Detection of PIT-Tagged Juvenile Salmonids Migrating in the Columbia River Estuary, 2016 Matthew S. Morris, Alexander J. Borsky, Paul J. Bentley, Lindsey N. Webb, and Benjamin P. Sandford Report of research by Fish Ecology Division, Northwest Fisheries Science Center National Marine Fisheries Service National Oceanic and Atmospheric Administration 2725 Montlake Boulevard East Seattle, Washington 98112-2097 for Division of Fish and Wildlife, Bonneville Power Administration U.S. Department of Energy P.O. Box 3621 Portland, Oregon 97208 Contract 46273 RL58 Project 1993-029-00 April 2017

ii

iii

iv

Executive Summary In 2016, we continued a multi-year study in the Columbia River estuary to detect juvenile Pacific salmon Oncorhynchus spp. marked with passive integrated transponder (PIT) tags. Fish were detected using a surface pair trawl with a matrix of rectangular detection antennas fitted into the cod-end. The matrix was configured with three parallel antennas in front and three in the rear for a total of six individual antennas. This configuration relied on trawl net wings to guide fish toward the cod end of the trawl, where they would come within detection range of the antennas. Entrained fish were able to exit the trawl safely, without capture or handling. We deployed the trawl in the Columbia River navigation channel between river kilometer (rkm) 66 and 84 and sampled for a total of 829 h. During this period, we detected 12,165 PIT-tagged juvenile salmon, of which 16% were wild, 4% were unknown, and 80% were of hatchery origin. Species composition of detected fish was 43% spring/summer Chinook, 3% fall Chinook, 43% steelhead, 2% sockeye, 6% coho, less than 1% cutthroat trout, and 3% unknown. Our sampling schedule began with a single daytime shift operating 3-5 d/week. We began sampling on 24 March to coincide with arrival in the estuary of spring-migrating juvenile salmon and steelhead. As numbers of juvenile migrants increased, we intensified the sample effort with two daily shifts: one operating 7 d/week during daylight and a second operating 6 d/week during darkness. Intensive sampling continued from 1 May through 9 June, when we returned to a single daytime shift. We ended sampling on 5 July after most spring migrants had passed. During the intensive sample period, average hourly detections of yearling Chinook were significantly higher in darkness than daylight hours (11 vs. 5 fish/h; P < 0.001). Conversely, average hourly detections of steelhead were higher during daylight than darkness hours (11 vs. 4 fish/h; P < 0.001). During intensive sampling, the trawl was deployed for an average of 15 h/d, the same as in 2015. Also during the intensive sampling period, we detected 2.0% of the yearling Chinook salmon and 3.4% of the steelhead detected at Bonneville Dam. These proportions were lower than those in 2015, when we detected 3.3% of the yearling Chinook and 4.6% of the steelhead detected at Bonneville Dam.

v

We also detected 1.6% of the yearling Chinook and 2.8% of the steelhead transported and released below Bonneville Dam. Again, these rates were lower than in 2015, when we detected 3.4% of transported yearling Chinook and 3.9% of transported steelhead. In 2016, we continued development of a flexible antenna array that could be towed behind two small vessels to detect juvenile salmon without a net. Objectives of this development effort are to simplify logistics, increase sample efficiency, and reduce the cost of sampling PIT-tagged fish in the estuary. For these tests, the flexible antenna array was configured with six rectangular antennas each housed in 1.9-cm diameter flexible PVC hose. The flexible antenna system can utilize up to 12 antennas, but six were used because of limited resources. Sampling cruises to test the flexible antenna array were conducted simultaneously with deployments of the matrix trawl system. We compared detection efficiency between the two systems to evaluate the feasibility of transitioning entirely to the flexible system in future years. The flexible antenna system detected a total of 549 fish across 21 d of operation in 2016 (78.2 h total). Of these total detections, 14% were Chinook, 79% steelhead, 4% coho, less than 1% sockeye, and 3% unknown species. We sampled the flexible and matrix systems simultaneously on 10 d. Daily ratios of mean detection rate (fish/h) between the two systems were calculated for each concurrent deployment. Over the 10 d of simultaneous deployment, the overall mean ratio of detection rates for the flexible vs. the matrix trawl systems was 60%. Juvenile steelhead were detected in disproportionately high numbers in the flexible antenna system, comprising 78% of total detections—more than double the proportion of steelhead detected in the matrix trawl (35%). This discrepancy was likely due to the shallow sample depth of the flexible antenna system compared to the matrix trawl system (3.0 vs. 5.0 m). We concluded that the relatively shallow sample depth of the flexible antenna system likely allowed Chinook salmon to pass below the array without being detected. After the spring migration season, we tested multiple orientations and configurations of a six-antenna flexible array with the objective of increasing sample depth. Testing confirmed the feasibility of re-configuring antennas to obtain a sample depth of 6.0 m without sacrificing electronic performance. We plan to use the new configuration in 2017 sampling and expect that it will increase detections of Chinook salmon.

vi

As in previous study years, we examined PIT-tag detection data collected in 2016 to evaluate potential factors related to detection probability and to compare passage performance metrics among fish groups by species, rearing type (hatchery or wild), and migration history (transported vs. inriver). For yearling Chinook salmon, we found a significant effect of migration history and arrival date on detection rate (P = 0.013 and P < 0.001, respectively), but no significant interaction between date and migration history (P = 0.957). We detected inriver migrants at a higher rate than transported fish, with detection rates for both groups increasing through the season. Similarly, for steelhead, there were significant effects of migration history and date-squared on detection rate (P = 0.029 and P = 0.021, respectively), but no interaction between date and migration rate (P = 0.429). As with Chinook salmon, we detected inriver migrant steelhead at a higher rate than transported steelhead. Detection rates for both steelhead groups increased through the season but began to decrease in early June. Over the years, we have observed an inverse relationship between river flow and detection rates in the trawl. Mean flow volumes at Bonneville Dam were 23% higher during the two-shift sample period of 2016 than during the two-shift period of 2015 (6,953 vs. 5,333 m3/s). However, mean flows during this period in 2016 were still 18% lower than the 14-year average for 2000-2014 (8,296 m3/s, excluding 2001). Mean daily river flows were above average for the first two weeks of the intensive sampling season (1-14 May 2016). After mid-May, flows decreased to drought-like conditions and remained low through the end of the season.

From the total combined species detected with the trawl system in 2016, 17% had been transported, while 20% had been detected passing Bonneville Dam. The remaining 63% had neither been transported nor detected at Bonneville Dam, although at least 94% of total detections had originated upstream from Bonneville. Mean migration rate to the estuary (rkm 75) was significantly faster for yearling Chinook salmon detected at Bonneville Dam than for those released from barges just below the dam (82 vs. 73 km/d, P < 0.001). Similar differences were noted for steelhead, with travel rates of 92 km/d for inriver migrants vs. 86 km/d for transported fish (P < 0.001). This trend was repeated in observations of mean migration rate between inriver migrant and transported sockeye (95 vs. 78 km/d) and subyearling Chinook salmon (72 vs. 58 km/d). However, for both species, detection numbers were insufficient for meaningful statistical analysis. Between Bonneville Dam and the sample reach overall,

vii

migration rates for both inriver and transported fish were faster in 2016 than in 2015 across all species. This was likely a function of high flow volumes during the first half of May. Trawl detections of subyearling Chinook salmon have decreased in recent years, commensurate with reduced tagging effort for these fish. In 2016, we detected 278 subyearling fall Chinook, with the majority of detections occurring from mid-May through the end of the sampling season. Of these 278 fish, 214 originated from the Snake River basin (63 inriver migrants and 151 transported). The remaining 64 were inriver migrants from Columbia River stocks, with 12 released above McNary Dam, 40 released between McNary and Bonneville Dam and 12 released below Bonneville Dam. We did not detect holdover subyearlings in 2016 (holdovers overwinter in freshwater and resume downstream migration in spring). Of the 234 sockeye salmon detected in the trawl, 47% had been released into the Snake River, 53% into the upper Columbia River above McNary Dam, and less than 1% into the middle Columbia River between McNary and Bonneville Dam. Of sockeye detected in 2016, 35% were hatchery, 12% were wild, and 53% were of unknown origin. Migration history of detected sockeye was 79% inriver migrant and 21% transported. As in previous years, detection data from the trawl in 2016 were essential for calculating annual survival probabilities to the tailrace of Bonneville Dam, the last dam encountered by migrating juvenile salmonids. Operation of a detection system in the estuary has provided data for estimates of survival to Bonneville Dam since 1998. These estimates are critical to research and management programs for endangered salmonids in the Snake and Columbia River basin and in other basins of the Pacific Northwest. Trawl detections also contribute data for estimates of travel time, migration rate, and other performance metrics between juvenile fish groups of differing migration history, origin, rearing type, and species. Annual releases of PIT-tagged fish in the Columbia River basin have been near 2 million for the past several years. Detections of these fish as they pass through the estuary continue to increase our understanding of behavior and survival during the critical smolt transition period.

viii

Contents Executive Summary ........................................................................................................... iv Introduction ......................................................................................................................... 1 Matrix Antenna Trawl System ............................................................................................ 2

Methods................................................................................................................... 2 Study Area .................................................................................................. 2 Study Fish ................................................................................................... 3 Sample Period ............................................................................................. 3 Trawl System Design .................................................................................. 4

Results and Discussion ........................................................................................... 7 Factors Affecting Detection Rate ................................................................ 7 Detection Rates ........................................................................................... 9 Species, Origin, and Migration History of Detected Salmonids ............... 10 Impacts on Fish ......................................................................................... 14

Development of a Flexible Antenna Detection System .................................................... 15

Background ........................................................................................................... 15 Methods................................................................................................................. 16 Results and Discussion ......................................................................................... 20

Detection Rates ......................................................................................... 20 Performance of the Flexible Antenna vs. Matrix Trawl System .............. 21 Post Season Testing .................................................................................. 25

Objectives for Future Development ...................................................................... 28 Analyses from Trawl Detection Data ............................................................................... 29

Travel Time and Migration Rate ........................................................................... 30 Methods..................................................................................................... 30 Results and Discussion ............................................................................. 31

Diel Detection Patterns ......................................................................................... 34 Methods..................................................................................................... 34 Results and Discussion ............................................................................. 34

Detection Rates of Transported vs. Inriver Migrant Fish ..................................... 37 Methods..................................................................................................... 37 Results and Discussion ............................................................................. 38

References ......................................................................................................................... 42 Appendix ........................................................................................................................... 46

ix

Introduction In 2016, we continued a multi-year study in the Columbia River estuary to collect data on migrating juvenile Pacific salmon Oncorhynchus spp. implanted with passive integrated transponder (PIT) tags (Ledgerwood et al. 2004; Morris et al. 2015). Data from estuary detections were used to estimate survival and downstream migration timing of these fish. As in previous years, we used a large surface pair trawl to guide fish through an array of detection antennas mounted in the cod-end. Target fish were those PIT-tagged by other researchers for various research projects at natal streams, hatcheries, collection facilities at dams, and other upstream locations (PSMFC 2016). When PIT-tagged fish are entrained in the trawl, they must pass through the detection field of the antennas to exit. Upon detection, the tag code, GPS position, and date and time of detection are electronically recorded. This study began in 1995 and has continued annually (except 1997) in the estuary near Jones Beach, approximately 75 river kilometers (rkm) upstream from the mouth of the Columbia River. Nearly 1.7 million Snake and Columbia River juvenile salmonids were PIT-tagged and released prior to or during the spring 2016 migration season (PSMFC 2016). A portion of these fish were detected at dams equipped with PIT-tag monitoring systems (Prentice et al. 1990a,b). These systems automatically upload detection information to the PIT Tag Information System database (PTAGIS), a regional database that stores and disseminates information on PIT-tagged fish (PSMFC 2016). We uploaded trawl detection records to PTAGIS and downloaded information on the fish we detected. We used data on the species, run, tagging/release time and location, and date/time of detection at interrogation sites downstream from release to evaluate migration performance metrics between Bonneville Dam and the estuary. Since 1998, these data have been used for annual estimates of survival for yearling Chinook salmon O. tshawytscha, steelhead O. mykiss, and sockeye salmon O. nerka from points of release to Bonneville Dam. Estuary detections are critical to complete estimates of survival through the entire hydrosystem (Faulkner et al. 2016). In 2016, 100,589 PIT-tagged fish were transported from dams on the Snake River and released below Bonneville, and over 110,000 PIT-tagged inriver migrants were detected at Bonneville Dam. Seasonal trends in estuary detection data continue to provide insight into the relationship between juvenile migration performance and smolt-to-adult return ratios (Marsh et al. 2008, 2012).

2

Matrix Antenna Trawl System Methods Study Area Trawl sampling was conducted in the upper Columbia River estuary between Eagle Cliff (rkm 84) and the west end of Puget Island (rkm 66; Figure 1). This is a freshwater reach characterized by frequent ship traffic, occasional severe weather, and river currents often exceeding 1.1 m s-1. Tides in this area are semi-diurnal, with about 7 h of ebb and 4.5 h of flood with a range of 1.9 m. During the spring freshet (April-June), little or no flow reversal occurs in this reach during flood tide, especially in years of medium-to-high river flow. The trawl was deployed adjacent to a 200-m-wide navigation channel, which is maintained at a depth of 14 m. Figure 1. Trawl sampling area adjacent to the navigation channel in the upper Columbia

River estuary between rkm 66 and 84.

3

Study Fish We continued to focus detection efforts on large release groups of PIT-tagged fish detected at Bonneville Dam or transported and released just downstream from the dam. The vast majority of these fish migrate through the tidal freshwater reach of the estuary from late April through late June. Release dates and locations of fish detected with the trawl were retrieved from the PTAGIS database (PSFMC 2016). Specific groups of tagged fish targeted for detection included over 194,000 fish released for a comparative survival study of hatchery fish, and 100,589 fish diverted to barges for NMFS transportation studies, as well as smaller groups released for other studies. Migrating juvenile fish released in the upper Snake River must traverse eight dams and reservoirs or be transported from one of three collector dams to reach the tailrace of Bonneville Dam. Transported fish can avoid passage of up to 7 dams and migration through approximately 461 km of river from the tailrace of Lower Granite Dam to the tailrace of Bonneville Dam (Marsh et al. 2005; 2008; 2010; 2012). Detection numbers in the pair-trawl were sufficient for analyses of timing and survival for yearling Chinook salmon and steelhead. Trawl detections of sockeye and subyearling Chinook salmon were fewer; therefore, analyses were limited due to smaller sample sizes for these fish. We also detected PIT-tagged coho salmon O. kisutch and coastal cutthroat trout O. clarki. Sample Period Spring sampling began on 24 March and continued through the summer migration period to 5 July 2016. Our sample effort varied commensurate with fish availability in the estuary. Early and late in the migration season, we sampled 2-5 d/week with a single shift, for an average daily sample effort of 6 h/d (sample effort was defined as full deployment of the trawl). During the peak of the spring migration from 1 May through 9 June, we sampled with two daily shifts, covering both daylight and darkness periods, for an average daily effort of 15 h/d. During the two-shift period, day shifts began before dawn and continued for 6-11 h, while night shifts began in early evening and continued through most of the night or until relieved by the day crew. Sampling was intended to be nearly continuous throughout the two-shift period except between 1400 and 1900 PDT, when we interrupted sampling for refueling and maintenance.

4

Trawl System Design Antenna Configuration—Configuration of the matrix antenna was similar to the design used since 2013 (Figure 2). A fish-passage corridor was formed using a front and rear antenna array, each consisting of three parallel antennas, for a total of six antennas. Inside dimensions of individual antennas measured 0.75 by 2.8 m. A 1.5-m length of net mesh connected front and rear components, and the overall fish-passage opening was 2.6 by 3.0 m. The matrix antenna array was attached to the cod end of the trawl and suspended by buoys 0.6 m beneath the surface. This design allowed fish collected in the trawl to exit through the antenna while remaining in the river. Both the front and rear antenna arrays weighed approximately 114 kg in air and required an additional 114 kg of lead weight to suspend in the water column (total weight of front and rear components was 456 kg in air). Figure 2. Basic design of the surface pair trawl used with the matrix antenna system to

sample juvenile salmonids in the Columbia River estuary (rkm 75), 2016.

5

Trawl Net—Basic configuration of the pair-trawl net has changed little through the years, despite considerable changes to the detection apparatus (Ledgerwood et al. 2004). The upstream end of each trawl wing was shackled to a 3-m-long spreader bar. The downstream end of each wing was attached to the 30.5-m-long trawl body, which was modified at the cod-end for attachment of the matrix array. The trawl mouth opening was 9 m wide by 6 m tall with a 6.3-m floor extending forward from the mouth. Sample depth was about 5.0 m due to curvature in the side-walls under tow. We towed the pair trawl with 73-m-long lines to prevent turbulence on the net from the tow vessels. After the trawl and antenna were deployed, one towline was passed to an adjacent tow vessel. Both vessels then towed the net upstream facing into the current, maintaining a distance of about 91.5 m between the distal ends of the trawl wings. Even though volitional passage through the trawl and antenna occurred while towing with the wings extended, we continued to bring the wings of the trawl together every 17 minutes to flush debris out of the system. The majority of fish were detected during these 7-minute net-flushing periods. Electronic Components and Data Transmission—For the matrix trawl detection system, we used essentially the same electronic components and procedures as in 2006-2015. In 2016 we again used a single FS1001M multiplexing transceiver, which was capable of simultaneously powering the detection field and recording and transmitting detection data for up to six antennas. Electronic components for the trawl system were contained in a water-tight box (0.8 × 0.5 × 0.3 m) mounted on a 2.4 by 1.5-m pontoon raft tethered behind the antenna. Data were transmitted from each antenna coil to specific transceiver ports via armored cable. A DC power source was used for the transceiver and antenna. Data were stored temporarily in the transceiver buffer and transmitted wirelessly in real-time to a computer on board a tow vessel. During the season, status reports generated by the transceiver were monitored in real time to confirm performance, and each antenna coil was tested periodically using a PIT tag attached to a telescoping pole. For each fish detected, the date and time of detection, tag code, coil identification number, and GPS location were received from the antenna and recorded automatically using the computer software program MiniMon (PSMFC 2016). Written logs were maintained for each sampling cruise noting the time and duration of net deployment, net retrieval, approximate location, and any incidence of impinged fish.

6

Detection data files were uploaded periodically (about weekly) to PTAGIS using standard methods described in the PIT Tag Specification Document (Marvin and Nighbor 2009). The specification document, PTAGIS operating software and user manuals are available from the PTAGIS website operated by Pacific States Marine Fisheries Commission (PSMFC 2016). Pair-trawl detections are designated in the PTAGIS database with site code TWX (towed array-experimental). Pre-Season Tests of Detection Efficiency—As in previous years, we used PIT tags attached to a test tape to evaluate performance of the matrix antenna system (Ledgerwood et al. 2005; Morris et al. 2013). For these tests, we positioned a 2.5-cm-diameter PVC pipe through the center of both the front and rear antenna arrays. The pipe was extended at least 0.5 m beyond reading range at both ends. The entire matrix was then deployed behind an anchored tow vessel without the trawl. Tests were conducted independently on port, middle, and starboard antennas. We attached PIT tags to a vinyl-coated tape measure, with tags spaced at intervals of 30, 60, and 90 cm, and at orientations of 45- or 90-degrees relative to the tape edge. The tape was then passed back and forth through the pipe and retrieved by a second vessel. Detection efficiency was evaluated based on the proportion of tags detected during a single pass of the tape. Estimated detection efficiencies from these tests were positively correlated with spacing between test tags, regardless of tag orientation (45 vs. 90 degrees). Of the 1,512 test tags passed through the matrix antenna, those spaced at 30-cm intervals were detected only 2% of the time when oriented at 45 degrees and less than 1% when oriented at 90 degrees. With spacing between tags increased to 60 cm, detection efficiency increased to 92 and 97% for tags oriented at 45 and 90 degrees, respectively. For test tags spaced 90 cm apart, respective reading efficiency was 97 and 94% for tags oriented at 45 and 90 degrees. Results in 2016 were similar to those in previous years and showed the matrix antenna was performing as expected.

7

Results and Discussion Factors Affecting Detection Rate Flow volumes—During the 2016 intensive sampling period of 1 May-9 June 2016, mean river flow volume at Bonneville Dam was 23% higher than during the same period in 2015 (6,953 vs. 5,333 m3/s; Figure 3). However, mean flow during intensive sampling in 2016 was still 18% lower than the 14-year average (8,296 m3s-1). During the first two weeks in May, mean river flows exceeded those of the 14-year average, but by mid-May, river flows declined substantially and were more similar to those seen in 2015. In general, the 2016 season was characterized by lower-than-average flows and light debris loads. Through years of sampling, we have observed an inverse relationship between river flow volumes and trawl detection rates. Higher flow volume has been consistently associated with lower detection rates of fish previously detected at Bonneville Dam. (Detection in the pair trawl of fish previously detected at Bonneville provides a rough index of estuary detection efficiency with the trawl). Figure 3. Columbia River flows (m3 s-1) at Bonneville Dam during the two-shift sample

periods in 2015 and 2016 compared to the average flow from 2000 to 2014 (excluding 2001). Drought-year flows for 2001 are also shown for comparison.

Flow

(m3 s

-1)

Seasonal Columbia River Flow Measured at Bonneville Dam, 2000-2016

29 Apr 7 May 15 May 23 May 31 May 8 Jun 16 Jun

avg 00-14 20162001 2015

10,000

8,000

6,000

4,000

0

12,000

2,000

8

A variety of factors contribute to the relationship between higher river flows and lower detection rates. First, higher flows carry fish downstream more quickly. This decreases the amount of time that a given fish is present in the sample reach and available for detection. Second, higher flows allow migrants to expand across a larger cross-sectional area of water. For fish present in the estuary during sampling, we expect that increased spatial dispersion of the passing population would decrease the likelihood of an individual fish entering the trawl. Higher flows also decrease actual sample time in three ways. First, they increase the transit time required for vessels to return to the upstream end of the sample reach, where the trawl is initially deployed. Second, they decrease the period during which the trawl is deployed by speeding transit to the downstream end of the sample reach, where the trawl must be retrieved. Finally, higher flows typically yield more debris accumulation in the trawl net, increasing sample time lost to debris removal. Presence of Tagged Fish in the Sample Reach—We estimated that intensive sampling in 2016 coincided with arrival time in the estuary of 75% of the yearling Chinook salmon and 80% of the steelhead passing Bonneville Dam (tagged and non-tagged). Our intensive sample period also coincided with 99% of the yearling Chinook salmon and steelhead transported for NMFS studies. The availability of in-river yearling Chinook salmon and steelhead passing Bonneville Dam was less in 2016 compared to 2015, when 80% of yearling Chinook salmon and 88% of steelhead were estimated in the estuary during our intensive sampling period. In contrast, the availability of transported yearling Chinook salmon and steelhead was greater in 2016 compared to 2015, when 97% of yearling Chinook salmon and 95% of steelhead were available during intensive sampling. In 2016, no transported fish were released before our intensive sampling period. After the intensive sampling period, the majority of fish detected at Bonneville Dam were subyearling Chinook salmon, although yearling Chinook, coho, sockeye, steelhead, and cutthroat were also detected. Fish transportation from upstream dams continued until the end of October (John Bailey, USCOR, Walla Walla Dist., personal communication). Since 2013, tagging effort for subyearling Chinook salmon has been reduced considerably, and these fish comprised only 32% of the detections at Bonneville after intensive sampling ended in 2016. In contrast, subyearlings comprised 64% of the detections at Bonneville after intensive sampling in 2015. Release dates for subyearling Chinook in 2015 and 2016 were comparable. However, the total number of tagged subyearlings released was considerably higher in 2015 than in 2016 (239,665 vs. 175,647) and may have accounted for the difference in detection rates.

9

Detection Rates We sampled with the matrix trawl system for 829 h during 2016 and detected 12,165 PIT-tagged fish. In contrast, we sampled for 813 h during 2015 and detected 19,889 fish (Figure 4). Higher-than-average flows early in the season likely contributed to lower average detection rates in 2016 (15 fish/h) than in 2015 (24 fish/h). Though considerably lower than in 2015, these detection rates were consistent with the average detection rate for the trawl over the past 15-years (19/hr SD 6.3). Figure 4. Daily sample effort in spring/summer 2015 and 2016 using a pair trawl fitted

with a "matrix" antenna for PIT tag detection. Sampling was conducted in a tidal freshwater reach of the Columbia River estuary near Jones Beach between rkm 66 and 84.

0

6

12

18

24

Sam

ple

Effo

rt (h

)

Two-Crew mean: 15 hours per day (4 May – 11 June)

2015

Detections: 19,889Total Hours: 813

Two-Crew mean: 15 hours per day (1 May – 9 June)

2016

Detections: 12,165Total Hours: 829

0

6

12

18

24

19 Mar 18 Apr 18 May 17 Jun

Sam

ple

Effo

rt (h

)

10

Species, Origin, and Migration History of Detected Salmonids In 2016, we detected a total of 11,634 juvenile salmonids of known species and origin (hatchery and wild) plus another 531 fish lacking release information in PTAGIS (Table 1; Appendix Table A1). For most fish, at least some release information was available; however, 378 detected fish had no release or species information associated with their respective tags. Overall Species Composition—Of detected fish with species information, 43% were spring/summer Chinook, 3% were fall Chinook, 43% were steelhead, 2% were sockeye, and 6% were coho salmon. Of the remainder, less than 1% were cutthroat trout, and 3% were unknown salmonid species. Total detections by origin were 16% wild, 80% hatchery, and 4% unknown origin at the time of this report. These numbers may change slightly as PTAGIS records are completed or updated. Table 1. Species composition and origin of PIT-tagged fish detected with the trawl

system in the upper Columbia River estuary near rkm 75 in 2016. Species/Run

Rear type Total Hatchery Wild Unknown

Spring/Summer Chinook salmon 4,514 703 0 5,217 Fall Chinook salmon 310 13 0 323 Coho salmon 641 53 3 697 Steelhead 4,198 1,088 27 5,313 Sockeye salmon 83 28 123 234 Sea-run Cutthroat 0 3 0 3 Unknown 0 0 378 378 Grand total 9,746 1,888 531 12,165 For all species except sockeye salmon, proportions of fish detected by origin in 2016 were similar to those in 2015. For sockeye salmon, the proportion of hatchery fish detected in 2016 (35%) was substantially lower than that detected in 2015 (73%). In addition, the proportion of wild sockeye salmon detected in 2016 (12%) was lower than that detected in 2015 (19%). These differences are likely attributed to an unusually high proportion of sockeye salmon with unknown-origin in 2016 (53%) compared to 2015 (8%). Greater proportions of sockeye with unknown origin were due to a lack of release information associated with tags originating from select hatcheries. Differences in PIT-tagging strategies, hydrosystem operations, and numbers of fish transported also contribute to annual variation in proportions of each species or rearing type detected in the estuary (Figure 5).

11

y N = 12,165

Not barged or detected at

Bonneville Dam, 63%

Lower Columbia2%, n = 275

Barged, 17%

Detected at Bonneville Dam,

20%

Middle Columbia9%, n = 1,104

Snake River60%, n = 7,220

Upper Columbia26%, n = 3,188

Unknown3%, n = 378

Juvenile subyearling fall Chinook salmon begin the downstream migration from late spring to fall, but some of these fish suspend migration and overwinter in freshwater, resuming migration in the following spring. These fish have been said to adopt a "holdover" or "reservoir-type" life-history strategy (Connor et al. 2005). In years with high numbers of tagged subyearling Chinook salmon, we commonly detect a few fish exhibiting this life history type. In 2013, we detected 53 holdover fish, the largest number to date. However, no overwintering fish were detected in 2016. Figure 5. Proportions of fish detected using the matrix trawl by source and migration

history, 2016. Upper and mid-Columbia River sources were defined relative to McNary Dam. Fish that originated in the Columbia River below Bonneville Dam could not be transported, nor could they pass Bonneville Dam.

12

Subyearling fall Chinook salmon—Fish considered subyearlings were only those that had measured less than 120 mm fork length (FL) at tagging and had been released after 11 April 2016. Most fall Chinook released prior to 30 April were yearling rather than subyearling fish, as indicated by a fork length greater than 120 mm at tagging. Based in these criteria, we detected 151 transported and 127 inriver-migrant subyearling fall Chinook in the matrix trawl system between mid-May and early July (Figure 6). Of the 278 inriver migrant and transported subyearlings detected by the trawl system, 77% originated in the Snake River, 4% in the Upper Columbia River at or above McNary Dam, 15% in the mid-Columbia River between Bonneville and McNary Dam, and 4% from the Lower Columbia River below Bonneville Dam. Figure 6. Temporal distribution for subyearling Chinook salmon detected in the

Columbia River estuary near rkm 75 after detection at Bonneville Dam (n = 127) or release from a barge below the dam (n = 151), 2016.

20

10

0

Daily Detections of Subyearling Fall Chinook Salmon, 2016

01 May 21 May 10 Jun

Det

ectio

ns (n

)

30 Jun

30

BargedIn-river40

11 May 31 May 20 Jun 10 Jul

13

12

8

4

0

Daily Detections of Sockeye Salmon, 2016

Det

ectio

ns (n

)

30 Apr20 Apr 10 May 20 May 30 May

10

6

2

BargedIn-river

14

20

16

18

9 Jun 19 Jun

Sockeye salmon—We detected 234 sockeye salmon between 1 May and 11 June (Figure 7). Of these, 35% were hatchery fish, 12% were wild fish, and the remaining 53% were of unknown origin. The disproportionate amount of unknown-origin sockeye salmon in 2016, compared to 2015 (8%), can likely be attributed to the high number of hatchery fish in 2015 (546 individuals) and the altered designation of “wild” to “unknown” at two key Upper Columbia release sites (OSOYOL and SKATAL) from 2015 to 2016. Fish released in the Snake River Basin made up 47% of our sockeye salmon detections, while fish released in the Columbia River Basin upstream from McNary Dam made up 53%. Sockeye salmon originating below McNary Dam comprised less than 1% of our sockeye salmon detections. Transported fish accounted for 49 of the 234 sockeye salmon detections. Of those transported, 28 were from Lower Granite Dam, 19 from Little Goose Dam, and 2 from Lower Monumental Dam. Figure 7. Temporal distribution of sockeye salmon detections in the Columbia River

estuary near rkm 75 during inriver migration (n = 185) or following release from barges below Bonneville Dam (n = 49), 2016.

14

Impacts on Fish We regularly inspected the cod-end of the trawl for debris accumulation near the antenna, which could affect fish. Other sections of the net were monitored visually from a skiff, and accumulated debris were removed periodically. During retrieval, the matrix antenna was hoisted onto a tow vessel while remaining attached to the pair-trawl. Debris that remained in the net was removed by hand through zippers in the top of the trawl body. During debris-removal activities, we recorded all impinged or trapped fish as mortalities, although most fish were released alive. In 2016, we recovered 207 such salmonids from the matrix antenna and trawl system (Appendix Table A2). In previous years, divers have inspected the trawl body and wing areas of the net while underway, and they have reported that fish rarely swim close to the webbing. Rather, fish tend to linger near the entrance to the trawl body and directly in front of the antenna, likely because the sample gear is more visible in these areas. Through the years, we have modified the net to minimize visible transition areas between the trawl, wings, and other components. Visible transition areas were found mainly at the seams joining net sections of different web size or weight. We now use a uniform color (black) of netting for the trawl body and cod-end areas. These modifications have reduced fish training and expedited passage out of the net. Although volitional passage through the antenna occurred with the wings extended, we continued to "flush" entrained fish out of the net by bringing the trawl wings together. To expedite fish passage, we flushed the net every 17 minutes. We kept the trawl wings together for 5 minutes during each flush, with a 1-minute transition between opening and closing the trawl wings. Flushing also helped to clear debris and may have reduced delay, and possible fatigue of fish pacing transition areas or lingering near the antenna. A majority of detections were recorded during these 7-minute periods. Fish appeared to move more readily through the system at night, probably because the trawl and antenna were less visible during darkness hours. Lower visibility at night also appeared to reduce the tendency of fish to pace near the entrance of the trawl body. A floor extending forward from the trawl body is meant to discourage fish from sounding to escape the trawl. However, fish likely sense the head rope and cork line that crosses between wings at the surface of the trawl body. Since we began using the larger matrix antenna system, detections during periods when the wings are held open have increased by about 10% (Magie et al. 2010).

15

Development of a Flexible Antenna Detection System Background In 2016, we continued development of a towed, flexible antenna system that does not require a net to concentrate and guide fish within range of the antennas. This new design is based on technology adapted for a stationary PIT tag monitoring system installed along a pile dike at rkm 70 (Magie et al. 2015). Our goals in developing this antenna were to reduce costs associated with sampling juvenile salmonids in the estuary and to improve sample efficiency using recent advances in PIT tag technology. Since 2013, we have used a new multiplexing transceiver (Biomark model IS1001MTS) that has allowed us to construct much larger rectangular antennas of 2.4 by 6.1 m. These antennas provide a detection field at least six times larger than that of the 0.8- by 3.0-m antennas used previously in our towed PIT-tag monitoring systems. In addition to increased reading range, the new transceiver allows antennas to be constructed from 1.9-cm-diameter flexible hose instead of the much larger 10.2-cm-diameter rigid PVC used previously. These changes have dramatically increased antenna utility and led to new applications. The primary configuration used for sampling in 2016 was an array of six flexible antennas. This system was deployed using two 6.4-m skiffs, with the first skiff deploying the array and the second supporting deployment. Smaller arrays were tested early in the season and occasionally when antennas were being repaired. In 2015, sampling with the new system focused on developing protocols for deployment, pinpointing electronics issues, increasing antenna read range, and reducing EMI (electromagnetic interference) while under tow (Morris et al. 2015). In 2016, we used the developed protocols and focused on concurrent sampling with the matrix trawl. Our goal was to determine the feasibility of using the flexible antenna system to replace the trawl system in future years. Our objective in 2016 was to compare the sample rate of the flexible system relative to the matrix trawl system. By simultaneously sampling with both systems during the peak juvenile migration period, we were able to compare species composition and sample efficiency during daylight and darkness periods between the two systems. After the migration period, we continued to test various antenna configurations to maximize detection efficiency of the six-antenna flexible system.

16

Methods Initial testing was conducted in March-April 2016, prior to the peak juvenile migration season. During this period, we focused on refining deployment protocols, training personnel to maintain maximum extension of the antennas, and further refine tow speed limits (RPM) related to EMI. During the peak of the juvenile migration (May), we conducted simultaneous tests of the flexible and matrix trawl systems. The matrix trawl system was deployed following its normal schedule, and the flexible antenna system was generally deployed nearby and 1 km upstream from the trawl. After these concurrent tests, and after the juvenile migration season had concluded, we conducted tests to evaluate the feasibility of different antenna orientations for the flexible antenna system. Antennas were oriented vertically in the water column rather than horizontally to increase sample depth. We also tested flexible antenna arrays with up to 11 antennas. Prior to testing in 2016, we redesigned the watertight capsule that housed the IS1001 reader and communications cables connecting different capsules. The new capsule was lighter, easier to deploy, and more streamlined than the previous design (Figure 8). The transceiver master controller and power source (two 12-volt batteries) were located in the deployment skiff, where EMI and detection data were monitored in real-time. Figure 8. Image of the electronic reader capsule used in 2014 compared to the new

capsule design used in 2015-2016. Photos are not to scale.

2014

2015-2016

17

Antenna Wire

Foam Backer Rod

Cellophane

We also redesigned the wire gallery within the antennas to reduce internal vibration and EMI (Figure 9). We installed foam backer rods to separate and support the three antenna wires. The foam rods also increased antenna current by reducing resistance from neighboring wires.

Figure 9. Cross-sectional area of the flexible antenna wire gallery. Antenna wires are placed around three glued foam backer rods, and then wrapped in cellophane to hold wires in position.

The flexible antenna system was designed to be modular, wherein antennas could be easily added to or removed from an array (Figure 10). For each flexible antenna added to the array, a controller area network (CAN bus) power and communications cable was routed along the towline. Cables extended from the deployment skiff to the watertight reader capsule on the first antenna. Additional short CAN bus cables were then connected consecutively between additional capsules until all antennas were connected. Figure 10. Basic configuration of the six-antenna array and modular rope frame system

with new reader (electronics) capsules tested in 2016. This array had a total reading range approximately 37 m wide by 3 m deep and was made up of six 2.4- by 6.1-m flexible antennas.

Spreader Bar

Weight 13.6 kg

Float (F-3)Towline68.5 m

RopeFrame

Antennas

2.4 m

Flow

Electronics Cable

Electronics Capsules

6.1 m

36.6 m

18

During testing in 2015, we discovered the need for additional distance between the CAN bus cable and antenna wires due to interference between the two. To address this need, we used CAN bus cables that were 1 m longer than the distance between antennas and attached them only at points near the reader capsules (Figure 10). This configuration allowed excess cable to drift behind and away from the antenna wires while under tow, and effectively eliminated interference. We first deployed the flexible system directly into the river in late 2015. For this initial deployment, we used one skiff to transport equipment to the sample area and deploy the system. A second skiff assisted deployment by towing the deployment vessel against the current, in an upstream direction, in order to maintain its position in the river. Batteries and electronics were housed in the primary transport skiff. Once the antenna array was fully extended, both skiffs motored upstream into the current while holding the array open. A radar range finder was used to maintain full extension of the array. Tow vessels were two 6.7-m-long skiffs powered by 135-hp outboard motors. Occasionally, a third skiff was used to monitor antenna configuration and to test system performance at different tow speeds. These tests were conducted using a "stick fish" comprised of a PIT tag attached to a pole. Similar to the matrix trawl system, the flexible antenna system moved downstream within the sample reach during deployment. While it was difficult to judge tow speed through the water against strong river currents, we estimated that approximately 1.7 knots was the maximum tow speed that would avoid excessive vibration-related EMI, which compromised read efficiency. This estimate represented the difference between a dead drift speed of 1.9 kn measured by the third skiff and a downstream movement rate of 0.2 kn based on GPS data for the main tow skiff. Skiffs were essentially idle (dead drift speed) at 900 rpm. By comparison, the matrix trawl system tows at about 1.2 kn with the net fully deployed. An array of six flexible antennas was chosen to evaluate performance in comparison with the six-antenna trawl system during simultaneous testing. The detection field generated by the flexible array was approximately 37 m wide by 3 m deep (Figure 10). Under tow, there was a slight curve to the array that varied with the distance between tow boats. The distance between boats is critical to maximize width of the detection field without over-straining the system. Measured depth of the flexible array was about 2.4 m, but the detection field extended about 0.3 m above and below the antennas. Thus, results from the six-antenna array theoretically represent one-half of the potential detection field, as the IS1001 transceiver system is capable of powering up to 12 antennas.

19

Rectangular antennas within the array were towed in an end-to-end, horizontal orientation at a pace that matched the trawl (1.2 kn). To provide additional stability and strength, all antennas were attached to a 13-mm-diameter non-stretch rope frame. The rope-frame extended 3 m beyond the antennas and attached to two 2.4-m aluminum spreader bars. Each bar had 13.6 kg of weight on one end and a buoy on the other to maintain the array in a vertical orientation in the water column and to keep it submerged about 0.5 m below the surface (Figure 10). Previous evaluations have shown a drop in detections when the top 1 m of water was not sampled, so these weights and buoys were used to maintain the system near the surface. The bars on each end of the array were bridled to tow lines and adjusted to maintain proper vertical orientation, width, and depth of the array under tow. During simultaneous testing, the matrix trawl system sample design, operation, and equipment were not modified in any way to accommodate comparative sampling. The flexible system was deployed based on the location of the trawl system. Simultaneous deployment did not affect the availability of fish for potential detection in the trawl system because the distance between systems was 1 km, and based on detections in the trawl system, we observed no migration delay or avoidance behavior of juvenile migrants after passing through the flexible system. Performance of the trawl system was meant to act as a reference from which to measure performance of the new system, so every effort was made to ensure standard operations were conducted. To evaluate performance differences between the matrix trawl and flexible antenna system, we compared average daily detection ratios, species composition ratios, and diel detection patterns. Data for comparisons were taken only from days and times when both systems were deployed and operational. Detection data obtained during tests of electronic settings, system equipment, and antenna configuration were omitted from any comparative analyses because these tests often altered the performance of the flexible system.

20

Results and Discussion Detection Rates In 2016, we detected 549 fish in 78.2 h of sampling across 21 d using the towed flexible antenna system (Table 2). This included 77 Chinook salmon (14%), 20 coho salmon (4%), 433 steelhead (79%), 3 sockeye salmon (less than 1%), and 16 fish without species data available in PTAGIS (3%). Table 2. Deployment dates of the flexible antenna system with total daily sample time,

total detections, and each days sampling objective conducted near rkm 75 in the upper Columbia River estuary, 2016.

Date Sampling objective Sample time (h) Detections (n) 19 April Testing 2.7 1 21 April Testing 3.2 6 25 April Testing 3 10 28 April Testing 3.5 12 3 May Simultaneous sampling 6.3 30 5 May Testing 4.8 29 10 May Simultaneous sampling 4 53 11 May Simultaneous sampling 5.4 145 12 May Simultaneous sampling 6.9 118 17 May* Simultaneous sampling 4 37 18 May* Simultaneous sampling 3.9 15 19 May* Simultaneous sampling 4.8 31 24 May Simultaneous sampling 3.3 23 25 May Simultaneous sampling 3.8 23 30 May Testing 4.4 0 31 May Testing 1.9 0 2 June Simultaneous sampling 4.1 15 9 June Testing 3 1 14 June Testing 2.2 0 29 June Testing 3.1 0 5 July Testing 2.6 0 Totals 78.2 549 * Deployment during darkness hours

21

Performance of the Flexible Antenna vs. Matrix Trawl System

In 2016, objectives of sampling with the flexible antenna system were to compare detection efficiency, species composition, and diel detection patterns to those of the matrix trawl system. These comparisons were used to determine the feasibility of replacing the matrix trawl with the flexible antenna in future years. We sampled both systems simultaneously on 21 dates from 19 April through 5 July; however, on 11 of these dates, performance of the flexible system was interrupted by antenna testing (Table 2). Data from deployments interrupted by testing were omitted from comparative analyses. There were 10 dates (43.1 h) when the towed flexible antenna system was fully operational and sampling was comparable to the trawl system (Table 3). Most sampling of the flexible system occurred during daylight hours (31.0 h), except during 17-19 May, when sampling was conducted during darkness hours (12.1 h). Over the 10 days of concurrent sampling, the average daily detection ratio between the flexible antenna and matrix trawl systems was 60%, with flex/matrix ratios ranging from 10 to 209%. Overall, the matrix trawl system detected more Chinook, coho, and sockeye salmon, while the flexible antenna system detected more steelhead. Species composition (proportion of total) in total detections from both systems was referenced to the species composition of fish detected at Bonneville Dam. When we compared species composition of total detections from the matrix trawl vs. the flexible antenna system, we found that the flexible antenna detected a disproportionally high number of steelhead (Table 4) This result was surprising, especially because for these tests, the flexible antenna system was configured for half its potential reading range (with 6 vs. 12 antennas). These results also suggest that the flexible system was considerably more efficient at detecting steelhead than the matrix trawl system. Disproportionately high numbers of steelhead passing through the flexible antenna system may have skewed proportions low for other detected species. However, even considering this potential bias, the flexible antenna clearly failed to detect Chinook, coho, and sockeye salmon at even half the rates detected by the trawl system.

22

Table 3. Daily detections (n) by species of the flexible and matrix trawl antenna systems during simultaneous sampling in the Columbia River estuary, 2016. Also shown are daily ratios between detection totals and mean ratio (*) for the season. Sampling was conducted during daylight except for 12.1 h of night sampling during 17-19 May.

Date (2016)

Flexible Antenna System Matrix Antenna Trawl System Flex/ matrix

ratio (%) Chinook Coho Steelhead Sockeye Unknown Total Chinook Coho Steelhead Sockeye Unknown Total 3 May 4 0 22 0 0 26 18 2 83 1 0 104 25 10 May 1 0 44 0 5 50 18 5 22 0 1 46 109 11 May 4 3 106 0 2 115 21 2 28 3 1 55 209 12 May 19 2 91 0 3 115 117 15 46 4 10 192 60 17 May 14 2 20 0 0 36 49 7 34 1 1 92 39 18 May 7 1 5 0 1 14 112 11 15 1 1 140 10 19 May 8 6 12 1 1 28 104 36 40 2 6 188 15 24 May 1 1 18 0 1 21 10 4 24 2 2 42 50 25 May 2 1 16 1 0 20 7 9 28 5 2 51 39 2 June 2 3 9 1 0 15 19 1 12 3 2 37 41 Total 62 19 343 3 13 440 475 92 332 22 26 947 *60

23

Table 4. Total detections and species composition for the flexible antenna vs. matrix trawl detection systems. Species composition of detections at Bonneville Dam during our intensive sample period are included as a reference for species composition of tagged fish present in the estuary during sampling.

Species

Flexible antenna system

Matrix trawl system

Detection at Bonneville Dam

(n) (%) (n) (%) (n) (%) Chinook salmon 62 14.1 475 50.2 35,747 43.2 Steelhead 343 78.0 332 35.1 36,448 44.1 Coho salmon 19 4.3 92 9.7 7,054 8.5 Sockeye salmon 3 0.7 22 2.3 1,103 1.3 Unknown 13 3.0 26 2.8 2,322 2.8 Total 440 100 947 100 82,674 100 Steelhead made up 78% of total detections on the flexible antenna system but only 35% of total detections on the matrix system. By comparison, steelhead made up 44% of total juvenile salmonid detections at Bonneville Dam. Thus the proportion of steelhead detected on the matrix trawl system was nine percentage points lower than that of the reference composition. In terms of total detections as well, the flexible antenna detected more steelhead than the matrix trawl system. Conversely, the flexible antenna system detected only 62 Chinook (14%) and 3 sockeye salmon (1%), while the matrix trawl detected 475 Chinook salmon (50%) and 22 sockeye salmon (2%). Both of these species are included in annual hydrosystem survival estimates (Faulkner et al. 2016). Diel detection patterns were also compared between the flexible antenna and matrix trawl systems during concurrent sampling efforts (Figure 11). Both systems exhibited similar trends in average hourly detection rates, with detection rates of yearling Chinook peaking during crepuscular periods, and detection rates of steelhead peaking near midday. For yearling Chinook salmon, average detections of fish per hour were higher in the matrix trawl than in the flexible antenna system for every hour sampled. However, during dusk, darkness, and dawn periods, detections in the trawl were dramatically higher than those in the flexible antenna system. Similar results were observed for steelhead except during mid-to-late morning hours (0800-1100), when the flexible system detected more fish per hour than the matrix system.

24

Figure 11. Average hourly detection rates of yearling Chinook salmon and steelhead

during concurrent sampling periods between the flexible and matrix trawl antenna systems near river kilometer 75 in the upper Columbia River estuary, 2016.

These observations may have been related in part to greater avoidance of the trawl by fish during daylight periods, when the net is more visible. They may also be related to the relatively shallow sample depth of the flexible antenna. Chinook salmon tend to occupy a lower position in the water column during migration (Beeman and Maule, 2006). Conversely, the flexible antenna may have detected more steelhead because of their higher position in the water column.

DarknessDarkness DaylightD

etec

tions

(n h

-1)

Yearling Chinook Salmon

Darkness Daylight Darkness

0 2 4 6 8 10 12 14 16 18 20 22Detection Hour

Det

ectio

ns (n

h-1

)

Steelhead

20

4

12

16

8

0

20

10

0

50

30

40

60

Flexible System, n = 62 Matrix Trawl System, n = 475

Flexible System, n = 343 Matrix Trawl System, n = 332

25

We concluded that the 3.0-m sample depth of the flexible system vs. the 4.9-m depth of the trawl system was likely the largest contributor to detection differences between systems. Observations in previous years indicate that steelhead resist passing through the matrix antenna more than any other species. Results from simultaneous sampling, which showed higher detection rates for steelhead in the flexible antenna than in the trawl system, supported the idea that steelhead avoidance behavior was less prevalent near the flexible antenna system. Therefore, we concluded that avoidance behavior was not likely responsible for the lower detection rates of Chinook, coho, and sockeye salmon observed in the flexible antenna system. It is more likely that those species were traveling below the array rather than avoiding it, and that the relatively shallow sample depth of the flexible antenna system was the primary reason for its lower detection totals. Post Season Testing After the 2016 sample period, we reviewed these results and tested new orientations and configurations to increase sample depth of the flexible system. Our objective was to attain a sample depth equal to or greater than that of the matrix trawl system. We also tested a newly designed and manufactured antenna cable during this period. Stacked horizontal array—We conducted preliminary tests of two new configurations to try to maximize sample depth of the flexible antenna system. The first was a three-by-two horizontal stacked array rather than the six-by-one horizontal configuration used during the 2016 sampling season. This configuration matched the sample depth of the trawl (4.8 m) but was only 18.3 m wide (Figure 12). Furthermore, the stacked configuration resulted in high EMI, which degraded electronic performance of antennas and resulted in reduced antenna current with minimal read ranges. Deployment and retrieval of the stacked array was also difficult to accomplish without tangling the CAN bus cable around the antennas and reader capsules, which resulted in additional interference while sampling. A significant benefit of this configuration would be the potential ability to differentiate the sample depth of individual detected fish. Further testing of this configuration will be done in early 2017 using rope spacers between antennas and a new CAN bus configuration to minimize tangling.

26

Figure 12. Configuration of the three-by-two, six-antenna stacked array and modular

rope frame system with new electronics capsules tested in 2016. This array measured a total distance of 18.3 m and consisted of six 2.4- by 6.1-m flexible antennas stacked in two rows of three antennas to sample to a depth of approximately 5 m.

Vertical array—The second test configuration also used six rectangular antennas, but each was rotated 90 degrees from the traditional horizontal position to a vertical orientation. This system sampled to a depth of 6.1 m (deeper than the trawl sample) but was only 14.4 m wide (Figure 13). This configuration did not compromise electronic performance of the antennas. Unfortunately, most juvenile PIT-tagged fish had already passed the sample reach by the time this configuration was tested, so detections were minimal. Testing of a 12-antenna array (sample width about 28.8 m) with vertically oriented antennas is planned for early 2017.

Spreader Bar

Towline68.5 m

Weight 13.6 kg

Float (F-3)Rope

Frame

Antennas

2.4 m

Flow

Electronics Cable

Electronics Capsules

6.1 m

18.3 m

27

Figure 13. Configuration of the vertically oriented six-antenna array and modular rope

frame system with new electronics capsules tested in 2016. This array measured a total distance of 14.4 m and consisted of six 2.4- by 6.1-m flexible antennas to sample to a depth of approximately 6.5 m.

Manufactured antenna cable—During post-season evaluations, we also tested a new manufactured antenna cable to replace our hand-assembled antennas. This cable was constructed in conjunction with a Sacramento Delta PIT feasibility study (Rundio et al. 2016). The objective of these tests was to determine whether we could use manufactured cables to 1) reduce the time and labor cost associated with assembling the flexible antennas by hand, and 2) increase the overall durability and electronic performance of system antennas. Early testing showed that the manufactured cable performed well both in dry-air tests and when deployed in the river. Overall, the manufactured cable antennas supported higher amperage (a stronger detection field) than the traditional flexible antennas with no difference in EMI. Further testing will be required to determine the feasibility of fully replacing our hand-assembled flexible-hose antennas with a manufactured antenna cable in future years. Rundio et al. (2016) includes a complete description of comparative tests between antenna types.

Spreader Bar

Towline68.5 m

Weight 13.6 kg

Float (F-3)

RopeFrame

Antennas

2.4 m

Flow

Electronics Cable

Electronics Capsules

6.1 m

14.4 m

28

Objectives for Future Development Sampling with the flexible antenna system in 2016 achieved our objective of determining the sample rate of the flexible antenna system relative to the matrix trawl system. We obtained sufficient data to compare species composition between the two systems and to assess differences in sample efficiency between daylight and darkness periods. These results highlighted the need to test alternate antenna configurations to optimize performance and improve detection rates for Chinook, coho, and sockeye salmon. We successfully conducted these tests immediately following the 2016 migration/sampling season. The one-by-six, horizontal orientation used for the rectangular antennas in most testing of the flexible array in 2016 detected numerous fish, particularly juvenile steelhead. This success represents a significant advancement for mobile PIT-tag detection systems. However, further modification will be needed to achieve the detection rates demonstrated by the matrix trawl system, which has routinely detected 2-3% of juveniles previously detected at Bonneville Dam. While further development is needed for the flexible antenna system, there are considerable advantages of moving away from use of the matrix trawl system, which requires fish to be guided and concentrated for detection. First, the transceiver for the new IS1001 sampling equipment is easier to maintain than the older FS1001M transceiver, which is no longer manufactured (Biomark Inc.). Second, the flexible antenna array is modular so that antennas and other system components can be easily swapped in to replace broken or malfunctioning parts, while the large trawl net is expensive and requires extensive maintenance. Third, a smaller vessel and support crew are sufficient to operate the flexible antenna system, while the matrix trawl system requires a larger, more expensive vessel and larger crew. Finally, the flexible system is highly versatile and can be used in shallower water and narrower areas, while the trawl system is limited to deployments in deep water with ample room for maneuvering. In past years, we attempted to monitor PIT-tagged fish in shallow-water areas using a downscaled trawl system, but the smaller trawl could not produce adequate detection numbers (Ledgerwood et al. 2007). The flexible system has potential to sample new habitats, and the technology is transferrable to small streams, reservoirs, and large open bodies of water. These are important objectives for future development of the flexible antenna system.

29

Analyses from Trawl Detection Data Detection data from the trawl are essential for calculating survival probabilities to the tailrace of Bonneville Dam, the last dam encountered by seaward juvenile migrants (Muir et al. 2001; Williams et al. 2001; Zabel et al. 2002). Operation of the trawl detection system in the estuary has provided data to calculate detection probabilities at Bonneville Dam since 1998. These detection data are necessary for unbiased estimates of survival because they provide the means to distinguish the probability of detection vs. the probability of survival. Unbiased estimates of survival in turn provide information that is critical to research and management programs for endangered salmonids in the Snake and Columbia River basins and in other basins of the Pacific Northwest (Faulkner et al. 2016). Trawl detections also allow comparison of relative detection percentages, travel speed, and other parameters between inriver migrant and transported fish groups after they comingle in the estuary. Annual releases of PIT-tagged fish in the Columbia River basin have approached or exceeded 2 million for the past several years. The ability to monitor these fish as they pass the estuary has increased our understanding of behavior and survival during the critical freshwater-to-saltwater transition period. In this section, we explore differences in travel time and detection rate by species and migration history, patterns in diel detection by origin and species, and detection rates by migration history using data collected by the matrix trawl detection system in 2016.

30

Travel Time and Migration Rate Methods We coordinated trawl system sample cruises with the expected estuary passage timing of yearling fish tagged and released for transportation and survival studies. During our sample period in 2016, fish were transported from Lower Granite, Little Goose, and Lower Monumental Dam on the Snake River. Our analysis included all transported fish detected in the trawl, regardless of the location from which they were transported. Groups of transported fish were compared to groups of fish detected at Bonneville Dam on the same day (24-h period) the transported group was released. At the transport dams, PIT-tagged fish were diverted to transport barges using separation-by-code (SbyC) systems, which automatically uploads detections of fish en route to transport raceways (Marvin and Nighbor 2009). We considered a fish to have been transported only if the last detection of that fish recorded in PTAGIS was at a dam on a route to a transport raceway. We created an independent database (Microsoft Access) using data downloaded from PTAGIS for this analysis. The U.S. Army Corps of Engineers provided individual barge-loading dates and times for each dam throughout the 2016 transportation season (John Bailey, USACE, personal communication). By comparing barge-loading times with the last detection time of fish diverted to transport raceways, we determined the individual barge-transport trip for each fish. With this information, we were able to derive the specific date, time, and release location of each individual transported fish. We then created paired comparison groups of fish either released from transported barges or detected at Bonneville Dam on the same date. For yearling Chinook salmon and steelhead, we plotted seasonal distributions of travel time for fish detected at Bonneville Dam vs. those transported and released just downstream from the dam. These distributions were plotted using median travel time by group. Travel time (in days) to the estuary was calculated for each fish on each date by subtracting time of barge release or detection at Bonneville Dam from time of detection in the trawl. A two-sample t-test was used to evaluate differences in the average daily median travel speed to Jones Beach between inriver migrants detected at Bonneville Dam and transported fish release just below the dam. Daily travel speeds (km/d) were calculated by dividing travel time by distance traveled from barge release or Bonneville detection to detection in the estuary. Daily median travel speeds were plotted against flow data, based on daily average discharge rates at Bonneville Dam (m3/s).

31

Results and Discussion Travel time for yearling Chinook and steelhead—For inriver migrants, median travel time to the estuary from Bonneville Dam was slightly longer than the 15-year average (1.8 d) for yearling Chinook salmon, and equal to the 15-year average (1.7 d) for steelhead (Table 5). However, travel times were generally shorter in 2016 than in 2015. For yearling Chinook salmon, median travel time was 1.9 d, compared to 2.1 d in 2015. For steelhead, median travel time was one-half day shorter in 2016 (1.7 d) than in 2015 (2.2 d). Similarly, for transported yearling Chinook, median travel time from just below Bonneville Dam to the estuary was shorter in 2016 than in 2015 (2.1 vs. 2.5 d), but equal to the 15-year average (2.1 d). For transported steelhead, median travel time was also shorter in 2016 than in 2015 (1.6 vs. 2.3 d) and consistent with the 15-year average (1.7 d). Table 5. Median time to detection on the trawl system during intensive sample periods

for yearling Chinook salmon and steelhead detected at Bonneville Dam or released from barges just downstream from Bonneville Dam, 2000-2016. Also shown are mean flow rates at Bonneville Dam from mid-April through June.

Year

Detected at Bonneville Dam (rkm 234)

Transported and released below Bonneville Dam (rkm 225)

Flow (m3 s-1)

Yearling Chinook Steelhead Yearling Chinook Steelhead

Travel time (d) (n)

Travel time (d) (n)

Travel time (d) (n)

Travel time (d) (n)

2000 1.7 479 1.7 296 1.9 495 1.6 301 7,415 2001 2.3 792 2.5 59 2.9 1,329 2.3 244 3,877 2002 1.8 1,137 1.7 156 2 1,958 1.6 296 8,071 2003 1.8 1,721 1.7 567 2.1 2,382 1.7 435 7,120 2004 1.9 672 2 110 2.2 2,997 1.9 333 6,663 2005 1.8 81 2 471 2.2 2,910 1.9 400 5,776 2006 1.7 888 1.6 131 2.1 1,315 1.6 170 9,435 2007 1.7 1,510 1.7 362 2.2 1,096 1.7 143 6,858 2008 1.7 749 1.6 830 2.1 1,884 1.6 788 8,714 2009 1.7 1,438 1.7 892 2.1 1,681 1.6 1,325 7,871 2010 2.0 3,258 1.9 2,188 2.2 1,149 2.0 1,068 6,829 2011a 1.8 240 1.6 216 2.1 673 1.6 831 7,911 2011b 1.5 39 1.3 47 1.6 418 1.5 275 13,462 2012 1.6 485 1.5 321 2.0 567 1.5 1,116 10,056 2013 1.6 645 1.6 745 2.2 1,029 1.6 1,333 7,470 2014 1.6 431 1.6 412 2.1 1,012 1.5 1,206 8,281 2015 2.1 1,065 2.2 1,885 2.5 714 2.3 611 5,333 2016 1.9 670 1.7 1,067 2.1 674 1.6 844 6,953 a Early migration period prior to the increase in river flow about 16 May. b Late migration period during the high flow event beginning about 16 May.

32

Trav

el S

peed

(km

/d)

Trav

el S

peed

(km

/d)

Release Date from Barge or Bonneville Dam Detection Date

Steelhead, 2016

Yearling Chinook Salmon, 2016

8,000

4,000

2,000

Flow

(m3 s

-1)

6,000

10,000

8,000

4,000

2,000

Flow

(m3 s

-1)

6,000

10,000

25 Apr 05 May 15 May 25 May 04 Jun

n = 674, mean = 73

n = 844, mean = 86

n = 670, mean = 82

n = 1,067, mean = 92

150

120

90

30

0

60

120

90

60

30

0

150

In-river Barged Flow

Migration rate for yearling Chinook and steelhead—We also compared daily differences in migration rate to the estuary between transported fish and inriver-migrants detected at Bonneville Dam relative to river flow volumes (Figure 14). For yearling Chinook salmon, average daily median travel speed to the estuary was significantly slower for transported fish than for inriver migrants detected at Bonneville Dam (73 vs. 82 km/d; P < 0.001). Likewise, for steelhead, average daily median travel speed was significantly slower for transported steelhead than for those detected at Bonneville Dam on the same day (86 vs. 92 km/d; P = 0.016) . These differences in travel speed by migration history were similar to observations from previous years. Figure 14. Daily median travel speed (km/d) to the estuary of yearling Chinook salmon

(top) and steelhead (bottom) following detection at Bonneville Dam or release from a barge to detection in the upper Columbia River estuary, 2016. Seasonal means are shown for comparison with flow.

33

Subyearling fall Chinook and Sockeye salmon—Of the 127 inriver-migrant subyearlings detected, only 11 had been previously detected at Bonneville Dam. For fall subyearling Chinook, mean migration rate from Bonneville Dam to the estuary trawl was slower for transported fish than for fish detected at Bonneville Dam (58 vs. 72 km/d), but there was not enough data for meaningful comparison. Analysis in prior years has consistently shown faster migration rates for subyearling fall Chinook detected at Bonneville Dam than for those transported and released below the dam (Morris et al. 2015). Of the 185 inriver migrant sockeye salmon detected, 41 had been previously detected at Bonneville Dam. Mean migration rate from Bonneville Dam to detection in the trawl was slower for transported sockeye (78 km/d) than for inriver migrant sockeye (95 km/d), but there was not enough data for a meaningful comparison.

34

Diel Detection Patterns Methods For analysis of diel detection rates, we compared detection numbers between darkness and daylight hours using a one-sample t-test (Zar 1999) of the daily ratios of darkness vs. daylight detections per hour. For this test, we used the natural log of detection ratios to improve normality, and estimated means were back-transformed. Fish included for this analysis were only those detected in the trawl system during the intensive sample period (1 May-9 June). We separated the number of detections and number of minutes that the system was operated into darkness and daylight hour categories for each date that fell within the this period. Daily darkness/daylight detections for each species were weighted by the number of minutes the detection system was operating on that date. We excluded dates when sample effort was reduced by missed or partially missed shifts. Detection numbers for this analysis were sufficient for yearling Chinook salmon and steelhead but not for sockeye salmon and subyearling Chinook salmon. Results and Discussion During the intensive sample period, we detected 5,079 yearling Chinook salmon and 4,422 steelhead, with the matrix trawl system operating an average of 15 h/d (Appendix Table A3). We generally stopped sampling each day between 1400 and 1900 PDT for crew changes and refueling. For hatchery yearling Chinook salmon, hourly detection rates were significantly higher during darkness than during daylight hours, at 10.8 vs. 4.0 fish/h (darkness/daylight ratio 2.7; P = 0.001). We assumed that the diel difference in hourly detection rates was constant through the season. However, for hatchery yearling Chinook during the first three weeks of intensive sampling, average detection rates were 2 to 20 times greater during darkness than during daylight hours. For the remainder of the season, average detection rates of hatchery yearling Chinook declined substantially during hours of darkness, and the ratio of darkness-to-daylight detection rates dropped below 1.0, with more detections occurring during the day than at night (Figure 15). For wild Chinook salmon, average detection rates were also significantly higher during darkness than during daylight hours, although the difference was not as large (1.5 vs. 0.9 fish/h; P = 0.011).

35

Night vs Day Ratio Catch per Hour

Yearling ChinookHatchery Wild

Steelhead

1 May 16 May 5 Jun

00.5

1.5

2.5

3.5

0

1.0

2.0

3.0

0

4

8

12

16

0

4

8

12

Nig

ht v

s Day

Rat

io

1May 16 May 5 Jun

For hatchery steelhead, hourly detection rates were significantly higher during daylight than darkness hours (8.2 vs. 3.2 fish/h, darkness/daylight ratio 2.6; P < 0.001). Similarly, for wild steelhead, there was a significant difference between daylight and darkness hours (2.3 vs 0.9 fish/h; P < 0.001). Differences in diel detection rates of steelhead remained constant throughout the intensive sampling period. In each year since 2003, hourly detection distributions have been similar between rear-types for both yearling Chinook salmon and steelhead. These numbers were similar again in 2016, so we pooled data by species and origin for a multi-year summary (Figure 16). Detection rates for yearling Chinook salmon have often been significantly higher during darkness than daytime hours. Detection rates of steelhead have generally been higher during daylight hours, with daylight detection rates often significantly higher in recent years. Figure 15. Daily ratios of darkness/daylight detection rates for combined wild and

hatchery yearling Chinook salmon and steelhead during intensive sampling (1 May-9 June). Ratios greater than 1.0 indicate a higher detection rates in darkness hours, and values less than 1.0 indicate a higher detection rate in daylight hours. Solid lines are estimated mean ratios, and dotted lines are estimated 95% confidence intervals.

36

DarknessDarkness Daylight

Det

ectio

ns (n

h-1

)

2003-2015, n = 120,344 2016, n = 5,151

2003-2015, n = 74,062 2016, n = 4,832

Darkness Daylight Darkness

0 2 4 6 8 10 12 14 16 18 20 22

Detection Hour

Det

ectio

ns (n

h-1

)

Steelhead

20

15

10

50

3530

25

25

5

15

35

4540

30

20

10

0

Yearling Chinook salmon Figure 16. Average hourly detection rates of yearling Chinook salmon and steelhead

during the two-shift sampling periods of 2003 through 2015, vs. 2016, using the matrix antenna system in the upper Columbia River estuary near river kilometer 75.

37

ii

ii

Xday

Xday

i eep

βββ

βββ

++

++

+=

10

10

1

Detection Rates of Transported vs. Inriver Migrant Fish Methods We compared daily detection rates in the trawl between transported fish and inriver migrants previously detected at Bonneville Dam. Fish included for this analysis were only those detected in the matrix trawl system during the two-shift sample period. These data were evaluated to assess whether differences in detection rate were related to migration history or arrival timing in the estuary. Detection rates of transported salmonids were compared to those of inriver migrants detected at Bonneville Dam using logistic regression (Hosmer and Lemeshow 2000; Ryan et al. 2003). Daily groups of inriver migrants detected at Bonneville Dam were compared with daily groups of fish released from a barge on the same day. For this comparison, we included only yearling fish released at, or upstream from McNary Dam. Fish released from a barge just after midnight were compared with fish detected the previous day at Bonneville Dam. Components of the logistic regression model were migration history (inriver or transport) as a "treatment" factor, with date and date-squared as covariates. The model estimated the log odds of detection for i daily cohorts (i.e., ln[pi/(1-pi)]) as a linear function of model components, assuming a binomial error distribution. Daily detection rates were estimated as: where β̂ was the coefficient of the components (i.e., 0β̂ for the intercept, 1β̂ for day i, andβ̂ for the set “Xi” of day-squared and/or interaction terms). The following stepwise procedure was used to select the appropriate model. First, we fit the model containing interactions between treatment and date and date-squared. We then determined the amount of overdispersion relative to that assumed from a binomial distribution (Ramsey and Schafer 1997). Overdispersion was estimated as “σ,” the square root of the model deviance statistic divided by the degrees of freedom. Overdispersion was the “difference” between expected and observed model variances, after accounting for treatment, date, and date-squared. If σ was greater than 1.0, we adjusted the standard errors and z-test of model coefficients by multiplying by σ (Ramsey and Schafer 1997). Finally, if interaction terms were not significant (likelihood ratio test P >0.05), these terms were removed, and we fit a reduced model.

38

The model was further reduced depending on the significance(s) between treatment and date and/or date-squared. The final model was the most reduced from this process. One constraint was that date-squared could not be included in the model unless date was included as well. Various diagnostic plots were examined to assess the appropriateness of the models. Extreme or highly influential data points were identified and included or excluded on an individual basis. Daily transported and inriver groups had similar diel distributions in the sampling area and presumably passed the sample area at similar times (Magie et al. 2011). Thus, we assumed these groups were subject to the same sampling biases (sample effort). If these assumptions were correct, then differences in relative detection rate between groups would reflect differences in survival from Bonneville Dam to the trawl. Results and Discussion A total of 47,360 yearling Chinook salmon and 37,098 steelhead were transported and released below Bonneville Dam during our intensive sample period. These included fish diverted to barges for NMFS transport studies and fish tagged and transported for other studies. Of these transported fish, we detected 727 yearling Chinook salmon and 919 steelhead in the upper estuary (Appendix Tables A4-A5). A total of 33,114 yearling Chinook salmon and 30,319 steelhead were released upstream from McNary Dam and detected at Bonneville Dam. We detected a total of 663 yearling Chinook and 1,011 steelhead released upstream from McNary and detected at Bonneville Dam (2.0 and 3.3%, respectively; Appendix Table A6). As in previous years, a portion of tagged fish from both the transport and inriver migrant groups passed through the estuary either before or after the trawl sampling period. We estimated the proportions of fish from these groups that were available in the estuary during our intensive sample period (1 May-9 June 2016), allowing 2 d for fish to reach the sample area from Bonneville Dam. For yearling Chinook, we estimated that 75% of inriver migrants and 99% of transported fish passed through the sample reach during our intensive sample period. For steelhead, we estimated that 80% of inriver migrants and 99% of transported fish passed through our sample reach during intensive sampling. These percentages were slightly higher than those estimated in 2015 for both migration-history groups.

39

During the intensive sampling period of 2016 average sample effort was 15 h/d, which was equal to the average effort in 2015. For both transported fish and those detected passing Bonneville Dam, rates of detection were lower in 2016 than in 2015 (Table 6). Table 6. Trawl detection rates of PIT-tagged fish released from barges or detected

passing Bonneville Dam during the intensive sample periods in the upper Columbia River estuary near rkm 75, 2015 and 2016.

Year/species

Barged fish originating upstream In-river fish detected from McNary Dam a at Bonneville Dam*

Released Detected % Released Detected % 2015

Chinook salmon 22,499 768 3.41 32,363 1,065 3.29 Steelhead 20,540 794 3.87 39,513 1,886 4.77 2016 Chinook salmon 47,360 727 1.54 33,114 663 2.00 Steelhead 37,098 919 2.48 30,319 1,011 3.33 * Inriver fish included only those released at or upstream from McNary Dam, although no fish were

transported from McNary Dam in 2016. For yearling Chinook salmon, logistic regression analysis showed a significant relationship between detection rate and date of arrival in the sample reach (P < 0.001) and between detection rate and migration history (P = 0.013). There was no significant relationship between detection rate and date-squared (P = 0.184), and no significant interaction between migration history and either date or date-squared (P = 0.957 and 0.197, respectively). Estimated detection rates for inriver migrant yearling Chinook increased steadily from 1.4% in early May to 4.8% by mid-June. Estimated detection rates for transported yearling Chinook also steadily increased from 1.1% in early May to 3.7% by mid-June (Figure 17, top panel). The adjustment for overdispersion was 3.83. For steelhead, logistic regression analysis indicated a significant relationship between detection rate and both date of arrival and date-squared (P < 0.000 and 0.021, respectively). The relationship between detection rate and migration history was also significant for steelhead (P = 0.029). We found no significant interaction between migration history and date or between migration history and date-squared (P = 0.429 and 0.817, respectively).

40

Throughout the season, estimated detection rates for inriver migrant steelhead fluctuated from 1.5% in early May to 5.2% near the end of May. Detection rates for inriver migrant steelhead decreased to 4.0% by mid-June (Figure 17, lower panel). Transported steelhead exhibited a similar fluctuation in estimated detection rates throughout the season, with detection rates beginning at 0.9% in early May, increasing to 3.3% at the end of May, and then declining to 2.5% by mid-June. The adjustment for overdispersion was 9.22. Figure 17. Estimated daily detection rates in the matrix trawl system for transported and

inriver migrant yearling Chinook salmon and steelhead. All fish were either detected at Bonneville Dam or released from a barge just downstream from the dam on the same date.

Tota

l Rel

ease

d (n

)

Det

ectio

n R

ate

(%)

Yearling Chinook Salmon, 2016n = 1,390

Tota

l Rel

ease

d (n

)

Det

ectio

n R

ate

(%)

28 Apr 8 May 18 May 28 May 7 Jun

Steelhead, 2016 n = 1,930

Barge Release In-river Release Barge %In-river % Barged Regr In-river Regr

0

1

2

3

4

5

6

7

0

2

4

7

1

3

5

6

0

3,000

4,000

2,000

5,000

6,000

7,000

1,000

0

2,000

2,500

1,500

3,000

3,500

500

1,000

41

Estimated daily detection rates in the trawl for both yearling Chinook salmon and steelhead were higher for inriver migrants than for transported fish throughout the season. In years where differences are present, it is possible that lower detection rates of one group represent higher mortality in that group between Bonneville Dam and the estuary. Over the last 12 years, there has been a general trend towards higher detection rates of inriver migrant Chinook salmon, but no apparent trend for steelhead (Morris et al. 2014). In summary, estuary detection rates were considerably lower in 2016 than in 2015 (an unusually low flow year), but similar to rates seen in 2013 and 2014. Detection rates of fish at Bonneville Dam were also lower in 2016 than in 2015 but were similar to those seen in other moderate-to-low flow years. Since 2012, the Bonneville Dam second powerhouse turbines have been operated at middle-1% efficiency. This mode of operation increases flow to the first powerhouse and spillway, where there are no PIT monitors. In 2016, operation near middle-1% of peak efficiency continued at the second powerhouse turbines; however, slightly higher river flows contributed to a decrease in the number of fish detected in the estuary this year. Estuary detections of fish previously detected at Bonneville Dam are required to estimate probabilities of survival of inriver migrants to the tailrace of Bonneville Dam, as well as estimates through the entire hydrosystem. Reduced rates of detection at Bonneville Dam may impair the accuracy of survival estimates, especially for species with fewer tagged fish.

42

References Axel, A. A., E. E. Hockersmith, B. J. Burke, K. Frick, B. P. Sandford, W. D. Muir, R. F.

Absolon, N. Dumdei, J. J. Lamb, and M. G. Nesbit. 2010. Passage behavior and survival of radio-tagged yearling and subyearling Chinook salmon and juvenile steelhead at Ice Harbor Dam, 2009. Report to the Army Corps of Engineers, Walla Walla, Washington.

Beeman, J. W. and A.G. Maule. 2006. Migration Depths of Juvenile Chinook Salmon

and Steelhead Relative to Total Dissolved Gas Supersaturation in a Columbia River Reservoir. Transactions of the American Fisheries Society 135:3, 584-594.

Connor, W. P., J. G. Sneva, K. F. Tiffan, R. K. Steinhorst, and D. Ross. 2005. Two

alternative juvenile life history types for fall Chinook salmon in the Snake River Basin. Transactions of the American Fisheries Society 134: 291-304.

Cormack, R. M. 1964. Estimates of survival from the sightings of marked animals.

Biometrika 51:429-483. Faulkner, J. R., S. G. Smith, D. L. Widener, T. M. Marsh, and R. W. Zabel. 2016.

Survival estimates for the passage of spring-migrating juvenile salmonids through Snake and Columbia River dams and reservoirs, 2016. Report of the National Marine Fisheries Service to the Bonneville Power Administration, Portland, Oregon.

Hockersmith, E. E., G. A. Axel, R. F. Absolon, B. J. Burke, K. E. Frick, J. J. Lamb, M.

G. Nesbit, N. D. Dumdei, and B. P. Sandford. 2010. Passage behavior and survival for radio-tagged yearling Chinook salmon and juvenile steelhead at Lower Monumental Dam, 2009. Report to the Army Corps of Engineers, Walla Walla, Washington.

Hosmer, D. W., and S. Lemeshow. 2000. Applied Logistic Regression, 2nd ed. John

Wiley & Sons, Inc., 375 pp. Jolly, G. M. 1965. Explicit estimates from capture-recapture data with both death and

immigration—stochastic model. Biometrika 52:225-247. Ledgerwood, R. D., F. P. Thrower, and E. M. Dawley. 1991. Diel sampling of migratory

juvenile salmonids in the Columbia River Estuary. U.S. Fishery Bulletin 68:203-217.

Ledgerwood, R. D. J. W. Ferguson, B. A. Ryan, E. M. Dawley, and E. P. Nunnallee.

2004. A surface trawl to detect migrating juvenile salmonids tagged with passive integrated transponder tags. North American Journal of Fisheries Management 24:440-451.

43

Ledgerwood, R. D., A. S. Cameron, B. P. Sandford, L. B. Way, and G. M. Matthews. 2005. Detection of PIT-tagged juvenile salmonids in the Columbia River estuary using a pair-trawl, 2002. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers, Walla Walla, Washington.

Ledgerwood, R. D., A. S. Cameron, B. P. Sandford, L. B. Way, and G. M. Matthews.

2007. Detection of PIT-tagged juvenile salmonids in the Columbia River estuary using pair-trawls, 2005. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers, Walla Walla, Washington.

Magie, R. J., M. S. Morris, R. D. Ledgerwood, A. L. Cook, B. P. Sandford, and G. M.

Matthews. 2010. Detection of PIT-tagged juvenile salmonids in the Columbia River estuary using pair-trawls, 2008. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers, Walla Walla, Washington.

Magie, R. J., M. S. Morris, R. D. Ledgerwood, A. L. Cook, and B. P. Sandford. 2011.

Detection of PIT-tagged juvenile salmonids in the Columbia River estuary using pair-trawls, 2010. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers, Walla Walla, Washington.

Magie, R. J., M. S. Morris, J. P. Bender, B. F. Jonasson, B. P. Sandford, and R. D.

Ledgerwood. 2015. Development of a passive integrated transponder (PIT) tag detection system for adult salmonids in the lower Columbia River, 2013. Internal report of the Fish Ecology Division, Northwest Fisheries Science Center. Available from www.nwfsc.noaa.gov/publications/scipubs/index.cfm (January 2015).

Marsh, D. M., J. R. Harmon, N. N. Paasch, K. L. Thomas, K. W. McIntyre, B. P.

Sandford, G. M. Matthews. 2005. Research related to transportation of juvenile salmonids on the Columbia and Snake Rivers, 2004: final report for the 2001 spring/summer Chinook salmon juvenile migration. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers Walla Walla, Washington.

Marsh, D. M., J. R. Harmon, N. N. Paasch, K. L. Thomas, K. W. McIntyre, B. P.

Sandford, G. M. Matthews, and W. D. Muir. 2008. Transportation of juvenile salmonids on the Snake River, 2005: Final report for 2003 steelhead juveniles with updates on other transport studies. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers Walla Walla, Washington.

Marsh, D. M. J. R. Harmon, N. N. Paasch, K. L. Thomas, K. W. McIntyre, B. P. Sandford,

and G. M. Matthews. 2010. Transport of juvenile salmonids on the Snake River, 2007: final report for the 2004 wild spring/summer Chinook salmon juvenile migration. Draft report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers, Walla Walla, Washington.

44

Marsh, D. M., W. D. Muir, B. P. Sandford, S. G. Smith, and Diane Elliott. 2012. Alternative barging strategies to improve survival of salmonids transported from Lower Granite Dam: final report from the 2006-2008 spring/summer Chinook salmon and steelhead juvenile migration. Report of the National Marine Fisheries Service to the U.S. Army Corps of Engineers Walla Walla, Washington.

Marvin, D. and J. Nighbor, editors. 2009. 2009 PIT tag Specification Document. Report of the Pacific States Marine Fisheries Commission to the PIT-tag Technical Steering Committee. Gladstone, Oregon (www.ptagis.org/resources/document-library/ptagis-program-documents).

Morris, M. S., R. J. Magie, B. P. Sandford, J. P. Bender, A. L. Cook, and R. D. Ledgerwood. 2013. Detection of PIT-Tagged Juvenile Salmonids in the Columbia River Estuary using a Pair-Trawl, 2012. Report of the National Marine Fisheries Service to the Bonneville Power Administration, Portland, Oregon.

Morris, M. S., R. J. Magie, J. P. Bender, B. P. Sandford, and R. D. Ledgerwood. 2014.

Detection of PIT-Tagged Juvenile Salmonids in the Columbia River Estuary using a Pair-Trawl, 2014. Report of the National Marine Fisheries Service to the Bonneville Power Administration, Portland, Oregon.

Morris, M. S., L. N. Webb, P. J. Bentley, A. J. Borsky, B. P. Sandford, and R. D.

Ledgerwood. 2015. Detection of PIT-Tagged Juvenile Salmonids in the Columbia River Estuary using a Pair-Trawl, 2015. Report of the National Marine Fisheries Service to the Bonneville Power Administration, Portland, Oregon.

Muir, W. D., S. G. Smith, J. G. Williams, E. E. Hockersmith, and J. R. Skalski. 2001.

Survival estimates for migrant yearling Chinook salmon and steelhead tagged with passive integrated transponders in the lower Snake and lower Columbia Rivers, 1993-1998. North American Journal of Fisheries Management 21:269-282.

Plumb, J. M., A. C. Braatz, J. N. Lucchesi, S. D. Fielding, A. D. Cochran, T. K. Nation, J.

M. Sprando, J. L. Schei, R. W. Perry, N. S. Adams, and D.W. Rondorf. 2003. Behavior and survival of radio-tagged juvenile Chinook salmon and steelhead and performance of a removable spillway weir at Lower Granite Dam, Washington, 2003. Annual Report by the U. S. Geological Survey to the U. S. Army Corps of Engineers, contract W68SBV00104592, Walla Walla, Washington.

Prentice, E. F., T. A. Flagg, and C. S. McCutcheon. 1990a. Electronic tags. American

Fisheries Society Symposium 7:317-322. Prentice, E. F., T. A. Flagg, C. S. McCutcheon, and D. F. Brastow. 1990a. PIT-tag

monitoring systems for hydroelectric dams and fish hatcheries. American Fisheries Society Symposium 7:323-334.

45

Prentice, E. F., T. A. Flagg, C. S. McCutcheon, D. F. Brastow, and D. C. Cross. 1990b. Equipment, methods, and an automated data-entry station for PIT-tagging. American Fisheries Society Symposium 7:335-340.

PSMFC (Pacific States Marine Fisheries Commission). 2016. The Columbia Basin

PIT-Tag-Information System (PTAGIS). Online interactive database, Pacific States Marine Fisheries Commission, Gladstone, Oregon (available via the internet at www.psmfc.org/pittag).

Ramsey, F., and Schafer, D. 1997. The Statistical Sleuth. Duxbury Press. Rundio, D. E., A. N. Montgomery, M. G. Nesbit, M. S. Morris, G. T. Brooks, G. A. Axel,

J. J. Lamb, R. W. Zabel, and S. T. Lindley. 2016. Central valley passive integrated transponder (PIT) tag array feasibility study. Report of the National Marine Fisheries Service to the California Department of Fish and Wildlife.

Ryan, B. A., S. G. Smith, J. M. Butzerin, and J. W. Ferguson. 2003. Relative

vulnerability to avian predation of PIT-tagged juvenile salmonids in the Columbia River estuary, 1998-2000. Transactions of the American Fisheries Society. 132:275-288.

Seber, G. A. F. 1965. A note on the multiple recapture census. Biometrika 52:249-259. Skalski, J. R., S. G. Smith, R. N. Iwamoto, J. G. Williams, and A. Hoffmann. 1998. Use

of passive integrated transponder tags to estimate survival of migrant juvenile salmonids in the Snake and Columbia Rivers. Canadian Journal of Fisheries and Aquatic Sciences 55:1484-1493.

Williams, J. G., S. G. Smith, and W. D. Muir. 2001. Survival estimates for downstream

migrant-yearling juvenile salmonids through the Snake and Columbia River hydropower system, 1966 to 1980 and 1993 to 1999. North American Journal of Fisheries Management 21:310-317.

Zabel, R. W. and J. G. Williams. 2002. Selective mortality for fish length and

migrational timing in Chinook salmon: what is the role of human disturbance? Ecological Applications 12:173-183.

Zar, J. H. 1999. Biostatistical analysis, 4th edition. Prentice Hall, Upper Saddle River,

New Jersey.

46

Appendix Appendix Table A1. Daily total sample time and detections for each salmonid species

using the matrix pair-trawl antenna system in the upper Columbia River estuary near Jones Beach (rkm 75), 2016.

Date

Total time underway

(h)

PIT-tag Detections (N)

Unknown Chinook Salmon

Coho Salmon Steelhead

Sockeye Salmon Cutthroat Total

24 Mar 3.67 0 0 0 0 0 0 0 25 Mar 0.00 -- -- -- -- -- -- -- 26 Mar 0.00 -- -- -- -- -- -- -- 27 Mar 0.00 -- -- -- -- -- -- -- 28 Mar 0.00 -- -- -- -- -- -- -- 29 Mar 4.73 0 0 0 0 0 0 0 30 Mar 0.00 -- -- -- -- -- -- -- 31 Mar 6.43 0 0 0 0 0 0 0 1 Apr 0.00 -- -- -- -- -- -- -- 2 Apr 0.00 -- -- -- -- -- -- -- 3 Apr 0.00 -- -- -- -- -- -- -- 4 Apr 4.98 1 0 0 1 0 0 2 5 Apr 4.93 0 1 0 0 0 0 1 6 Apr 4.52 0 0 0 0 0 0 0 7 Apr 6.58 0 1 0 1 0 0 2 8 Apr 4.55 0 0 0 0 0 0 0 9 Apr 0.00 -- -- -- -- -- -- -- 10 Apr 0.00 -- -- -- -- -- -- -- 11 Apr 6.07 0 1 0 1 0 0 2 12 Apr 4.35 0 1 0 0 0 0 1 13 Apr 5.33 0 0 0 2 0 0 2 14 Apr 5.55 0 1 0 1 0 0 2 15 Apr 6.23 2 1 0 1 0 0 4 16 Apr 0.00 -- -- -- -- -- -- -- 17 Apr 0.00 -- -- -- -- -- -- -- 18 Apr 4.20 0 4 0 20 0 0 24 19 Apr 6.33 1 8 0 31 0 0 40 20 Apr 5.63 2 1 0 39 0 0 42 21 Apr 5.20 1 7 0 33 0 0 41 22 Apr 5.62 1 3 0 35 0 0 39 23 Apr 0.00 -- -- -- -- -- -- -- 24 Apr 0.00 -- -- -- -- -- -- -- 25 Apr 5.07 0 0 0 40 0 0 40 26 Apr 6.23 2 8 0 40 0 0 50 27 Apr 5.10 0 16 1 37 0 0 54 28 Apr 5.35 0 10 0 33 0 0 43 29 Apr 5.40 1 13 0 58 0 0 72

47

Appendix Table A1. Continued.

Date

Total time underway

(h)

PIT-tag Detections (N)

Unknown Chinook Salmon

Coho Salmon Steelhead

Sockeye Salmon Cutthroat Total

30 Apr 0.00 -- -- -- -- -- -- -- 1 May 11.37 2 97 3 120 2 0 224 2 May 14.25 5 178 5 228 4 0 420 3 May 14.38 4 180 6 191 7 0 388 4 May 14.85 5 190 3 102 6 0 306 5 May 14.40 7 256 2 163 8 0 436 6 May 15.62 7 308 9 192 7 0 523 7 May 15.38 8 275 7 206 7 0 503 8 May 12.22 10 145 10 184 8 0 357 9 May 19.27 14 333 15 108 7 0 477 10 May 15.53 13 334 22 117 5 0 491 11 May 16.33 18 322 19 217 4 0 580 12 May 17.87 16 347 27 215 7 0 612 13 May 15.40 5 158 18 154 7 0 342 14 May 13.05 3 109 6 220 5 0 343 15 May 13.05 9 169 19 183 1 0 381 16 May 17.57 9 207 30 140 9 1 396 17 May 18.55 5 197 26 107 6 0 341 18 May 15.77 10 173 25 140 2 0 350 19 May 18.98 14 276 73 196 8 0 567 20 May 20.35 12 260 80 373 8 0 733 21 May 14.62 12 122 50 184 9 0 377 22 May 14.17 8 78 34 149 11 0 280 23 May 14.70 6 41 31 84 4 0 166 24 May 17.25 8 76 24 107 5 0 220 25 May 17.55 9 54 43 134 20 0 260 26 May 17.23 6 44 20 90 13 0 173 27 May 14.38 5 42 13 65 12 0 137 28 May 8.92 2 12 1 29 4 0 48 29 May 11.92 2 35 5 52 4 0 98 30 May 15.70 6 50 4 105 8 0 173 31 May 13.12 4 25 4 35 7 0 75 1 Jun 13.48 2 27 5 56 6 1 97 2 Jun 13.65 2 36 11 50 5 0 104 3 Jun 12.77 4 21 12 33 3 0 73 4 Jun 7.65 1 20 7 13 0 0 41 5 Jun 11.45 5 15 4 25 3 0 52 6 Jun 13.20 2 17 2 13 0 0 34 7 Jun 14.07 2 32 3 49 1 0 87 8 Jun 13.43 2 12 2 7 0 0 23 9 Jun 13.58 3 11 1 23 0 0 38 10 Jun 8.35 1 21 0 12 0 0 34 11 Jun 7.08 1 1 0 4 1 0 7 12 Jun 0.00 -- -- -- -- -- -- -- 13 Jun 6.58 22 10 0 32 0 0 64 14 Jun 7.05 3 33 0 8 0 0 44 15 Jun 6.20 1 6 4 2 0 0 13

48

Appendix Table A1. Continued.

Date

Total time underway

(h)

PIT-tag Detections (N)

Unknown Chinook Salmon

Coho Salmon Steelhead

Sockeye Salmon Cutthroat Total

16 Jun 6.92 0 22 0 5 0 0 27 17 Jun 6.23 1 4 1 1 0 0 7 18 Jun 0.00 -- -- -- -- -- -- -- 19 Jun 0.00 -- -- -- -- -- -- -- 20 Jun 6.58 1 25 1 2 0 0 29 21 Jun 6.28 0 11 3 3 0 0 17 22 Jun 6.70 0 8 1 3 0 0 12 23 Jun 6.23 0 8 0 3 0 0 11 24 Jun 6.75 0 10 0 1 0 0 11 25 Jun 0.00 -- -- -- -- -- -- -- 26 Jun 0.00 -- -- -- -- -- -- -- 27 Jun 6.57 3 4 1 1 0 0 9 28 Jun 6.92 5 4 1 1 0 1 12 29 Jun 6.95 31 3 2 1 0 0 37 30 Jun 6.92 12 3 1 2 0 0 18 1 Jul 4.97 13 3 0 0 0 0 16 2 Jul 0.00 -- -- -- -- -- -- -- 3 Jul 0.00 -- -- -- -- -- -- -- 4 Jul 0.00 -- -- -- -- -- -- -- 5 Jul 6.28 6 4 0 0 0 0 10 Total 828.64 378 5,540 697 5,313 234 3 12,165

49

Appendix Table A2. Combined daily totals of impinged or injured fish resulting from the matrix pair-trawl antenna system used in the upper Columbia River estuary (rkm 75), 2016.

Date

Yearling Chinook

Subyearling Chinook Coho Steelhead Sockeye Chum

24 Mar 0 0 0 0 0 0 25 Mar -- -- -- -- -- -- 26 Mar -- -- -- -- -- -- 27 Mar -- -- -- -- -- -- 28 Mar -- -- -- -- -- -- 29 Mar 0 0 0 0 0 0 30 Mar -- -- -- -- -- -- 31 Mar 0 0 0 0 0 0 1 Apr -- -- -- -- -- -- 2 Apr -- -- -- -- -- -- 3 Apr -- -- -- -- -- -- 4 Apr 0 0 0 0 0 0 5 Apr 0 0 0 0 0 0 6 Apr 0 0 0 0 0 0 7 Apr 0 0 0 0 0 0 8 Apr 0 0 0 0 0 0 9 Apr -- -- -- -- -- -- 10 Apr -- -- -- -- -- -- 11 Apr 0 0 0 0 0 0 12 Apr 0 0 0 0 0 0 13 Apr 0 0 0 0 0 0 14 Apr 0 0 0 0 0 0 15 Apr 0 0 0 0 0 0 16 Apr -- -- -- -- -- -- 17 Apr -- -- -- -- -- -- 18 Apr 0 0 0 1 0 0 19 Apr 0 0 0 0 0 0 20 Apr 0 0 0 0 0 0 21 Apr 9 0 0 6 0 2 22 Apr 6 0 0 3 0 0 25 Apr 0 0 0 1 0 0 26 Apr 2 0 0 1 0 0 27 Apr 6 0 2 3 0 1 28 Apr 0 0 1 0 0 0 29 Apr 3 0 0 0 0 0 30 Apr -- -- -- -- -- -- 1 May 0 0 0 1 0 0 2 May 1 0 1 0 0 0 3 May 2 0 0 0 0 0 4 May 5 0 0 2 0 1 5 May 3 0 1 4 0 0 6 May 0 0 0 0 0 0 7 May 0 0 0 0 0 0 8 May 5 0 0 1 0 0 9 May 0 0 1 0 0 0 10 May 0 0 0 0 0 0 11 May 9 0 2 3 0 1 12 May 2 0 0 3 1 0 13 May 0 0 0 0 0 0

50

Appendix Table A2. Continued. Date

Yearling Chinook

Subyearling Chinook Coho Steelhead Sockeye Chum

14 May 0 0 0 0 0 0 15 May 1 0 0 0 0 0 16 May 0 1 0 0 0 0 17 May 2 1 2 1 0 0 18 May 0 0 0 0 0 0 19 May 0 0 0 0 0 1 20 May 1 0 0 0 1 0 21 May 1 1 0 1 0 0 22 May 1 1 2 0 0 0 23 May 0 0 0 0 0 0 24 May 0 0 3 0 0 0 27 May 4 0 2 1 0 0 28 May 0 0 0 0 0 0 29 May 0 0 0 1 0 1 30 May 0 0 0 0 0 0 31 May 0 0 0 0 0 0 1 Jun 0 0 0 0 0 0 2 Jun 0 1 1 1 0 0 3 Jun 3 1 0 0 0 1 4 Jun 0 0 0 0 0 0 5 Jun 0 0 0 1 0 0 6 Jun 0 0 0 0 0 0 7 Jun 0 0 1 0 0 0 8 Jun 0 0 0 1 0 0 9 Jun 0 0 0 0 1 0 10 Jun 0 0 0 0 0 0 11 Jun 0 0 0 1 0 0 12 Jun -- -- -- -- -- -- 13 Jun 0 0 0 0 0 0 14 Jun 0 0 0 0 0 0 15 Jun 0 0 0 0 0 0 16 Jun 0 0 0 0 0 0 17 Jun 0 1 1 1 0 0 18 Jun -- -- -- -- -- -- 19 Jun -- -- -- -- -- -- 20 Jun 0 0 0 0 0 0 21 Jun 0 0 0 0 0 0 22 Jun 0 0 0 0 0 0 23 Jun 0 0 0 0 0 0 24 Jun 0 0 0 0 0 0 25 Jun -- -- -- -- -- -- 26 Jun -- -- -- -- -- -- 27 Jun 0 0 0 0 0 0 28 Jun 0 0 0 0 0 0 29 Jun 0 0 0 0 0 0 30 Jun 0 1 0 0 0 0 1 Jul 0 0 0 0 0 0 2 Jul -- -- -- -- -- -- 3 Jul -- -- -- -- -- -- 4 Jul -- -- -- -- -- -- 5 Jul 0 0 0 0 0 0 Total 85 12 30 61 10 9

51

Appendix Table A3. Mean diel catch by hour for yearling Chinook salmon and steelhead during during intensive sampling (1 May-9 June) with a PIT-tag detector surface pair trawl at Jones Beach in the upper Columbia River estuary (rkm 75), 2016. Total effort was rounded to the nearest tenth of an hour.

Diel hour Total effort

(h)

Yearling Chinook salmon Steelhead (n) Mean detections/h (n) Mean detections/h

Hatchery Wild Hatchery Wild Hatchery Wild Hatchery Wild 0 35.0 317 37 9.06 1.06 93 16 2.66 0.46 1 33.9 354 47 10.46 1.39 81 36 2.39 1.06 2 23.3 336 34 14.45 1.46 53 14 2.28 0.60 3 17.2 226 27 13.18 1.57 37 11 2.16 0.64 4 14.7 231 23 15.68 1.56 51 11 3.46 0.75 5 24.8 303 37 12.21 1.49 52 22 2.10 0.89 6 37.1 248 55 6.68 1.48 162 58 4.36 1.56 7 39.9 175 28 4.39 0.70 236 78 5.91 1.95 8 40.0 119 21 2.98 0.53 235 68 5.88 1.70 9 40.0 100 18 2.50 0.45 281 60 7.03 1.50 10 40.0 129 26 3.23 0.65 353 76 8.83 1.90 11 38.8 125 25 3.23 0.65 429 101 11.07 2.61 12 25.5 99 15 3.88 0.59 353 83 13.83 3.25 13 14.2 61 8 4.30 0.56 271 55 19.11 3.88 14 5.9 27 6 4.56 1.01 204 36 34.48 6.08 15 2.0 5 -- 2.46 0.00 118 8 58.03 3.93 16 0.0 -- -- -- -- -- -- -- -- 17 0.0 -- -- -- -- -- -- -- -- 18 0.0 -- -- -- -- -- -- -- -- 19 16.5 26 5 1.58 0.30 73 43 4.43 2.61 20 33.8 170 66 5.02 1.95 225 102 6.65 3.01 21 35.0 615 88 17.57 2.51 242 77 6.91 2.20 22 35.0 482 64 13.77 1.83 164 46 4.69 1.31 23 35.0 332 41 9.49 1.17 99 19 2.83 0.54 Total 587.5 4,480 671 3,812 1,020

52

Appendix Table A4. Number of PIT-tagged yearling Chinook salmon loaded for transport at dams and numbers detected in the upper Columbia River estuary (rkm 75). Transport dates were 3 May-3 July; trawl operations 24 March-5 July, intensive sampling 1 May-9 June 2016. Season totals are shown.

Release date (2016) and time

Numbers loaded by dam (n) Total fish loaded (n)

Detections by transport dam (%) Total trawl detections Lower

Granite Little Goose

Lower Monumental

Lower Granite

Little Goose

Lower Monumental n (%)

3 May, 9:50 pm 1,646 2,271 1,116 5,033 1.70 2.47 1.88 105 2.09 4 May, 10:05 pm 912 1,069 516 2,497 1.54 1.78 2.52 46 1.84 5 May, 11:50 pm 1,112 946 362 2,420 0.72 0.53 0.55 15 0.62 7 May, 12:30 am 2,330 991 786 4,107 1.37 1.61 1.78 62 1.51 7 May, 7:10 pm 3,347 935 324 4,606 1.17 1.18 3.09 60 1.30 8 May, 9:50 pm 2,836 1,199 481 4,516 1.34 1.00 0.83 54 1.20 10 May, 12:50 am 2,767 1,138 1,119 5,024 2.02 1.67 1.16 88 1.75 11 May, 12:05 am 2,947 1,494 1,584 6,025 1.59 1.20 1.01 81 1.34 11 May, 7:25 pm 1,859 617 790 3,266 0.91 0.65 0.76 27 0.83 12 May, 7:30 pm 1,219 565 325 2,109 0.25 0.53 0.31 7 0.33 13 May, 9:00 pm 515 488 224 1,227 2.14 1.84 2.23 25 2.04 14 May, 8:45 pm 237 320 181 738 2.53 1.56 1.10 13 1.76 15 May, 7:50 pm 136 265 166 567 4.41 1.13 1.20 11 1.94 16 May, 7:20 pm 273 194 78 545 4.03 3.09 1.28 18 3.30 17 May, 8:00 pm 174 219 71 464 4.02 4.11 2.82 18 3.88 18 May, 8:45 pm 114 291 57 462 1.75 3.78 0.00 13 2.81 19 May, 7:25 pm 70 150 40 260 0.00 2.00 0.00 3 1.15 20 May, 8:10 pm 393 81 37 511 2.29 3.70 2.70 13 2.54 21 May, 10:45 pm 420 84 27 531 6.67 3.57 0.00 31 5.84 22 May, 9:10 pm 248 72 31 351 2.82 1.39 3.23 9 2.56 23 May, 8:00 pm 49 70 18 137 0.00 0.00 0.00 0 0.00 24 May, 7:50 pm 75 88 13 176 0.00 2.27 0.00 2 1.14 25 May, 8:20 pm 547 37 14 598 2.19 2.70 0.00 13 2.17

53

Appendix Table A4. Continued.

Release date and time

Numbers loaded by dam (n) Total fish loaded (n)

Detections by transport dam (%) Total trawl detections Lower

Granite Little Goose Lower

Monumental Lower Granite Little Goose

Lower Monumental n (%)

26 May, 8:00 pm 357 23 2 382 0.84 0.00 0.00 3 0.79 28 May, 7:40 pm 319 57 4 380 2.19 1.75 25.00 9 2.37 30 May, 7:50 pm 124 39 2 165 2.42 2.56 0.00 4 2.42 1 June, 7:30 pm 23 19 0 42 0.00 5.26 -- 1 2.38 3 June, 8:10 pm 62 25 8 95 4.84 0.00 0.00 3 3.16 5 June, 7:15 pm 54 36 0 90 0.00 2.78 -- 1 1.11 7 June, 8:00 pm 14 14 8 36 0.00 0.00 12.50 1 2.78 9 June, 8:00 pm 14 19 5 38 0.00 0.00 0.00 0 0.00 11 June, 7:05 pm 23 15 6 44 0.00 0.00 0.00 0 0.00 13 June, 7:30 pm 14 19 7 40 0.00 0.00 0.00 0 0.00 15 June, 7:40 pm 8 14 1 23 0.00 0.00 0.00 0 0.00 17 June, 7:05 pm 9 12 0 21 0.00 0.00 -- 0 0.00 19 June, 7:00 pm 5 24 3 32 0.00 0.00 0.00 0 0.00 21 June, 8:00 pm 3 13 0 16 0.00 0.00 -- 0 0.00 23 June, 7:15 pm 3 4 0 7 0.00 0.00 -- 0 0.00 25 June, 8:40 pm 4 9 0 13 0.00 0.00 -- 0 0.00 27 June, 8:10 pm 4 1 0 5 0.00 0.00 -- 0 0.00 29 June, 7:15 pm 3 2 2 7 0.00 0.00 0.00 0 0.00 1 July, 7:45 pm 3 3 1 7 0.00 0.00 0.00 0 0.00 3 July, 8:10 pm 5 6 1 12 0.00 0.00 0.00 0 0.00 Totals/means 25,277 13,938 8,410 47,625 1.57 1.60 1.38 736 1.55

54

Appendix Table A5. Number of PIT-tagged steelhead loaded for transport at dams and numbers detected in the upper Columbia River estuary (rkm 75). Transport dates were 3 May-3 July; trawl operations 24 March-5 July, intensive sampling 1 May-9 June 2016. Season totals are shown.

2016 release date and time

Numbers loaded by dam (n) Total fish loaded (n)

Detections by transport dam (%) Total trawl detections Lower

Granite Little Goose Lower

Monumental Lower Granite Little Goose

Lower Monumental n (%)

3 May, 9:50 pm 729 580 896 2205 1.78 1.38 1.34 33 1.50 4 May, 10:05 pm 578 301 471 1350 0.87 0.33 0.85 10 0.74 5 May, 11:50 pm 1055 228 303 1586 3.03 2.19 1.32 41 2.59 7 May, 12:30 am 785 365 349 1499 0.76 1.37 0.86 14 0.93 7 May, 7:10 pm 553 619 359 1531 1.81 0.81 3.06 26 1.70 8 May, 9:50 pm 891 723 463 2077 0.90 1.38 0.86 22 1.06 10 May, 12:50 am 1237 571 633 2441 2.51 3.85 3.32 74 3.03 11 May, 12:05 am 1654 508 490 2652 1.75 2.56 4.08 62 2.34 11 May, 7:25 pm 1077 530 491 2098 2.88 5.28 3.67 77 3.67 12 May, 7:30 pm 998 441 386 1825 7.52 5.90 5.96 124 6.79 13 May, 9:00 pm 907 371 246 1524 3.31 2.43 5.69 53 3.48 14 May, 8:45 pm 855 191 189 1235 1.99 3.14 2.12 27 2.19 15 May, 7:50 pm 539 274 273 1086 1.30 1.82 1.10 15 1.38 16 May, 7:20 pm 819 184 178 1181 3.17 7.61 4.49 48 4.06 17 May, 8:00 pm 425 170 180 775 5.65 4.12 2.78 36 4.65 18 May, 8:45 pm 315 357 108 780 6.67 5.60 7.41 49 6.28 19 May, 7:25 pm 393 167 105 665 5.09 2.40 5.71 30 4.51 20 May, 8:10 pm 1068 107 75 1250 5.24 2.80 2.67 61 4.88 21 May, 10:45 pm 1691 177 106 1974 2.01 1.13 1.89 38 1.93 22 May, 9:10 pm 930 97 116 1143 3.01 5.15 0.86 34 2.97 23 May, 8:00 pm 182 116 68 366 2.20 3.45 0.00 8 2.19 24 May, 7:50 pm 160 98 67 325 1.88 1.02 0.00 4 1.23 25 May, 8:20 pm 963 163 43 1169 1.87 0.00 4.65 20 1.71 26 May, 8:00 pm 843 61 46 950 0.83 3.28 2.17 10 1.05

55

Appendix Table A5. Continued.

2016 release date and time

Numbers loaded by dam (n) Total fish loaded (n)

Detections by transport dam (%) Total trawl detections Lower

Granite Little Goose Lower

Monumental Lower Granite Little Goose

Lower Monumental n (%)

28 May, 7:40 pm 1239 84 55 1378 4.12 5.95 5.45 59 4.28 30 May, 7:50 pm 490 61 9 560 3.06 1.64 11.11 17 3.04 1 June, 7:30 pm 63 70 12 145 4.76 2.86 8.33 6 4.14 3 June, 8:10 pm 506 62 13 581 2.96 3.23 0.00 17 2.93 5 June, 7:15 pm 485 40 12 537 5.77 7.50 8.33 32 5.96 7 June, 8:00 pm 82 104 24 210 4.88 2.88 0.00 7 3.33 9 June, 8:00 pm 100 46 20 166 2.00 0.00 0.00 2 1.20 11 June, 7:05 pm 93 69 8 170 6.45 8.70 0.00 12 7.06 13 June, 7:30 pm 39 57 24 120 0.00 0.00 0.00 0 0.00 15 June, 7:40 pm 10 14 5 29 0.00 0.00 0.00 0 0.00 17 June, 7:05 pm 5 2 2 9 0.00 0.00 0.00 0 0.00 19 June, 7:00 pm 8 11 3 22 0.00 0.00 0.00 0 0.00 21 June, 8:00 pm 4 3 2 9 25.00 0.00 0.00 1 11.11 23 June, 7:15 pm 1 1 0 2 0.00 0.00 -- 0 0.00 25 June, 8:40 pm 1 2 0 3 0.00 0.00 -- 0 0.00 27 June, 8:10 pm 1 2 0 3 0.00 0.00 -- 0 0.00 29 June, 7:15 pm 2 3 0 5 0.00 0.00 -- 0 0.00 1 July, 7:45 pm 2 3 0 5 0.00 0.00 -- 0 0.00 3 July, 8:10 pm 2 1 1 4 0.00 0.00 0.00 0 0.00 Totals/means 22,780 8,034 6,831 37,645 2.90 2.83 2.66 1,069 2.84

56

Appendix Table A6. Trawl system detections of PIT-tagged juvenile Chinook salmon and steelhead previously detected at Bonneville Dam, 2016.

Date detected at Bonneville (2016)

Detections

Bonneville Dam (n) Jones Beach (n) Bonneville and Jones Beach

(%) Chinook Steelhead Chinook Steelhead Chinook Steelhead

24-Mar 0 1 0 0 -- 0.00 25 Mar 0 4 0 0 -- 0.00 26 Mar 2 1 0 0 0.00 0.00 27 Mar 0 1 0 0 -- 0.00 28 Mar 1 4 0 0 0.00 0.00 29 Mar 1 2 0 0 0.00 0.00 30 Mar 1 1 0 0 0.00 0.00 31 Mar 0 6 0 0 -- 0.00 1 Apr 0 7 0 0 -- 0.00 2 Apr 1 11 0 1 0.00 9.09 3 Apr 6 10 0 0 0.00 0.00 4 Apr 13 12 0 0 0.00 0.00 5 Apr 34 9 0 1 0.00 11.11 6 Apr 73 6 0 0 0.00 0.00 7 Apr 80 15 0 0 0.00 0.00 8 Apr 100 16 0 0 0.00 0.00 9 Apr 96 35 0 0 0.00 0.00 10 Apr 70 13 0 0 0.00 0.00 11 Apr 75 29 0 1 0.00 3.45 12 Apr 632 37 0 0 0.00 0.00 13 Apr 178 22 0 0 0.00 0.00 14 Apr 302 25 1 0 0.33 0.00 15 Apr 492 208 4 0 0.81 0.00 16 Apr 480 228 3 12 0.63 5.26 17 Apr 547 212 1 3 0.18 1.42 18 Apr 502 302 1 4 0.20 1.32 19 Apr 521 530 5 6 0.96 1.13 20 Apr 607 549 0 9 0.00 1.64 21 Apr 834 788 2 3 0.24 0.38 22 Apr 1,004 818 6 0 0.60 0.00 23 Apr 723 789 4 5 0.55 0.63 24 Apr 864 565 5 4 0.58 0.71 25 Apr 1,076 580 5 3 0.46 0.52 26 Apr 800 431 3 4 0.38 0.93 27 Apr 736 881 5 13 0.68 1.48 28 Apr 1,047 1,050 2 0 0.19 0.00 29 Apr 1,372 1,859 11 11 0.80 0.59 30 Apr 1,698 1,492 27 47 1.59 3.15 1 May 1,922 1,605 36 53 1.87 3.30 2 May 1,880 1,338 32 39 1.70 2.91 3 May 1,697 2,013 29 42 1.71 2.09 4 May 2,024 3,516 29 75 1.43 2.13 5 May 2,442 2,304 37 85 1.52 3.69 6 May 1,778 1,519 20 28 1.12 1.84 7 May 2,194 1,181 26 19 1.19 1.61 8 May 2,506 1,072 29 26 1.16 2.43

57

Appendix Table A6. Continued.

Date detected at Bonneville (2016)

Detections

Bonneville Dam (n) Jones Beach (n) Bonneville and Jones Beach

(%) Chinook Steelhead Chinook Steelhead Chinook Steelhead

9 May 1,918 813 41 28 2.14 3.44 10 May 1,465 692 38 24 2.59 3.47 11 May 1,128 653 18 19 1.60 2.91 12 May 1,462 826 23 38 1.57 4.60 13 May 1,360 1,272 11 34 0.81 2.67 14 May 1,458 1,204 49 56 3.36 4.65 15 May 1,328 978 35 27 2.64 2.76 16 May 585 840 21 43 3.59 5.12 17 May 1,137 1,067 35 36 3.08 3.37 18 May 752 1,930 43 167 5.72 8.65 19 May 765 1,386 31 82 4.05 5.92 20 May 429 817 7 22 1.63 2.69 21 May 367 523 10 16 2.72 3.06 22 May 213 551 7 31 3.29 5.63 23 May 216 554 5 22 2.31 3.97 24 May 204 816 3 36 1.47 4.41 25 May 228 517 11 23 4.82 4.45 26 May 126 421 4 13 3.17 3.09 27 May 92 407 3 17 3.26 4.18 28 May 112 364 4 19 3.57 5.22 29 May 120 330 6 9 5.00 2.73 30 May 99 293 6 11 6.06 3.75 31 May 100 316 4 19 4.00 6.01 1 Jun 44 163 2 10 4.55 6.13 2 Jun 60 134 0 3 0.00 2.24 3 Jun 27 114 0 4 0.00 3.51 4 Jun 32 96 1 2 3.13 2.08 5 Jun 48 102 0 6 0.00 5.88 6 Jun 31 183 1 4 3.23 2.19 7 Jun 34 151 1 6 2.94 3.97 8 Jun 30 114 0 6 0.00 5.26 9 Jun 32 129 0 3 0.00 2.33 10 Jun 35 135 1 0 2.86 0.00 11 Jun 20 97 1 10 5.00 10.31 12 Jun 54 46 1 1 1.85 2.17 13 Jun 70 25 1 3 1.43 12.00 14 Jun 46 23 1 1 2.17 4.35 15 Jun 128 25 0 1 0.00 4.00 16 Jun 141 36 0 0 0.00 0.00 17 Jun 116 42 0 1 0.00 2.38 18 Jun 138 20 5 0 3.62 0.00 19 Jun 42 9 1 0 2.38 0.00 20 Jun 91 38 0 2 0.00 5.26 21 Jun 79 45 2 0 2.53 0.00 22 Jun 71 38 1 0 1.41 0.00 23 Jun 50 27 0 0 0.00 0.00 24 Jun 24 19 1 0 4.17 0.00

58

Appendix Table A6. Continued.

Date detected at Bonneville (2016)

Detections

Bonneville Dam (n) Jones Beach (n) Bonneville and Jones Beach

(%) Chinook Steelhead Chinook Steelhead Chinook Steelhead

25 Jun 45 24 0 0 0.00 0.00 26 Jun 40 28 0 0 0.00 0.00 27 Jun 15 14 0 1 0.00 7.14 28 Jun 16 17 0 1 0.00 5.88 29 Jun 24 18 0 0 0.00 0.00 30 Jun 16 22 0 0 0.00 0.00 01 Jul 23 17 0 0 0.00 0.00 02 Jul 22 8 0 0 0.00 0.00 03 Jul 7 2 0 0 0.00 0.00 04 Jul 19 8 0 0 0.00 0.00 05 Jul 24 13 0 0 0.00 0.00

Totals 48,769 45,649 758 1,351 1.55 2.96