivity analysis was performed to model drawdowns to 901 feet to mobilize small annual amounts of sediment. The 2014 Alden Report concluded that an adaptive management plan could be developed to draw down the reservoir level for controlled release of sediment on a periodic basis.
Sediment exclusion structure in Connecticut River intake/tailwater
Alden Research Laboratory was contracted by FirstLight to create a 3-dimensional (3-D) Computational Fluid Dynamic (CFD) model of the Connecticut River Intake/Tailwater to better understand the mobilization of Connecticut River sediment determine if physical modifications to the intake/tailwater could help to reduce future sediment accumulation in the Upper Reservoir t. Their work was summarized in a report called “Engineering Studies of Sediment Uptake at the Northfield Mountain Connecticut River Intake/Tailwater” and dated September 12, 2016 (“September 2016 Alden Report”). It was included in Appendix C to Study 3.1.3.
Table 7 in the September 2016 Alden Report summarizes the model runs, looking at sediment transported during five different representative river flow levels (5,000; 15,000; 25,000; 35,000; and 50,000 cfs), representative sediment concentrations for each flow, scenarios for 1-4 pumps, and weighted against actual operational patterns with these flows. It shows an uptake of 24,155 tons of sediment material transported into the upper reservoir over the course of a year. If you add up the sediment transported under all pumping scenarios for the three lowest flows (approximately 85% of the year, according to the flow duration curve), 13,719 tons of sediment is transported to the upper reservoir during these times of year, or 56.8% of all the sediment in the year.
The September 2016 Alden Report concluded that plant operations and pumping rates have an influence on the amount of sediment uptake to the Upper Reservoir. The model runs showed consistently higher amounts of sediment uptake when 3 and 4 pumps were running. For example, when the river flow was 5,000 cfs and 4 pumps were running, this transported 2,878 tons of material, which was 12% of the annual sediment transport to the upper reservoir. In fact, this scenario estimated more sediment transport than any flow scenario using fewer pumps, except for 3 pumps at 50,000 cfs,
The September 2016 Alden Report concluded that a sediment exclusion structure could be expected to decrease sediment mobilization to the Upper Reservoir by 10-20%.
Physical model testing of exclusion structure
Alden Research Laboratory was contracted by FirstLight, based on a request for proposals in 2015, to conduct field data collection and create a scaled physical model of the Connecticut River Intake/Tailwater with the “main objective” to design and test proposed new civil works to be constructed at the existing Connecticut River intake structure in order to reduce the intake of sediment during the pumping cycle of the plant. The model upstream end is 3.2 km upstream of the Northfield Mountain intake. The downstream model boundary is 0.8 km from the intake, for a total length of 4.0 km. Their work was summarized in a report called “Connecticut River Physical Modeling Project” and dated October 12, 2016 (“October 2016 Alden Report”). It was included in Appendix C to Study 3.1.3.
In Section 2.5.3 of the October 2016 Alden Report, it states, “Based on discussions with GDF Suez the target model sediment concentrations shown in Figure 2-12 were identified. At a river flow of 70,000 cfs, a suspended sediment concentration of about 400 mg/L was targeted.” As the report points out, a flow of 70,000 cfs has a recurrence interval of 5-10 years (page 7) and the target suspended sediment concentration of 400 mg/L is significantly higher than observed values in the river (page 12). CRC is unclear why so much effort was put into modeling a high flow event that occurs relatively rarely, and at a concentration not representative of typical conditions. The motive was not stated. Most of the physical model test runs were based on 70,000 cfs. When they ran the test at 40,000 cfs, they used an SSC concentration much lower than observed results. One wonders what the other objectives of the project were.
Section 6.3 of the October 2016 report indicates that the contribution of sediment to the reservoir during periods of low flow relative to periods of high river flow remains unknown. Based on the physical model results, Alden recommended further analysis to estimate the amount of sediment transported to the reservoir during periods of low river flow. It also recommended exploring the constructability of a moving weir and to conduct the physical model tests during generation.
The Background section in the October 2016 Alden Report states, “The upper reservoir has experienced chronic sediment accumulation; however, the rate of accumulation appears to have increased in part due to an operational change in the reservoir management… Historically, the reservoir level varied between a high of about 1,000 feet and a low of about 920 feet. More recently the reservoir low water level was increased to 938 feet.” CRC is curious about this statement, since the original license mentioned a low of 938 and this, to our knowledge, has been the license limit of lower reservoir level for the history of the project, other than during temporary amendments.
Pilot Dredging of Upper Reservoir
FirstLight conducted a pilot dredging project between April and November of 2015, during which approximately 45,000 cubic yards of sediment were successfully removed by deep water dredging. The study indicates that hydraulic dredging was found to be a viable sediment management measure.
Proposed Sediment Management Measures
After six years of study, FirstLight’s plan is the following:
- Conduct bathymetric surveys of the Upper Reservoir and intake channel at least every two years.
- Excavation of the intake channel and/or other target areas will be planned and initiated “as needed to minimize entrainment of sediment into the Project works during dewatering.
- FirstLight will develop protocols to be followed during 1) an emergency dewatering and 2) during a maintenance or other type of dewatering.
- No operational or physical modifications are proposed.
Absence of protocols
CRC was surprised that protocols were not included in the final report. This is what EPA and FERC requested in 2010.
CRC recommends FirstLight prepare an addendum to Study 3.1.3 that contains the following details:
- protocol to follow during future dewatering events.
- levels (or approximate levels) of sediment accumulation in the upper reservoir that would trigger either targeted hydraulic dredging or dewatering and maintenance. If this is not possible, state a maximum number of years between maintenance activities. Note: CRC asked about this at the study report meeting, and we were told that the information was in the Alden 2014 report. After close read of this report, we believe sediment accumulation trigger levels were not discussed in this report.
- FirstLight should specify which technology will be consistently used for future bathymetry surveys (i.e., multi-beam echosounder surveying) so that years can be compared.
- FirstLight should identify how much storage capacity it has in the upper reservoir area for dredged sediments, and what the plan is for future disposal/storage options.
- Schedule for implementation of plan.
- FL should clarify what operational change in upper reservoir management has increased the rate of sediment accumulation, as mentioned in the October 2015 Alden Report.
3.3.1 Instream Flow Study in Bypass Reach and below Cabot
CRC has been participating in meetings related to this study, and we will be reviewing subsequent filings. At a meeting held on December 2, 2016, FirstLight and stakeholders agreed on additional runs and mapping of currently completed runs.
3.3.2 Evaluate Upstream and Downstream Passage of Adult American Shad
Study 3.3.2 is an important study, with the goal of the study, “to identify the effects of the Turners Falls and Northfield Mountain Projects on adult shad migration.” There are 9 study objectives listed in the RSP, aimed at understanding upstream and downstream migration delays, route selection and behavior, passage rates, and effectiveness of the existing fish ladders. Don Pugh wrote a letter to FirstLight dated March 25, 2016 that summarized all the telemetry information he and other stakeholders wished to see in the report. Don’s letter is attached at the end of this letter. Much of that information was not in the report. In addition, clarification and additional data are necessary to understand whether or not FirstLight followed the RSP and whether or not the study objectives were met. Moreover, stakeholders want to have a good understanding of the results so as to enter into future upcoming license discussions. CRC is recommending that FirstLight revise and re-submit the study report for Study 3.3.2, so that stakeholders can adequately review the methods, results, and conclusions of this important study. We prefer a revised report to an addendum to avoid confusion as to what information is in what report.
- The report identifies the Montague receiver as part of the lower river spoke. Discussions with all stakeholders, prior to the release of the report, noted, with agreement of all, that Montague was within the project. That fish detected at Montague had arrived at the project.
- Tabular data should be presented for the numbers of fish at important locations (Montague, tailrace, entering Cabot ladder, exiting ladders, at base of dam, at upper end of the canal, in the Cabot forebay, etc.) and associated passage statistics including total project passage efficiency.
- Additional tabular data should include but not be limited to: number of hours fish are in the tailrace and number of entries by hour, time spent by fish at the base of the dam, time spent in the canal prior to Gatehouse passage, time spent in the impoundment prior to downstream passage (both routes), time spent in the canal during downstream passage.
- The terms survival, transition, recapture, attraction are at times confusing (e.g. Cabot attraction, Spillway attraction are unclear as to whether that means entry of close proximity). These should be clarified.
- For all state tables and charts, the number of fish represented should be included.
- Adjust numbers of fish detected by eliminating ‘fish’ that were spurious detections (e.g. Holyoke released fish only detected in the canal or with few detections at a single location that has a high proportion of spurious detections).
- Heat maps that include bypass flows should be limited to 7,500 cfs to understand in better detail the potential effects of manageable flows as opposed to flows beyond control by the project
3.2 Study Design and Methods
- The report states that the fishways were operated in the normal method with one foot differentials at the entrance. Does this apply to the old entrance to the Gatehouse ladder?
- The number of detections at T1, T2, and T3 should be included. T3 is considered as the arrival at project location.
4.4 Mobile Tracking and Evaluation of Mortality
- Mortality determined from receiver detections for a prolonged period of time with no subsequent location up- or downstream should be included.
- Route specific mortality should be reported (e.g. x number of fish pass through the units and y number died).
4.5 Data Reduction
- How were single detections evaluated for validity? Fig 4.5-2 would seem to indicate that all single detections were false.
4.6.1 Holyoke to Montague
- It is not clear how hours in the state table relate to individual fish movement as it appears to group all fish. For example, there were 514 hours when fish remained at the Canoe Club, but was this one fish or 25 fish, and if it was many fish, what was the distribution of hours?
- The report needs to define the “project”. Previous discussions, with all stakeholders, defined arrival at the project as detection at Montague. It should not be included with T1 or T2 as lower river.
4.6.2 Montague Spoke
- Tabular numbers of fish at Montague, moving to Smead Island, the tailrace and the bypass will enhance the understanding of fish behavior at this location.
- Fish likely made multiple attempts moving upriver from Montague and this should also be noted. Time spent at Montague should be evaluated including the effect of flow and diel period.
- Some fish moved from Montague to the bypass (Conte discharge) without being detected at Smead Island or the tailrace but are not discussed in the text. Others were detected at the tailrace far field antenna for a brief period indicating movement through that area, but not a “stop” at the tailrace. It is not clear which category these fish fell in: tailrace or bypass. They should be included in the Montague to bypass category. For these fish, the time of detection at the tailrace that is determinative of route/location and the justification for that time should be provided.
4.6.3 Cabot Ladder Attraction
- Tabular data for numbers of fish in the tailrace, entries into the ladder, range/distribution of entries by fish, time from first detection in the tailrace to entry, number of hours fish were in the tailrace would provide context to movement probabilities.
- “The state table counts 137 forays into Cabot ladder, with 120 from the tailrace, 8 from downstream receivers and 9 from the bypass reach. This number of forays differed from the sum of the number of forays per fish according to the raw recapture data.” (Quote from 4-55) How and why they are different should be explained.
4.6.5 Bypass Reach
- Again, tabular data would assist understanding movement in the bypass reach: how many fish moved to either side of Rawson Island, how many fish passed each side and how many failed, what were the times fish spent in the east side, how many fish passed Rawson Island undetected and what were the flows associated with success or failure in passing.
4.6.6 Spillway ladder attraction
- Figure 4.6.6-2 describes fish approaching the spillway ladder by time of day. It is unclear if approach means ‘in the proximity of’ or ‘enters the ladder.’ For the limited number of fish in the pool below the dam, 144 ‘approaches’ seem high.
- Paragraph 5 says 11 dual tagged fish made at least one attempt on Spillway ladder (entered?) and in the next paragraph it says, “In total, 34 dual tagged fish made at least 17 successful attempts into the spillway ladder from the spillway,…”. This seems contradictory or confusing.
- A more complete analysis of the time from the first Montague detection to the first detection at the dam T-19 & T20 should be done to assess delay associated with finding and entering the spillway ladder.
- During different periods of the study Station #1 operated or did not. Bypass flow is earlier described as spill plus Station #1. For the Spillway ladder attraction model it is not clear what flow is used in Tables D-1.6-1, D-1.6-2 and the histogram of bypass flow. For analysis of entry into the bypass or passage at Rawson Island including flow from Station No. 1 is appropriate. For entry into the Spillway ladder it is not.
4.6.7 Spillway Ladder Efficiency
- Spillway ladder entrance efficiency for both dual and PIT only tagged fish is stated as 91.5%. It is unclear how this is derived. Ladder entrance efficiency is generally the number of fish that enter divided by the number available/in close proximity (detected in this case by T19 or T20). Since PIT tagged fish cannot be detected in the pool, they cannot be used in calculating a measure of entrance efficiency. A table of fish in the pool below the dam the fish that entered, and the number of attempts would be appropriate.
- Overall ladder efficiency is the number of fish that pass, divided by the number of fish available. Again, as PIT tagged fish are not ‘available’, overall ladder efficiency can only be calculated with dual tagged fish as opposed to internal efficiency which can use PIT tagged fish. Tabular data of PIT and dual tagged fish would better describe the performance of the ladder.
4.6.8 Upstream Migration through the Canal
- An analysis of the telemetry database shows only 6 dual tagged fish passed the Cabot ladder making a total of 56 dual tagged fish for canal upstream movement.
- Of those 22 were detected at T22 (downstream of gatehouse). Four fish that came up the Spillway ladder were also detected on T22. It is unclear if the detections at T22 were if the fish dropped into the canal or when they were in the Spillway ladder not, but they cannot be considered as fish that moved from the lower canal to the head of the canal
4.6.9 Gatehouse Ladder
- Gatehouse ladder efficiency should be calculated as the number of fish passed / number of fish available (detected at T22).
4.6.10 Upstream Migration through the TFI Impoundment
- Movement and delay times at Northfield Mountain Pumped Storage Facility(“Northfield”) for all operating conditions should be more fully detailed. The report states for upstream movement past Northfiel,d there was some delay and milling at the intake. The numbers of fish, project operation status, and their respective delays should be provided along with analysis.
4.6.11 Downstream Migration through the TFI Impoundment
- Impoundment-released fish that were detected at Shearer or Stebbins Island and then moved downstream should be included in the analysis.
- A more complete analysis of delay at the Northfield intake, including specifics as to delay and the effect of operations on that delay, is needed.
31 4.6.12 Downstream Migratory Route Choice at Turners Falls Dam
- In addition to impoundment-released fish and TransCanada fish, FirstLight fish that passed into the impoundment should be included in the evaluation of downstream route.
- Downstream route choice and delay by route should be analyzed in relation to flow to the canal, to spill and the number of fish available under different conditions.
- Delay at the dam should not include impoundment released fish that move downstream within 24 hours of release.
4.6.13 Downstream Migration through the Canal
- Multi-state Markov has 86 fish “moving through the telemetry subnetwork” and Time to Event uses 98 fish. An explanation as to why different numbers were used would be helpful.
- The second paragraph states that 28 fish transitioned from the Cabot forebay to the tailrace but the fifth paragraph says 37 passed from the forebay through the turbines (pg. 4-88). This is confusing.
- The sixth paragraph on page 4-88 states, “Fish at the Downstream Bypass were most likely to be detected next at the Cabot Forebay, though the probability of next detection decreased with increases in flow (93% at 25th percentile flow decreasing to 97% at 100th percentile flow; …” It seem like the probability is increasing when it goes from 93 to 97%.
- Last paragraph on page 4-88 says that fish passed quickly through the downstream bypass while the text on page 4-89 (2nd paragraph) says that there was a large delay for fish using the downstream bypass. How are these reconciled?
Discussion and Conclusions
- The 2nd paragraph says, “Fish preferred to move into the Cabot Tailrace during times of low flow,” while the 3rd paragraph says, “Attraction to the Cabot ladder increased as Cabot discharge increased, suggesting the discharge from the powerhouse provides attraction flow.” These statements appear to be contradictory.
- Though just over half of the fish released at Holyoke were detected, a better metric for fish reaching the project is the number of dual tagged fish detected at the project (94) divided by the number detected (154) or detected at Rt. 116 or above (116) to account for post tagging effects. In that case, movement to the project would be 61.0% or 81.0%, respectively.
3.3.3 Evaluate Downstream Passage of Juvenile American Shad (Interim Report)
Problems associated with milling at the hydroacoustic installations other than Cabot, and the erratic swimming and poor survival of radio tagged fish, negate the results of this study. As such, no information on entrainment at Northfield or route choice of juvenile shad at the Turners Falls dam or in the canal is available. FirstLight made a good faith effort to re-do the telemetry portion of this study in the fall of 2016, but river conditions (low flow due to a serious drought all summer) would have made any results of little value, and the study was not done. CRC recommends that the study be repeated in 2017.
Our additional comments on specific parts of the Study Report follow.
4.1 Run Timing, Duration, and Magnitude (Hydroacoustics)
- For the Northfield Mountain and Power Canal locations that had significant milling, a review of the data to evaluate the amount of milling may provide a general picture of the timing of shad movement.
- Section 4.1.3 in the report fails to note that the study was conducted during a season in which only three of the possible four turbine pumps/generators were in operation.
4.2.2 Routes of Passage
- The report says 129 fish were released above Shearer. Of the 129, 77 were detected at Shearer and 24 detected in the Northfield forebay and 3 in the upper reservoir. This does not provide sufficient information as to whether fish different fish were detected down river from Shearer. A table with individual fish ID’s at each location would be very helpful to better understand movement.
- The description of the fate and routes of fish released above the Turners Falls dam is inadequate.
- A calculation of a 3.9% entrainment requires that the 77 fish detected at Shearer all pass Northfield.
- Based on the fish loss from upstream station to downstream station, it seems overly optimistic to assume that no fish were lost between Shearer and the project. As such any calculation of entrainment would be bias.
- Three fish were entrained and 21 last detected at the intake of the Northfield station. It is likely that some of the 21 fish were entrained, and that they lost their tag during the pumping cycle (highly likely as tags were lost in the control tank) or that they were not detected in the upper reservoir. If all 21 were entrained and 77 did pass, the entrainment rate would be 31.2% (24/77). And if only the fish last detected during pumping were entrained, the entrainment rate would be 22.1%. Again, this assumes that all 77 Shearer fish passed the project. While this is somewhat speculative, the potentially high entrainment rate reinforces the need to repeat the study.
- CRC agrees that the problems with the study call into question the telemetry results. A similar study done at the Vernon project (FERC # 1904) was successful in assessing routes of passage. While monitoring juvenile shad is difficult, it is not impossible, and not a justification to not repeat the study. The potential entrainment of more than 30% of juvenile shad passing the project and the high mortality of shad spilling over the dam would severely impact restoration efforts in the Connecticut River.
3.3.6 Addendum: Impact of Project Operations on Shad Spawning, Spawning Habitat and Egg Deposition in the Area of the Northfield Mountain and Turners Falls Projects
We spent some time comparing Table 1 of the Addendum (2005-2009) with Table 4.1-1 of the Study Report (2010-2014) side by side. Each shows the discharge changes/generation changes over 5 years for the hours of 8 PM to 2 AM for May and June. For 2010-2014, the total number of decreases = 130 and total number of increases = 216. During this 5-year period, there were more increases in generation than decreases. For 2005-2009, the total number of decreases = 483 and total number of increases = 242. For this earlier 5-year period, there were many more decreases in generation than increases.
For 2010-2014, 33% of changes were increases of 0-10 MW, followed by 22% decreases of 0-10 MW and 20% increases of 10-20 MW. For 2005-2009, 46.3% of changes were decreases of 0-10 MW, followed by 19% increases of 0-10MW and 15.7% decreases of 10-20 MW. Generation change decreases in the highest category went from happening almost never (only 1 time total from 2005-2008) to something happening at least 1 time per “season” up to 5 times/season.
It seems that operations in one 5-year period, 2005-2009, was quite different than the following 5-year period, 2010-2014. What that means for spawning, or any other study, we don’t know. We also are not sure what this means for future operations or for using 2002 as a “typical” year in other studies, since it we don’t have information for 2002 and it appears that operations have not stayed similar over the past 10 or so years.
3.3.7 Fish Entrainment and Turbine Passage Mortality Study
The last paragraph of the Executive Summary states that, “Operation of the Northfield Mountain Project may impact fishes due to entrainment. However, pumping operations generally only occur over a few hours between midnight and 6:00 a.m., thereby limiting impacts to a 6 hour period each night.” No further information is included in the body of the report regarding this statement. CRC does not have access to operational data, other than the data filed with FERC during temporary amendment periods during the winter. That data show that pumping often lasts until 7 AM, 8 AM, or 9 AM in the morning during the winter months. Also, in March of 2016, the data show a few afternoons when pumping occurred. We have no way of verifying whether or not that is true during the later spring and summer; however, we question the validity of FirstLight’s statement until more information is provided. Moreover, there are no restrictions on times of day that the facility can pump or generate.
CRC offers the following additional comments on the study report:
- The adult shad telemetry study (Study Report 3.3.2) notes that six fish detected were detected at the Station No. 1 forebay with 7 successful escapes (pg. 4-72). The Kleinschmidt Associates (contractor to FirstLight) database has 7 fish including 2 released in the impoundment including one which did not escape. Clarification is needed.
- The estimate of 3.9% entrainment at the Northfield project of juvenile shad represents the absolute minimum, as not all 77 fish detected at Shearer likely passed the Northfield project and more than 3 fish were likely entrained.
3.3.13 Impacts of the Turners Falls Project and Northfield Mountain Project on Littoral Zone Fish Habitat and Spawning Habitat
The RSP to this study listed one of the specific objectives as, “delineate, qualitatively describe…, and map shallow water habitat types.” The report shows a map with dots for nesting spots identified, but no delineation of habitat types.
The RSP had many maps showing the study area for this study (Figures 3.3.14-1, Pages 1 through 23[we think the numbering should have been 3.3.15-1]). It appears that the entire study area could not have been covered in the two field days devoted to this effort in May and then June.
Task 2 of the RSP indicates that FirstLight was supposed to observe tributariess identified in Study 3.3.17 as accessible during spawning seasons. The report says they looked at “major” tributaries and list a few by example, but it is unclear if Study 3.3.17 was consulted in any way.
The second paragraph of Section 3.1 in the Study Report said the littoral zone was considered to be the area extending from the edge of the water line at the shore of the time of survey to 6 ft in depth. Relying on 6 ft of water during the field visit is a little odd, since water level fluctuated by 2 ft or so during a field day. Also, the RSP maps show some potential littoral areas in the middle of the river. Did the field crew confirm that those areas were not good candidates?
The raw data sheets in Appendix A only contain sites 001-006 for the early spring surveys. Sites 8-17 were not included.
The literature review section 4.1 is very paltry. Did the Fish Assemblage Study 3.3.11 results have any bearing?
All the figures showing the sites and the unsteady and steady state flows (Figures 4.3.2-1 through 15) indicates that many sites are very susceptible flow fluctuations. This seems to be glossed over in the text.
We believe that the flow duration curves for the Turners Falls dam from the PAD dated October 30, 2012 (Figure 188.8.131.52-19 from the PAD copied below) are essential to the analysis of the “steady state” graphs provided in this report. In this study, steady state is a simulated condition under a Turners Falls elevation of 176, 181.3, and 185 ft. For example, in Figure 4.3.2-3, Site 10, if you brought the Turners Falls pool level down to 176 ft, you would need a flow of 25,000 cfs just to have the water level match that of the spawning habitat. Looking at the flow duration curve, that only happens 30% of the time in May. That seems like a high impact, if the facility operated under its licensed conditions.
3.3.15 Assessment of Adult Sea Lamprey Spawning within the Turners Falls Project and Northfield Mountain Project
CRC offers the following comments on the study report:
- The FERC Study Plan Determination dated February 21, 2014, stated on page B-71, …”we recommend FirstLight not limit its detailed monitoring to only 25 redds, but utilize all survey data, including the location and depth of suitable habitat and redds, for comparison with results of the hydraulic model in study 3.2.2. FirstLight should then determine if spawning areas/redds are subject to dewatering and describe the degree of project-related water level fluctuation at each spawning site.” The hydraulic model was used only to confirm that during the period from June 19 to July 10 that the observed redds were not dewatered. This is a single year of data when flows during this period were generally high. Redd locations observed during the study and all suitable habitat should be evaluated using a low flow year.
- Because FirstLight did not adequately use the hydraulic data, we do not concur wioth the conclusion that there is no project effect at all spawning sites in the study. For comparison, Table 5.3-3 in TransCanada’s Study 16 showed that some of the sea lamprey nests near Stebbins Island were exposed 5% of the time. There was nothing like that in the FirstLight observations.
- Depths and velocities from field notes when sea lamprey were observed on redds should be used to revise the Habitat Suitability Index (HSI) curves.
- Revised HSI curves should then be used to revise habitat mapping.
3.3.16 Habitat Assessment, Surveys and Modeling of Suitable Habitat for State-Listed Mussel Species in the Connecticut River below Cabot Station
CRC did not review this study report in detail, but we would like to point out that having the consultant who was hired to prepare the mussel report (Ethan Nedeau of Biodrawversity) also sit on the Delphi panel compromises the objectivity of the results.
3.6.6 Assessment of Effects of Project Operation on Recreation and Land Use
This study used results from multiple other relicensing studies to analyze the effects of project operation on project recreation facilities and land use. The study did not assess the effects of project operation on the ability to recreate in certain areas. For example, the current minimum flows in the bypass reach prevent the use of boats in that stretch, but that kind of effect was not assessed in this study.
Overall, the study preparers expended little effort to produce this report and the data involves little meaningful analysis. Below are our comments.
4.2.2 Pauchaug Boat Launch. As we discussed at the Study Report meeting held on November 1, 2016, the presentation of water level data in this report leaves much to be desired. The analysis involves median monthly water elevations and water surface elevation curves for each summer month. What is most important is the daily fluctuation below a minimum level. What happens in the middle of the night when people aren’t boating is irrelevant.
CRC looked at the actual logger data provided as part of Relicensing Study 3.2.2 in Excel format. The report states that 3 feet of water at the end of the boat ramp is necessary for launching and/or retrieving boats on trailers, so a water surface elevation (WSEL) of 181 ft is necessary for the boat ramp to be usable for power boats. From this, we see that it’s not unusual at all for river levels to drop below the 181 ft level during the night-time early morning hours, making it difficult to launch boats until mid morning or noon, or even later. In Sept. of 2014, there were even a couple of stretches where the river level was too low for the better part of two entire days. This happened twice. September of 2014 was more typical of dry summer conditions than the rest of the summer, which was on the wet side.
See graphs below.
River elevations from FirstLight logger located downstream Pauchaug
Same data, but shorter time span showing 8/8 to 8/13 in 2014.
River elevations from FirstLight logger located downstream Pauchaug
Same data, shorter time span showing 9/1 to 9/4 in 2014, showing typical dry weather summer flow conditions. Note that river elevations tend to dip below 181 ft around 3:00 AM and then rise above 181 between noon and 3:00 PM. This would tend to make the river unusable to motor boats all morning into the early afternoon on a typical summer day.
4.2.3 Munn’s Ferry Boat Camping. The results from Study 3.3.9, showing conditions when the river flows upstream and strange eddies, do not seem to have been considered.
4.2.4 Boat Tour and Riverview Picnic Area. This section evaluated the use of power boats at this location only. Study Report 3.6.4 listed the Riverview Picnic Area as a formal river access site. If that is what FirstLight considers this site to be, and dismisses the need for additional water trail access points, then Study 3.6.6 needs to assess project effects for paddlers in this location, including operational impacts that cause the river to flow upstream as shown in Study 3.3.9.
4.2.5 Cabot Camp Access Area. Study Report 3.6.4 listed the Cabot Camp Access Area as a formal river access site. If that is what FirstLight considers this site to be, and dismisses the need for additional water trail access points, then Study 3.6.6 needs to assess project effects for paddlers in this location, including operational impacts that cause the river to flow upstream, or create eddies, as shown in Study 3.3.9.
4.2.10 Poplar Street Access Site and 4.2.11 Sunderland Bridge Boat Launch. Our comment from Study 3.3.2 was that based on the August 11-16, 2012 graph downstream of the Turners Falls dam (Appendix C of Study 3.3.2), peaking flows out of the Cabot units can result in 5-ft sub-daily fluctuations in Montague and 4-foot subdaily fluctuations at the Sunderland Bridge in the middle of the summer. Flows rapidly decrease at midnight until mid morning or mid day, then steadily increase during the latter half of the day. At Poplar Street, our experience and anecdotal stories indicate that higher water levels can make launching a boat more dangerous and difficult. As for the Sunderland Boat launch, the graphs provided do not tell us the full story, and no user surveys were conducted in that area, despite CRC’s request during the review of the RSP (see RSP page 3-352).
The whole point of conducting these expensive studies is to better inform all involved in the relicensing effort so that we understand the project effects. When we are given flow duration curves, that obscure the true issue of the subdaily fluctuations of the river, we are little better off than we were at the beginning of this process. We knew then that river users complain about low river levels in the morning at Pauchaug and at the Barton Cove state boat ramps, for example. This study has done little more to add to the understanding.
We appreciate the opportunity to provide comments on the studies submitted on October, 2016.
Princeton Hydro peer review memorandum
Letter from Don Pugh to James Donahue of FirstLight dated March 25, 2016, regarding shad telemetry data presentation for Study 3.3.2
 “Fluvial Geomorphology Study of the Turners Falls Pool on the Connecticut River Between Turners Falls, MA and Vernon, VT. Prepared for Northfield Mountain Pumped Storage Project. Prepared by Field Geology Services, Farmington, ME. November, 2007.