Stocking Brook Trout in Virginia’s reservoirs:
dispersal,
naturalization, and introgression
p
g
with native trout populations
R. Humston, K.A. Hemminger, N.D. Adkins (W&L ‘10) , R.J. Elsey (W&L
‘11), J. Huss (W&L ‘10), B.A. Meekins (W&L ‘11), P.R. Cabe, and T.L. King†
Washington & Lee University, Department of Biology
† U.S. Geological
Survey, Leetown Science Center, Kearneysville WV
Stocking Hatchery Trout
Stocking Hatchery Trout
y
Positives:
◦ Reduces fishing pressure on wild (native)
populations.
◦ Restore populations where extirpated.
◦ More cost effective than translocation of fish
f
from
intact populations.
l
y
Negatives:
◦ Non-native species / strain introductions
◦ Suppression of native stocks and genetic
impacts
Genetic impacts of hatchery
supplementation
y
Poor fitness of hatchery lineages
y
Homogenization of genetic structure
among populations.
y
Loss of functional diversity in native /
wild stocks (e.g. local adaptation).
y
Potentially can hinder recovery,
persistence of wild stocks
- e.g. Marie et al. 2010 Molecular Ecology v.19
Direct / Indirect Genetic Impacts
y
Direct
◦ Results of interbreeding
◦ Reduced genetic variability among populations
(homogenization)
◦ Outbreeding depression and reduced fitness of
subsequent generations
y
Indirect:
◦ Reduced effective population size (Ne)
◦ Increased inbreeding,
inbreeding genetic drift
Brook trout stocking in VA
mountain reservoirs
Brook trout stocking in VA
mountain reservoirs
Put-and-grow
g
stockingg program
p g
y Unique, trophy fisheries in wilderness
sett gs
settings
y Open questions:
y
◦ Dispersal into feeder streams
◦ Reproduction / naturalization of stocked strains
◦ Hybridization with native populations
Study Objectives:
1))
Determine if stocked brook trout
disperse into and inhabit tributaries.
2)
Determine if natural reproduction of
hatchery-origin fish occurs in tributaries.
3)
Assess degree
A
d
off introgression
i
i
(interbreeding) occurring between
h h
hatchery
and
d native
i strains.
i
Study Sites - 2009
Moore Creek Reservoir
• Historic water supply impoundment
for Lexington.
• Stocked with brook trout since 1984.
• Moore Creek (feeder tributary)
did not hold brook trout prior to
stocking.
g
• Code name: STO
• Historic water supply & flood
control impoundments.
• Stocked with brook trout since
1989
1989.
• Wild (putative native) brook trout
populations in feeder streams.
• Code named: MIX-1 & MIX-2
Coles Run Reservoir
(MIX-2)
Mills Creek Reservoir
(MIX-1)
Kennedyy
Creek
•U
Unstocked
t k d wild
ild ((native)
ti ) ttroutt
stream.
• “Reference” native population for
subwatershed.
b t h d
• Code name: NAT
Stocking History
y
All fingerlings stocked in reservoirs from
a single brood stock lineage (Paint Bank
hatchery,VA).
y
Only exception:
◦ Fingerlings stocked in 2008 derived from
Rome Lab strain.
◦ This cohort would not have spawned in time
to contribute offspring to population by
summer 2009.
Study Objectives:
Determine if stocked fish disperse
p
into and
inhabit tributaries.
◦ Sample Moore Creek (STO; no previous wild
population) for trout.
◦ Use genetic data to distinguish hatchery
hatchery-strain
strain
from native trout in Mills Creek (MIX-1) and
Coles Run (MIX-2).
(
)
Study Objectives:
Determine if natural reproduction
p
of
hatchery-origin fish occurs in tributaries.
◦ Compare genotypes of 2008 year class with
known hatchery strains (Paint Bank, Rome Lab) to
determine if Paint Bank strain naturally produced
offspring in 2008.
Study Objectives:
Assess degree
g
of introgression
g
occurringg
between hatchery and native strains.
◦ Use “assignment
assignment tests”
tests to identify potential
‘hybrids,’ and examine individual genotypes to
determine likelihood of hybrid status.
Field collections
J
June-July
J y 2009
y Single-pass electrofishing
y Pelvic fin snips
y Measure in field (TL)
y
Hatchery samples
Paint Bank brood stock
and 2009 young-of-year
y Pelvic
e v c fin snips
s ps
y Rome Lab genotyped in 2005 and 2009
(TL King
(T.L.
King, unpublished data)
y
DNA analysis
y
Genotyped
yp at ten (10)
( ) established,
polymorphic microsatellite loci (T.L.
King,
g unpublished
p
data))
◦ Non-coding regions of DNA
p
sequences
q
of 1-6 specific
p
base ppairs.
◦ Repeated
◦ Alleles vary by length (number of repeats in a
row).
◦ Highly versatile genetic markers
Data Analysis – Genetic Data
y
Program
g
Structure
◦ Groups individuals into “clusters” that best
approximate distinct populations based on
allele frequencies.
◦ Provides posterior probability of population
membership for individuals (population
assignment).
How to read Structure output
The p
probabilityy of membershipp to each
population is represented in a ‘stacked’ bar
chart with a bar for each fish.
How to read Structure output
Paint Bank Hatchery strain
Rome Lab Hatchery strain
Native strain
33.3% Paint Bank
33% Paint Bank
100% Native
33.3% Native
67% Native
33.3% Rome Lab
MIX1-108
MIX1-109
MIX1-110
Assignment “Rules”
Rules
Paint Bank Hatchery strain
Rome Lab Hatchery strain
Native strain
8% PB
90%
18% PB
94% PB
92% Native
82% Native
6% Native
MIX1-111
MIX1-112
MIX1-113
Results:: STRUCTURE
Results
y
Best solution = 3 p
populations
p
in collection
Results:: STRUCTURE
Results
Hybrids
Results:: Dispersal into Tributaries
Results
Sample
Location
STO
MIX-1
MIX-2
NAT
Paint Bank
Rome Lab
Total
Size Range
Native
(mm TL)
38
66 180
66-180
0
54
47-235
43
44
46-245
33
40
41 216
41-216
100%
30
74-321
45
NA
251
41-321
N
No Native Pop’n:
STO
Paint
Bank
38
7
10
Hybrids
0
4
0
100%
100%
Mixed Native / Stocked:
MIX-1
MIX-2
Rome Lab
Paint Bank
Rome
Lab
1
0
1
Rome Lab
Native
Paint Bank
Rome Lab
Admixed
Native
Native
Results:: Hatchery Strain Spawning
Results
All age-1 stocked fish should be Rome Lab strain.
Age-0
40
Age-2
Age-1
Moore Crk -> Paint Bank Hatchery
STO Æ Paint Bank
Coles / Mills -> Paint Bank Hatchery
Coles / Mills -> Native
MIX Æ Paint Bank
Frequency
30
MIX Æ Native
20
10
0
0
50
100
150
Length (mm TL)
200
250
Conclusions
y
Hatcheryy fish do disperse
p
into streams
from reservoirs, even where native
populations
p
p
exist.
◦ Not a big surprise.
Conclusions
y
Hatcheryy fish do reproduce
p
in streams
even where native populations exist, and
can establish ‘naturalized’ populations.
p p
◦ Promising for re-establishment of extirpated
stocks.
◦ Suggests fitness of hatchery stocks may not be
an impediment to interbreeding.
Conclusions
y
Introgression
g
has been limited over the
20+ year stocking history in these
systems.
y
◦ Assortative mating?
Richards et al. 2008
Conclusions
y
Good news scenario:
◦ Direct impacts to “native” genotype has been
minimal.
y
Potential issues:
◦ Naturalized hatchery strains could lead to
future introgression
g
in native populations.
p p
New Questions:
Stream residents or lacustrine migrants?
g
y Timing of upstream dispersal?
y Ratio of dispersers : stayers?
y Exclusionary processes?
y
Indirect genetic effects?
y Environmental factors, e.g. stream pH?
y
Acknowledgements
Larryy Mohn, Paul Bugas,
g Jason
J
Hallacher,
Aaron Coffman (VA DGIF)
y Dawn
aw Kirk (USFS)
(US S)
y Barb Lubinski (USGS)
y Eric Hallerman,
Hallerman Joanne Davis
Davis, Jamie
Roberts (VA Tech Fisheries & Wildlife)
y Funding
F di ffrom Hess
H
Scholars
S h l Program
P
(W&L), Mellon Foundation
y
Results: Hatchery / Native
Results:
Introgression
Paint Bank
0
10
100
90
20
80
30
70
40
60
50
50
60
Native
Paint Bank
Rome Lab
Hybrid
70
80
40
30
20
90
10
100
Rome Lab0
0
10
20
30
40
50
60
70
80
90
100
Native
Data Analysis – Length Data
y
MIXDIST p
package
g in R
◦ Age structure and age-specific size
distributions.
◦ Estimates average size-at-age and standard
deviation from collective length data.
◦ Assumed 4 age classes in collection (0 – 3)
Histogram of Lengths in Collections
10
F
Frequency
8
6
4
2
0
0
50
100
150
Total Length (mm)
200
250
300
Results
from MIXDIST
Histogram
of Lengths
in Collections
2008 Year Class
Power Analysis
y
Use data from known (NAT, PB) strains to
generate hypothetical populations:
◦
◦
◦
◦
y
Parent generation
F1, F2 generations
back-crossings
HYBRIDLAB Nielson
HYBRIDLAB,
Ni l
et al.l 2001
Analyze
A
l
simulated
i l d populations
l i
using
i
Structure, determine frequency of missed
assignments.
assignments
Power Analysis
y
Correct identification of parent
p
generations (NAT, PB strains): 100%
y
Correct identification of F1, F2, 1st B-C
generations as hybrids: ~100%
100%
y
Withh our sample
Wi
l sizes,
i
we would
ld d
detect
introgression as low as 3% of population
with
ith 95% percentt accuracy.
Hatchery supplementation and
native brook trout
y
Previous studies mixed, but generally
suggest levels of introgression between
native and hatchery brook trout are low.
◦ Possible exception: hybridization between
northern / southern strains, Great Smokey
Mountains
How to explain low introgression
despite widespread stocking?
y
Low fitness of hatchery stocks
Fraser 1989
How to explain low introgression
despite widespread stocking?
y
Genetic differences between native and
hatchery lineages may be small (difficult to
distinguish strains)
y
Technology may be inadequate to detect
introgression (poor “markers”)
y
Context-dependence;
p
; different strains,,
different systems