Transcript Presentation ()

Estimates of Steelhead Harvest
Rates Based on PIT Tag Sampling in
Columbia River Fisheries
Steve VanderPloeg &
Dan Rawding
Outline
•
•
•
•
Introduce fishery sampling for PIT tags
Summarize detection study
Report on basic sampling statistics
Example for harvest rate calculations
– Preliminary steelhead harvest rates based on
Zone 6 sampling and BON (Bonneville Dam)
detections for selected tag groups
– More work to be done
Fishery Sampling
• BPA funded WDFW to sample mainstem
Columbia River fisheries to fill a gap in the PIT tag
recovery program.
• Sampling program includes sport, commercial,
and treaty fisheries in the Columbia River.
• Initiated in August 2010 with commercial and
treaty fishery sampling.
• Sport sampling was added in Feb 2011.
Fishery Sampling Goals
• Report fishery recovery
information to PTAGIS
• Expand PIT tag recoveries
to estimate harvest of PIT
tag groups
• Estimate harvest rates of
PIT tag groups
Data for Harvest Estimates
• Build on existing CWT
recovery program
• # of harvested fish with
PIT tags
• Detection rate for
presence of PIT tags
• # of harvested fish
sampled for PIT tags
• Total number of
harvested fish
2010 Hatchery Study for determining
detection rates for presence of PIT tags.
Models (#)
Psion Teklogix
data logger with
RFID – LF module
(1)
Destron Fearing
FS2001F-ISO
'Cheese Block' w/
Racket with Racket
antenna (4)
Biomark Pocket
Reader (1)
All Flex RS601-3 (4)
Models (#)
Destron Fearing
FS2001F-ISO
'Cheese Block' w/
Biomark Flat Plate (1)
Destron Fearing
FS2001F-ISO
'Cheese Block' w/
Biomark 24” square (1)
Study Design
• PIT tag 130 - 200 adult steelhead,
coho, and Chinook salmon at
Skamania and Kalama Hatcheries
• Hold salmon for 7 days and
steelhead for 30 days, then
sacrifice fish
• Follow PIT tag sampling protocols
Scanning a Fish for a PIT tag
Hand-held detectors PIT sampling protocol
• Sacrifice fish and lay
groups of 50 fish out
on a series of tables.
• Space between fish ~
18 inches to avoid tag
collision.
• At the end of sampling, the 50 fish were
passed over the flat
plate and through the
24 “ square in the same
order (1,2,…,50)
• First sampler uses 'Cheese
Block' w/ Racket (>99%
detection rate) & recorder
takes notes on presence /
absence of PIT tag for each fish
and records PIT tag #.
• Tag detection can now be
referenced for other samplers
with different models and
serial #.
• Multiple passes ensures
detection probability ~ 100%
for each fish.
24 “ Square antenna
• Variable detection rates; rates dependent on fish
orientation, distance from the edge of the antenna,
and speed of fish passing through.
• When sampling protocols were standardized;
a slide was used and fish
were passed head first
through the square, next to
the antenna: PIT tag
detection rates were 99.5%
and 100% for coho and
steelhead, respectively.
Flat Plate Detector
• Detection rates were high
but sometimes variable.
• Variability occurred when
samplers were in a hurry
(coho) & did not pass the
entire fish over the plate;
detection rate: 93.3%.
• When fish were passed
within range, detection
rates were 99.4% and 100%
for Chinook and steelhead
respectively.
Single Model Results for
Chinook
• Older/broken All Flex model had a detection rate
of 63%.
• Older/broken model Pocket Reader had a
detection rate of 6%.
• Psion Teklogix data logger with RFID had a
detection rate of 98.9%
Salmon
Coho PIT Tag Detection Rates with
95% CI
Chinook PIT Tag Detection Rates
with 95% CI
100%
Detection Probability
99%
98%
97%
96%
99%
98%
Unit
Unit
SQ
4
CB
#1
CB
#2
CB
#3
CB
#4
AF
AF
AF
FL
CB
#1
CB
#2
CB
#3
CB
#4
#4
AF
#3
AF
#1
1
95%
3
97%
AF
Detection Probability
100%
Steelhead
• Chinook and coho test ~ 100% of the fish
in the sample were tagged. For
steelhead less than 50% of the sample
was tagged which is more similar to field
conditions.
• Found similar detection rates (> 99%) as
observed for Chinook and coho salmon
with All Flex, 'Cheese Block' w/ Racket,
24” square, and flat plate.
Detection Conclusions
• Standardization of sampling protocols allowed for
repeatable results between samplers and units.
• Field Protocol = two pass method
• Other results
– No difference in detection rates based on tag location
– No difference in detection rates between species
– No difference in detection rates between samplers
Fall Treaty Sampling
• 2010 only includes fall
fishery, not C&S, sockeye, or
spring, summer, and winter
fisheries
• Fall harvest rates underestimates harvest for
steelhead because they are
also caught in other
fisheries
• Assumes commercial fish
sales are representative
(same as CWT assumption)
• Assumes PIT tag groups are
representative of untagged
fish
Fall Commercial and Treaty Sampling
• Fall Commercial and Treaty (Zone 6) sampling in 2010
• 296 tags reported to PTAGIS
• Most with bio-data length, species, mark status
Zone 6 Summary
Species
BON Count BON PIT Z6 Catch Z6 Sampled % Sampled
F Chinook
530971
5807 138836
19350
13.94%
Coho
124484
660
9555
1318
13.79%
Steelhead
410390
9973
21308
3663
17.19%
Direct Acyclic Graph (DAG) for Estimation of Harvest Rates
#estimate number of fish based on tickets by period
a v g _ wt
mu _ wt
we i g h t s [ i ]
No .
Fi s h
t a u _ wt
l bs
f or ( i
I N 1 :
b i o s a mp l e s )
m_ s a mp l e s
Su m_ Gr o u p [ i ]
p r o p _ ms mp l
Ha r v Ra t e
No .
BON_ Gr o u p
Estimate Harvest
Rate by Group
Gr o u p [ i ]
Ad j _ GT a g s [ i ]
p_det
PI T _ Gr o u p [ i ]
#Estimate expanded PIT
Tags by sampling period
n
r
#estimate detection rate
Of PIT tag readers
Steelhead PIT tags at BON, 2010
• From Apr 1 to Oct 31 there were over 7000
individual adult steelhead detections at BON
• 141 tag sites contributed to these adults
passing at BON
• Adult returns by release site ranged from a
low of 1 to a high of 1,658 from LWG release
• For analysis, individual tag groups were
grouped together and analyzed individually
Recommendations
• WDFW successfully built on CWT sampling
infrastructure to add other sampling (genetic, PIT
tags, etc.).
• Recommend ODFW get funded for similar effort,
which should increase PIT tag fishery recoveries.
• Recommend Zone 6 catch to be sampled and
reported by pool.
• Coordination with Accord treaty PIT tag sampling
proposal because C&S and OTB sales are not
currently sampled for PIT tags.
Summary
• Individual tag groups are often small, which can lead to
imprecise & possible unrepresentative harvest rates
• Framework to estimate harvest rates by group:
– Pooled Analysis DPS & run timing at BON
– Alternate Approach – Hierarchical Model
• All data are preliminary & additional work is needed
– Zone 6 over the bank sales variance was not available
(2010)
– In 2011, sampling included Sport, commercial and
treaty
Acknowledgements
• Various WDFW, ODFW, IDFG, and PSMFC biologists for sharing
experiences with portable PIT tag sampling equipment
• Joe Hymer and Vancouver PSMFC & WDFW sampling staff for
the implementation of commercial and treaty sampling.
WDFW, PSMFC, and NOAA staff that assisted in this PIT tag
detection study
• Michelle Groesbeck (WDFW) and Bob Woodard (WDFW) for
detection study and for recommendation on data collection
and ultimately data logger programming and database design.
• Dave Marvin (PSMFC) and Rick Golden (BPA) for support of
this project.
• US v. OR TAC (Robin, Stuart, & Rodger) for supplying
preliminary Zone 6 Fall harvest data. Doug Case (ODFW) &
Ron Roler (WDFW) for landings & biosample data. Alan Byrne
& Jay Hesse for help with Salmon & Clearwater groups