{"Bibliographic":{"Title":"Juvenile radio-tag study : Lower Granite Dam","Authors":"","Publication date":"1986","Publisher":""},"Administrative":{"Date created":"08-16-2023","Language":"English","Rights":"CC 0","Size":"0000061535"},"Pages":["uvenile Radio-Tag Study:\nSH153\nUn5511\n1988\nLower Granite Dam\nlibrary\nOAA,\nAlaska\nNational\nFisheries\nMarine\nCenter\nBoulevard,\nFisheries\nWA\n98112\nService\nE.\nFinal\nReport\n1985-6\nDepartment of Energy\nU.S. Bonneville Power Administration\nDivision of Fish & Wildlife\nDepartment of Commerce\nU.S. National Oceanic & Atmospheric\nNational Administration Marine Fisheries Service\nNorthwest & Alaska FIsheries\nCoastal Center Zone & Estuarine\nStudies Division\nMarch 1988","This report was funded by the Bonneville Power Administration (BPA), U.S. Department of\nEnergy, as part of BPA's program to protect, mitigate, and enhance fish and wildlife affected by\nthe development and operation of hydroelectric facilities on the Columbia River and its\ntributaries. The views in this report are the author's and do not necessarily represent the views\nof BPA.\nFor copies of this report, write to:\nBonneville Power Administration\nDivision of Fish and Wildlife - PJ\nP.O. Box 3621\nPortland, OR 97208","SH\n153\nUn5511\nJUVENILE RADIO-TAG STUDY:\nLOWER GRANITE DAM,\n1988\nLibrary\n2725 NOAA, Northwest NationAlaska & MariFisheries\n1985-86\nWA Fishericenter E\nService\nFinal Report\nby\nAlbert E. Giorgi\nLowe11 Stuehrenberg\nJohn Wilson\nFunded by\nU.S. Department of Energy\nBonneville Power Administration\nDivision of Fish and Wildlife\nContract DE-A179-85BP21237\nProject 85-35\nand\nCoastal Zone and Estuarine Studies Division\nNorthwest and Alaska Fisheries Center\nNational Marine Fisheries Service\nNational Oceanic and Atmospheric Administration\n2725 Montlake Boulevard East\nSeattle, Washington 98112\nMarch 1988\nMAR 19 1997","ABSTRACT\nIn 1985 and 1986, research was conducted at Lower Granite Dam to assess\nthe feasibility of using a miniaturized radio tag for estimating spill\neffectiveness, fish guidance efficiency (FGE), collection efficiency (CE), and\nsurvival at the dam. The results indicate that the tag can provide acceptable\nestimates of powerhouse and spillway passage, that FGE and CE estimates may be\naffected by the chinook salmon smolts inability to compensate for the tags\nweight, and that survival estimates could be frustrated by an inability to\ncompletely separate dead fish bearing live tags from live tagged fish moving\ndownstream. The passage model developed for Lower Granite Dam is applicable\nto other dams that have similar smolt passage configurations, and it can be\nadapted to situations with more passage routes.","CONTENTS\nPage\nINTRODUCTION\n1\nPART I:\n1986 FIELD TESTS\n2\nMethods and Materials\n2\nStudy Area\n2\nEqui pment\n4\nResults\n13\n1986 Passage and FGE Evaluation\n13\nForebay Monitor Evaluation\n15\nTailrace Release\n16\nSpillway Release\n17\nPART II:\nASSUMPTION TESTS\n19\nMethods and Materials\n20\nResults\n21\nImpact/Turbulence Effects on Radio Tag\n21\nBuoyancy Compensation\n21\nPART III:\nSPILL EFFECTIVENESS PROBABILITY MODEL\n24\nSUMMARY 1985-86 TESTS\n28\nCONCLUSIONS AND RECOMMENDATIONS\n32\nACKNOWLEDGMENTS\n33\nLITERATURE CITED\n35\nAPPENDIX A\n37\nAPPENDIX B\n49","INTRODUCTION\nUsing group releases of radio-tagged smolts represents a new and\npotentially powerful research tool that could be effectively applied to\njuvenile salmonid passage problems at dams on the Columbia and Snake Rivers.\nA system of strategically located radio monitors could automatically detect\nand record individually tagged juvenile salmonids as they pass through the\nspillway, powerhouse, bypass system, or tailrace. Estimation of\nspill\neffectiveness, fish guiding efficiency (FGE), collection efficiency (CE),\nspillway survival, powerhouse survival, and bypass survival may be possible\nwithout handling large numbers of unmarked fish. Because nearly all tagged\nfish arriving at the dam can potentially be sampled, the numbers of marked\nfish required for individual experiments could be reduced to a small fraction\nof those required with conventional marking techniques.\nA prototype juvenile radio-tag system was developed and tested by the\nNational Marine Fisheries Service (NMFS) and Bonneville Power Administration\n(BPA) at John Day Dam in 1984 (Giorgi and Stuehrenberg 1984). Results\nindicated that the system could provide acceptable estimates of powerhouse and\nspillway passage.\nResearch at Lower Granite Dam in 1985 (Stuehrenberg et al. 1986)\nindicated that measures of spillway effectiveness were probably attainable,\nbut acceptable measures of FGE and estimates of survival may be difficult to\nachieve.\nResearch in 1986 continued testing of the tag system to further define\nits application and limitations. Field work included 1) releases in the\nforebay and tailrace under a no-spill environment and 2) testing of new\nsystems to improve tag detection. Laboratory tests included 1) the response","2\nof the tag in hostile environmental conditions (spillway passage) and 2) the\neffects of the radio tag on fish buoyancy compensation. This report provides\nresults of the work along with a summarization of the combined 1985-86 field\nand assumption testing.\nPART I: 1986 FIELD TESTS\nThe objective of the 1986 field studies was to continue assessment of the\njuvenile radio-tag system's ability to measure spillway and powerhouse\npassage; FGE; CE; and survival through spillways, bypasses, and turbines.\nTo achieve this objective we 1) released tagged fish in the forebay and\ntailrace and monitored their passage through the dam under a no-spill\nenvironment, 2) tested the effectiveness of underwater antenna systems and a\nrecently designed microprocessor-based tag monitor, and 3) determined whether\ncriteria could be established which would enable us to distinguish tagged live\nfish from tagged dead fish.\nMethods and Materials\nStudy Area\nLower Granite Dam is located at Snake River Kilometer 173. It is the\nfourth dam upstream from the confluence of the Snake and Columbia rivers. The\ndam has six turbines and eight spill gates. The turbines are on the south end\nof the dam, the spillway is on the north end of the powerhouse, and the\nnavigation lock and earthen fill portion of the dam are north of the spillway\n(Fig. 1). Smolts passing through the powerhouse may pass through the turbines\nor the juvenile bypass system. If they enter the bypass system, they can exit\nback through the turbines, fall out of the overflow on the north end of the\nbypass gallery into the spillway tailrace, or travel through the bypass pipe","Ledder\nFish\nCollection\nFacilities\nJuvenile\nSpillway Powerhouse\nFigure 1. -- Radio telemetry monitor locations at and downstream from Lower Granite Dam, 1985-86.\n000000\n*\nFLOW\n* MONITOR LOCATIONS\nLOWER GRANITE LOCK AND DAM\nNavigation\nLock\nEarthen Fill\nLOWER GRANITE\nLOCK AND DAM\n0\nMARINA\nBOYER\nDOWNSTREAM STUDY AREA\nMONITOR LOCATIONS\n0\n0\nU\n0","4\nto the separator and into the collection facility downstream from the dam.\nOther passage routes at the dam are the spillway where smolts may pass under\nthe spill gates (when there is spill), through the navigation lock, or down\nthe fish ladder.\nEquipment\nThe juvenile radio tag was developed by NMFS electronics personnel to\nmonitor movements of individual salmonid smolts. The tags are battery powered\ntransmitters that operate on a carrier frequency of approximately 30 megahertz\n(MHz). The transmitter and batteries are coated with Humiseal 1/ and a mixture\nof paraffin and beeswax to form a flattened cylinder 26x9x6 which weighs\napproximately 2.9 g in air (Fig. 2). A 127-mm flexible whip antenna is\nattached to one end of the tag. Each tag transmits pulses of information on\none of nine frequencies spaced 10 kilohertz (kHz) apart (30.17 to 30.25) . The\npulse rate was set at two per second to provide a minimum tag life of\n4 days. The width characters of each pulse provide individual identification\n(codes) for each tag. Detection range of the tag varied from 12 to\n1,000 meters depending primarily on the depth of the fish and the type of\nantenna used on the monitor. Underwater antennas have the shortest detection\nrange.\nThe juvenile radio-tag system utilizes a series of strategically located\nsignal monitors. Each monitor is made up of a broadband radio receiver, a\n1 Reference to trade names does not imply endorsement by the National Marine\nFisheries Service, NOAA.","5\n(Actual Size)\nGround band and\npower switch\nTransmitter\n(Magnified 1.5 times)\nBattery package\nAntenna\nFigure 2. --NMFS, -- juvenile radio tag used at Lower Granite Dam, 1985-86.","6\npulse decoder, a digital printer, and a cassette tape recorder. Monitors\noperate on 12-volt DC current.\nThe location of the monitors was essentially the same as in 1985\n(Fig. 1). Monitors were arranged so that it was possible to isolate various\npassage locations including the powerhouse, spillway, gatewells, and\nseparator. Additionally, three sets of monitors were located 1.4, 3.2, and\n6.1 km downstream from the dam. These three transect sites were the primary\nrecovery sites for the radio-tagged fish. Two auxilliary sets of monitors\nwere also tested in 1986 -- one near the powerhouse and one near the Central\nFerry Bridge, 22 km downstream from the dam.\nTwo types of antennas were used. Underwater antennas were suspended in\nall gatewells, along the face of the dam in front of the powerhouse, in each\nspill opening, and in the juvenile separator. Three-element beam antennas\nwere used at the downstream transect sites and the powerhouse tailrace. The\npowerhouse and spillway antenna systems were ganged together with line\namplifiers. Each amplifier boosted the signal to a level equal to the signal\nlost in the line between underwater antennas. This effectively produced equal\ntag signals at the monitor for radio tagged smolts at both ends of the\npowerhouse and spillway. All of the monitors were operated with single\nantenna input with the exception of the gatewell monitors.\nEach of the\ninputs (2) for the gatewell monitors was capable of monitoring three gatewells\nand thus gatewell activity was definable to a given turbine unit.\nTest fish were collected from the bypass population at Lower Granite and\nMcNary dams. Fish from McNary Dam were used to augment the limited number of\nlarge chinook salmon available at Lower Granite Dam. Yearling chinook salmon\nsmolts (>150 mm FL) with minimal descaling were separated from the sample and","7\nheld for radio tagging. Fish were collected 1 or 2 d prior to tagging and\nheld in 1.3-m diameter (open system) tanks at Lower Granite Dam. Smolts\ncollected at McNary Dam were transported to Lower Granite Dam and held at\nleast 1 d prior to tagging. Fish were identified as to source at the time of\ntagging (Table 1).\nFish were tagged in accordance with procedures described by Stuehrenberg\net al. (1986). The tagged smolts recovered in the circular tanks for at least\n10 h prior to release. Tags were decoded just prior to release, and fish were\nthen placed on two boats and transported to the release site 5 km upstream\nfrom the dam. Half of the fish were simultaneously released on each side of\nthe river about 100 m from shore. Following the upstream release, separate\ngroups of live and dead fish were released into the tailrace frontroll of the\nturbine boil near the center of the powerhouse. Sample sizes for forebay and\ntailrace releases are detailed in Tables 2.\nIn addition to the forebay and tailrace releases, another release was\nmade in 1986. The additional release utilized a few of the Dworshak Hatchery\nspring chinook salmon which were dedicated to a spill/turbine survival study\nconducted at Lower Granite Dam. The purpose of this trial was to examine the\nfeasibility of utilizing the radio tag in a survival study of this nature. On\n30 March 1986 at 1930 h, two groups of radio-tagged smolts (spillway and\ntailrace) were included with the branded fish released for the survival\nstudy. Thirty-three and twenty-nine fish each were released via a 10.25-cm\ndiameter hose into the spillbay and at the barge loading dock into the\ntailrace, respectively. Tag recoveries were monitored at the downstream\ntransects including the one at Central Ferry.\nIn 1985, approximately 15% of the radio-tagged fish entering the\npowerhouse were not detected at the face of the dam. In 1986, we attempted to","8\nTable 1.\nSource of yearling chinook salmon smolts radio tagged at Lower Granite Dam,\n1986\nFish source\nRelease date\nMcNary\nLower Granite\nTotal tagged\nReleased\n9 April\n61\n50\n111\n104\n18 April\n84\n47\n131\n124\n26 April\n70\n71\n141\n139","9\nheld for radio tagging. Fish were collected 1 or 2 d prior to tagging and\nheld in 1.3-m diameter (open system) tanks at Lower Granite Dam. Smolts\ncollected at McNary Dam were transported to Lower Granite Dam and held at\nleast 1 d prior to tagging. Fish were identified as to source at the time of\ntagging (Table 1).\nFish were tagged in accordance with procedures described by Stuehrenberg\net al. (1986). The tagged smolts recovered in the circular tanks for at least\n10 h prior to release. Tags were decoded just prior to release, and fish were\nthen placed on two boats and transported to the release site 5 km upstream\nfrom the dam. Half of the fish were simultaneously released on each side of\nthe river about 100 m from shore. Following the upstream release, separate\ngroups of live and dead fish were released into the tailrace frontroll of the\nturbine boil near the center of the powerhouse. Sample sizes for forebay and\ntailrace releases are detailed in Table 2.\nIn addition to the forebay and tailrace releases, another release was\nmade in 1986. The additional release utilized a few of the Dworshak Hatchery\nspring chinook salmon which were dedicated to a spill/turbine survival study\nconducted at Lower Granite Dam. The purpose of this trial was to examine the\nfeasibility of utilizing the radio tag in a survival study of this nature. On\n30 March 1986 at 1930 h, two groups of radio-tagged smolts (spillway and\ntailrace) were included with the branded fish released for the survival\nstudy. Thirty-three and twenty-nine fish each were released via a 10.25-cm\ndiameter hose into the spillbay and at the barge loading dock into the\ntailrace, respectively. Tag recoveries were monitored at the downstream\ntransects including the one at Central Ferry.\nIn 1985, approximately 15% of the radio-tagged fish entering the\npowerhouse were not detected at the face of the dam. In 1986, we attempted to","10\nTable 2. Release data for radio tagged yearling chinook salmon smolts,\nLower Granite Dam, 1986.\nRelease\nRelease\nRelease\nRelease\ndate\ntime (h)\nlocation\nnumber\n9 April\n0920\n5 km upstream\n68\n0949\nTailrace\nLive 20\nDead 16\nTotal 104\n18 April\n1318\n5 km upstream\n86\n1350\nTailrace\nLive 23\nDead 15\nTotal 124\n26 April\n1730\n5 km upstream\n99\n1803\nTailrace\nLive 25\nDead 15\nTotal 139","11\nimprove this recovery rate of tags at the dam by employing a new antenna/radio\nreceiver system. On 9, 18, and 26 April 1986, a total of 68, 86, and 99\nradio-tagged yearling chinook salmon, respectively, were released 5 km\nupstream from Lower Granite Dam, and their passage was monitored at the dam in\nan effort to evaluate the antenna/receiver systems.\nPrior to the first forebay release, a system of underwater antennas was\ndeveloped for both the powerhouse and spillway forebay monitors. Using a\njuvenile radio tag suspended on a downrigger at various depths and towed\nacross the upstream face of the dam, we were able to define the detection zone\nat the turbine intake (Fig. 3a).\nTwo monitors were placed on the powerhouse, each covered half of the\npowerhouse. Underwater antennas, 30 m long, were suspended from the deck into\nthe trashrack (three antennas per turbine intake). A monitor was also placed\non the spillway at the start of the first test, but it was moved to the\npowerhouse tailrace when flow projections indicated that water would not be\navailable for spill (Fig. 3a). Before the second release, we changed the\nconfiguration of the underwater antenna system (Fig. 3b) to increase the\ndetection range for the fish that sound near the face of the dam upon entering\nthe turbine intakes (the area of shortest exposure, Fig. 3a). In addition\nto\nthe 30-m long antennas that were left in place, another set of short antennas\nwere suspended down to the top of the trashrack. The resultant detection zone\nis depicted in Figure 3b. Prior to the third release, changes were again made\nto the underwater antenna systems in an effort to increase the tag detection\nzone (Fig. 3c). . The monitor with antennas suspended to the top of the\ntrashracks was not changed, but the antennas for the deep system were moved\nupstream from the metal trashracks. To support the antennas, a rope was","Ne\n3. --Areas Three antenna configurations: (a) to the top of the trashrack, (c) framework. cable\n-- bounded by curved lines indicate coaxial detection antenna zones cable for suspended miniaturized into original radio trashrack,\ntag.\nST M.\n27 Ma\n(b)\n(b) an additional suspended in cable front is of suspended trashrack avoiding contact with metal\n42 M.\n87 M.\n27 M.\n(c)\n42 M.\n28 n.\n22 M.\n(a)\nis\n(a)\nFigure","13\nstretched across the powerhouse roughly 10 m upstream from the face of the\ndam. Inner tubes were tied to the rope, and the underwater antennas were run\nfrom the intake deck through the inner tubes and down to a depth of 24 m.\nPrior to the field studies the monitors and cassette tape recorders were\nchanged from integrated circuitry to microprocessor based circuitry. This\nchange reduced the time required to detect and record (less than 1 second) the\ncoded juvenile radio tags. For tagged smolts passing through the powerhouse,\ntag exposure was about 6 seconds. The operation of the monitors in the field\nwas observed continuously. Prior to the third forebay release, we tuned all\nreceivers to maximize detection sensitivity throughout the total band width of\nour nine channels and incorporated a design change to stabilize individual\nchannel frequency windows for the forebay monitors. The design change was\ndeveloped before the field season, but due to lack of parts could not be\ninstalled in time for the first two releases. The late arriving parts were\ninstalled following the second release.\nResults\n1986 Passage and FGE Evaluation\nMigration routes observed for radio-tagged chinook salmon smolts released\nat Lower Granite Dam in 1986 are summarized in Figure 4. There was no spill,\nso all passage was through the powerhouse. Results indicated that\napproximately 66% (range 62-75%) passed through the turbines and 34% (range\n25-38%) were intercepted by the submersible traveling screen and/or diverted\ninto the collection system. Fish identified as turbine passage were 1) tags\nlast heard in the forebay that were not detected by the gatewell and separator\nmonitors plus 2) those tags not heard in the forebay but detected by the","14\nRELEASE 1 (9-12 APRIL)\nTRANSECTS (25)\n[\nTURBINES (28)\nLOST\n( 3)\nREMAIN* (1)\nREMAIN ( 1)\nI\nHEARD IN FOREBAY (41)\nPASSED (40)\nGATEWELL (10)\nSEPARATOR ( 6)\nTRANSECTS ( 0)\nLOST\n( 2)\nSEPARATOR ( 2)\nTRANSECTS (11)\nREMAIN* ( 1)\nRELEASED UPSTREAM (68) NOT HEARD FOREBAY (13)\nGATEWELL ( 1)\nSEPARATOR ( 0)\nTRANSECTS ( 0)\nTAILRACE ( 1) LOST ( 0)\nSEPARATOR ( 8)\nNOT HEARD POST RELEASE (14)\nTAGGED (104)\nTRANSECTS\n(18)\nLIVE (20)\nTAILRACE MONITOR ( 0)\nNOT HEARD\n( 2)\nRELEASED DOWNSTREAM (36)\nTRANSECTS\n( 4)\nDEAD (16)\nTAILRACE MONITOR ( 1)\nNOT HEARD\n(11)\nTRANSECTS (29)\nRELEASE 2 (18-21 APRIL)\nI\nTURBINES (34)\nLOST\n( 5)\nREMAIN* ( 1)\nREMAIN* ( 5)\nHEARD IN FOREBAY (63)\nPASSED\n(62)\nGATEWELL (27)\nSEPARATOR (21)\nTRANSECTS ( 0)\nLOST\n( 1)\nSEPARATOR ( 1)\nTRANSECTS (11)\nREMAIN* ( 0)\nRELEASED UPSTREAM (86) NOT HEARD FOREBAY (12) GATEWELL ( 0)\nSEPARATOR ( 0)\nTRANSECTS ( 0)\nTAILRACE ( 1)\nLOST\n( B)\nSEPARATOR ( 0)\nNOT HEARD POST RELEASE (11)\nTAGGED (124)\nTRANSECTS\n(23)\nLIVE (23)\nTAILRACE MONITOR ( 0)\nNOT HEARD\n( 8)\nRELEASED DOWNSTREAM (38)\nTRANSECTS\n( 2)\nDEAD (15)\nTAILRACE MONITOR ( 2)\nNOT HEARD\n(11)\nTRANSECTS (25)\nRELEASE 3 (26-29 APRIL)\nI\nTURBINES (81)\nLOST\n( 6)\nREMAIN ( 1)\nREMAIN* ( 1)\nHEARD IN FOREBAY (58) PASSED\n(58)\nGATEWELL (23)\nSEPARATOR (20)\nTRANSECTS ( 0)\nLOST\n( 2)\nSEPARATOR ( 4)\nTRANSECTS (19)\nREMAIN* ( 0)\nRELEASED UPSTREAM (99) NOT HEARD FOREBAY (20)\nGATEWELL ( 1)\nSEPARATOR ( 1)\nTRANSECTS ( 0)\nTAILRACE ( 0)\nLOST\n( 0)\nSEPARATOR ( 0)\nNOT HEARD POST RELEASE (20)\nTAGGED (139)\nTRANSECTS\n(19)\nLIVE (25)\nTAILRACE MONITOR ( 0)\nNOT HEARD\n( 6)\nRELEASED DOWNSTREAM (40)\nTRANSECTS\n(10)\nDEAD (15)\nTAILRACE MONITOR ( 0)\nNOT HEARD\n( 5)\nFigure 4. --Migration routes observed for radio-tagged chinook salmon\nsmolts released at Lower Granite Dam, 1986.\n*\nIndicates tags remaining in the study area at the end of the test period.","15\ntailrace and/or downstream monitors. Measures of absolute FGE could not be\nmade because of tags unheard in the forebay and the tag's impairment of swim\nbladder gas exchange which could affect the fish's vertical distribution in\nthe water column (an assumption based on prior testing).\nForebay Monitor Evaluation\nForebay releases were made on 9, 18, and 26 April 1986 to evaluate the\neffectiveness of the forebay monitor system. From the first upstream release,\n24% (13 of 54) of the detected tagged fish were missed at the face of the dam\n(Fig. 4). Results from the second forebay release indicated that detection\nimproved; only 16% of the total detected population passing the dam were\nundetected at the turbine intake (Fig. 4). . Data from the third release\nindicated that 25% (20 of 79) were not detected before passage (Fig. 4).\nThese rates may be considered minimum figures as fish not detected while they\npassed through the study area are not included in the rates presented.\nThe failure to improve detection from 1985 to 1986 was not attributable\nto the antenna system alone but also to detector sensitivity and to lack of\nstabilization of the individual tag channel frequencies. The failure to\nimprove detection in the third release was caused by an error in tuning the\nchannel frequencies (below the frequency band transmitted by the tags). The\nerror was not discovered until most of the fish from the third release had\npassed the dam.\nReceiver tests conducted at the electronics shop following the field\nseason indicated that the mistuned receivers probably caused the lower\ndetection rate observed in the third and final release. Changes made in\ntuning the channel windows stabilized the receiver within the ranges of\ntemperature and humidity experienced on the Columbia River system.","16\nFurther post field season testing was conducted by the electronics shop\nto establish the proper channel window width. Most of the tags tested\ntransmitted on a frequency 1 kHz above the channel center frequency. Several\nwere 2 kHz above center frequency, one was 4 kHz above, and one was 2 kHz\nbelow. With the 1985-86 receivers, the tag 4 kHz above center frequency would\nnever have been recorded and those 2 kHz from center frequency would only\nrarely be recorded. In situations with long time exposure to the antenna,\nsufficient records would be obtained from tags 2 kHz off frequency to\nsubstantiate a tag's presence, but in short time exposure situations detection\nwould be unlikely. Because of the limited amount of electronic components\nthat can be placed on the juvenile tag substrate, the loading of the radio\nantenna can change the output of the tag. The loading of the antenna is\naffected by the relationship of the antenna to the fish's body and can change\nsomewhat as the fish moves. Based on this information, the monitor channel\nwindows will be set to plus and minus 5 kHz for future research.\nIn summary, for fish arriving at the dam, a detection rate of 85% is\nachievable assuming equipment problems that occurred in 1986 are eliminated.\nDetection rates of 85% were more than adequate to generate the estimates of\npowerhouse and spill passage proportions presented in Stuehrenberg et al.\n1985, where the 95% C.I. around the spill passage estimates for 20 and 40%\nspill were 28.7 to 49.0% and 50.5 to 71.1%, respectively.\nTailrace Release\nRecoveries of tagged fish at the downstream transects indicated that live\nfish could not always be discriminated from dead fish. Generally, we expected\nthat live fish would always move downstream at a faster rate than dead fish.","17\nThis was not the case. Inspection of Figure 5 shows there is some overlap in\nthe travel times of live and dead fish at every transect. Furthermore, nine\nlive fish (13% of all released) were never detected anywhere following\nrelease. We assume these fish failed to migrate through the detection zone\nduring the battery life of the tag, or that the tags or the fish died (both\nlow probability based on laboratory tests). It is unlikely that any tagged\nfish could traverse all three transects without being detected.\nThese data would indicate that dead radio-tagged fish cannot be\nconsistently differentiated from live ones in the tailrace. Therefore, it\nwould appear that accurate measures of passage survival of chinook salmon will\nnot be possible with the juvenile radio-tag system on the main-stem Snake and\nColumbia River dams.\nSpillway Release\nThe fish released into the spillway on 30 March 1986 during a special\ntest spill condition (Park 1987) were recovered at a higher rate than those\nreleased in the tailrace. Of all spillway fish, 82% (n=27) were detected on\nat least one transect station following release, compared with 59% (n=17) for\nthe tailrace release. The net result is survival rate of 139% clearly an\nunreasonable estimate.\nTwo factors are believed to have greatly affected this test. First, the\ntest was run before the normal spring chinook salmon outmigration. With fish\nnot willing to move in the river, differences between the flows that the\nradio-tagged smolts were released into could significantly affect movement to\nthe transects closer to the dam and vulnerability to predation. None of the\nsmolts reached the monitor at Central Ferry. Secondly, smaller fish were used","18\n50\nD\nTransect #1\n#\n40\ne\nt\n30\nLive\nT e\n20\na C\n10\nt\ng\n0\nS\ne\nDead\nd\n10\n3\n5\n6\n7\n8\n9\n10\n>\n10\n1\n2\n4\nHours Post Release\n50\nD\nTransect #2\n#\n40\ne\nt\n30\nLive\nT e\n20\na C\n10\nt\ng\n0\nS e\nDead\nd\n10\n5\n6\n7\n8\n9\n10\n10\n1\n2\n3\n4\n>\nHours Post Release\n50\nD\nTransect #3\n# e\n40\nt\n30\nLive\nT e\n20\na C\n10\nt\ng\n0\nS e\nDead\nd\n10\n6\n7\n8\n9\n10\n1\n2\n3\n4\n5\n>\n10\nHours Post Release\nFigure 5. - Number of tagged fish which were detected following their\nrelease in the tailrace. The time scale indicates elapsed\ntime (h) from release to passage through each transect's\ndetection zone. Data for both live and dead fish bearing\nactive radio tags appear in the upper and lower portion of\neach histogram, respectively.","19\nthis test than in later releases. Small fish size increases tagging\nin\nmortality, decreases buoyancy equilibration rates, and decreases tag\nreliability.\nPART II: ASSUMPTION TESTS\nIn 1985, we conducted a variety of tests to address tag regurgitation,\ndelayed mortality, tag effects on buoyancy and swimming performance, duration\nof tag life, and response of the tag to hostile environmental conditions. Of\nthose items we examined in 1985, two required further scrutiny in 1986.\nIn 1985, we assessed the effects turbulence/impact on tag operation by\ndischarging 51 subyearling chinook salmon through a water cannon at our field\nstation at Pasco, Washington. The cannon nozzle is directed toward the pond\nsurface at a 45 degree angle with the tip approximately 1.5 m above the\nsurface. Fish exit the nozzle at approximately 17 ft/s. These conditions\nwere intended to approximate the conditions a tag-bearing fish encounters when\npassing through the spillway. In our test, 16% of the tags failed. However,\nthe fish were quite small (<140 mm fork length), and considerable effort was\nrequired to push the tag into the esophagus. We suspect that this difficulty\nmay have caused tag failure by cracking the water-tight wax seal during\ninsertion. Consequently, we repeated this test in 1986 employing yearling\nchinook salmon of the larger size used in field studies.\nAlso in 1985, we observed that the radio tag impaired a fish's ability to\nregulate its buoyancy. Yearling chinook salmon displayed responses that\nindicated that the tag was interfering with swim bladder inflation by either\noccluding the duct leading from the esophagus to the bladder or occupying so","20\nmuch space that the bladder could not expand sufficiently. There was enough\nconcern regarding this effect that we felt it necessary to continue this line\nof investigation in 1986.\nMethods and Materials\nOn 5 May 1986, yearling chinook salmon were acquired from the collection\nfacility at McNary Dam and transported to NMFS' Pasco Field Station. Fifty- -\nfour fish (>155 mm) were anesthetized and tagged according to the procedures\ndetailed in Stuehrenberg et al. (1986). Tag function and fish condition were\nchecked at 12 h post-tagging and just prior to testing at 24 h. Radio-tagged\nfish were then discharged through the water cannon, recaptured in accordance\nwith the procedures in Stuehrenberg et al. (1986), and tag operation assessed.\nBuoyancy compensation tests were carried out on 6 and 7 May with 67\nyearling chinook salmon collected at McNary Dam and transported to the Pasco\nfacility. Fish were anesthetized and individually placed in the chamber\ndescribed by Stuehrenberg et al. (1986). A partial vacuum was applied, and\nthe pressure was reduced until the fish just rose off the bottom. The\npressure of neutral buoyancy (Pnb) was determined by subtracting the reduction\nin pressure necessary to float the fish (Pr) from the atmospheric pressure\n(Pa). The P nb approaches atmospheric pressure as buoyancy nears neutrality\nand is thus an indirect measure of bladder volume (Saunders 1965). After\ninitial measurements of Pnb were made, the control fish were returned to\nholding tanks for 24 h to recover. Test fish were similarly anesthetized and\ndecompressed, but were tagged prior to being returned to their holding area.\nA second buoyancy measurment was made 24 h later on all control and test\nfish. Post-treatment Pnb values were expressed as a percent of pre-treatment\nvalues as follows:","21\nPercent recovery of = P initial final )\n(100)\nX\n(Fried et al. 1976). . Percent recovery values for controls should fluctuate\naround 100%. Tagged fish should approach 100% as the bladder is inflated as\ncompensation for the weight of the tag and initial buoyancy is regained.\nResults\nImpact/Turbulence Effects on Radio Tag\nThe impact tests indicated that such conditions can cause tags to\nmalfunction but at a very low rate. Of the 54 fish initially tagged, 46 were\nactually tested and evaluated in a 24-h post test observation period. The\nremainder either died during the holding period, were consumed by predators,\nsuffered tag failure immediately following insertion, or were entrained in the\ncannon. Only 1 of the 46 (2.2%) test fish exhibited tag failure\nwater\nfollowing the test. The failure resulted from a broken switch mechanism which\nwe attributed to the impact the fish experienced.\nBuoyancy Compensation\nThe P nb values could not be measured for 22 of the 67 yearling chinook\nsalmon tested (33 control and 34 tagged). During decompression, 13 fish (11\ncontrols and 2 tagged) never rose off the bottom of the test chamber but\nemitted gas through their mouth. The remaining nine fish (all tagged) floated\nat the surface at ambient pressure (Tables 3 and 4). . Thus 26% of all tagged\nfish (9 of 34) exhibited a response never observed for any control fish. This\nindicates that the tag does affect buoyancy.\nThere is further evidence that the radio tag affects buoyancy.\nThe\npercent recovery to initial nb was measured for 45 fish (23 controls and 22\ntagged fish) which did not exhibit gas emission or floating. The mean percent","22\nTable 3.\nBuoyancy compensation data for radio-tagged yearling chinook\nsalmon (N=33), 1986.\nLength\nWeight\n% recovery of\n(mm)\n(g)\ninitial P nb\nComments\n173\n51.7\n187\n164\n41.7\n121\n202\n83.7\n129\n190\n67.7\n138\n198\n78.7\n117\n183\n57.7\n144\nFloating fish\n180\n51.7\n139\n170\n48.7\n<74\nGas emitted\n171\n54.7\n36\n177\n49.7\n129\n178\n53.7\n152\n194\n69.7\n108\n180\n59.7\n198\n175\n49.7\n>114\nFloating fish\n192\n66.7\n>116\nFloating fish\n192\n71.7\n126\n170\n50.7\n>120\nFloating fish\n175\n50.7\n>131\nFloating fish\n182\n56.7\n85\nGas emitted\n160\n39.7\n156\n185\n61.7\n118\n187\n65.7\n95\n185\n63.7\n>112\nFloating fish\n194\n74.7\n118\n191\n68.7\n86\n185\n59.7\n>122\nFloating fish\n195\n77.7\n>144\nFloating fish\n182\n55.7\n>159\nFloating fish\n179\n53.7\n130\n182\n60.7\n106\n188\n59.7\n281\n170\n49.7\n106\n170\n45.7\n65","23\nTable 4. Buoyancy compensation data for control yearling chinook\nsalmon (N=34), 1986.\nLength\nWeight\n% recovery of\n(mm)\n(g)\ninitial Pnb\nComments (Pr = in Hg)\n225\n108.0\n101\nGas emitted Pr>15.0\n193\n77.5\n118\n177\n53.5\n84\n187\n66.5\n82\n192\n67.5\n<96\nGas emitted Pr>7.0\n195\n72.5\n106\nGas emitted Pr>3.0\n177\n52.5\n<92\nGas emitted Pr>4.0\n174\n52.5\n75\n170\n45.5\n100\n172\n49.5\n92\n170\n46.5\n90\n178\n55.5\n84\n190\n67.5\n100\n216\n97.5\n109\n188\n65.5\n<97\nGas emitted Pr>4.0\n196\n77.5\n122\nGas emitted Pr>10.0\n177\n51.5\n61\n189\n61.5\n<100\nGas emitted Pr>8.0\n174\n48.5\n<103\nGas emitted Pr>8.5\n196\n69.5\n96\n186\n63.5\n131\nGas emitted Pr>7.0\n185\n55.5\n112\n193\n62.5\n100\n213\n92.5\n154\nGas emitted Pr>14.0\n170\n43.5\n118\n198\n74.5\n100\n163\n42.5\n108\n203\n87.5\n113\n177\n48.5\n115\n180\n63.5\n93\n196\n75.5\n90\nGas emitted Pr 0.0\n181\n57.5\n118\n156\n37.5\n130\n173\n48.5\n110","24\nrecovery values were 100.0% for controls and 123.5% for tagged fish. Data are\ndetailed in Tables 3 and 4. Using a Mann-Whitney U test, we found the percent\nrecovery of the two groups significantly different (U statistic = 121,\nP = 0.003).\nThese results are considerably different from those observed in 1985\n(Table 5). In 1985, 35% of the tagged fish exhibited either gas emission or\nflotation at ambient pressure, but only 2% of the controls exhibited such\nresponses. Furthermore, in 1985, tagged fish had difficulty entraining a\nsufficient volume of air to regain their pretagging Pnb values, and the mean\npercent recovery was 85.4% (Stuehrenberg et al. 1986). In contrast, tagged\nfish in 1986 entrained excess air in their gas bladders and apparently had\ndifficulty discharging it; the mean percent recovery for tagged fish was\n123.5% (Table 5). The reason for these interannual differences in percent\nrecovery is not certain. Even so, in evaluating both years of data, it\nappears that the tag impairs swim bladder gas exchange which could affect the\nvertical distribution of tagged fish in the water column.\nPART III: SPILL EFFECTIVENESS PROBABILITY MODEL\nSpill effectiveness estimates were calculated for data collected in 1985\nat two spill levels, 20 and 40%. For details of the estimation procedure, see\nAppendix A. The levels of discharge were maintained for a 48-h period, during\nwhich the radio-tagged fish were passing the dam. For both spill conditions,\nyearling chinook salmon passed over the spillway at a rate in excess of the\nproportion of the total flow discharged through the spillway. During the time\n20% of the river flow was discharged through the spillway, an estimated 40.5%\n+11.8 (95% C.I. = 28.7 to 52.3%) of the tagged chinook salmon passed the\nspillway. At 40% spill, spillway passage was 60.6% +13.8 (95% C. I. = 46.8 to","25\nTable 5.\nComparison of buoyancy data from 1985 and 1986 tests of\nradio tag effects on yearling chinook salmon.\n1985\n1986\nAverage length (mm)\n176.0\n182.0\nTagged\n37\n33\nGas emitted (n)\n11\n2\nFloating (n)\n1\n9\n% recovery (x)\n85.4\n123.5\nControl\n39\n34\nGas emitted (n)\n1\n11\nFloating (n)\n0\n0\n% recovery (x)\n107.3\n100.0","26\n74.4%). We then tested the null hypothesis that the observed spill\neffectiveness was equal to the prevailing spill level, using standard normal\ndeviates (Sokal and Rohlf 1987, p. 105)\nThe test statistics were calculated at 3.41 and 2.80 for the 20 and 40%\nconditions, respectively. For both cases, we rejected the null\nspill\nhypothesis (P<0.01).\nSpill effectiveness estimates are plotted in Figure 6, and a straight\nline is extrapolated through the origin. These data suggest that for yearling\nchinook salmon at Lower Granite Dam, the relationship between spill passage\nand the percentage of water spilled may be a curvilinear function rather than\na straight line relationship.\nBased on the relationship between migration routes of radio-tagged smolts\nand purse seine catches in the John Day forebay (Giorgi 1984), migration\nroutes of large radio-tagged chinook salmon smolts accurately reflect those of\nthe untagged population. The direct effect of fish buoyancy on the spill\neffectiveness estimates is reduced by the fact that spill water is taken from\nthe same depth as the entrance of the turbine intakes.\nThe previously mentioned effect of tagging on fish buoyancy leads to a\nquestion in using our model to estimate spill effectiveness. If tagged and\nuntagged fish differed in buoyancy and vertical distribution during the 1985\nfield experiments, they would have been guided into the bypass system in\ndifferent proportions. As a result, spill effectiveness estimates made using\ntagged fish might not apply to all migrating fish. The simulation exercise in\nAppendix A shows that under a wide range of vertical distribution bias\nconditions, our spill effectiveness estimates apply to untagged as well as\ntagged fish. We therefore believe that our estimates accurately represent the\nchinook salmon smolts migrating during the time period of our experiments.","27\nSPILL EFFECTIVENESS\nWITH 95% C.I.\n80\n70\n60\nESTIMATED 50\n0-20 Extrapolated\nSPILLWAY 40\nPASSAGE (%)\n30\n20\n10\n0\n0\n20\n40\n%TOTAL FLOW DISCHARGED\nOVER THE SPILLWAY\nFigure 6. -- Estimated spill effectiveness at Lower Granite Dam based on the\npassage of radio-tagged chinook salmon smolts through the spillway\nat spill levels of 20 and 40% of the total river flow (1985).","28\nSUMMARY 1985-86 TESTS\nMiniaturized radio tags which are inserted into the stomachs of yearling\nchinook salmon may cause unacceptable rates of mortality in host fish or may\nimpair their swimming performance. Effective tag loss can result from\nregurgitation of the tag or operational failure of the device. Furthermore,\nthe requirement of tagging smolts large enough to accommodate the tag may\nprovide data unrepresentative of the general population. All of these factors\nare important considerations when evaluating the feasibility of using the\nradio tag to estimate FGE, survival, or spill efficiency. In\n1985\n(Stuehrenberg et al. 1986) and 1986, we conducted investigations to address\nthese concerns (Table 6)\nThese tests indicated that the effects of radio tags on yearling chinook\nsalmon were minimal and acceptable. Tagged fish did not incur higher\nmortality than untagged individuals. Whether tagged or not, fish exposed to\npressure changes simulating those experienced during turbine passage died at\nthe same rate (0.7 to 1.6% mortality) (Stuehrenberg et al. 1986). Tagged fish\nappear to be representative of the general population with respect to\nsurvival.\nTag regurgitation was minimal, ranging from 0 to 2.7% Regardless of the\ntreatment (simulated turbine passage, simulated spill passage, or ambient\nconditions), regurgitation rates were about the same (Stuehrenberg et al.\n1986) (Table 6). e Thus we would expect no differential tag loss due to\nregurgitation resulting from passage through a particular conduit (e.g.,\nspillway or powerhouse).\nIn our field studies, we selected the largest fish available since they\ncould better accommodate the tag. There was some concern that these fish were","29\nTable 6: Summarization of tests to evaluate the various effects of the radio tags on\nyearling chinook salmon and the effects of passage conditions on the radio\ntag. Tests were conducted over 2 years, 1985 (Stuehrenberg et al. 1986) and\n1986.\nTest/objective\nResults\nConclusions\n1) Compare survival of\nH accepted\nWhen exposed to\ntagged VS. control\nconditions, tagged fish\nfish exposed to\nexhibit the same\npressure changes\nsurvival as untagged\nsimulating turbine\nfish.\npassage.\n2) Determine tag\n1) Under ambient holding\nTag regurgitation\nregurgitation rate\nconditions, all volitional\nassociated with either\nunder three conditions:\nregurgitation occurs within\nturbine or spill passage\nambient and simulated\n4 h post tagging.\nis negligible.\nspill and turbine\n2) Turbine condition =\na\n0.8 - 1.4% tag\npassage\nregurgitation.\n3) Spill condition = 0%\ntag regurgitation.\n3) Determine if large\nAccept H :\nLarge (taggable smolts\n(taggable) smolts\nguided = u unguided\nare representative of the\nexhibit passage\ngeneral population with\nbehavior different\nrespect to guidance\nfrom the general\nbehavior.\npopulation, using\nfish guidance as\nthe response. a\n4) Compare tag failure\nAccept H0: u turbine =\nPressure changes associated\nrate under three\nu ambient\nwith turbine passage and\nconditions: ambient\nspill-like impact does not\nand simulated spill\naffect tag performance.\na/b/\nand turbine passage\n5) Determine if the tag\nThe tag impaired the\nImpaired gas exchange\ninterferes with the\nhosts ability to entrain\nmay affect vertical\nregulation of air\nand discharge air from\ndistribution. Therefore,\nbladder volume a/b/\nthe gas bladder.\nrecommend against using\nradio tag for FGE work.\n6) Determine if the tag\nAccept HO:\nRadio tags do not\nimpairs swimming\nu tagged = u controls\ndecrease swimming\nperformance, using\nperformance.\nswimming stamina as\na\nthe response.\na / Tests were conducted in 1985. Details regarding tests can be found in Stuehrenberg\net al. (1986).\nb Tests were conducted in 1986. Details regarding tests can be found in this document.","30\nnot representative of the general population, especially with respect to their\nguidability by submersible traveling screens (STS). However, when examined,\nthe size composition of guided and unguided fish were the same, indicating\nthat the screens were not size selective (Stuehrenberg et al. 1986).\nOverall, radio-tag performance was acceptable. Most failures observed\nwithin the 72-h test period for field studies occurred within 10 h following\nactivation and insertion, and we recommend this as a minimum holding time\nprior to release. During the potential detection, or tag recovery, period (10\nto 72 h) for field studies, the tag decay or failure rate was only 4.3%. When\nactive tags were subjected to simulated turbine pressure conditions\n(Stuehrenberg et al. 1986) and spill-like impact, they exhibited the same\nfailure rate as control tags held under ambient conditions. Thus, passage\nroute should not affect the rate of tag failure.\nRadio tags apparently interfered with some fish's ability to adjust swim\nbladder volume. Impaired fish were unable either to entrain or discharge the\namount of air necessary to attain pretagging buoyancy levels. It is possible\nthat this condition may to some extent perturb their normal vertical\ndistribution in the water column which in turn may affect FGE.\nRadio tags did not reduce swimming capability of yearling chinook salmon\nin tests conducted in 1985 (Stuehrenberg et al. 1986). Fish fitted with radio\ntags exhibited levels of swimming stamina which were slightly lower than those\nobserved for control fish, with mean Ucrit values of 4.04 and 4.43 BL/S,\nrespectively. However, the means were not statistically different. On this\nbasis, we conclude that the radio tag does not significantly impact the\nswimming performance of yearling chinook salmon and that tagged fishes\nmigrational behavior is representative of the general population in that\nrespect.","31\nBased on results from this 2-year study, we do not recommend that the\nminiaturized radio tag be employed in estimating absolute FGE. Since the tag\ndoes impair swim bladder gas exchange, this could affect the vertical\ndistribution of tagged fish in the water column and potentially, the fish's\nsusceptibility to guidance by STS. However, the tag system could provide\nrelative week to week or year to year differences in FGE. This could be\nuseful to verify the net and hydroacoustic data at dams such as Lower Granite\nDam where there is considerable variability in FGE.\nIn this program we also evaluated the feasibility of using the radio tag\nin survival studies. We found that it was not possible to definitively\ndiscriminate between live and dead fish bearing active tags. Some dead fish\nwere observed to drift to the downstream monitor transects at the same rate as\nlive fish. In a river situation where high velocities prevail, it is unlikely\nthat an absolute criteria for identifying live fish can be developed.\nConsequently, we recommend against using the current radio tag for survival\nstudies in river situations of this nature. However, in smaller tributaries,\nthese criteria may not be so hard to define. Stier and Kynard (1986)\nsuccessfully employed a miniaturized radio tag to estimate survival of\nAltantic salmon, Salmo salar, smolts passing through a turbine at Holyoke Dam\non the Connecticut River. In that study, investigators were able to readily\ndistinguish dead from live fish based on rate of downstream movement.\nConsidering their success in a relatively small river system, we could expect\nthat the NMFS radio tag may be successfully employed in survival studies at\nsmaller rivers within the Columbia-Snake River Basin.\nThe most promising use for the radio tag in passage research in the\nColumbia and Snake rivers is for estimating the proportion of the yearling\nchinook salmon population which passes a dam via either the spillway or","32\npowerhouse. Research conducted at John Day Dam demonstrated that radio-tagged\nfish approaching the dam exhibited the same migration patterns as the general\npopulation (Giorgi et al. 1985). In that study, radio-tagged yearling chinook\nsalmon were tracked through the same areas in the forebay where purse seine\nsampling indicated fish were concentrated. Also, the diel passage patterns\nwitnessed for radio-tagged fish were consistent with observations made for the\ngeneral population (Giorgi et al. 1985). .\nIn 1985, Kuehl (1986) also estimated spill effectiveness at Lower Granite\nDam using hydroacoustic techniques. She found that 11, 19, and 35% of the\nfish population passed over the spillway when 4, 20, and 40% of the river flow\nwas discharged through the spill way, respectively. These estimates are\nconsiderably different from our measures of 41% at 20% spill and 61% passage\nat 40% spill. There may be several reasons for this, Kuehl's (1986) estimates\nare not species specific whereas ours pertain only to yearling chinook\nsalmon. Also Kuehl generated her estimates at different times. In one case,\nthe estimate was based on only 4 h of sampling. We suggest that in the\nfuture, hydroacoustic and radio tag studies be complementary and that\nindependent estimates be generated simultaneously, on the same population, and\nsame flow conditions. Such an approach would permit us to evaluate\nunder\nthe\nthe merits and deficiencies of both techniques in an efficient manner.\nCONCLUSIONS AND RECOMMENDATIONS\n1. The miniaturized radio tag system is an effective tool for estimating\nthe proportions of yearling chinook salmon populations passing a dam via\neither the spillway or powerhouse and for estimating spill effectiveness\n(proportions passing over the spill at varying levels of spill).","33\n2. With respect to migration routes and passage location, there is no\nevidence to indicate that radio-tagged smolts exhibit passage behavior\ndifferent from untagged fish.\n3. We recommend that concurrent radio-tag and hydroacoustic spill\neffectiveness studies be conducted. This direct comparison would permit\nthe merits and deficiencies of both techniques to be efficiently\nevaluated.\n4. We do not recommend that the radio tag be used to estimate mortality\nassociated with dam passage in large, swift rivers such as the Snake or\nColumbia. However, based on radio-tag survival studies conducted in a\nsmaller river (Stier and Kynard 1986), its use in tributaries within the\nColumbia Basin warrants investigation.\n5. We do not recommend that the juvenile radio-tag system be employed to\nestimate absolute FGE for chinook salmon. Host fish exhibited difficulty\nadjusting swim bladder volume which could potentially perturb their normal\nvertical distribution and guidance.\nACKNOWLEDGMENTS\nSupport for this research came from the region's electrical ratepayers\nthrough the Bonneville Power Administration.\nWe thank the National Marine Fisheries Service personnel who provided\nsupport and assistance crucial to the execution of this project, especially","34\nDave Brastow, Cheryl Buck, and Jay Wilson for programming, data analysis, and\nstatistical analysis. Chuck Barlett, John Govig, and Mark Kaminski provided\nthe design, construction, and maintenance of all electronic equipment employed\nduring this study.\nWe gratefully acknowledge the assistance of the U.S. Army Corps of\nEngineers and Fish Passage Center personnel for obtaining the flow conditions\nnecessary for this study and the project personnel at Lower Granite Dam for\ntheir assistance during a very busy period at the dam.","35\nLITERATURE CITED\nBrownie, C., D. R. Anderson, K. P. Burnham, and D. S. Robson.\n1985. Statistical inference from band recovery data - A hand book. U.S.\nFish and Wildlife Service, Resource Publication 156, 305 p.\nFried, S. M., J. D. McCleave, and K. A. Stred.\n1976. Buoyancy compensation by Atlantic salmon (Salmo salar) smolts\ntagged internally with dummy telemetry transmitters. J. Fish. Res.\nBoard Can., , 33: 1377-1380.\nGiorgi, A. E. , L. C. Stuehrenberg.\n1984. Smolt Passage Behavior and flow-net Relationships in the forebay\nof John Day Dam. Coastal Zone and Estuarine Studies, Nat. Marine Fish.\nServ., , Seattle, WA.\nGiorgi, A. E., L. C. Stuehrenberg, D. R. Miller, and C. W. Sims.\n1985. Smolt passage behavior and flow-net relationship in the forebay of\nJohn Day Dam. Coastal Zone and Estaurine Studies Division, Nat. Marine\nFish. Serv., , Seattle, WA.\nKuehl, S.\n1986. Hydroacoustic Evaluation of Fish Collection Efficiency at Lower\nGranite Dam in Spring 1985. Biosonics, Inc., Seattle, WA.","36\nPark, D. L. and S. Achord.\n1987. Evaluation of Juvenile Salmonid Passage through the Bypass System,\nTurbine, and Spillway at Lower Granite Dam - 1986. Coastal Zone &\nEstuarine Studies, Nat. Marine Fisheries, Seattle, WA.\nSaunders, R. L.\n1965. Adjustment of buoyancy in young Atlantic salmon nad brook trout by\nchanges in swim bladder volume. J. Fish. Res. Board Can., 22: 335-352.\nSokal, R. R., , and F. J. Rohlf.\n1981. Biometry. W. H. Freeman and Company. San Francisco. p. 105.\nStier, D. J. and B. Kynard.\n1986. Use of Radio Telemetry to Determine the Mortality of Atlantic\nSalmon Smolts Passed through a 17-MW Kaplan Turbine at a Low-Head\nHydroelectric Dam. Trans. Am. Fish. Soc., 115: 771-775.\nStuehrenberg, L. C., A. E. Giorgi, C. W. Sims, J. Ramonda-Powell, and\nJ. Wilson.\n1986. Juvenile Radio-Tag Study: Lower Granite Dam. Coastal Zone and\nEstuarine Studies Division, Nat. Marine Fish. Serv., Seattle, WA.\nWilson, J. W.\n1987. A method of estimating spill effectiveness and fish guidance\nefficiency for fish passage through hydroelectric dams in the Columbia\nRiver Basin. Master of Science Thesis, University of Washington,\nSeattle, 44 p.","37\nAPPENDIX A\nMODELING AND ESTIMATION","38\nUsing data recovered from the radio receiver monitors, each released fish\nwas assigned to one of eight categories referred to below as detection fates:\n1) Detected passing via spillway and again downstream.\n2) Detected passing via spillway but not downstream.\n3) Detected passing via turbine and again downstream.\n4) Detected passing via turbine but not downstream.\n5) Detected passing both into powerhouse and in bypass system.\n6) Detected only in bypass system.\n7) Detected downstream but not at the dam.\n8) Not detected after release\nEach fish released during the experiment underwent exactly one of these\ndetection fates. We assumed that the probability of experiencing a particular\nfate was the same for each fish released and that each fish's fate was\nindependent of all others.\nIf N1, N8 are the numbers of fish observed in each category and\n\"1, TT the probabilities of the fates, then the N are\nmultinomially distributed with\n2,***,\"8\"\nP(N1, N\nNR\n\"1\"\nwhere is the number of fish\nreleased.","39\nThe probabilities 11 were reexpressed in terms of the following\nparameters:\nPd = probability that a fish migrated to the dam with a functional\ntag.\nPs = probability that a fish reaching the dam passed via the\nspillway.\nPg = probability that a fish entering the powerhouse was guided into\nthe bypass system.\nPfs = probability that a fish passing via the spillway was detected by\nspillway intake monitors.\nPft = probability that a fish passing via the powerhouse was detected\nby powerhouse intake monitors.\nP1s = probability that a fish was lost to downstream detection after\npassing via the spillway.\nPit = probability that a fish was lost to downstream detection after\npassing via the turbines.\nFor the purpose of estimation, spill effectiveness was considered equivalent\nto P S and FGE equivalent to Pg\nAn example illustrates the process of reexpression. If a fish underwent\nthe first detection fate, it reached the dam with a functional tag, passed\nthrough the spillway, was detected by the spillway intake monitors, and\nreached the downstream monitors with a functional tag. If each of the events\nin this series was independent of the others, then the probability of\nundergoing the first fate was the product of the probabilities of these\nevents:\n-","40\nThe remaining 11 were reexpressed in a similar manner. Appendix Figures A1\nto A4 present schematically the series of events corresponding to each of the\ndetection fates and may be used in verifying the reexpressions of \"2-\"\"8\"\nThe reexpressed \" i are as follows:\n-\n= -\nPS) (1 - Pg) (1- Pft) (1 P1t)]\n-\n(1 - P S ) 1 Pg) (1 Pft)\nThe maximum likelihood estimator (MLE) for P = (Pd, PS, .... P1t)' was\nobtained using the invariance property of maximum likelihood estimation (Mood\net al. 1974). The MLEs for the parameters are:\nfs15536\ni","41\nFish Enter Bypass System\nFish Enter Turbines\nFish Enter Spillway\nrelease.\nAppendix Pd' Figure Pt' A1. and -- Ps Possible are as dam passage routes of radio tagged fish following\n1 - Ps - Pt\nPt\nPs\nFish Reach Dam\ndefined in the text.\nPd\nFish Released","Downstream Detection\nDownstream Detection\nDownstream Detection\nDownstream Detection\nAppendix Figure A2. Possible detection fates of fish entering the spillway. Pft and P1s\nNot Lost to\nNot Lost to\nLost to\nLost to\n1- P1s\n1 - P1s\n1s\nP1s\nNot Detected in Forebay\nDetected in Forebay\nare as defined in the text.\n1 - Pfs\nPfs\nFish Enter Spillway","Downstream Detection\nDownstream Detection\nDownstream Detection\nDownstream Detection\nAppendix Figure A3.--Possible detection fates of fish entering the turbines. Pft and P1t\nNot Lost to\nNot Lost to\nLost to\nLost to\n1- - P1t\n1 - P1t\nP1t\nP1t\nNot Detected in Forebay\nDetected in Forebay\nare defined in the text.\n1 - Pft\nPft\nFish Enter Turbines","Appendix Figure A4. --Possible detection fates of fish entering the bypass system. Pft is as\nNot Detected in Forebay\nDetected in Forebay\n1 - Pft\nPft\nFish Enter Bypass System\ndefined in the text.","45\nwhere C = (N1 + N2) (N1 N5 + N5 N7 - N3 N6) and\nD = N1 (N3 + N4 + N5) (N5 N6).\nSampling variances were estimated numerically for each parameter using\nthe delta method (Brownie et al. 1985, p. 214). For further details on MLE\nderivation and sampling variance estimation, see Wilson (1987).\nThe afore mentioned effect of tagging on fish buoyancy leads to a\nquestion in using the modeling and parameter estimation process proposed\nherein. If tagged and untagged fish differed in buoyancy and vertical\ndistribution during the 1985 field experiements, they were guided into the\nbypass system in different proportions. As a result, those parameter\nestimates depending on N3N7 (the observed quantities directly affected by\nvertical distribution and fish guidance) were biased relative to untagged\nfish. Using the Monte Carlo experiment outlined below, the following question\nwas addressed: Does FGE bias alter parameter estimates sufficiently to be of\npractical importance in making management decisions?\nExperimental releases of radio-tagged fish were simulated under various\ncombinations of spill effectiveness and FGE. Each release was considered a\nsample from a multinomial population with parameters NR = 100, \"1\"\nThe 1 are functions of P. For simulation purposes, was calculated\nspecifying the following values for the P (.)","46\n1) PS was 0.4 or 0.6, similar to the estimates obtained in the field\nexperiments.\n2) P was 0.25, 0.50, or 0.75, representing low, medium, and high FGE\nlevels.\n3) Pd' Pfs' Pft' P1s' and Pit were assigned the values estimated during\nthe field experiments. For example, Pd was 0.788 when Ps was 0.4 and\n0.754 when PS was 0.5 (Appendix Table A1).\nThe simulation of these populations was conducted under the six\ncombinations of spill effectiveness and FGE presented in Appendix Table A2.\nFor each combination, 1,000 releases were simulated, assigning 100 \"released\"\nfish to the N using pseudorandom number generation based on the P obtained\nas above. For each release, P was estimated, P , the mean P estimate over\n1,000 releases was then calculated. For each parameter, I calculated P(.) -\nP(.), the deviation of the mean estimate from the value used in the\nsimulation. The larger the magnitude of the deviation, the greater the effect\nFGE had on the parameter estimates of spill effectiveness and the other\nparameters.\nThe deviations of the mean estimates from the simulation values are\npresented in Appendix Table Al. The mean parameter estimates showed\nnegligible deviations from the true value when FGE ranged from 0.25 to 0.75.\nThe deviation of mean spill effectivenes (PS) estimates from the parameter\nvalue was less than or equal to 0.01 for all simulation conditions. We\ntherefore believe that it is unlikely that FGE bias would seriously affect\nspill effectiveness estimates in radio-tagging experiments. Pd' Pft' and P1s'\nPfs' and Pft showed no discernible bias over a range of FGE values.","47\nAppendix Table Al: Results of 1985 field experiment.\na)\nNumbers of fish released and observed for Detection Fates 1-7. -\nN\nFate a/\nSpill\nlevel\nReleased\n1\n2\n3\n4\n5\n6\n7\n(%)\n20\n101\n21\n5\n18\n8\n19\n1\n6\n40\n100\n31\n7\n12\n7\n7\n1\n8\nb) Maximum likelihood estimates and standard deviations (SD) of model\nparameters.\n20% spill\n40% spill\nParameter\nMLE\nSD\nMLE\nSD\nPc\n0.405\n0.0601\n0.606\n0.0702\nS\nP\n0.422\n0.0736\n0.269\n0.0870\ng\nPd\n0.788\n0.0399\n0.754\n0.0417\nPt\n0.344\n0.0580\n0.288\n0.0673\nP fs\n0.806\n0.0887\n0.831\n0.0800\nPft\n0.950\n0.0487\n0.875\n0.1169\nP\n0.192\n0.0773\n1s\n0.134\n0.0629\nP\n0.308\n1t\n0.0905\n0.368\n0.1107\na / Fates are defined on the first page of Appendix A.","48\nAppendix Table A2: Sensitivity of various probability estimates (P S , Pd, Pft'\nP1s' Pfs' Po, and P1t) at specified values of FGE and\nspill effectiveness. The values are the mean MLE - true\nparameter value, for the specified parameter estimate and\nFGE level.\nActual spill effectiveness\nParamenter estimate\nActual FGE\n0.4\n0.6\n0.25\n0.001\n0.002\nPd\n0.50\n-0.001\n0.000\n0.75\n0.003\n0.000\n0.001\n0.25\n0.002\nPft\n-0.003\n0.50\n0.001\n0.75\n0.001\n-0.002\n0.001\n-0.001\nP 1s\n0.25\n0.50\n-0.003\n-0.001\n-0.001\n0.75\n-0.002\n0.25\n0.007\n-0.001\nPfs\n0.008\n0.50\n0.004\n0.75\n0.002\n0.003\n-0.003\n0.25\n-0.004\nP 1t\n0.50\n0.000\n-0.007\n-0.004\n0.75\n0.002\n0.000\n-0.006\n0.25\nP\nS\n0.50\n0.001\n-0.003\n0.75\n-0.002\n0.000","49\nAPPENDIX B\nBudgetary Summary","50\nSummary of expenditures\nA.\n$321,938\n1) Labor\n14,620\n2) Travel persons\n18,591\n3) Transportation of things\n7,907\n4) Rent, communication, and utilities\n67\n5) Printing and reproduction\n2,501\n6) Contract services\n227,104\n7) Supplies, materials, and equipment\n6,808\n8) SLUC\n123,208\n9) NOAA and DOC overhead\n722,744\nTOTAL\nB. Major property items\n$2,613\n1) Graphics plotter\n3,175\n2) Microcomputer Compaq Deskpro\n515\n3) Printer, Epson FX-286"]}