Overview of Study Design

Design of this study was based on a meeting held on April 18, 2019, in Florence, Oregon, with a group of stakeholders concerned with actions to improve habitat and population status of coho salmon in Lake Creek, a major tributary of the Siuslaw River (fig. 1). The meeting was hosted and facilitated by the Confederated Tribes of the Coos, Lower Umpqua, and Siuslaw Indians, with Tribal staff in attendance, along with staff representing the Bureau of Land Management (BLM), National Marine Fisheries Service, Oregon Department of Fish and Wildlife, Siuslaw Watershed Council, and U.S. Geological Survey (USGS). At the meeting, stakeholders discussed fish passage over existing structures (fish ladders) installed by BLM and methods for monitoring fishway effectiveness for upstream passage of adult coho salmon. The overall management objectives identified were to (1) maximize the window of opportunity for upstream passage of adult coho salmon over Lake Creek Falls; and (2) identify cost-effective investments over short and long-term horizons (timeframes not specified) for maintenance, modification, or replacement of structures to improve passage. In the framework of structured decision making (Conroy and Peterson, 2013), the first objective can be considered a fundamental objective (what stakeholders want) and the second a means objective (the means to attain the fundamental objective). Once these objectives were identified, methods for evaluation were discussed, leading to the collective formulation of the study to meet management objectives Based on this discussion, stakeholders indicated that the preferred method of evaluating passage was tracking individual adult salmon using radio telemetry during their upstream migration passage over the three fish ladders in autumn. We initiated this work with a strong participatory element to ensure that collective objectives associated with this study were clear and based on a consensus among stakeholders responsible for managing coho salmon in the study system.

In addition to the basic question of upstream passage, stakeholders also expressed interest in the potential influences of stream discharge (river flows), time of year (for example, stream temperature, fish motivation), responses to each ladder (for example, physical features of each structure), and specific fish characteristics (for example, length, sex, stage of maturation). Because these factors were often naturally confounded and sample sizes with telemetry were limited (for example, by available funding and permitting limitations), we did not expect results of this study to resolve each of these influences. These limitations aside, we included a descriptive assessment of river flows, which are believed to limit upstream passage, particularly when flows are low or exceptionally high. In addition to passage efficiency (number of fish that successfully passed over a barrier divided by the number that approached it), we also documented duration for each upstream crossing at each ladder (for example, number of days to cross), and downstream passage owing to fallback (involuntary or unintentional crossing downstream) or emigration (directed downstream movement motivated by a proximal stimulus). We also mapped the final location of every fish to determine whether they used tributaries that could serve as spawning locations upstream and downstream from the falls. We did not assess pre-spawning survival in this study.

River Flows

Stakeholders identified river flows as a major factor of interest with respect to upstream passage of adult coho salmon over Lake Creek Falls. We selected two USGS streamgages near Lake Creek Falls to represent river flows in the study area—streamgages USGS 14307620 on the Siuslaw River near Mapleton, Oregon, and USGS 14306340 on East Fork Lobster Creek Near Alsea, Oregon. Although the latter streamgage is not within the Siuslaw River Basin, it is close to the study site and represents a basin area that is more comparable to Lake Creek. River flows in the Oregon Coast Range are directly and immediately impacted by rainfall events. Triangle Lake likely modifies the hydrographic response at Lake Creek Falls after upstream rain events; however, we assumed that nearby river flows reasonably approximated river flow conditions at Lake Creek Falls.

Radio Telemetry Receiving Systems

Gastrically implanted radio tags and telemetry were used to track adult coho salmon, a method that has been shown to be effective in other applications (for example, Liedtke and Wargo-Rub, 2012; Kock and others, 2019; Chapman and others, 2020; Liedtke and others, 2020). Radio telemetry antennas and data-logging receivers were installed throughout the study area. Antennas were either aerial Yagi-Uda or underwater stripped coaxial cable. Signal acquisition at each antenna was logged using either Lotek SRX_400-W16 receivers (Lotek Wireless, Inc., Newmarket, Ontario, Canada) or Orion receivers (Sigma Eight Inc., Aurora, Ontario, Canada). Receivers and 12-volt marine deep-cycle batteries were secured in locked, protective metal boxes with an informative pamphlet attached to the top that provided study information (fig. 3).

Arrays to monitor fish movement were deployed at the lower and upper end of each ladder as well as downstream and upstream from the three ladders to measure passage and egress from the vicinity of Lake Creek Falls. Aerial arrays of single Yagi-Uda antennas were deployed 0.2 miles (mi) downstream from the lower ladder (station 1), directly downstream and upstream from the upper ladder (stations 6 and 7), and 0.25 mi upstream from the upper ladder (station 8; table 1). Four sites deployed downstream and upstream from the lower and middle ladders each consisted of two stripped coaxial cable antennas positioned underwater and monitored by an Orion receiver (stations 2–5, fig. 1). Antennas were placed in areas where overlap between sites was minimal, and the signal was attenuated to further reduce overlapping detections. All receiver systems with multiple antennas were amplified with a 12-volt, 25-decibel-milliwatt (dBm) signal amplifier as soon as the coaxial cable was either out of the water, or if it was located 20 feet (ft) or more from an aerial Yagi-Uda antenna. Single antennas were connected directly to the receiver if the coaxial run was shorter than about 30 ft; otherwise, they also were amplified. Antenna runs were amplified to compensate for any (1) signal loss through the coaxial cable, (2) combiners used to combine multiple antennas into one signal, or (3) splitters used to split a single signal into multiple receivers. Fixed-attenuators then were added at the receiver to adjust the signal down to balance the system. A balanced antenna system resulted in a net gain or loss of 0-dBm of signal strength from the point of amplification to the receiver, regardless of how long the coaxial run was or how many combiners/splitters were used. Balancing was accomplished by inputting a known strength signal into the amplifier and adding attenuators at the receiver to reduce the signal down to the original strength.

Receiver sites were visited every 3–4 days throughout the study period for data download, maintenance, and troubleshooting. The 12-volt batteries were switched out on each visit and charged prior to the next visit.

Radio Transmitters

We used digitally encoded radio transmitters that operated at frequencies ranging from 166.320 to 166.700 megahertz. The radio transmitters used (Sigma Eight, Inc. model TX-PSC-I-450, weight 8.5 grams in air, 28-centimeter [cm] antenna, 46.0-millimeter [mm] length × 12.0-mm width) emitted a radio signal every 4.5, 4,7, 4.9, 5.1, or 5.3 seconds. The expected battery life was 2,501 days. Each tag was programmed to produce the radio signal at a single interval and the intervals were varied for the tags used on each tagging day to reduce the chance of radio signal collisions if multiple tags were in the same area for an extended time. Tags were fitted with a band (about 1.5 mm) of surgical tubing to increase roughness to reduce the likelihood of transmitter regurgitation (Keefer and others, 2004).

Fish Tagging

Adult coho salmon migrating upstream from the main-stem Siuslaw River and into Lake Creek were captured and tagged as they moved through the first two ladder structures on Lake Creek Falls. All fish were captured using large rubber-meshed dip nets and immediately transferred to large plastic tubs for tagging (fig. 4) or if not immediately tagged, they were held in large (2-ft × 4-ft) mesh-sided enclosures within the river prior to tagging (figs. 5 and 6). Selection of fish for tagging was opportunistic and dependent on river flows sufficient to allow fish to enter the ladder structures. We tagged most fish and only rejected very small individuals, or those in obviously poor condition.

Tagging was initiated by placing fish into an anesthetic tank (figs. 4 and 5), and a timer was started to ensure that no individual was in the tank for longer than 5 minutes. All fish were tagged in 4 minutes or less. Fish were anesthetized 1–3 minutes in a concentration of 15–20 milligrams per liter (mg/L) of Aqui-S 20E (Aqua Tactics, Kirkland, Washington) in river water, prior to tagging. A cover was immediately placed over the tank to prevent fish escape during initial stages of anesthesia. Prior to insertion, transmitters were disinfected with chlorhexidine diacetate solution (diluted to 30 mg/L) for at least 15 minutes and rinsed in distilled water, then placed in glycerin to help ease the friction of tag insertion. Sex, condition, fork length in centimeters, and maturity were all recorded for each captured specimen. A ratio of tag weight to fish weight was calculated by dividing the tag weight in air with an estimated fish weight using fork length-weight relationships described by Schneider and others (2000). The maturity category (Theresa Liedtke, U.S. Geological Survey, personal commun.) was defined as B, Bright (silvery coloration, with little evidence of external damage such as injuries and deteriorated fin margins; see fig. 4); M, Mature (spawning pigmentation, with minor external injuries or erosion of fin margins; see fig. 6); and R, Ripe (major external injuries or erosion of fin margins, including fish that appeared to have spawned). These categories were unavoidably subjective, but there were clear and obvious differences between what we called bright individuals compared to fish in more advanced stages of maturation and spawning condition (see figs. 4–7).

Tag implantation followed methods described in Keefer and others (2004). Tagging involved insertion of the tag antenna into a 7-inch long and ¼-inch wide polyvinyl chloride (PVC) pipe. The anesthetized fish was held by an assistant as the tagger rotated the fish dorsal-ventrally, using one hand to open the mouth slightly and the other to carefully insert the PVC pipe into the stomach to avoid rupture of the stomach lining (fig. 8). Once the tagger determined the correct position of the tag, the PVC guide tube was carefully removed, and the antenna was bent to run parallel to the axis of the tagged fish’s body. Bending the antenna reduced the likelihood that it would interfere with the natural movement and navigation of the fish.

Fish Movements

Movements of radio-tagged fish were tracked with stationary receivers and mobile tracking. Each of the eight stationary receivers behaved as gates through which each tagged fish passed, moving either upstream or downstream at Lake Creek Falls (fig. 1). In addition to tracking fish at these stationary locations, we used opportunistic mobile tracking. Mobile tracking was intended to confirm what the stationary receivers detected and to extend the reach of tracking outside the vicinity of Lake Creek Falls, including tracking fish to tributaries potentially used for spawning. For mobile tracking, a single Yagi-Uda and a Lotek SRX_400-W16 receiver was used near rivers and spawning tributaries to locate tagged fish. The mobile receivers used in this study were capable of scanning one frequency at a time and multiple trackers were deployed multiple times each week to confirm results among trackers. Fish locations were recorded and later entered into a database to confirm the exact position of each tag. Fish were tracked immediately after the first release on October 21, 2019, through February 20, 2020.

River Flow

Mean daily river flows in available years from 1991 to 2022 were downloaded (U.S. Geological Survey, 2019a, 2019b) and plotted from September 1, 2019, to February 28, 2020. Daily river flows during the September–February period in previous years also were plotted for reference.

Tagged Fish Characteristics

For each fish and tagging event, we summarized sex, fork length (in centimeters), maturity category, and relative tag weights. Relative tag weights were derived by estimating weights of fish based on a length-weight regression (Schneider and others, 2000) and dividing the tag weight by the weight of the individual tagged.

Fish Movements

Most of our efforts were focused on tracking movements of fish upstream or downstream across the three ladders at Lake Creek Falls. Data downloaded from fixed receivers in the Lake Creek Falls area were converted from text files to comma-separated value (csv) files. These files were combined into one or two csv files per receiver, depending on file size, and sorted chronologically by each tag code. Noise codes (tag codes not used in this study) and data prior to release were removed from the data files. Each fish record was reviewed separately by two trained technicians to determine arrival and departure at each receiver detection area upstream and downstream from the released location. If a gap in detections was determined for a fish moving upstream, for example, it was interpreted that the fish was somewhere between receivers and would be confirmed with mobile tracking equipment. The date and time of arrival and departure for each fish were recorded in a spreadsheet. A fish was determined to have passed a ladder if it was confirmed not only by the stationary receivers at the ladder, but also by the next upstream receiver. This redundancy helped reduce false detections that cause inaccurate passage rates for each ladder. This entire data entry and interpretation process was completed twice. Any discrepancies were reconciled in a final arrival and exit dataset. This process reduced human error and provided the most accurate passage rates for the three ladders. Additional mobile tracking allowed us to document entry into tributary streams potentially used for spawning. These locations are reported herein, but we did not confirm spawning activity. We were unable to rigorously track individual fish to determine their ultimate fates, but last known locations were reported.

River Flows

A graph of river flows in 2019 relative to previous years at both streamgages indicated relatively low flows with delayed onset of higher flows until December 2019 compared to the 24- and 30-year periods of record available for USGS streamgages on the Siuslaw River and Lobster Creek, respectively (figs. 9 and 10).

Fish Tagging

Capture of fish followed peaks in river flow (figs. 9 and 10), with most fish tagged late in November and in December (table 2). We tagged and tracked a total of 87 naturally produced adult coho salmon, including 48 females and 39 males. A total of five tagging sessions occurred—one in October, two in November, and two in December 2019 (table 2). Median fork lengths of fish tagged were 69.1 and 70 cm, respectively, whereas the range of fork lengths was 54–82 cm (fig. 11). Using estimated fish weights, the tag-weight-to-fish-weight ratio (tag burden) ranged from 0.13 to 0.54 percent and the median was 0.22 percent. Fish matured throughout the season with a greater number of tagged fish observed in categories Bright and Mature in October and categories Mature and Ripe in December (table 2). No fish were observed to regurgitate tags, and zero mortalities occurred during capture, tagging, or release.

Fish Movements

The total number of fish passing over each ladder varied by tagging group (fish tagged on a specific date), with the percentage of tagged fish passing ladders decreasing over time (table 2, fig. 12). Of the 15 fish tagged in October, 93.3 percent passed the lower ladder within the first 2 days and 100 percent passed by December 15, 2019. Passage over the lower ladder was slower for the fish released in November and December. More than 60 percent of fish passed the lower ladder by December 7 for the November tagged group and by December 14 for the December tagged group. A total of 83.3 percent of November fish and 97.6 percent of December fish passed over the lower ladder and rates of passage between males and females were similar. All the fish tagged in October eventually passed over the middle ladder and 80.0 percent did so by October 30, with the remaining fish passing at a rate of about one fish per week. A total of 50.0 percent of tagged fish released in November passed over the middle ladder. These fish were composed of more than 50 percent Ripe fish. Of the November fish that passed over the middle ladder, 50.0 percent passed by December 10, and 100 percent within 23 days after tagging. A total of 23.8 percent of December fish passed over the middle ladder and did so within 12 days after tagging. The December group was composed of 50.0 percent mature fish and 45.2 percent ripe fish. The trends were similar with upper ladder passage. October fish steadily passed (> 93 percent), and the November (46.7 percent) and December (19.0 percent) groups passed within a few weeks of release.

Movement behaviors through Lake Creek and Lake Creek Falls varied throughout the study period. We considered whether fish moved upstream or downstream relative to their original tagging location (that is, detected at receivers upstream or downstream from an individual’s original tagging location), and locations of final detections for each fish. Nearly twice as many males moved only upstream compared to those who moved only downstream, whereas an equal number of females moved exclusively upstream or downstream (table 3). About 43 percent of both sexes stayed in a single location for a minimum of 5 days, predominantly from the October and November release groups when flows were lower (fig. 11). From 9 to 10 percent of coho salmon stayed in a single location for at least 5 days after the December tagging events. More than one-half of all fish in each November and December release group moved upstream and downstream during the study period, except for the males in the November 27 release group, 40 percent of which moved in both directions.

Many of the radio-tagged coho salmon were last detected in a tributary downstream from Lake Creek Falls, and locations of final detections were similar between the sexes (table 3). By tag date, the predominant last known location for fish tagged in October was in upstream tributaries for males, whereas locations for females were mostly uncertain. Males and females tagged in November were widely distributed in the study area at the last known location. More than one-half of the fish tagged in December were last found in tributaries downstream from Lake Creek Falls (see Appendix 1, table 1.1 for details).