Transcript Slide 1

Biological and environmental
factors influencing recruitment
success of North Sea demersal
and pelagic fish stocks
Alan Sinclair
Fisheries and Oceans Canada
Pacific Biological Station, Nanaimo, BC
Laurence Kell and Georgi Daskalov
CEFAS Lowestoft Laboratory, UK
Motivation
North Sea stocks are assessed on a single stock basis
• However fishing fleets exploit a range of species
– For example cod are taken by many gears and as
a bycatch in various non-target fisheries.
• It is important therefore to look at whether stocks
vary together and how environmental factors
influence the main commercial fish stocks
• Since this has important implications both for yields
to the various fishery sectors and for the
management of the North Sea fisheries
The Main Question
Clearly there must be spawners (S) to have recruits (R)
– However, inter-annual variability in recruitment far outweighs
variability in spawners
How Do Environmental Conditions Affect Recruitment?
• Do environmental conditions determine
– the number of recruits, regardless of spawning stock size?
– or the juvenile survival rate (R/S)?
• Is recruitment affected by biological processes such as predation,
competition or spawning condition?
• Is recruitment affected by physical processes such as
temperature, salinity …?
Main North Sea Commercial Stocks
herring
1.0
0.5
2.5
2.0
3.0
3.0
1.5
sandeel
1.5
2.0
2.0
1.0
0.5
0.0
0.0
0.0
0.5
0.5
1.0
1.0
1.5
1.5
1.0
0.5
2.5
2.5
saithe
2.0
2.5
plaice
0.0
3
2
1
0
1.5
1.0
0.5
3.0
0.0
0.5
1.5
1.0
2.0
4
0.0
1.5
1.0
0.5
0.0
2.0
w hiting
1.0
3
1.5
sole
0.5
2
0.0
Recruitment
0.0
0
0.0
2
0.5
1.0
4
1.5
1.5
6
2.0
2.0
2.5
haddock
8
cod
1
Plaice
Sole
Cod
Haddock
Saithe
Whiting
Sandeel
Herring
0
•
•
•
•
•
•
•
•
2.5
North Sea Species Stock/Recruitment
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
Spawning Biomass
0.0
0.5
1.0
1.5
Species Likelihood Ratio
cod
16.93
haddock
0.05
herring
32.43
plaice
1.36
saithe
1.66
sandeel
0.07
sole
-0.06
whiting
3.78
2.0
Stepwise Analysis of Environmental
Effects Using Likelihood Ratio Test
Null: Recruitment varies around a mean
R  e
Recruitment is density dependent
Following a BH relationship
S 
R
e
 S
2
R  e
Recruitment is density dependent
With an Environmental Component
2
Recruitment is density independent
With an Environmental Component
E  2 / 2
/2
S E 
R
e
 S
 2 / 2
/2
R = recruitment
S = spawning stock biomass
E = environmental covariate
α = maximum recruitment
β = biomass at ½ max recruitment
κ = environmental parameter
σ = residual standard deviation
Biotic Hypotheses
• Competition / Predation
– Recruitment of one species is negatively affected by R or S
of another species in the year of spawning.
• Juvenile feeding
– Recruitment of a pisciverous juvenile (cod, plaice, saithe,
whiting) is positively affected by R of a suitable prey
species (herring, sandeel) in the year of spawning.
• Feeding and spawning fitness
– Recruitment of cod, haddock, plaice, saithe, sole or whiting
is positively affected by R of any species in the year prior to
spawning.
– Recruitment of cod, saithe or whiting is positively affected
by S of herring or sandeel in the year prior to spawning.
Abiotic Hypotheses
• North Atlantic Oscillation (NAO)
– A single annual mean NAO index was used
– Positive or negative effect of NAO on recruitment
• Temperature
– May act on different part of life history, therefore
temperature variables were created from monthly
time series and from North Sea sea surface
temperature (SST)
– Effect of temperature variables may be positive or
negative
Temperature Variables
• SSTY
– Sea surface temperature annual mean anomaly
• Q1Y, Q2Y, Q3Y, Q4Y
– Quarterly mean anomalies in year of spawning
• Q1Y-1, Q2Y-1, Q3Y-1, Q4Y-1
– Quarterly mean anomaly in the year prior to spawning
• SSTDJF:
– Mean winter anomaly, Dec Y-1, Jan Y, Feb Y;
• SSTFJ
– Mean anomaly Feb – July
• PC1, PC2, PC3
– First 3 principle components
PCA of Monthly Sea Surface
Temperature
PC A Monthly Sea Surface Temperature
Fi rst 3 C omponents
0.75
Y
0.50
PC1
0.25
PC2
0.00
PC3
-0.25
-0.50
Month
Sep
Oct
Nov
Dec
May
Jun
Jul
Aug
-0.75
Jan
Feb
Mar
Apr
Eigenvector
• PC1 ~ 54% of Variance
– annual signal
• PC2 ~ 16% of Variance
– contrast between
first and second
half of year
• PC3 ~ 10% of Variance
– contrast between
summer and winter
temperature
Results
A preliminary first cut at an analysis of this
type
A broad look at how the North Sea
commercial fish species vary together
and the possible mechanisms
Cod 1963-2002
Model
Base
Sigma Alpha Beta
0.668 1.012
BH
0.541 2.929 1.765
BH + PC1
0.436 1.053 0.142
Term Estimate
PC1
-0.170
Chi
p
16.926
0.0000
17.135
0.0000
• Recruitment density dependent
• Negative effect of SSTemp (PC1) on recruitment
– this has been noted by Planque and Frédou 1999
among others
α = maximum recruitment
β = biomass at ½ max recruitment
Cod 1983-2002
Model
Base
Sigma Alpha Beta
0.610 0.674
BH
0.535 4.636 3.325
BH + R1San
0.356 3.066 4.421
BH + PC1
0.450 0.867 0.018
BH + R1San + PC1
0.335 0.873 0.586
Term Estimate
Chi
p
5.266
0.0218
αR1San
= maximum
recruitment
0.582
16.277 0.0001
β = biomass at ½ max recruitment
PC1
-0.188
6.862
0.0088
R1San
PC1
0.497
-0.086
11.869
2.453
0.0006
0.1173
• Positive effect of Sandeel Recruitment (R1San) and negative
effect of SST (PC1)
• The PC1 term not significant in model with both terms
• The stock/recruitment parameters are very sensitive to the
environmental effect
• Resolving which environmental effect is operating is important
for interpreting stock/recruitment dynamics
Haddock 1963-2002
Model
Base
Sigma Alpha
1.079 0.981
Term Estimate
Chi
p
R1Sol
0.953 1.592
R1Sol
-0.617
9.925
0.0016
S1Her
0.978 1.600
S1Her
-0.646
7.826
0.0052
R1Sol + S1Her
0.910 1.979
R1Sol
S1Her
-0.474
-0.435
5.785
3.686
0.0162
0.0549
• Cannot reject density independent recruitment hypothesis (i.e.
no evidence of a significant stock recruitment relationship)
• Negative effect of sole recruitment (R1Sol) and herring
spawning biomass (S1Her)
• The S1Her term not significant in model with both terms
• High residual standard deviation (Sigma) regardless of model
Sole 1957-2002
Model
Base
Sigma Alpha
0.770 1.001
Term Estimate
Chi
p
PC2
0.681 0.954
PC2
0.276
11.254
0.0008
R1Pla
0.694 0.483
R1Pla
0.673
9.596
0.0020
PC2 + R1Pla
0.634 0.549
PC2
R1Pla
0.224
0.518
8.207
6.549
0.0042
0.0105
• Cannot reject density independent recruitment
hypothesis Positive effect of contrast in SST
between winter and summer (PC2)
• Positive effect of plaice recruitment (R1Pla)
• Both effects significant in 2-parameter model
Herring 1967-2002
Model
Base
Sigma Alpha Beta
0.924 1.077
BH
0.632 2.568 0.962
BH + S1Sai
0.530 3.092 0.323
Term Estimate
S1Sai
-0.785
Chi
p
27.396
0.0000
12.626
0.0004
• Recruitment density dependent
• Negative effect of Saithe spawning biomass (S1Sai)
on recruitment
Plaice
Model
1957-2002
Base
Sigma Alpha
Term Estimate
Chi
p
0.417 0.992
SSTFJ
0.371 1.053
SSTFJ
-0.385
10.673
0.0011
1983-2002
Base
0.471 1.003
R1San
0.401 0.680
R1San
0.359
6.380
0.0115
SSTFJ
0.370 1.275
SSTFJ
-0.617
9.642
0.0019
R1San + SSTFJ
0.336 0.938
R1San
SSTFJ
0.240
-0.496
3.815
7.077
0.0508
0.0078
• Cannot reject density independent recruitment hypothesis
• Negative effect of SSTemp Feb-Jun (SSTFJ) for 1957-2002 and
1983-2002 periods
• Positive effect of Sandeel recruitment (R1San)for 1983-2002
period
• Sandeel recruitment not significant (barely) in model with both
Summary of Biological Effects
Hypothesis
Comp/Pred
Juv Surv feeding
Covariate
Recruits cod
haddock
herring
plaice
saithe
sandeel
sole
whiting
Spawners cod
haddock
herring
plaice
saithe
sandeel
sole
whiting
Recruits
Dependent Variable
cod
haddock
-
herring
sandeel
+
+
cod
haddock
herring
plaice
saithe
sandeel
sole
whiting
Spawners herring
sandeel
+
+
+
+
+
+
+
+
+
Spawning Fitness
Recruits
year before spawning
herring
-
plaice
saithe
sandeel
sole
whiting
A very large number of plausible biological
hypotheses
were
tested
involving
competition,
predation,
- and feeding of
feeding
of- juveniles
- It was- surprising
spawners.
how -little
evidence
to support
any
-was found
- of
these.
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Summary of Temperature and NAO
Effects
Hypothesis
Temperature
NAO
Covariate
pc1
pc2
Feb-Jun
Q1
Q4
Q3
Q4
Dec-Feb
Annual
Q1 y-1
Q2 y-1
Q3 y-1
Q4 y-1
NAO
Dependent Variable
cod
haddock
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
herring
±
±
±
±
±
±
±
±
±
±
±
±
±
plaice
±
±
±
±
±
±
±
±
±
±
±
±
saithe
±
±
±
±
±
±
±
±
±
±
±
±
±
sandeel
±
±
±
±
±
±
±
±
±
±
±
±
±
sole
±
+
±
±
±
±
±
±
±
±
±
±
±
whiting
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
Temperature effects may be important for cod, plaice and sole recruitment. For cod and
plaice the effects are negative and related to annual or seasonal temperature values. For
sole, recruitment was best in years with high seasonal contrast in temperatures.
The NAO did not enter in any of the ‘best’ models. However, there is a strong correlation
between the NAO and SST, especially SST in the first quarter. Thus, colinearity may be
masking important relationships with NAO.
Summary and Additional Questions
• Why do there appear to be so few significant
relationships
• Is there ancillary information to support the findings
through diets, laboratory study, earlier publications, etc.?
• Are some of the ‘significant’ relationships obviously
spurious or at odds with accepted conditions in the North
Sea?
• Are there other mechanisms that should be investigated?
– For example, is it temperature or Sandeel that affects
cod recruitment?
• What are the implications of specific ‘environmental’
relationships for management targets and limits