Previous Page Table of Contents Next Page



Howard L. Raymond and Gerald B. Collins
Northwest Fisheries Center, National Marine Fisheries Service
National Oceanographic and Atmospheric Administration
2725 Montlake Boulevard East, Seattle, Washington 98112


Techniques used in the northwestern United States to study the effect of dams on the migration and survival of juvenile Pacific salmon, Oncorhynchus spp., and steelhead trout, Salmo gairdneri, are described. Populations are estimated by a mark and recapture programme that includes collection of fish samples in free-flowing rivers, in impoundments, in an estuary, and at dams.


On décrit les techniques utilisées dans le nord-ouest des Etats-Unis pour étudier l'effet des barrages sur la migration et la survie du saumon juvénile du Pacifique, Oncorhynchus spp., et de la truite arc-en-ciel, Salmo gairdneri. Les populations sont évaluées au moyen d'un programme de marquage et recapture qui comprend le rassemblement d'échantillons de poisson des cours d'eau courante, des étangs de barrage, des estuaires et des barrages.




2.1 In Free-flowing Rivers
2.2 In Impoundments and Estuaries
2.3 At Dams



4.1 Sampling Efficiency
4.2 Population Estimates
4.3 Timing and Migration Rate
4.4 Survival
4.5 Evaluation of the Effect of Dams



The measurement of populations of juvenile Pacific salmon, Oncorhynchus spp., and steel- head trout, Salmo gairdneri, in the Columbia River of the northwestern United States has been the focus of large-scale research efforts for the past decade. The Columbia River has been changed from a swift, free-flowing stream to a series of long, slow-flowing impoundments by the construction of many dams (Figure 1). The effect of this river development on valuable stocks of fish has been of special concern to fishery agencies responsible for managing the Columbia River fishery.

Losses to juvenile seaward migrants passing through a total of 8–10 dams and impoundments could be substantial. Losses from passage through the turbines at McNary Dam averaged 11 percent (Schoeneman, Pressey and Junge, 1961). Long (1968) measured a 33 percent loss of fish passing through turbines of Ice Harbor Dam. Assuming a loss through turbines of 11–33 percent per dam, total losses over eight dams could range from 60–96 percent of those fish passing through turbines.

There are problems in passage through impoundments also. Downstream passage of juveniles takes longer in the slower moving flows in impoundments, exposing the fish for a greater length of time to the adverse effects of predation, disease, and supersaturation (Ebel, 1971). Delays of six days during passage through McNary Reservoir and eight days through John Day Reservoir have been measured (Raymond, 1968 and 1969).

Anticipating the need for more data on migration and survival as the environment becomes altered by new dams, scientists of the National Marine Fisheries Service (NMFS) initiated a major study in 1969 to measure the effects of the new dams and reservoirs on the migration and survival of juvenile chinook salmon, O. tshawytscha, and steelhead trout. Estimates of juvenile populations were made for many sites and involved collection of fish under a variety of conditions in free-flowing sections of the river, in impoundments, at dams, and in the estuary. The basic method for appraisal of migrating populations of young salmonids was a mark and recapture programme. At each of the sampling sites, the juvenile salmonids were collected, counted, separated by species, examined for marks, and then marked and released at designated sites. Subsequent recovery of the marked fish at downstream sites provided data for assessment of the effect on juvenile salmonids of new dams and impoundments.


2.1 In Free-flowing Rivers

In creeks and rivers at sites where water velocities exceed 1.2 m/s, floating, self-cleaning scoop traps (Humphreys, 1969) proved to be the most efficient method for collection (Figure 2). We have successfully fished single 1.2 /times/ 1.8 m traps on smaller streams and arrays of two or three traps measuring 3.0 /times/ 3.2 m on the larger ones. Traps can be fished up to a depth of 2.1 m depending on water velocity and stream profile. The self-cleaning principle eliminated most of the debris problems that plagued operations using scoop traps equipped only with stationary screens. We replaced the stationary screen or the inclined plane with a travelling screen to carry debris, such as bark, loaves, small sticks and algae up and into the live box where it is removed by a rotary drum screen (Figure 3). In a normal year we are able to fish these traps throughout most of the outmigration except for limited periods when large logs are passing the trap site. Collection efficiencies have ranged from 3–15 percent, depending on river flows. By contrast, the earlier (stationary screen) models often could not be fished during high water due to debris. Even when fishing, collection efficiency was lower than at a self-cleaning trap because any debris on the screen created a velocity barrier in front of the trap which the fish (especially juvenile steelhead trout) could detect and thus avoid.

“Migrant Dippers” (Mason, 1966) were effective collectors on slower-moving rivers where water velocities were less than 1.2 m/s (Figure 4). Krcma and Raleigh (1970) used a modified “dipper” in the Snake River. The basic unit consisted of a trap section 12.2 m long by 7.6 m wide by 1.8 m deep, with fixed louver leads that extended 9.8 m upstream at a 10° angle to the flow. A self-cleaning travelling screen formed the rear of the trap and a metal screen floor extended upstream to the two fixed louver sections. The louvers guided the fish into the trap area where a continuously rotating scoop removed the fish from the trap and deposited them into a trough. The fish were then flushed into a holding pen at the side of the trap. Depending on river flows, the “dipper” with louver leads collected between 2.3 and 16.7 percent of the juveniles passing the trap site.

2.2 In Impoundments and Estuaries

In impoundments and sections of the river in which velocities are minimal, “Merwin Traps”, gillnets, beach seines and purse seines have proven effective.

“Merwin” floating traps with shore leads, similar to that used by Hamilton et al. (1970) are most efficient in the shallower portions of impoundments. The floating trap (Figure 5), adapted from designs formerly used for commercial fishing in Alaska, intercepts fish migrating relatively close to shore.

Although gillnets could not be used in our programme to provide samples of live fish for marking, they were effective for determining vertical and horizontal distribution of finger-lings (Figure 6) in reservoirs (Smith, Pugh and Monan, 1968).

Beach seining is used wherever there are beaches of sand, hard mud or gravel; a technique described by Sims and Johnsen (in press) has been effective in rivers, reservoirs and in the estuary. The variable mesh net is 101 m long, is designed for a two- or three-man operation, and can be set with a small boat (5.5–6.1 m) powered by an outboard engine (40 hp minimum). To set the net, the anchor wing is attached to an anchor or log on the beach and the net is laid out along the water's edge in the direction of the current. The tow line is picked up by the boat and the net is towed into the current as close to the beach as possible without grounding the motor. During the sweep, just enough power is used to keep an arc in the net. At the end of the sweep, the lead wing is closed and the net is worked along the beach; one man to the floating line and one to the leadline until the catch is forced into the bag. After the catch is removed from the bag, the net is in position to be reset without further handling. Collection efficiency varies depending on species and size of the river or reservoir. In the Columbia River estuary, beach seining effectively collects “O”-age autumn chinook salmon which tend to migrate near shore but very few yearling chinook and steelhead trout that normally migrate in mid-channel. An excellent method for obtaining migrants in mid-channel areas is to utilize a purse seine just offshore from the operational limits of the beach seine.

Purse seining (Durkin and Park, 1967; Johnsen and Sims, 1973) is effective in the deeper offshore portions of reservoirs and in the estuary. The original equipment, developed by Durkin and Park, seined with a 3 × 6 m raft powered by a 28 hp outboard motor and a skiff. An A-frame and double gypsy winch were mounted on the raft to facilitate pursing the net and lifting or lowering the depressor weight. The net is set by the seine skiff pulling the bunt end from the raft as it circles away in a clockwise direction. When the entire net is in the water, we either purse immediately by countinuing to circle until the skiff and barge meet and the bunt end of the net is returned to the barge, or tow, by powering both the barge and skiff for a period of 5–15 minutes before pursing the net. A second function of the skiff is to hold the purse open to facilitate removal of the fish with a dipnet. Purse seines vary in size from 183–279 m long by 6.1–10.7 m deep. The major disadvantages of the barge are lack of freeboard and slow speed which restrict its use to calm days or areas generally protected from winds. For work in an estuary, Johnsen and Sims needed a more stable platform and sufficient mobility to reach widely separated sites within a short period. A boat of the type used by local fishermen (gillnetters) was modified for purse seining; and it seems to have met the above requirements. The high shear bow section provided additional forward freeboard necessary in rough water and for high speed running (up to 25 kn), and provided a safe, effective work platform while seining. Techniques of purse seining are generally the same for the modified gillnetter and the barge. In more recent years we have had good success with the technique in the ocean and upriver in the large impoundments behind major dams, even in waves of 1–2 metres.

2.3 At Dams

The large size of the Columbia River made the collection of significant numbers of juvenile salmon and trout at many sites a difficult, dubious, and expensive undertaking until a technique was developed that uses the dams as collectors. Research revealed that juvenile migrants concentrate in the upper levels of turbine intakes (Long, 1968) and enter turbine gate-wells in large numbers. Each turbine unit has three gate-wells in which slots are provided for gates to seal off the turbine from the forebay during maintenance. Juvenile salmon and trout may be collected in gate-wells in large numbers by means of a specially designed dipnet (Figure 7) (Bentley and Raymond, 1968). A crane lowers the net in closed position through the slot to the desired depth. The net is then opened, slowly raised to the surface, and closed again to ease its withdrawal through the gate opening. An isometric view of one Kaplan turbine at Ice Harbor Dam showing the open net in one gate-well is given in Figure 8. Gate-wells are natural collectors; during low river flows, when the entire river passes through the powerhouse, up to 25 percent of the seaward migrants enter the gate-wells on their own volition. With recently developed devices, such as travelling screens (Figure 9) in the turbine intakes (Farr, in press) up to 80 percent of the migration can be diverted into the gate-wells and become available for collection. Such availability ensures adeqúate sample size for studies to determinc the timing, migration rates, survival and relative abundance of the seaward migrations.


Marking of fish was by thermal branding with various symbols, letters and numbers. Either boiling water (Groves and Novotny, 1965), alcohol and dry ice (Park, 1969), or liquid nitrogen (Mighell, 1969) was used as the medium for transferring the brand to the fish. We found no differences between brand retention among the three mediums used, either in the field or in the laboratory where a retention period of 6–8 months was required between release and recapture of branded fish. Liquid nitrogen proved to be the most convenient for mass marking requirements in our studies.

Generally, each symbol provided 16 combinations (four rotations of the brand placed on four different areas of the fish). Use of this technique minimized the numbers of symbols required; an important consideration since marks were changed every three days at many sites.


4.1 Sampling Efficiency

In order to accurately assess survival and relative abundance of fish stocks and to determine timing of runs, it was necessary to determine the proportion of the total downstream migration being sampled by the collection equipment. This was accomplished by marking and releasing fish upstream from the primary assessment sites. The percentage of marked fish subsequently recovered determined the sampling efficiency. Sampling efficiency was simply the number of marks recovered at the assessment site divided by the number released upstream for each marking period. Continuous releases were made throughout the downstream migration each year. Since efficiency varied considerably with varying river flows, a valid measure of sampling efficiency requires that the marked fish released upstream of the primary assessment sites are mixed with the unmarked population. We determined that there was mixing by comparing returns of fish released on each bank. It was assumed that, since the proportion of marked fish recapture was consistent from both sides of the river, mixing did take place.

4.2 Population Estimates

Estimates of the magnitude of the population passing each site, made on a weekly and seasonal basis, were determined by Chapman's (1948) general formula N= nt/s where N is the estimated number of fish, n = number of fish marked and released upstream, t = total collected and s = number of recaptured marks.

4.3 Timing and Migration Rate

Timing from various tributaries of the Snake River was based on the arrival of the marked fish from each stream at Ice Harbor Dam. The median of the respective migrations from the various tributaries was used to compare the timing of these populations. The duration of migrations - as determined by recoveries of marked fish - was also obtained.

Estimates of timing at Ice Harbor and The Dalles Dams were based on the median date of arrival of each of the species as determined from the estimated proportion of the migration passing each week, rather than actual numbers collected. The latter does not account for variation in sampling efficiency from week to week which, if significant, would bias our estimates.

Migration rate (travel time) is based on the difference between the median date of release and median date of recapture for each marked release group. For example, if the elapsed time between median dates of release and recapture was 20 days and the distance travelled was 300 km, the migration rate was 15 km per day. Each release group was weighed according to the proportion of the total outmigration represented by that group in determining average migration rate for the entire outmigration.

4.4 Survival

Assessment of survival was based on the difference in recovery rate at dams (e.g., Ice Harbor Dam and The Dalles Dam) of test fish marked and released at an upriver sampling station (e.g., Whitebird, Idaho, on the Salmon River) and control fish marked and released a short distance upstream from each dam. For example, if test fish were recovered at a rate of 2 percent and control releases at Ice Harbor were recovered at a rate of 4 percent, then survival from the Salmon River to Ice Harbor Dam would be 2 percent, 4 percent or 50 percent. Measurements of survival through other stretches of river or through a dam were calculated in a like manner.

4.5 Evaluation of the Effect of Dams

Values obtained each year for timing, migration rate, and survival from the Salmon River to Ice Harbor Dam and to The Dalles Dam were then summarized for the predam period and for the postdam period. Differences obtained between the two periods for the variables measured provided our assessment of the effects of the new dams on juvenile salmonids.

Figure 1

Figure 1 Location of dams in the Columbia and Snake Rivers and sites where marked fish were released or later recaptured

Figure 2
Figure 2 Self-cleaning scoop traps in fishing position on the Salmon River

Figure 3

Figure 3 Plan and side view of scoop trap
(Oregon State Game Commission)

Figure 4

Figure 4 “Migrant Dipper” floating traps with louver leads for sampling smolts in Snake River

Figure 5

Figure 5 Lake Merwin floating trap for sampling smolts in reservoirs and lakes (from Hamilton et al., 1970)

Figure 6

Figure 6 System for suspending gillnets in reservoirs

Figure 7

Figure 7 Dipnet used to remove fish from turbine intake gatewells at dams

Figure 8

Figure 8 Isometric view of Ice Harbor Dam showing:

  1. Area unavailable for fishing
  2. Operating gate and cylinder in stored position
  3. Gatewell net in open position
  4. Warehouse crane
Figure 9

Figure 9 Section of a typical turbine showing travelling screen in the intake for diverting fish into gatewells

Previous Page Top of Page Next Page