PWS_ PIT Interrogation system_Attachment C - 2016 Report .pdf

Willamette Project Juvenile Salmonid PIT Interrogation System Development – South Santiam River at Lebanon Dam Technology Development (TD-10-08): Willamette Project Juvenile Salmonid PIT Interrogation

System Development

Report covers work performed under Work Order #(s) W9127N-10-2-0008

Report was completed under USACE contract #(s) W9127N-10-2-0008-0032

4/01/2015 - 09/30/2016

Brian M. Alfonse

Derrek M. Faber

Oregon Department of Fish and Wildlife

The Dalles, OR

Report Created 8-2016

This report should be cited as follows:

Alfonse, B. M & D. M. Faber. Willamette Project Juvenile Salmonid PIT Interrogation System Development – South Santiam River at Lebanon Dam, 4/1/2015 - 9/30/2016. USACE Annual Technical Report, Project # W9127N-10-2-0008-0032.

Abstract

We installed robust and efficient passive integrated transponder (PIT) arrays at Lebanon Dam site on the

South Santiam River, Oregon, to remotely transfer tag‐detection information on migrating adult and juvenile salmonids to the Pacific States Marine Fishery Commissions’ (PSMFC), PTAGIS database. This detection site is 27 kilometers downstream of Foster Dam Dam on the North Santiam River. This placement was selected in order to facilitate fishery research needs in the Sorth Santiam River, a fourth order tributary to the Willamette River. We constructed durable antennas and deployed associated system components to last multiple years under harsh environmental conditions. At the Lebanon dam site 25 individual antennas were istalled in order to monitor fish moving through every feature of the dam. These antennas were divided into four unique interigation sites. The first of these sites was located on the south fish ladder and consisted of four antennas (LD1). The next site was on the north fish ladder and also consisted of four antennas. Both of these sites were designed to monitor adult and juvenile fish moving through the ladders (LD2). The next consisted of two antennas located on the irrigation canal bi‐ pass downstream from the V‐screen, with the primary purpose to detect juvenile salmonid outmigrants

(LD3). The final site consisted of 15 antennas mounted along the dam spillway and was intended to monitor juvenile outmigrants passing over the weir (LD4). Detection interrrigation site was configured to detect both Full Duplex (FDX) and Half Duplex (HDX) tags concurrently, using the Biomark™ IS1001‐MTS with a IS1001‐ACN located at each antenna. We calculated the detection efficiency for each fully operating site.

The antennas at the Lebanon Dam on the the South Santiam River are divided into four different groups.

Each of these groups are registered with PTAGIS as an individual ‘interrogation’ site. They were given the preambles ‘LD1’ (Lebanon Dam 1) for the south fish ladder, ‘LD2’ (Lebanon Dam 2) for the north fish ladder, ‘LD3’ (Lebanon Dam 3) for the fish bypass pipe, and ‘LD4’ (Lebanon Dam 4) for antenna located on the dam spillway. Detailed information on fish detections and site metadata can be found at the following url’s:

LD1 ‐ Lebanon Dam South Ladder:

http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD1

LD2 ‐ Lebanon Dam North Ladder:

http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD2

LD3 ‐ Lebanon Dam Diversion Bypass:

http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD3

LD4 ‐ Lebanon Dam Spillway:

http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD4

Contents

Abstract

Figures

Tables

Introduction

Methods

Antenna Array Design

Antenna Construction

Mounting the Antennas and Hardware

Estimating Detection Efficiency

Results

References

INFORMATION ON THE SOUTH SANTIAM WATERSHED. South Santiam River Watershed Council

(Dec 2016). Availible: http://www.sswc.org/about/watershed/

Figures

Figure 1. Study site location at the Lebanon Dam on the South Santiam River

Figure 2: The four interrogation sites that make up the Lebanon Pit tag detection complex on the South

Santiam River. Interigation sites LD1 and LD2 consisted of four antennas each and were designed to detect both adult and juvenile fish mving through either the North (LD2) or South (LD1) fish ladders. The

LD3 site consisted of two antennas desigigned to detect fish moving through the v‐screen bypass pipe located in the irrigation ditch and leading back into the South Santiam river downstream of the dam. The

LD4 site consisted of 15 antennas installed along the spillway designed to detect any juvenile fish passing over the dam. (http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD1)

Figure 3: The PIT tag antenna configuration in the south fish ladder at Lebanon dam on the South

Santiam River (LD1). All four antennas were pass‐through style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently to the cement.

(http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD1)

Figure 4: The PIT tag antenna configuration in the north fish ladder at Lebanon dam on the South

Santiam River (LD2). All four antennas were pass‐through style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently to the cement.

(http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD2)

Figure 5: The PIT tag antenna configuration in a diversion pipe at the Lebanon dam site on the South

Santiam River (LD3). This diversion pipe provided fish passage back to the South Santiam River downstream of the screened irrigation channel. Both antennas were pass‐through style and attached permanently around the outside of the pipe. (http://www.ptagis.org/sites/interrogation‐site‐ metadata?IntSiteCode=LD3)

Figure 6: The PIT tag antenna configuration along the Lebanon dam spillway on the South Santiam River

(LD3). All 15 antennas were pass‐over style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently the cement platform directly upstream from the hydrolic weir gates. (http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD4)

Figure 7: The detection efficiency at four interigation sites located at the Lebanon Dam on the South

Santiam River. Estimates at the South Fish Ladder (LD1), North Fish Ladder (LD2), and Diversion Bypass

(LD3) were all caluculated using detections form efficiencies between the antenna arrays at these sites.

To determine efficiency over the Spillway (LD4) three groups of PIT tagged juvenile spring Chinook

Salmon were used

Figure 8: The detection efficiency at the fish ladders on Lebanon Dam on the South Santiam River.

Estimates include differences between adult and juvenile detections

Tables

Table 1: PIT tag antennas built to collect data on migrating steelhead and salmon at the Lebanon Dam on the South Santiam River. Multiple anntennas were installed at both the South (LD1) and North (LD2) fish ladders as well around the southern fish bypass pipe and along the spillway (LD4). The type of antenna built was dependent on the mounting site and took into account metal interference and river flow. The number of wire loops was based on reaching the target inductance of 130‐180 µH and was variable with the antenna size

Table 2: The salmon and steelhead detected at Lebanon Dam PIT tag array sites between April 1, 2015 and November 30, 2016, divided by year tagged. No distinction was made between fish tagged as adults or juveniles

Introduction

The South Santiam River is a 106 km tributary to the Santiam River, which flows another 16 km before feeding into the Willamette River. The South Santiam headwaters flow from the Willamette

National Forest in the North Cascades and drains approximately 1,674 km2 . Two main stem dams are located on the South Santiam River, Foster dam located at river kilometer 38 and Green Peter dam at river kilometer 5. Both of these structure pose substantial barriers to mirating salmon and Steeelhead.

Since neither facility have juvenile bypass structure young salmon migrating out to the ocean must pass over the spillway or through the power turbines in order to geth past these facilities. Adult fish are able to pass over Foster dam via a fish passage facility but Green Peter dam is a complet barrier to adult salmon and Steelhead.

Since installation of the these dams in the 1960s, native salmon and steelhead population in the

South Santiam River have declined.

Figure 1. Study site location at the Lebanon Dam on the South Santiam River.

Placement of PIT‐detection arrays were selected to take advantage of existing infrastructure on the South Santiam River and where migrating fish were known to concentrate.

The Lebanon Dam is located near the town of Lebanon Oregon, and owed by the City of Albany, and was designed to divert water to the Willamette Valley for irrigation. Our goal was to install durable

PIT tag arrays at the site that effectively monitored migrating juvenile and adult salmon and steelhead to address objectives outlined in the 2008 BiOP.

Methods

Antenna Array Design

The Lebanon Dam site was selected as a PIT‐tag array site due to its ability to withstand high flows and funnel adult fish through confined areas where antennas can be mounted. In order to monitor all the potential routes that fish could pass through this dam site a total of 25 antennas were installed. These antennas were separated into four interrigaion site to allow tag detections information to be more easily analysed based on the location of the detection. These for interrogation sites are the :

South Fish Ladder (LD1), North Fish Ladder (LD2), V‐screen Diversion Bypass (LD3), and the Spillway

(LD4; Figure 2).

Figure 2: The four interrogation sites that make up the Lebanon Pit tag detection complex on the South Santiam River.

Interigation sites LD1 and LD2 consisted of four antennas each and were designed to detect both adult and juvenile fish mving through either the North (LD2) or South (LD1) fish ladders. The LD3 site consisted of two antennas desigigned to detect fish moving through the v‐screen bypass pipe located in the irrigation ditch and leading back into the South Santiam river downstream of the dam. The LD4 site consisted of 15 antennas installed along the spillway designed to detect any juvenile fish passing over the dam. (http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD1).

Figure 3: The PIT tag antenna configuration in the south fish ladder at Lebanon dam on the South Santiam River (LD1). All four antennas were pass‐through style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently to the cement. (http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD1).

Figure 4: The PIT tag antenna configuration in the north fish ladder at Lebanon dam on the South Santiam River (LD2). All four antennas were pass‐through style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently to the cement. (http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD2).

Flow

Flow

Figure 5: The PIT tag antenna configuration in a diversion pipe at the Lebanon dam site on the South Santiam River (LD3). This diversion pipe provided fish passage back to the South Santiam River downstream of the screened irrigation channel. Both antennas were pass‐through style and attached permanently around the outside of the pipe.

(http://www.ptagis.org/sites/interrogation‐site‐metadata?IntSiteCode=LD3).

Figure 6: The PIT tag antenna configuration along the Lebanon dam spillway on the South Santiam River (LD3). All 15 antennas were pass‐over style, shielded with ferrite tiles to reduce interference from metal rebar in the ladder, and were attached permanently the cement platform directly upstream from the hydrolic weir gates. (http://www.ptagis.org/sites/interrogation‐ site‐metadata?IntSiteCode=LD4).

Antenna Construction

We used Biomark IS‐1001 (ACNs) in combination with the IS‐1001‐MTS (master controller). This system has the ability detect both Full Duplex and Half Duplex tags, self‐tune up to 12 antennas, which allows the antennas to adjust to changing flow levels and environmental noise in‐situ and does not require personal to tune antennas.

Two different types of antennas were constructed dependent upon the placement at the site.

The first type was built using reinforced flexible hose (1 inch clear polybraid PVC) and was designed to be mounted around the outside of a pipe. The second type of antenna was constructed out of fiberglass reinforced plastic (FRP) to accommodate ferrite tiles used to shield the antenna from metal interference within its detection range. The initial construction of the antenna types was identical, only the outer encasement differed.

North

Bank

South

Bank

The number of wire loops was first determined for the antenna based on the antenna size. The

Biomark ACN’s can tune an antenna with an inductance in the range of 120 µH to 180 µH. We aimed for an inductance in the middle of that range to account for minor changes during antenna construction and the effect of the placement in the river. The inductance was measured with an LCR (inductance, capacitance and resistance) meter and was influenced by the number of wire loops within the antenna.

The larger the perimeter of the antenna, the fewer loops it took to achieve the target inductance. We used Litz™ 10AWG‐magnet wire that was manufactured to best achieve inductance in our target frequency range (New England Wire Technologies).

In order to provide an adequate air‐gap, and for durable waterproof antennas, we used three layers of protection around the antenna wires. First, we incased the length of the antenna wires with flame‐retardant heat‐shrink tubing (3M ITCSN‐0800 heavy wall heat shrink tubing) using a heat gun.

Second, we fed the heat shrunk wires through reinforced flexible hose (1 inch clear polybraid PVC).

Using both the heat shrink tubing layer and the flexible hose protected the antennas from a breach in either of the layers. At this stage, the antenna wires were fed into an FRP frame.

To build the FRP, ferrite shielded antennas, we cut the FRP to length with a miter cut for the corners. Ferrite tiles (EMI suppression ferrite tiles, Kitagawa Industries) were cut to size using a tile saw.

Only the sides of the antenna directly touching cement or metal was shielded with the ferrite tiles. The ferrite tiles were glued to the inside of the FRP and the antenna was then fed into the FRP pieces. Then the antenna wire was soldered together to create a continuous loop, and attached to the antenna cable for routing to the ACN. We used fiberglass epoxy to join the FRP corners and strengthened the corners with FRP angle pieces.

Recommendations for operating ACN’s include achieving capacitance settings in the 400‐600 range (Steve Anglea, personal communication, August 27, 2014). Therefore we used capacitors (CDE film capacitors ranging from 1.0nF‐3.3 nF) to tune the antenna within that range and to lower the voltage across the antenna cable. To find the correct number and configuration of capacitors we connected the negative end of the antenna cable to an electronic bread‐board and tested various capacitor configurations through the antenna tuning cycle (full‐tune). This usually required an added capacitance within the range of 400µF‐1600µF. We then balanced the capacitors to ease the strain on each capacitor thus extending capacitor life. This was done by doubling the desired amount of capacitors, connecting them in parallel to the original set, then creating an identical bundle and connecting it in series to the first set (Prentice 2008). This capacitor bundle was then wrapped with electrical tape and soldered in‐series to the antenna cable at the antenna end.

Air space within the antenna is necessary to maximize the performance of the antenna, but subject to breach under water. Therefore, we filled extra space within the antenna with urethane expansion foam to offer another degree of protection. The liquid foam was poured through access holes and was allowed to expand throughout the antenna volume. We sealed the capacitors and antenna wire‐ends within the head‐box with pourable polyurethane plastic that hardened to ensure water could not reach the exposed soldered wires and capacitors.

Table 1: PIT tag antennas built to collect data on migrating steelhead and salmon at the Lebanon Dam on the South Santiam River. Multiple anntennas were installed at both the South (LD1) and North (LD2) fish ladders as well around the southern fish bypass pipe and along the spillway (LD4). The type of antenna built was dependent on the mounting site and took into account metal interference and river flow. The number of wire loops was based on reaching the target inductance of 130‐180 µH and was variable with the antenna size.

Site Antenna Number (s)

Location Type Dimensions (cm)

Wire Loops

Install Date

LD1 1‐3 Ladder South Shielded 245x163 4 27‐May

LD1 4 Ladder South Shielded 213x152 4 13‐Nov

LD2 1 Ladder North Shielded 109x81 7 16‐Apr

LD2 2 Ladder North Shielded 142x107 7 16‐Apr

LD2 3 Ladder North Shielded 61x46 9 16‐Apr

LD2 4 Ladder North Shielded 142x107 7 28‐May

LD3 1‐2 Bypass Flex 302 circ. 8 30‐Apr

LD4 1‐5 Spillway North Shielded 610x20 5 2‐Oct

LD4 6‐10 Spillway Center Shielded 610x20 5 2‐Oct

LD4 11‐15 Spillway South Shielded 610x20 5 2‐Oct

Mounting the Antennas and Hardware

Antennas were mounted using methods appropriate for the site and the antenna style. In the case of FRP antennas, FRP angle was epoxied to the sides of the antenna and used as mounting brackets. For the conduit and flexible antennas, we purchased conduit mounting brackets. In both cases concrete wedge‐anchors were used with cement epoxy. For the flat antenna, holes were drilled through the flat sheeting and then anchored into place.

Each ACN was placed as close to an antenna as possible to reduce noise and increase current through the antenna. At all antennas placed at LD1, LD2, and LD3 locations, water resistant lockers were mounted to house the ACNs to allow easy access. For antennas installed in the spillway (LD4), ACNs were either potted in polyurethane and submerged along side of the antenna or mounted on top of cement pillers in the middle of the spillway.

The ACNs were synchronized by an IS‐1001 Master Controller connected to a 24V‐DC system. A single master controller was used at each interrogation site except for at the spillway where two master controllers were used since the number of antennas at this site (n‐15) more than the number of antenna a single master controller can operate (n‐12). The batteries at these sites were charged off AC grid power. We used a battery switcher to isolate the system from grid power. Battery banks were alternated between charging and power so as to isolate external electrical noise. For antennas located on the both adult fish ladder and on the spillway data is offloaded through a B&B Electronics 3G‐wireless router to an offsite location and then uploaded to PTAGIS daily. Data from the antennas located on the diversion bypass were downloaded manually since no wireless signal was in range from this site.

Estimating Detection Efficiency

Individual PIT tag antennas can miss tag‐detections for a number of reasons such as noise interference, tag orientation or multiple tag interference. Different antenna designs may also have varying detection ranges. Therefore we calculated the probability of tag detection at each antenna and then calculated the overall detection efficiency of each array, for the most commonly used 12mm PIT tag. We queried data from PTAGIS to use for analysis from the installation date until September 30, 2016.

Since antennas were installed in different arrangement at each interigations sties, the calculation used to determine PIT tag detection efficiency also varied veried between site. For instance, at the South Ladder antennas 1, 2, and 3 were side by side and antenna 4 was downstream of all three.

Since these first three antennas were adjacent, a fish moving through this ladder could have passed through either of these antennas. So, when determining detection efficiency these three antennas were grouped together. Antenna 4 had no antennas adjecnt to it and affectively monitored all water passing through that section of the ladder. Therfore any tagged adult fish passing through the south latter would have had to pass through antenna 4 as well as either antennas 1, 2, or three. At the north ladder interrogation site the antennas were arranged opposite that of the south ladder with antenna 2,3, and 4 being adjacent and representing one group of antennas and antenna 1 being upstream from these three. At the fish bypass diversion pipe, fish were funneled through both pass‐through antennas.

Therefore we made the assumption that once a fish entered the array it had an equal opportunity to be detected by both these antennas as the fish could not turn around or go around any of the antennas.

We adapted the formula in Connolly et al. (2008) for a two antenna system by first finding the following summed (S) detection histories:

South Ladder (LD1) antenna group 1

S123 = fish detected on antenna 1, 2, or 3 but not 4

S4 = fish detected on antenna 4 only

S1234 = fish detected on antenna 1, 2, or 3 and also on 4

North Ladder (LD2) antenna group 1

S1 = fish detected on antenna 1 only

S234 = fish detected on antenna 2, 3, or 4 but not 1

S1234 = fish detected on antenna 1 and also on either 2, 3, or and 4

To calculate the probability of detection we then found the following four values for each antenna, the following is for:

South Ladder (LD1) antenna group 1

N1 = All fish detected on antenna 1, 2, or 3 (S123+S1234)

N14 = Fish detected on antenna 1, 2, or 3 and also on 4 (S1234)

N4 = Fish detected on antenna 4 only (S4)

M1 = Fish estimated to missed by antenna 1 ((N1xN4)/N14)

North Ladder (LD2) antenna group 1

N1 = All fish detected on antenna 1 (S1+S1234)

N12 = Fish detected on antenna 1 and also on either 2, 3, or and 4 (S1234)

N2 = Fish detected on antennas other then 1 (S234)

M1 = Fish estimated to missed by antenna 1 ((N1xN4)/N14)

The detection efficiency (𝑃 ) was then calculated as:

𝑃

After this formula was adapted to the other antennas, the overall detection efficiency was calculated as:

P = 1‐[(1‐P1)(1‐P2)(1‐P3)]

We could not use the same method to find the detection efficiency for antennas installed along the Lebanon Dam spillway since there was only one row of antennas. Instead, a known number of tagged fish were realeased upstream of the dam and these detection histories were used to estimate detection efficiency at this side. These tagged fish were all juvenile spring Chinook Salmon released in three batches on 11/3/2015 (n‐489), 11/16/2015 (n‐515), and 11/23/2015 (n‐506). Since fish released from these group also had the opertunity to pass through both fish ladders and the bypass pipe instead of going over the spillway, estimates for each of these site were removed from the total fish available when estimating efficiency for the spillway. After removing these estimates of fish that passed through the other interigation sites a final estimate for detection efficiency was calculated for the spillway by deviding the number of fish detected at the spillway by the number of fish released from these tag groups.

Results

Efficiency of each system is displayed in the following tables, and was influenced by locatikon and life stage of fish that was interrogated.

Figure 7: The detection efficiency at four interigation sites located at the Lebanon Dam on the South Santiam River. Estimates at the South Fish Ladder (LD1), North Fish Ladder (LD2), and Diversion Bypass (LD3) were all caluculated using detections form efficiencies between the antenna arrays at these sites. To determine efficiency over the Spillway (LD4) three groups of PIT tagged juvenile spring Chinook Salmon were used.

Figure 8: The detection efficiency at the fish ladders on Lebanon Dam on the South Santiam River. Estimates include differences between adult and juvenile detections.

64.5 66.7

59.4

97.5

LD1 Juvenile LD1 Adult LD2 Juvenile LD2 Adult

D e te ct io n E ff ic ie n cy

Table 2: The salmon and steelhead detected at Lebanon Dam PIT tag array sites between April 1, 2015 and November 30, 2016, divided by year tagged. No distinction was made between fish tagged as adults or juveniles.

Number of Fish Detected Year Tagged

LD1 LD2 LD3 LD4

Unknown 45 19 0 0 2012 2 0 0 0 2013 4 2 0 0 2014 134 74 0 0 2015 530 104 26 66 2016 197 6 1 27 Total 912 205 27 93

Efficieny should be measured on a per study basis, given the effects of life stage and flow conditions that may influence the in‐situ effieiency of the interrogation system.

References

National Oceanic and Atmospheric Administration, National Marine Fisheries Service West Coast Region, 2016 5‐Year Review: Summary & Evaluation of Upper Willamette River Steelhead Upper Willamette River Chinook.

Prentice, Earl 2008. PIT‐Tag Antenna Construction Guide. Prentice & Associates.

Green‐Peter. US Army Corp of Engineers, Portland District (Dec 2016). Availible:

http://www.nwp.usace.army.mil/Locations/Willamette‐Valley/Green‐Peter/

Foster. US Army Corp of Engineers, Portland District (Dec 2016). Availible:

http://www.nwp.usace.army.mil/Locations/Willamette‐Valley/Foster/

INFORMATION ON THE SOUTH SANTIAM WATERSHED. South Santiam River Watershed Council

(Dec 2016). Availible: http://www.sswc.org/about/watershed/