a productivity-susceptibility risk analysis for the
TRANSCRIPT
A productivity-susceptibility risk analysis for
the effects of fishing on Arctic Char (Salvelinus
alpinus) stocks from the Nunavut Territory
Marie-Julie Roux* and Ross F. Tallman
Fisheries and Oceans Canada
Central and Arctic Region
*Present address :Current Address: National Institute of Water
and Atmospheric Research , 301 Evans Bay Parade, Greta
Point, Wellington, New Zealand
paucity of data
too large a place
too many stocks
Background
Problems
(�) “what needs to be done to address priority gaps”
195 Arctic Char350 FW and Anadromous
paucity of data
too large a place
too many stocks
prohibitive time and financial costs to gather standard
assessment data
$
Background
Problems
(�) “what needs to be done to address priority gaps”
paucity of data
too large a place
too many stocks
prohibitive time and financial costs to gather standard
assessment data
$charr
diversity
makes it difficult to generalize scientific information for the
species and related management plans/actions
Background
Problems
(�) “what needs to be done to address priority gaps”
charr diversity
Objective
(�) “for addressing of gaps for the scientific understanding and management of Arctic charr fisheries
charr modelling
solutions
from Gaps to Solutions
Life history traits determine stock productivity
life history traits
age at maturity (R) fecundity (R)growth rate (G)survival (M)
R = reproductive capacityG = growth
M = mortality = 1-survival
elements of productivity
(productivity ≈ resilience)
Conceptual approach
life history traitsstable population:
R + G = M
traits are in balance and the population persists
R G M
population trend = nil (0)
age at maturity (R) fecundity (R)growth rate (G)survival (M)
R = reproductive capacityG = growth
M = mortality = 1-survival
elements of productivity
(productivity ≈ resilience)
Conceptual approach
Life history traits determine stock productivity
Conceptual approach
Charr modelling using life history traits
life history traitsAM = age at maturity F = fecundityGR = growth rateZ = instantaneous mortality
R = reproductive capacityG = growth
M = mortality
elements of productivity
R MG
F AM GR Z
Conceptual approach
life history traitsAM = age at maturity F = fecundityGR = growth rateZ = instantaneous mortality
R = reproductive capacityG = growth
M = mortality
elements of productivity
R MG
F AM GR Z
H
H = harvest rate(management history and population response information)
Charr modelling using life history traits
Conceptual approach
life history traitsAM = age at maturity F = fecundityGR = growth rateZ = instantaneous mortality
R = reproductive capacityG = growth
M = mortality
elements of productivity
R MG
F AM GR Z H
H = harvest rate(management history and population response information)
explore and identify sustainable H rangeR + G – M > 0
a model using life history traits and harvest rates as input parameters to
predict population trends
Charr modelling using life history traits
Conceptual approach
• demonstrate the extent of inter-population variability in life-history traits and how this can be used for evaluating the relative vulnerability of Arctic charr stocks to fishing activities (SEMI-QUANTITATIVE RISK ANALYSIS)
• model development and testing (FULL QUANTITATIVE ANALYSIS)
phase 1 :
phase 2 :
Charr modelling using life history traits
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
exposure
Risk assessment using life history traits
Productivity-Susceptibility AnalysisPSA
a relative (as opposed to absolute) risk assessment method
• its ability to recover from increased mortality(productivity≈resilience) (more productive = less vulnerable)
• the extent of the impacts due to fishing activities(susceptibility≈exposure) (more exposed = more vulnerable)
based on the principle that population vulnerability to fishing activities depends on:
PSA introduction
Productivity-Susceptibility AnalysisPSA
PSA introduction
•Australian Northern Prawn fishery by-catch spp.(see Stobutzki et al. 2001)
• recently proposed as a semi-quantitative step in a hierarchical ecological risk assessment for the effects of fishing (ERAEF) framework by the Marine Stewardship Council (MSC eco-certification body) (see Hobday et al. 2007)
• currently being evaluated for use in determining the vulnerability of U.S. fisheries by NOAA (see Patrick et al. 2009)
not a new method:
�but has never been experimented on a single-species basis
Risk assessment using life history traits
Productivity-Susceptibility risk assessmentPSA methods
STEP 1: define and estimate productivity and susceptibility attributes
Productivityattributes
Susceptibilityattributes
Productivityattributes
Susceptibilityattributes
STEP 1: define and estimate productivity and susceptibility attributes
historical data base for charr from across Nunavut (1971-2003)
HUDSON
BAY
BAFFIN SEA
ARCTIC OCEAN
PACIFIC
OCEAN
ATLANTIC OCEAN
CANADA
Nunavut
HUDSON
BAY
BAFFIN SEA
ARCTIC OCEAN
PACIFIC
OCEAN
ATLANTIC OCEAN
CANADA
Nunavut
97 anadromous stocks (5.5” (139.7 mm) gillnet studies)
Productivity-Susceptibility risk assessmentPSA methods
PSA methods
Productivityattributes
age at first maturityyounger age at first maturity =
more productive
individual growth ratefaster growth = more productive
lifespanshort-lived = more productive
mortalityhigher mortality = more productive
STEP 1: define and estimate productivity and susceptibility attributes
estimation
first age group with mature fish (all data pooled)
or approximated from a linear model using max
age as predictor
Brody (K) growth coefficient calculated by fitting von Bertalanffy growth curves to fully recruited age groups* (all 139.7mm gillnet data pooled)
maximum age observed in the sample(all 139.7mm gillnet data pooled)
annual mortality estimated from catch curves (catch at age plots)(all 139.7mm gillnet data pooled)
Productivity-Susceptibility risk assessment
PSA methods
STEP 1: define and estimate productivity and susceptibility attributes
Susceptibilityattributes
availability
selectivity
likelihood that the stock will encounter fishing gear
overlap between fishing effort and stock distribution
potential for the gear to capture/retain fish
estimation
distance from nearest human community
ranked high for anadromous stocks
ratio of mean fork length over mesh size (in mm)
encounterability
Productivity-Susceptibility risk assessment
PSA methods
STEP 2: rank individual productivity attributes for data quality
Estimate was approximated using life-history information.
Sample size is <100 AND other conditions a priori determined for an estimate to be reliable are not met.
5
OR – Sample size is >100 BUT other conditions a priori determined for an estimate to be reliable are not met.
- Sample size is <1004
OR – sample size is >300 BUT other conditions a priori determined for an estimate to be reliable are not met.
- Sample size is <300 but >1003
- Data is not recent and unlikely to reflect current state environmental conditions.
- All conditions a priori determined for the estimate to be reliable are met.
- sample size is large (≥300).2
- Data is recent and is likely to reflect current state environmental conditions.
- All conditions a priori determined for the estimate to be reliable are met.
- sample size is large (≥300).1
ConditionsData Quality Score
Estimate was approximated using life-history information.
Sample size is <100 AND other conditions a priori determined for an estimate to be reliable are not met.
5
OR – Sample size is >100 BUT other conditions a priori determined for an estimate to be reliable are not met.
- Sample size is <1004
OR – sample size is >300 BUT other conditions a priori determined for an estimate to be reliable are not met.
- Sample size is <300 but >1003
- Data is not recent and unlikely to reflect current state environmental conditions.
- All conditions a priori determined for the estimate to be reliable are met.
- sample size is large (≥300).2
- Data is recent and is likely to reflect current state environmental conditions.
- All conditions a priori determined for the estimate to be reliable are met.
- sample size is large (≥300).1
ConditionsData Quality Score
data quality scores = a measure of uncertainty
Productivity-Susceptibility risk assessment
PSA methods
STEP 3: rank productivity and susceptibility attributes into tiers corresponding to different levels of risk
Productivityattributes
Susceptibilityattributes
age at first maturityindividual growth rate
lifespanmortality
encounterabilityavailabilityselectivity
*DQ
Productivity
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Susceptibility
Risk
1=Low 3=High2=Moderate
1=Low 2=Moderate 3=High
determine a range of values for each attribute
divide this range into thirds:
Productivity-Susceptibility risk assessment
PSA methods
STEP 3: rank productivity and susceptibility attributes into tiers corresponding to different levels of risk
Productivityattributes
*DQ
0.51-0.660.35-0.500.17-0.340.660.17Mortality
12-1718-2324-313112Lifespan (Agemax)
0.128-0.1750.080-0.1270.028-0.0790.1750.028Growth rate (K)
3-56-89-12123Age at first maturity
HIGH = 3MODERATE = 2LOW = 1MaxMin
Productivity Tiers and ScoresObserved RangeAttribute
0.51-0.660.35-0.500.17-0.340.660.17Mortality
12-1718-2324-313112Lifespan (Agemax)
0.128-0.1750.080-0.1270.028-0.0790.1750.028Growth rate (K)
3-56-89-12123Age at first maturity
HIGH = 3MODERATE = 2LOW = 1MaxMin
Productivity Tiers and ScoresObserved RangeAttribute
*only stocks with values corresponding to data quality scores ≤3 were used to define the observed range
Productivity-Susceptibility risk assessment
PSA methods
STEP 3: rank productivity and susceptibility attributes into tiers corresponding to different levels of risk
Susceptibilityattributes
Productivity-Susceptibility risk assessment
encounterability scores
≥ 201151-160101-11051-60≤ 10Distance from community (km)
1 = Low1.52 = Mod2.53 = HighEncounterability scores
≥ 201151-160101-11051-60≤ 10Distance from community (km)
1 = Low1.52 = Mod2.53 = HighEncounterability scores
PSA methods
STEP 3: rank productivity and susceptibility attributes into tiers corresponding to different levels of risk
Susceptibilityattributes
selectivity scores
4.91-5.554.26-4.903.60-4.255.553.60772501
HIGH=3MODERATE=2LOW = 1MAXMINMAXMIN
Gear Selectivity Tiers and ScoresObserved range(Mean FL / Mesh size)
Observed range(mean FL)
4.91-5.554.26-4.903.60-4.255.553.60772501
HIGH=3MODERATE=2LOW = 1MAXMINMAXMIN
Gear Selectivity Tiers and ScoresObserved range(Mean FL / Mesh size)
Observed range(mean FL)
Productivity-Susceptibility risk assessment
PSA methods
STEP 3: rank productivity and susceptibility attributes into tiers corresponding to different levels of risk
Susceptibilityattributes
lacustrine -medium size lake fisherylacustrine -large lake fishery
anadromous - river mouth fisheryanadromous - medium size lake fishery anadromous - large lake fishery
high=3moderate=2low=1
Availability tiers and scores
lacustrine -medium size lake fisherylacustrine -large lake fishery
anadromous - river mouth fisheryanadromous - medium size lake fishery anadromous - large lake fishery
high=3moderate=2low=1
Availability tiers and scores
availability scores
Productivity-Susceptibility risk assessment
PSA methods
STEP 4: plot average productivity-susceptibility scores and estimate vulnerability as the Euclidean distance from the origin
average Productivity score (P) average Susceptibility score (S)
Vulnerability (V)
V = √(P - 3)2 + (S – 1)2)
Productivity attributes Susceptibility attributes
Susceptibility
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Susceptibility
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Productivity
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Productivity
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
(Euclidean distance from origin)
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Su
sce
ptibili
ty
(High) (Low)
(Low
)(H
igh)
High ProductivityLow Susceptibility= least vulnerable
Low ProductivityHigh Susceptibility= most vulnerable
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Su
sce
ptibili
ty
(High) (Low)
(Low
)(H
igh)
High ProductivityLow Susceptibility= least vulnerable
Low ProductivityHigh Susceptibility= most vulnerable
*DQ
Productivity-Susceptibility risk assessment
PSA methods
STEP 5: average data quality scores into a data quality index
average Productivity score (P) average Susceptibility score (S)
Productivity attributes Susceptibility attributes
Susceptibility
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Susceptibility
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Productivity
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
Productivity
Risk
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
1=High 3=Low2=Moderate
1=Low 2=Moderate 3=High
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Su
sce
ptibili
ty
(High) (Low)
(Low
)(H
igh)
High ProductivityLow Susceptibility= least vulnerable
Low ProductivityHigh Susceptibility= most vulnerable
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Su
sce
ptibili
ty
(High) (Low)
(Low
)(H
igh)
High ProductivityLow Susceptibility= least vulnerable
Low ProductivityHigh Susceptibility= most vulnerable
*DQ
High data quality
Medium data quality
Low data quality
≤22.1-3.5≥3.6
HIGH (H)MEDIUM (M)LOW (L)
overall DATA QUALITY Index
≤22.1-3.5≥3.6
HIGH (H)MEDIUM (M)LOW (L)
overall DATA QUALITY Index
Productivity-Susceptibility risk assessment
PSA results
Productivity-Susceptibility risk assessment
1.0
1.5
2.0
2.5
3.0
Susc
ep
tibili
ty(L
ow
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
1.0
1.5
2.0
2.5
3.0
Susc
ep
tibili
ty(L
ow
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
High data quality
Medium data quality
Low data quality
Kitikmeot Region
Less Vulnerable(1.27) Arrowsmith River(1.27) Kellett River(1.28) Nakyoktok River(1.28) Perry River(1.47) Becher River
More Vulnerable(2.06) Freshwater Creek(2.07) Paliryuak River(2.08) Hayes River(2.24) Jayco River(2.30) Abernethy River(2.46) Lord Lindsay Lake
Less Vulnerable stocks
Vulnerable stocks
More Vulnerable stocks
PSA results
Productivity-Susceptibility risk assessment
High data quality
Medium data quality
Low data quality
More Vulnerable(2.05) Ross Inlet
Kivalliq Region
Susc
eptib
ility
1.0
1.5
2.0
2.5
3.0
(Low
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
Susc
eptib
ility
1.0
1.5
2.0
2.5
3.0
(Low
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
1.0
1.5
2.0
2.5
3.0
(Low
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
(1.09) Stony Point Area(1.10) Barbour Bay(1.10) Robinhood Bay(1.15) Lorillard River(1.18) Gore Bay Area(1.18) Hanway River(1.19) Steep Bank Bay
(1.23) Pistol Bay(1.36) Chesterfield Inlet(1.36) Rankin Inlet Area(1.36) Ranger Seal Bay
Less Vulnerable
Less Vulnerable stocks
Vulnerable stocks
More Vulnerable stocks
PSA results
Productivity-Susceptibility risk assessment
High data quality
Medium data quality
Low data quality
More VulnerableLess Vulnerable
1.0
1.5
2.0
2.5
3.0
Susc
eptib
ility
(Low
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
1.0
1.5
2.0
2.5
3.0
Susc
eptib
ility
(Low
)(H
igh)
1.01.52.02.53.0
Productivity(High) (Low)
Baffin High Region
(1.23) Ikalukjuaq Lake and River(1.30) Tasirululik Lake(1.41) Kipisa Lake(1.41) Tugaat River(1.42) Ava Inlet(1.42) Nettilling Lake
(2.14) Ijaruvung Lake(2.14) Kukaluk River(2.15) Ikikeesarjuq Lake(2.15) Stanwell-Fletcher Lake(2.20) Tarsiujuak Arm(2.23) Magda River(2.29) Gifford River
(2.30) Ivisarak Lake(2.30) Koluktoo Bay(2.30) Naulingiavik Lake(2.35) Hall Lake(2.39) Rowley River(2.41) Ikpikiturjuaq Lake(2.44) Neergaard Lake
(2.44) Ravn River(2.50) Hall Beach Area(2.55) Windless River(2.56) Fury and Hecla Strait
Less Vulnerable stocks
Vulnerable stocks
More Vulnerable stocks
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
PSA results
Productivity-Susceptibility risk assessment
Less Vulnerable stocks (more sustainable fisheries)
Vulnerable stocks (sustainable fisheries)
More Vulnerable stocks (less sustainable fisheries)
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
PSA results
Productivity-Susceptibility risk assessment
Less Vulnerable stocks (more sustainable fisheries)
Vulnerable stocks (sustainable fisheries)
More Vulnerable stocks (less sustainable fisheries)
• 15% of fisheries examined• 45% of more sustainable fisheries
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
Rankin Inlet
Gjoa Haven
Cambridge BayKugluktuk
Kugaaruk
Igloolik
Hall Beach
Taloyoak
Arviat
Kimmirut
Iqaluit
Pangnirtung
Pond Inlet
Clyde River
Arctic Bay
Repulse Bay
Chesterfield Inlet
PSA results
Productivity-Susceptibility risk assessment
• 22% of fisheries examined• 60% of less sustainable fisheries
Less Vulnerable stocks (more sustainable fisheries)
Vulnerable stocks (sustainable fisheries)
More Vulnerable stocks (less sustainable fisheries)
PSA implications
Productivity-Susceptibility risk assessment
over a broad geographic landscape:
• PSA analysis allows us to distinguish between regions characterized predominantly by less or more vulnerable Arctic charr populations
prevalence of more vulnerable stocks = greater risk of overfishingprevalence of less vulnerable stocks = lower risk of overfishing
Regional differencesin stock vulnerability
PSA results
Productivity-Susceptibility risk assessment
High data quality
Medium data quality
Low data quality
Cambridge Bay Area
Less Vulnerable stocks
Vulnerable stocks
More Vulnerable stocks
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Susc
eptib
ility
(High) (Low)
(Low
)(H
igh)
Jayco
Paliryuak
Ellice
Perry
Halovik
LauchlanEkalluk
Kulgayuk
1.01.52.02.53.0
Productivity
1.0
1.5
2.0
2.5
3.0
Susc
eptib
ility
(High) (Low)
(Low
)(H
igh)
Jayco
Paliryuak
Ellice
Perry
Halovik
LauchlanEkalluk
Kulgayuk
Yield Calculationslog L∞= 0.044 + 0.9841 log Lmax (Froese & Binohlan2000)
K = -ln (1-Lm/L∞) / (tm - t0) (Froese & Binohan 2003)
M = 1.5K F = Z – M – Z calculated from catch curve (Jensen 1996)
Biomass = Yield/F
Max Sustainable Production = 0.33 x Z x Biomass
General conclusion
Charr modelling using life history traits
• Among population Variability in life history traits is a characteristic of Arctic charr populations in Nunavut
General conclusion
• Among population Variability in life history traits is a characteristic of Arctic charr populations in Nunavut
• Life history traits for a stock can be determined from a single sample of ≈ 200 fish (intermediate data quality) or approximated using life history theory
Charr modelling using life history traits
General conclusion
• Among population Variability in life history traits is a characteristic of Arctic charr populations in Nunavut
• Life history traits for a stock can be determined from a single sample of ≈ 200 fish (intermediate data quality) or approximated using life history theory
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
• Life history traits can be used together with a measure of exposure (susceptibility attributes) to determine the relative vulnerability of Arctic charr stocks to fishing activities over a broad geographic landscape and within smaller areas.
exposure
Charr modelling using life history traits
General conclusion
• Among population Variability in life history traits is a characteristic of Arctic charr populations in Nunavut
• Life history traits for a stock can be determined from a single sample of ≈ 200 fish (intermediate data quality) or approximated using life history theory
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
soon:
• Life history traits can be used in a quantitative model to explore and identify sustainable harvest rates for a stock and thereby facilitate the formulation of scientific advice for arctic charr in the CNA region.
• Life history traits can be used together with a measure of exposure (susceptibility attributes) to determine the relative vulnerability of Arctic charr stocks to fishing activities over a broad geographic landscape and within smaller areas.
exposure
Charr modelling using life history traits
charr diversity
(�) “for addressing of gaps for the scientific understanding and management of Arctic charr fisheries
life history traits
solutions
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
R MG
F AM GR Z
exposure
relative risk assessment full quantitative analysis
General conclusion
Gaps Solutions