ML042790400
| ML042790400 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry  |
| Issue date: | 05/26/2003 |
| From: | Barnthouse L LWB Environmental Services |
| To: | Dixon D Electric Power Research Institute, Office of Nuclear Reactor Regulation |
| References | |
| EP-P821O/C4139 | |
| Download: ML042790400 (19) | |
Text
EF'121 APPENDIX A EPRI COMMENTS ON THE EPA PHASE II PROPOSED 316(b) RULE AND NOTICE OF DATA AVAILABILITY LW Environmental
- Services, Inc.
May 26, 2003 Doug Dixon Manager, Fish Protection Research Rlectric Power Research Institute 7905 Berkeley Drive Gloucester Point VA 23062
Dear Dr. Dixon:
At your request,I have reviewedthe Pisces, Ltd. commentson the benefits case study EPA perfonnedto supportthe proposed316(b)rule for existingfacilities. In addition,I havereviewedEPA's March2003Noticeof DataAvailability(NODA),whichdescribed severalsignificantproposedchangesto themethodsusedto calculatebenefitsof reducing impingement and entrainment losses.
My detailed comments are provided in the Attachmentto thisletter;summariesareprovidedbelow.
In its comments on the benefits case study, PISCES claimed that EPA had underestimatedentrainmentand impingementloss rates, underestimatedthe impact of impinging age-l and older fish, used inappropriatelylow survival rates to scale entrainmentlossesto age-l equivalentlosses,and underestimatedthe economicvalueof entrained and impingedfish. In addition,PISCESargued that habitat restorationand replacementprojects shouldnot be used to satisfythe requirementsof Section 316(b) becausein-kindreplacementof entrainedandimpingedfishcannotbe guaranteed.
In reviewing Pisces' comments, I identified several significant errors and misinterpretations,andI findthatmostof Pisces' majorconclusionsareincorrect.
In its analysis of entrainment and impingementloss rates, Pisces used an inapplicabledata set and misinterpretedcooling water withdrawaldata for the Salem GeneratingStation. I found no evidencethat EPA had underestimated theselosses.
Pisces' assertionconcerningthe impactof impingingfisholderthanone year-of-age is partially correct. Howeverthe conclusionthat impingementloss rates, whenexpressedas age-l equivalentlosses,are nearlyas largeas entrainmentloss ratesis basedon invalidreasoningandis incorrect.
EPRI APPENDIX A
EF'121 Pisces' analysisof variabilityand bias in estimatesof naturalsurvivalrates for early life stages of fish is invalidand the conclusionthat survivalrates for all speciesshouldbe increasedby 25%is incorrect.
Pisces' "reproductivevalue" approachto estimatingthe economicvalue of fish that die of naturalcausesignoresdensity-dependenceand would be expectedto greatlyoverstatetheactualeconomicvalueof unharvestedfish.
Pisces' critiqueof habitatrestorationand replacementprojectsis one-sidedand substantiallyunderstatesthepotentialenvironmentalbenefitsoftheseprojects.
In the NODA,EPA announceda changein the assumptionsmade concerningthe age distributionof impingedfish and concerningthe fractionof foragefish biomassthat is converted to harvestablepredator biomass (termed "trophic transfer efficiency"). In reviewing the NODA itself and the supportinginformationprovided in the docket, I foundsignificantproblemswithbothof thesemethodologicalchanges.
The new assumptionconcerningimpingementage distributionsis clearlywrong, is contradictedby data alreadyin the docket, and would greatly overstatethe benefitsof reducingimpingement.
The new assumptionconcerningtrophic transferefficiencyis inconsistentwith the most recent scientificliteratureand wouldoverstatethe benefitsof reducing entrainmentandimpingementof foragefish.
I was greatlyimpededin my reviewby the poor documentationprovidedin the NODA and in the docket. I was unableto reproduceany of the age-l equivalentloss, foregone yield, or productionforegone estimatesprovided in the North Atlantic and Northern Californiacase studies. For this reason,I cannotevaluatethe quantitativeimportanceof the errorsandoverlyconservativeassumptionsI identified. In additionto the problemsI foundin my review,someof thevaluesprovidedinthe benefitstablesfor thesenewcase studiesappearsuspicious,(e.g., the extremelyhigh value of tautogproductionforegone in Table X-7, FR page 13553),however,the informationneeded to confirm whether errorshavebeenmadeisunavailable.
Please contact me if you have any questions concerning this review.
Sincerely, Lawrence W. Bamthouse President
!O5 Wc~:cy I.:wl' Onk Kidg"'. T~I1I1~~~"'~
Phone: (XI15}4~;\\-(J: 00: hlX' (\\(1:'\\i 4~n-(J I ;5: Email: Iwb.env(a>.aUglobal.net 2
EPRI APPENDIX A
EPI2I APPENDIX A
EPRI COMMENTS ON THE EPA PHASE II PROPOSED 316(b) RULE AND NOTICE OF DATA AVAILABILITY ATTACHMENT Reviewof Commentsby Pisces,Ltd. On EPA' 316(b)Case Study Analysis Reviewof Modificationsto the Case Study MethodologyDocumentedin EPA's Noticeof Data Availability Lawrence W. Barnthouse, Ph. D.
LWB Environmental Services, Inc.
Oak Ridge, Tennessee Prepared for the Electric Power Research Institute under Agreement No. EP-P821O/C4139 May2003 3
EPRI APPENDIX A
EPf21 On behalf of the Riverkeeper Organization,Pisces, Ltd. prepared an analysis of biological issues relating to the benefits case study EPA perfonned to support the proposed3l6(b) rule forexistingfacilities.Pisces' analysiswassubmittedto EPA during the commentperiodfor theproposedrule. In March2003 EPA issueda Noticeof Data Availability(NODA)relatedto the rule. The NODAprovideddocumentationof several significantchangesin the case study methodology. In this reportI reviewboth Pisces' analysisandthechangesannouncedintheNODA.
Review of Comments bv Pisces. Ltd. on EPA's 316(b I have reviewedPisces' analysiswithrespectto technicalaccuracyand relevanceto the rulemakingprocess. My commentsare organizedaround five major issues raised by Piscesthat directlychallengethe basisforEPA'sproposedrule. Briefcommentsarealso providedconcerningsomesecondaryissuesthatdonot directlychallengeEPA's analyses or conclusionsbut that supportthe RiverkeeperOrganization'scontentionsconcerning the importanceof minimizingentrainmentandimpingementlosses.
Major Issue I: Underestimation of lossesdue to underestimation of expected future water withdrawal rates.
Pisces argued (Section 1.4) that impingementand entrainmentvary nonlinearlywith flow, and that, because of recent increases in flow rates, EPA's estimates of future entrainmentand impingementlossratesarebiasedlow.
Pisces cited a paper by Kelso and Milburn(1979) as support for the propositionthat impingementand entrainmentare nonlinearlyrelatedto flow. These authorsexamined entrainmentand impingementdata for 37 powerplantslocatedon the GreatLakes, and developedempiricalequationsrelatingflow rate, entrainment,and impingement. Both equationsarebothpowerfunctions,implyingthatentrainmentand impingementincrease exponentiallywith flow. Clearly,sincethe datarelate to the GreatLakesand are more than 25 years old, the equations themselves are irrelevant to evaluation of EPA's analysis. Moreover,as is shownin Figure I, the quantitativesignificanceof the non-linearity described by Kelso and Milburn (1979) is trivial. Figures la and Ib plot entrainmentand impingementrates calculatedfromthe equationsprovidedin the Pisces analysisover an arbitraryrange of 500 to 1600gallonsper second(the actual rangeof flowsused is irrelevant-theplotswouldlookthe sameoveranyrangeof flows). Along with curvescalculatedusing the power functionsfrom Kelsoand Milburn(1979),each plot shows a linear approximationcalculatedby drawinga straightline connectingthe twoendsof the curves. It is clearthatthedegreeof non-linearityis small. Foranygiven flow rate, the differencesbetweenthe curvesand the lines are probablymuch smaller than the uncertainty(not discussedby Pisces)in the impingementor entrainmentrate expectedat any given flow. Moreover,at every flow rate, the linear approximations overestimatethe lossestimatesobtainedusingthepowerfunctions.
Pisces also statedthat EPA underestimatedlossesby failingto accountfor increasesin the mean flowat manyplants. The reportillustratesthis contentionthroughan analysis 4
EPRIAPPENDIXA
51=121 of withdrawalrate data for Salem. Accordingto Pisces, EPA should have used the estimatedwithdrawalrate for 1998,whichwas veryhigh,ratherthana long-tennaverage withdrawalrate. This particularexampleis clearlyinappropriate,becausethe principal use for thesedata in EPA's analysiswas to calculateentrainmentand impingementloss rates, measuredas fish per unit flow. The loss rates werethen scaledto other facilities using estimatedwithdrawalrates for thosefacilities. In any case, 1998withdrawalsare probablynot representativeof future operationsat Salem, becausethis particularyear immediatelyfolloweda two-yearshutdownof both units for a major facilityupgrade.
Representativefuturewithdrawalrates wouldhave to be adjustedfor periodicrefueling and maintenanceshutdowns.
Major IssueNo.2: Underestimationofimpingement due to age 1assumptions Piscesclaimedthat EPA underestimatedeffectsof impingementlosses"bothin termsof their impact on the populations and relative to entrainment" by assuming that all impingedfish are age 1. Pisces supportedthis argumentwith a series of calculations (sections 1.5 and 1.6) in which, insteadof assumingthat all impingedfish are age 1, actual age distributionof the impingementcounts provided in the Salem filing (as reproducedin EPA's inputspreadsheetfor Salem)wereused to convertage 1and older impingementlosses inta age-l equivalents. The results of these calculations,were especially dramatic for white perch, because white perch up to eight years old are impingedat Salem.
Piscesis correctin statingthatthe appropriateapproachfor calculatingage-l equivalents is to scaleall of the age groupsto age 1usingestimatesof the fractionof fishexpectedto survive from age 1 to the age at which impingementoccurred(e.g., for eight-year-old fish, the fraction expectedto survive from age I to age 8). I was able to reproduce Pisces's calculations,and indoingso I foundthatPisces's survivalrates forage groups 1 through 7 are incorrect. These values were apparentlytaken from EPA's input data spreadsheetfor Salem. This spreadsheetcontainsan erroneousfonnula for calculating totalmortalityrates fromestimatesof age-specificnaturalmortalityand fishingmortality rates. The errorinflatesthe totalmortalityrateestimatesfor adultfish,and consequently inflatestheestimatednumbersof age-l equivalentsfor fishthatareage 1andolder.
Although the calculationspresentedin Pisces's analysis are correct in principle, the interpretationof the resultsby Piscesis incorrect. For the purposeof impactassessment and benefitsanalysis,scalingolder fish backwardsto age I is not equivalentto scaling youngerfishforwardto age 1. For age 0 fish,the scalingadjustslossesof eggs, larvae, andjuvenilesto a commonfutureage,priorto theage at whichthe fishmaybe expected to reproduceor to be harvested. Theseestimatescan then be usedto calculateexpected harvestor reproductionthat wouldhaveoccurredat futureages,had these fishnot been entrained or impinged. The same procedure cannot be applied to age-I-equivalent estimatesderivedfrombackward-scalingof lossesthat occurat olderages. The reason for this is thatno reproductivepotentialor opportunityfor harvestis lostpriorto theage at which a fish is actuallyimpinged. For example,each eight-year-oldwhiteperch is, accordingto Pisces, equivalentto 13,572age-I whiteperch (the correctvalue is 5,961 5
EPRIAPPENDIXA
4 h.
epf21 age-l equivalents). Thismeansthat, for every 13,572(5,961)whiteperchaliveon their first birthday,only 1wouldbe expectedto smviveto age 8 years. The 13,571(5,960) fish that did not smvive would have died of natural causes (most likely due to consumptionby predators)or wouldhave been harvestedprior to the age at which the impingedfish was lost. No foregoneyield or reproductionwould accrue due to the deathsof thesefish.
In Table 6 of its analysis,Piscescalculatednumbersof equivalentI-year olds, by age group,for the total numbersof all RlS fishspeciescollectedover all availableyears of samplingat Salem. Figure2 of my commentscomparesestimatesof equivalentl-year-olds (recalculatedusingcorrectmortalityrates),yieldforegone,andproductionforegone for each white perch age group. The followingequations were used to make these calculations:
(Eq. 1)
Where Yj = yield foregone for fish impinged at age i Lj = number of fish impinged at age i Sij = fraction offish smviving from age i to agej (Su = J)
~ =average weight of fish harvested at age j Fj = instantaneous rate of fishing mortality at age j Z]= instantaneous rate of total mortality at agej Ri = LiSoISIJEj j=i (Eq. 2)
Where Rj = reproduction foregone, expressed as age-l equivalents E} = expected egg production at age j, adjusted for sex ratio and % mature So=probability of survival from egg to age 1 All parametersinthe aboveequationsexceptfor the lossratesweretakenfromAppendix L, Tab 18of the Salemfiling(pSEG1999). Figure2 showsthatwhiteperchimpingedat an age of eight years, which accountmore than 75% of the age-l equivalentlossesas calculatedusingPisces' method,accountfor only 7%of the totalyield foregoneand for less than 2% of the totalreproductionforegone. It is true that assumingthat age 1 and older fishare impingedat age 1underestimatesthe yieldand reproductionforegonedue to impingement,however,the magnitudeof the bias is muchsmallerthan is impliedin Pisces' analysis. Pisces's assertionthat impingementlosses-whenmeasuredin terms 6
EPRI APPENDIX A
EI=I2I that arerelevantfor impactassessmentandbenefitsanalysis'-are similarinmagnitudeto entrainmentlossesis erroneous.
Major Issue No.3: Effect of survival rate on age 1equivalent calculations In Section 1.6.2, Pisces argued that survival rates used by EPA in scaling losses of early life stages to age-l equivalents were too low. Support for this argument included (I) a comparison of striped bass and cunner survival rates used by EPA to other published values (2) a sensitivity analysis demonstrating that increasing the assumed survival fractions increases the estimates numbers of age-l equivalents, (3) an assertion that EPA's survivorship estimates already include effects of power stations and therefore are probably biased high, and (4) an assertion that for this reason all of the survivorship estimates should be increased by 25%.
The factthatsurvivalratesof earlylifestagesarehighlyvariableanddifficultto measure is wellknown. Table I comparesempiricalestimatesof survivalratesfromfivedifferent studiesof bay anchovy,one of the most frequentlystudiedof all fish speciesvulnerable to entrainmentand impingement. This tableshowsthat the actualrange of variationin estimatedvalues is much greater even than is suggestedin Pisces's comments. This variability does not imply, however, that EPA systematically underestimated or overestimatedsurvivalratesas comparedtopublishedstudies.
It is definitely not true that the estimates used by EPA in general include station mortality. This could be the case only for survivalestimatesderivedfromsite-specific studiesof populationssusceptibleto stationimpacts.However,survivalestimatesusedin 316(b)demonstrationsareonlyrarelybasedon site-specificdata. In thegreatmajorityof studies,survivalestimatesarederivedfromavailablescientificliterature.
Even if the estimated survival rates did include station mortality,it is not true that conditionalmortalityrates are "often in the 10%-25%range." Such rates have been observedonly at a few sites (most notoriously,the DelawareEstuaryand the Hudson River)and for onlythe mostsusceptiblespeciesat thosesites. Valuesthis highare not representativeof all sites or species nationwide. Hence, there is no justificationfor increasingthe survivalrateestimatesusedintheage-l equivalentcalculations.
Major Issue No.4:
Calculating the worth of commercial species impinged and entrained Pisces asserted that EPA had underestimatedthe economic value of entrained and impingedfish speciesby neglectingto valuethose fish that wouldhavedied of natural causesratherthan beingharvested. Pisces'sargumentis basedon "reproductivevalue,"
definedas the expectedcontributionof a fishat any givenage to futuregenerationsof fish. If the reproductivevalueof eachegg is dermedto be 1.0(sinceat equilibriumone egg will be produced in each generation for each egg produced in the previous generation),thenthereproductivevalueofa fishat anygivenageis givenby:
7 EPRIAPPENDIXA
EPf21
°mn I(x)
V(a) =L ~(x)
""01(0)
(Eq. 3)
Where:
I(a) = fraction of eggs expected to survive to age a l(x) = fraction of eggs expected to survive to age x (x~)
m(x) = fecundity of a fish at age x For age a = 1 year, Equation (3) calculates the number of eggs expected to be produced over the lifetimeof each age-l equivalentfish. Whenappliedto age-l equivalentfish, reproductivevalue as defined in Equation(3) is identicalto reproductivepotentialas defined in Equatior@iivided by the agepsurvival rate.
As noted by Pisces, the economicbenefitsmodelused by EPA calculatesthe expected lifetimeyield fromeach l-year-oldequivalentfish, and then assignsan economicvalue to that yield. Piscesused Equation(3) to assignvaluesfor those fishthat die of natural causesratherthanbeingharvested. Piscescalculatedthenumberof eggsthatwouldhave been produced by the unharvestedfish, multipliedthis value by the fraction of eggs expectedto surviveto age 1,and thencalculatedthe valueof thesesecond-generationt-year-oldsusing EPA'smodel.
Piscesprovidednumericalcalculationsfor stripedbass and for 11 speciesentrainedand impingedat Pilgrim. The survivaland fecundityvaluesfor stripedbass appearto have been taken fromSetzler-Hamiltonet at. (1980)and do not matchvaluesused by EPA or by PSEG(1999). It is not clearwhetherthe age-specificfecundityvalueswereadjusted to accountfor sexratio(whichthey shouldhavebeen). However,the principalproblem with the approachis with the validityof the multi-generationalextrapolation. Pisces's approachassumesthatrecruitmentis directlyproportionalto eggproduction,i.e.,thereis no density-dependence. Numerousrecent studies,as reviewedby Rose et aI. (2001),
have shownthat density-dependentrecruitmentin marinefish speciesis the rule rather than the exception;evidencefor density-dependencein stripedbass is especiallystrong.
For this reason, Pisces's approach should overestimatenext-generationreductions in harvest. In a populationthat is relativelystable from generationto generation,there wouldbe littleor no net lossto thenext generationbecausereducedegg productiondue to the losseswouldbe balancedby improvedreproductionor survivalof those fish that were not entrained or impinged. Even if a reduction in recruit production due to entrainment and impingementdid occur, any reduction in value assigned to these foregonefuturefishwouldhaveto be convertedto netpresentvalueusingan appropriate discountrate.
r-Pisces's comment simply reflects an alternativeand highly conservativeassessment approach,not an erroron thepartof EPA.
8 EPRIAPPENDIXA
51='121 Major Issue No.5: Biologicalissuesimplicitin habitat replacement Section5.1 of Pisces' commentsconsistsprimarilyof a one-sidedvalue-basedargument (non-scientific) against PSEG's Estuary Enhancement Program and other similar restorationprojects. However,three importantand arguablyvalid points are raisedon page 55:
- 1. Habitat equivalency analysis is primarily aimed at offsetting past losses or damage,ratherthancontinuingloss
- 2. Considerableuncertaintyexists as to whether equivalencecan be focused on actualspeciesharmed;and
- 3. Sufficienthabitatto offsetlossesor damagemayoftenbeunavailable.
The firsttwopointsraisedby Piscesaretechnicallycorrectbut irrelevant. Sincecooling water withdrawalsdo not affect the abilityof habitat to performits normal ecological function,the conceptof habitatequivalencyis inapplicableto 316(b)issues,regardlessof whetherthe damageis past or continuing. In addition,it shouldbe obviousthat habitat restorationprojectscannotpossiblybe designedto providespecificnumbersof specific fish species. Successfulprojects can enhancethe productivityand diversityof entire ecosystems,however,the numbersor biomassof individualspeciesthatwillbe produced by any given project cannot be confidentlypredicted. Even the most aggressive proponentsof habitat restorationmake no such claims. In raising the issue of in-kind replacementas a defect in EPA's proposedrule, Pisceshas simplyerecteda convenient strawmanto knockdown.
With regardto point no. 3, Pisces is probablycorrectthat lack of suitablequantitiesof habitatwill oftenpreventcompaniesfromusingrestorationas a meansof satisfyingthe rule. However,this does not mean that these activitiesshouldnot be pursued where feasible.
Regardlessof the objectionsraisedby Pisces,habitatrestorationis a worthwhileactivity that canprovidea widevarietyof tangibleenvironmentalbenefits. The mostobviousof these benefits include enhancedproductionof all types of aquatic biota, provision of habitat for wildlife,and increasedopportunitiesfor aestheticenjoymentand education.
Benefitsof restorationcan be expectedto continuelongafterthe retirementof all of the facilitiessubject to the proposedrule.
In contrast,long-termmonitoringstudieshave providedat best equivocalevidencethatfishpopulationshavebeenadverselyaffectedby entrainmentand impingementlosses. Reducingthoselossesmayproduceno measurable environmentalbenefits.
Other issues Impacts on threatenedand endangeredspecies. The need for additionalreductionsin lossesto protectT&E speciesis raised in Section7.1 of Pisces' comments. However, 9
EPRl APPENDIX A
epl2l these species are already protectedby the EndangeredSpeciesAct. Operatorswhose facilitieshave the potentialto entrainor impingeT&E species are alreadyrequiredto have consultationswiththe appropriateagencies,and to obtaincertificationthat they are not harmingthesespecies.
Problems in calculating age-i equivalents.
Section 1.5(page 8) correctly notes that the validity of the calculations is limited by the quality of available data on stage-specific losses and survival rates. As noted above, there is no indication that EPA systematically overestimated or underestimated these values.
Trendsinabundanceoffishes. Section1.7(pages22-26)notesthat,becauseof improved waterqualityand in somecasesimprovedfisheriesmanagement,the abundanceof some fish species has increased.
Any such increases would likely result in increased entrainmentand impingementlosses. This,accordingto Pisces,meansthatthe economic benefitsof reducingthe lossesmighthavebeenunderestimatedifthe availabledatawere collectedin the 1970s. On the otherhand,somespecieshave declinedsincethe 1970s (Atlantictomcodin the HudsonRiveris citedasan example). In thesecases,usingolder data meansthat the potentialimpactsof entrainmentand impingementon the declining speciesmay have been underestimated. This commentis clearlyan exampleof "spin" and not a technicalcommenton the benefitsanalysis. Sincethe benefitsanalysisis a national aggregate,as long as the increasesand decreasesare roughly balancedthere wouldbeminimaleffectsonthenetresults.
increased species richness andfISh/crustaceanabundancefollowing plant closure. I havepreviouslyreadthroughthe Hendersonet a1.HinkleyPointmonitoringreportthat is cited in Section 1.7.2as supportfor the propositionthat reducinglossesthroughplant closuresis beneficialto fishpopulations.Effectsofreducedwithdrawalof coolingwater from the Bristol Channel are confoundedwith effects of improvedwater quality and regional oceanictemperatureincreasesthat have occurredover the same time period.
Althoughthe authorsclaimthat withintwo moreyearsthey will be ableto test whether stationclosureshavecontributedto theobservedincreases,they provideno indicationof how theywillperformthetest.
Comment on EPA's Noticeof Data Availabilityfor the 316(b)Phase II Existing FacilitiesRule My commentson the NODA addresschangesin the benefits assessmentmethodology that were announcedin SectionX. Both of thesechangeswouldsignificantlyincrease the estimates of benefits resulting from reductions in impingementand entrainment losses. One of these changes, a revisionin the assumptionmade concerningthe age distributionof impinged fish, is consistent with a recommendationmade by Pisces (reviewed above). The new assumptionclearly conflicts with readily availabledata concerningthe typicalages of impingedfish and resultsin greatlyinflatedestimatesof the economic impact of impingementlosses. The second change, a revision in the assumptionmade concerningtrophictransferefficiency(i.e., the fractionof foragefish biomass that is converted to harvestable predator biomass) likely also leads to an 10 EPRl APPENDIX A
EPf21 overestimationof benefits,but the quantitativeimportanceof the changeis impossibleto calculatefromthe informationprovidedby EPA.
Inflatedestimatesof foregoneyieldduetoimpingement In the NODA, EPA states that:
In the case studies prepared for proposal, EPA determined that all impingedfish are age 1 becauseof a lack of data on the actual ages of impingedfish. As severalcommenterspointedout, this biases estimates lowbecauseimpingedfishmayincludeolderindividualsthat arecloserto harvestableage. This is confirmedby data on the agesof impingedfish presented in studies conducted at Salem (pSEG, 1999) and Millstone (NortheastUtilities EnvironmentalLaboratory,1992). To address this concern,the currentstudiesrelaxthe assumptionthatall impingedfishare age 1,and assumeinsteadthat the ages of impingedfish are I and older, and followan age distributionthat is impliedby the associatedsurvival rates. Thisapproachtakesintoconsiderationthe commonobservationthat relatively few older, larger fish are impinged.
The effect of this adjustmentis that a higher proportionof impingedfish are assumed to survive until harvest. As a result of this adjustment,the estimate of foregoneyieldassociatedwith impingementincreasesby a factorranging from about three to ten, dependingon a species' age-specificsurvival rates. [NODA,sectionX.B.3.b(4),68Fed.Reg. 13,546,col.2]
The adjustmentmade by EPA is based on erroneousassumptionsconcerningthe age compositionof typicalimpingementcollections.Becauseof theseerrors,EPA's adjusted estimatesgreatlyoverstatethe expectedforegoneyield due to impingement. Studiesof the agesof impingedfishhaveconsistentlyshownthat:
I) Most impinged fish are younger than one year of age, and not one year old or older as assumed by EPA.
- 2) The vulnerability of most species to impingement decreases with age, so that EPA's use of survival rates to estimate the age composition of impinged fish usually overstates the relative contributions of older fish to impingement losses.
The importanceof theseerrorsis demonstratedbelow,usingthe data for Salemprovided in Docket No. 4-2051 to the Proposed Rule.
The Salem data are used for this demonstration because, unlike the other input data files provided by EPA, the impingementdata for Salemincludea breakdownby life stageand age class. Figure3 plots age distributionsof fish impingedat Salem from 1990 through 1998 for three representative species: weakfish, striped bass, and white perch.
The actual age distributionsare comparedto the distributionsimpliedby EPA's original(proposedrule)
) )
EPRI APPENDIX A
epf21 and revised(NODA)assumptionsconcerningthe agedistributionsof impingedfish. For all threespecies,the impingementtotals arecomprisedprimarilyof fishin thejuvenile 1 and juvenile 2 life stages. Only 0.2% of weakfish, 15%of striped bass, and 26 % of white perchwere age 1 or older. For onlyone of the specieslisted in the Saleminput file, bay anchovy, do age 1 and older fish make up more than half of the total impingementlosses. Figure3 showsthat, contraryto EPA's assumptionin the NODA, no weakfish older than age 1 and no striped bass older than age 2 were reportedin impingementcollectionsat Salemfrom 1990through1998. Whiteperchup to age 8 are impinged at Salem, however, even for this species EPA's assumed age distribution greatlyoverstatestheproportionof fishimpingedat agesolderthanage I.
As in my commentson the Pisces report(above),I used Equation(1) to calculatethe foregoneyielddue to fishimpingedat eachageor stage. Thisisthe sameyieldequation used by EPA [see NODA, section X.B.3.b(I), 68 Fed. Reg. 13,545-46].The total foregoneyield due to impingementis obtainedby summingthe stage and age-specific valuesover all stagesand age classes. Resultsof thesecalculationsare shownin Figure 4.
For all three species, foregone yield estimates calculated using the NODA assumptionsare inflatedcomparedto estimatescalculatedusing actualage distributions.
For weakfish, the NODA value is inflated by a factor of 70 times over the value calculatedusingactualagedistributions.Forstripedbassand weakfish,assumingthatall impingedfish are age 1 alsogreatlyoverestimatesforegoneyield.Only for whiteperch doesthe age-I assumptionunderestimateforegoneyield,andthedifferencein thiscaseis onlyabout20%.
The age distributionof fish impingedat Salemis probablytypicalof estuarinefacilities, and perhapsmost facilities. Verysmallfishhavelowerswimspeedsand smallerenergy reserves than larger fish, and are therefore more vulnerable to being trapped and impinged. Moreover,a largefractionof the speciesimpingedin highnumbersat Salem and otherestuarinefacilitiesspendmostof their lifecyclesat sea. Thesespeciesinclude anadromousspeciessuchas stripedbass,Americanshad,alewife,and bluebackherring; and estuarine-dependentspecies such as weakfish,spot, Atlantic croaker,and Atlantic menhaden. For all of these species, EPA's originalapproachto calculatingforegone yield due to impingementalmostcertainlywouldhave overstatedthepotentialreduction inharvest;EPA's revisedapproachgreatlyoverstatesthisreduction.
Relativelyfew of the species addressedin the case studies(e.g., white perch and bay anchovy)are estuarine-residentthroughouttheir lifecyclesand, therefore,vulnerableto impingementat all ages.
However, the revised approach still would overestimate foregoneyieldbecauseit assumesthatall impingedfishareat leastone yearold.
The extent to which the above commentsapply to EPA's estimates of production foregonedue to impingementis unknown,becauseEPA has providedno documentation of the method used to calculateproductionforegonefor harvestedspecies. Although Chapter5 of the originalcase studyreportstatesthat theproductionforegonemodelwas appliedonly to forage species,both the originalcase study and the new regionalcase studies documentedin the NODA (Tables X-6, X-8, X-20, and X-22, 68 Fed. Reg.
12 EPRI APPENDIX A
EPf21 135,52-53, 13561-62) include estimates of production foregone for impinged fish belongingtoharvestedspecies. If the sameassumptionsusedto calculateyieldforegone for these species were also used to calculateproductionforegone,then the production foregoneestimateswouldbe similarlybiased.
EstimationofTrophicTransferEfficiency In the case study performedto supportthe proposed316(b)PhaseII ExistingFacilities Rule, EPA used a trophic transfer model to estimate the yield of harvested species foregonedue to entrainment'andimpingementofforage species. EPA's modelassumed that 20% of forage species biomassis directlyconsumedby harvestedspecies. EPA assumedthat the remaining80%is consumedby intermediatepredators,whichare then consumedby harvestedspecies. EPA assumedthatthe trophictransferefficiencyfor the direct pathwayis 9%, and that the transferefficiencyfor the indirectpathwayis 0.9%.
These values imply a net transfer efficiency, consideringboth direct and indirect pathways,of 2.5%.
In the NODA (sectionX.B.3.b(2),68 Fed. Reg. 13,546,col. 1), EPA statedthat it had revisedthe trophictransfermodeland wasnowassuminga net trophictransferefficiency of 20%. Thechange,accordingto theNODA,was"basedon an additionalreviewof the scientificliterature." Thechangereflectsa questionablereviewof the scientificliterature and could lead to overestimationof the estimatedbenefitsof reducingentrainmentand impingementof foragefish.
The followingissuesarerelevant:
- 3) Thetrophictransferefficiencyis derivedfroman unrefereedsourceandis at thehigh endof the rangeof acceptedvalues.
- 4) The modifiedapproachassumesthat 100%of forage fish biomassis consumedby economicallyvaluablespecies.
Withregardto the firstissue,the onlycitationprovidedin the NODAto supportthenew value is to "Reedet af. (1994)." Thereferencecitedby EPAis thedocumentationreport for NOAA's Type A NaturalResourceDamageAssessmentmodelfor the GreatLakes.
Thismodelwasdevelopedto facilitatecalculationof naturalresourceinjuriesand service lossescause by spills of oil or hazardoussubstances. Ratherthan providinga detailed reviewof the literatureon trophictransferefficiency,theNOAAreportsimplystatesthat a rangeof valuesbetween10%and30%hasbeenestimatedby variousauthors,provides a brief list of citations,and statesthat the"preferred"valueis 20%. The mostrecentof thepaperscitedinthe reportwaspublishedin 1987. Paulyet al. (1995)publisheda more recent and more thoroughreviewof the literatureon trophictransferefficiency. These authorscompiled140estimatesof trophictransferefficiencyfrom48 trophicmodelsof aquatic ecosystems. They foundthat, althoughthe range of valueswas very wide, the meanvaluewas 10%and onlya fewof the valueswere20%or higher. It appearsfrom Pauly and Christensen'sstudythat the value chosenby EPA is at the upper end of the 13 EPRI APPENDIX A
eF'f21 range of accepted values and probably overstates the average trophic efficiency across all aquatic ecosystems.
Withregardto thesecondissue,althoughthe actualpercentageof foragespeciesthat are consumedby harvestedspeciesis unknown,it is certain that a large fractionof forage speciesproductiongoes to unharvestedspecies,includinginvertebratessuchasjellyfish.
In the LakeTurkanafoodweb describedby Paulyand Christensen(1995),for example, approximately60% of pelagicforagefishbiomassis directlyconsumedby the top-level predators (tigerfishand Nile perch). The remainingforage biomass is consumedby catfish,whichare then consumedby Nile perch. EPA's originalassumptionconcerning the fractionof foragefish biomassdirectlyconsumedby economicallyvaluablespecies may be eitheran underestimateor an overestimateof the actualaveragevalue;the new assumptionclearlyisan overestimate.
EPA claimedin the NODAthat the effectof the changein trophictransferassumptions insignificantbecauseforegoneyield attributableto lossesof foragefish is only a small componentof thetotalforegoneyieldduetoentrainmentandimpingement.EPA hasnot providedsupportinganalysesto verifythis claim.However,the changein assumptions results in a factor-of-eightincreasein all estimatesof yield foregonedue to losses of forage species.
The difference might well be significant for facilities at which entrainment and impingementlosses consist primarily of!orage species. ef ~
References Kelso, J. R. M., and G. S. Milburn. 1979.Entrainment and impingement offish by power plants in the Great Lakes which use the once-through cooling system. Journal of Greal Lakes Research 5:182-194 Pauly, D., and V. Christensen. 1995. Primaryproductionrequiredto sustain global fisheries.Nature374:255-257.
PSEG 1999. Permit Renewal Application,NJDPES Permit No. NJ0005622,Salem GeneratingStation,March4, 1999.
Rose, K. A., J. H. Cowan, Jr., K. O. Winemiller, R. A. Myers, and R. Hilborn. 2001.
Compensatory density-dependence in fish populations:
importance, controversy, understanding, and prognosis. Fish and Fisheries 2:293-327.
Setzler-Hamilton, E. M., W. R. Boynton, K. V. Wood, H. H. Zion, L. Lubbers, N. K.
Mountford, P. Frere, L. Tucker, and 1. A. Mihursky. 1980. Synopsis of biological data on striped bass, Morone saxati/is (Walbaum). NOAA Technical Report Circular 433, FAO Synopsis 121.
14 EPRIAPPENDIXA
"assuming a stage duration of 30 days.
15 EPRIAPPENDIXA EPI2I Table I. Publishedmortalityratesandsurvivalfractionsforbayanchovypostyolk-saclarvae(FromTable 10-10ofPSEG 1999,Appendix1Il,AttachmentC-5)
Study System M (l/day)
Fraction survivin CastroandCowen,1991 GreatSouthBay,NY Min 0.20 0.25%
Max 0.48
<0.001%
Purcelletat, 1994 ChesapeakeBay Min 0.41
<0.001%
Max 4.25
<0.001%
CowanandHoude, 1989 ChesapeakeBay Min 0.08 9.07%
Max 0.23 0.10%
Houde,1989 BiscayneBay.FL Min 0.30 0.01%
Max 0.45
<0.00I%
PSE&G,1984 DelawareRiver Min 0.07 12.25%
Max 0.10 4.98%
epl2l (a)
PowerFunction
- - Linear Approximation PowerFunction
- - Linear Approximation 700 900 1100 1300 1500 Flow (GPS)
Figure 1. Power functions fitted to Great Lakes entrainment and impingement loss data (tTomKelso 1979),
compared to linear approximations over the same flow range.
16 EPRIAPPENDIXA 30000 25000 C/)
CD C/)
C/)
20000 0-I c
15000 CD E
CD 10000 C) c "5-E 5000 0
500 700 900 1100 1300 1500 Flow (GPS)
(b) 600000 500000 C/)
CD C/)
C/)
400000 0-I c
300000 CD Ec "e
200000
.....cW 100000 0
500
EPI2I
. Age-1 Equivalents (I Foregone Yield 1mForegone ReDroduction 1
3 8
4 6
7 2
5 Figure2. Relative contribution of different age groups to total losses of white perch at Salem, over all available years, expressed as age-l equivalents, foregone yield, and foregone reproductive potential.
17 EPRlAPPENDIXA 80%
70%
60%
50%
40%
30%
20%
10%
0%
J1 J2
EPf21 Figure3. Comparisonof actual age distributionof impingementlossesto distributions assumedby EPA (1) in the originalcase study report and (2) in the new case studies documented in the NODA.
This graph was constructed using estimates of total impingementlossesat Salem,by stageandage,from1990through1998.
18 EPRlAPPENDIXA a) weakfish 10000000 8000000 6000000 8 Aclual
.. Case Study 4000000 ICNODA 2000000 0
J1 J2 1
2 3
4 5
6 7
8 b) striped bass I
I 200000 150000 100000 50000 0
J1 J2 1
2 3
4 5
6 7
8 c) white perch 6000000 5000000 4000000 3000000 2000000 1000000 0
J1 J2 1
2 3
4 5
8 7
8
UJ.c
- =, 15,000,000 "0
Q)
>=
Q) 10,000,000 c:oC)
~o LL.
UJ
.c
- =,800,000 "0
Q)
>= 600,000 Q) c:o C) 400,000
~o LL. 200,000 UJ.c-
- 2
<D
>=
8.000
<D c:
o 6.000 C)
~
4,000 oLL.
o EPf2l a) Weakfish b) striped bass c) white perch Figure4. Comparisonof foregoneyieldestimatescalculatedusingtheimpingementloss estimatesfromFigure1.
19 Actual NODA Case Study o
Actual Case Study NODA 2,000 o
Actual Case Study NODA EPRIAPPENDIXA