Spatial variation in landlocked Atlantic Salmon smolt survival associated with dam passage, avian predation, and stocking location

Heim, Kurt 1 ; Ardren, William1 ; Withers, Jonah1 ; Eisenhauer, Zachary1 ; Mensinger, Matthew2; Castro-Santos, Theodore3

Published Dec 10, 2024 on Dryad. https://doi.org/10.5061/dryad.vt4b8gv25

Data files

Dec 10, 2024 version files 1.45 GB

20241016_DetectionData_2021.csv

1.20 GB
20241016_DetectionData_2022.csv

250.45 MB
20241016_DetHist2021_V2.csv

22.13 KB
20241016_DetHist2022_V2.csv

18.09 KB
20241016_RecieverLocationsV2.csv

1.12 KB
README.md
22.05 KB

Abstract

Objective: Managers of Atlantic Salmon populations (Salmo salar) often stock smolts at downstream locations to minimize within-river mortality. However, downstream stocking may limit imprinting potential and homing ability of returning adults relative to stocking further upstream. We evaluated the survival differences between upstream and downstream stocking for landlocked Atlantic Salmon in a tributary to Lake Champlain.

Methods: We radio tagged hatchery-reared smolts, monitored these for a 7-8 day holding period, then stocked them concurrently with 22,000 smolts at two release sites in two years. The downstream location (DS, river kilometer 16, no dam passage) was a historically used site in a dam tailrace, whereas the upstream site (US, river kilometer 27, two dams) was in a side-channel and stocked for the first time. We estimated survival, performed avian counts during stocking, and searched avian nesting colonies for transmitters.

Results: Within their respective stocking reaches, survival per-kilometer was markedly lower for the DS release group compared to the US release group in both years (US 2021 and 2022 = 0.98, 0.98; DS 2021 and 2022 = 0.82, 0.69). At the DS stocking site, we documented a tenfold increase in avian predators (predominantly Larus delawarensis) in response to stocking, whereas no increase was detected US. Dam passage rates were generally high (> 0.96), but post-passage survival was lower at the second dam encountered (first dam per km survival = 0.96,0.97 vs. second dam = 0.78,0.82). Surprisingly, cumulative survival to Lake Champlain was higher for the US release group in 2021 (0.43 vs. 0.32), despite a longer migration and dam passage. In 2021, the downstream group performed better (DS = 0.55, US = 0.34). Based on transmitter detections at a nearby nesting colony, at least 20% (2021) and 7% (2022) of fish that reached the lake (i.e., successful migrants) were later consumed by birds.

Conclusion: Upstream stocking did not consistently result in lower cumulative survival, likely due to predators habituated to annually re-occurring stocking at the DS site in a dam tailrace that increased stocking-related mortality. We highlight the importance of critically evaluating historically used stocking sites, as substantial loss of smolts could be avoided by simple changes to stocking practices. Additionally, avian predation is a major source of smolt mortality, necessitating further studies to understand and address within Lake Champlain.

README: Spatial variation in landlocked Atlantic Salmon smolt survival associated with dam passage, avian predation, and stocking location

https://doi.org/10.5061/dryad.vt4b8gv25

Description of the data and file structure

Tagging fish

Fish were tagged at Dwight D. Eisenhower National Fish Hatchery, observed for a period of 7-8 days, and those surviving the holding period were mixed with the full lot of hatchery smolts into trucks for stocking (Table 1, Heim et al. in press). Fish were tagged with the methods described in Liedtke et al. (2012) using transmitters (TX-PSC-I-27-D, 1.5 grams, 16.5 x 8.0 x 6.5 mm, Sigma Eight, Ontario CAN) transmitting on 164.340 and 164.480 MHz at a 4 second ping rate. Only fish weighing more than 30g were tagged to ensure a tag burden of less than 5%, and the realized tag burden averaged 3.2% (max = 4.9%, min = 1.6%, Heim et al. 2024).

Mobile tracking fish

Fish were tracked in the lower Winooski River from a boat, using a Lotek SRX800 connected to a 3-element Yagi Antenna.

Fixed telemetry stations

Fixed stations were set up at 9 sites along the river, each included a power source, an antenna, and either a SRX800 (Lotek) or an Orion (Sigma Eight) that continuously scanned for transmitters. Data were downloaded regularly during the study period and archived as raw .TXT files in a central file repository and combined into CSV files for later use.

Files included

20241016_DetectionData_2022.csv

This file contains all detection data from 2022. It has data from mobile tracking , fixed SRX800 receivers, and fixed Orion Sigma Eight receivers. This includes many false positives that need to be removed - see methods in manuscript for approach to dealing with false positives. We generally used hitRatio (defined in Nebiolo, K., and T. Castro-Santos. "MAST: Movement Analysis Software for Telemetry data. Part I: the semi-automated removal of false positives from radio telemetry data (vol 10, 2, 2022)." ANIMAL BIOTELEMETRY 12.1 (2024) ) to remove false positives from Orion receiver detections. Very few false positives were generated via Lotek receivers so these were treated manually by viewing detections of movement plots.

Variable	Description
recID	The identification code of the receiver. This is an integer value with NA for mobile tracking detections.
Date	The date of the detection. Units are month/day/year
Site	The identification code of the receiver - this is the same as recID. This is an integer value with NA for mobile tracking detections.
Freq	Frequency of detection. Units are MHz with NA for mobile tracking detections.
Code	The unique ID code of the transmitter. Units are integer with NA for mobile tracking detections.
Power	The RSS for mobile tracking records (received signal strength, unitless measure of signal power). For Orion detections this value is decibels of detections reported by the unit with units of decibels.
ID	The unique ID of the tagged fish which is a combination of year of release transmitter frequency and CODE. Unit is a character / factor level variable.
DT	Date Time with following units e.g. "2022-04-05 11:28:43"
nHitz	Number of detections (of that particular freq-code) counted in a time span looking forward and backwards 1 minute from the present detection date time. Units is a count (integer)
recType	Either mobile srx , orion, or SRX . Units is character / factor level variable.
Name	A written description / name for each site. Unit is character
RKM	Meters upstream (from the river confluence) to the detection. Units are meters.
recID_NAJFM	Receiver ID used in manuscript in NAJFM which ranges from 1 (most downstream) to 9 (most upstream) and NA for mobile detections. Units are integer.
HitRatio	Number of detections observed during a 120 second interval centered on the detection at hand/divded by the expected number of the transmitter was present and all detections were recorded. Units are ratio scale (i.e., proportion).

20241016_DetectionData_2021.csv

This file contains all detection data from 2021. It has data from mobile tracking , fixed SRX800 receivers, and fixed Orion Sigma Eight receivers. This includes many false positives that need to be removed - see methods in manuscript for approach to dealing with false positives. We generally used hitRatio (defined in Nebiolo, K., and T. Castro-Santos. "MAST: Movement Analysis Software for Telemetry data. Part I: the semi-automated removal of false positives from radio telemetry data (vol 10, 2, 2022)." ANIMAL BIOTELEMETRY 12.1 (2024) ) to remove false positives from Orion receiver detections. Very few false positives were generated via Lotek receivers so these were treated manually by viewing detections of movement plots.

Variable	Description
recID	The ID of the fixed receiver. NA for mobile tracking data. Unit is character / factor.
FreqCode	Frequency and Code of detection. Unit is character / factor
timeStamp	Date Time , units e.g. "2022-04-05 11:28:43"
Power	Signal strength (RSS) of detection for those made using SRX800s, or , decibel of detection for detections made using Orion receivers. Units are RSS or decibel depending on the observation type.
RecType	Either mobile SRX for mobile tracking from boat, SRX for fixed station using SRX800, or Orion for fixed station using an Orion receiver. Factor / character variable.
HitRatio_A	HitRatio calculated by BIOTAS - is NA for Lotek receivers and for Orion detections not classified as REAL detections by MITAS software. Unit is ratio scale.
FishID	The Unique ID of the fish, unit is character.
DT	Date Time unit is eg. "2022-04-05 11:28:43"
Name	Name of the fixed telemetry station. Unit is character.
RKM	River Kilometer, unit is meters upstream (from the river confluence) to the detection. Units are meters.
recID_NAJFM	Receiver ID used in manuscript in NAJFM which ranges from 1 (most downstream) to 9 (most upstream) and NA for mobile detections. Units are integer.

20241016_DetHist2021.csv

This is a processed version of the dataset. Each individual fish detection history was scrutinized for false positives, and irregular movements associated with fish or bird predation. False positives were removed and detection history truncated if the fish was believed to be dead (in a bird or predatory fish). Following this, we determined if the fish was seen at each of the 9 stations (see last columns S1 - S9). These variables were used to build the detection histories.

Variable	Description
TagID	Unique ID of the fish. Character.
OutlierSession	Was this fish part of an outlier tagging session for surgeon B and thus should be removed prior to survival analysis (Y/N). Character.
TaggingDate	Date tagged. MM/DD/YY
ReleaseDateTime	The date and time of fish release into river unit is eg. "2022-04-05 11:28:43"
TL	total length (mm)
Weight	weight of fish (g)
ClipType	The mark applied to fish (LV/AD = left ventral fin and adipose fin, LV/AD/CWT = left ventral fin and adipose fin and coded wire tag). This is a character.
Tagger	The identify of the surgeon tagging the fish (A or B). Character.
SurgTime	Minutes and seconds of the surgery duration (e.g. ,MM:SS)
BS	Broodstock (MAX or TT). Two level factor/character.
Location	release location (W1D = the downstream location, E19 = the upstream location). Two level factor/character.
Comment	Any comments about the fish tagging procedure. text.
ReleaseLat	Latitude of release
ReleaseLong	Longitude of release
RKM_release	River kilometer of release (units are meters upstream from confluence of river).
S1	Seen at station 1 (0 or 1)
S2	Seen at station 2 (0 or 1)
S3	Seen at station 3(0 or 1)
S4	Seen at station 4 (0 or 1)
S5	Seen at station 5 (0 or 1)
S6	Seen at station 6 (0 or 1)
S7	Seen at station 7 (0 or 1)
S8	Seen at station 8 (0 or 1)
S9	Seen at station 9 (0 or 1)

20241016_DetHist2022.csv

This is a processed version of the dataset for 2022. Each individual fish detection history was scrutinized for false positives, and irregular movements associated with fish or bird predation. False positives were removed and detection history truncated if the fish was believed to be dead (in a bird or predatory fish). Following this, we determined if the fish was seen at each of the 9 stations (see last columns S1 - S9). These variables were used to build the detection histories.

Variable	Description
TagID	Unique ID of the fish. Character.
OutlierSession	Was this fish part of an outlier tagging session for surgeon B and thus should be removed prior to survival analysis (Y/N). Character.
TaggingDate	Date tagged. MM/DD/YY
ReleaseDateTime	The date and time of fish release into river unit is eg. "2022-04-05 11:28:43"
TL	total length (mm)
Weight	weight of fish (g)
ClipType	The mark applied to fish (LV/AD = left ventral fin and adipose fin, LV/AD/CWT = left ventral fin and adipose fin and coded wire tag). This is a character.
Tagger	The identify of the surgeon tagging the fish (A or B). Character.
SurgTime	Minutes and seconds of the surgery duration. MM:SS
BS	Broodstock (MAX or TT)
Location	release location (W1D = the downstream location, E19 = the upstream location). Two level factor/character.
Comment	Any comments about the fish tagging procedure. Text.
ReleaseLat	Latitude of release
ReleaseLong	Longitude of release
RKM_release	River kilometer of release. Units in meters upstream from river confluence.
S1	Seen at station 1 (0 or 1)
S2	Seen at station 2 (0 or 1)
S3	Seen at station 3(0 or 1)
S4	Seen at station 4 (0 or 1)
S5	Seen at station 5 (0 or 1)
S6	Seen at station 6 (0 or 1)
S7	Seen at station 7 (0 or 1)
S8	Seen at station 8 (0 or 1)
S9	Seen at station 9 (0 or 1)

20241016_RecieverLocations.csv

This provides the information on where each station was located and how it was used for generating encounter histories for the CJS model.

Variable	Description
recID	The ID of the fixed receiver. Character.
recID_NAJFM	Name of receiver ID for use in CJS model. Integer.
SiteName	Name of the site. text
UTM_X	UTM X coordinate (zone 18). Units are meters - rounded to nearest KM as required by dryad for publication of animal locations that may be harvested. Contact authors if exact locations needed.
UTM_Y	UTM Y coordinate (zone 18). Units are meters - rounded to nearest KM as required by dryad for publication of animal locations that may be harvested. Contact authors if exact locations needed.
riverKM	River KM of the receiver. Units are meters upstream from river confluence.