219
and impact metrics (FPMs and FIMs) were
calculated from 13 temperate shelf sea communities from the Western and Eastern North Atlantic
and the Mediterranean. Catch statistics were used
to calculate FPM. Bottom trawl survey data were
used to estimate FIMs. Time series of FPMs were
examined to identify ~10-year time periods with
consistent levels, or at least consistent trends, in
fi shing pressure. FPMs were averaged over these
periods; FIMs were averaged across the subsequent 10-year period, allowing for a 10-year lag
between pressure and impact. The relationship
between pressure and impact metrics was examined by a canonical correlation analysis with 27
data points (number of ecological units × number
of time periods). Although stark contrast was
found between FPMs across places and/or time
periods, only a weak link (22 % of total variance)
was found between fi shing selectivity and the
community biodiversity 10 years later: communities from which a more diverse catch was taken
had higher biodiversity, while communities from
which more predators were extracted had a higher
total biomass. Although we examined fi shing
impacts with a reasonable time lag after fi shing
pressure metrics, there is still a suspicion that
these results mostly reveal that fi sheries extract
from a community what is available – if there are
more species in a community, then more species
may have some value and be targeted and caught.
6
Evaluating Management
Performance Indicators
In this example, we used a simulation approach
to study the properties of fi sheries indicators in
view of selecting robust and relevant indicators
for the evaluation of the performance of management measures. A simulation approach allows
disentangling the effects of several factors and
identifying the sensitivity to a particular pressure
(Lehuta et al. 2013b ). Our methodology has two
steps: step 1, review and select metrics for evaluating the long-term performance of management
measures using management strategy evaluation,
and step 2, simulate management strategies
using selected management performance metrics
to assess their capacity of reaching objectives
and their robustness to uncertainties. To carry
out the second step, we used the ISIS-Fish model
(Pelletier et al. 2009 ), a list of management strategies to be assessed with explicit management
objectives and a list of uncertainties and a simulation design to run effi ciently the operating
model (Mahévas and Iooss 2013 ; Mahévas and
Lehuta 2013 ), to compute the selected performance metrics and fi nally to assess their properties. Sensitivity analysis was used to derive
sensitivity indices of model outputs to model
inputs based on variance decomposition.
The approach was applied to the pelagic fi shery in the Bay of Biscay. The population model
included a spatially explicit description of
anchovy dynamics (Lehuta et al. 2010 ) and
global surplus production models for four other
stocks also targeted by the fi shery. The fi shing
behaviour model provided spatial predictions of
fi shing time allocation (Vermard et al. 2008 ).
The overall model was validated for the period
2000–2004 (Lehuta et al. 2013a ). We then simulated seven management measures as defi ned in
the long-term management plan proposed for the
anchovy fi shery and considered six management
performance metrics: (1) P1, the number of years
when biomass drops below Blim; (2) P2, the
number of years of fi shery closure (should not
excess one in 10 years); (3) P3, the number of
years with catch of anchovy higher than 7,000 t
which is the threshold of economical profi tability expressed by the sector; (4) P4, the linear
trend in SSB (should show no degradation to
guaranty reproductive capacity); (5) P5, the
interannual variability in landings (which should
be as low as possible as a proxy of fi shery stability, to guaranty continuous supply to the industry); and (6) P6, the interannual variability in the
proportion of recruits in the population (as low
as possible).
Management scenarios signifi cantly infl uenced the fi shery’s dynamics as far as landings
values and stability, prevention of fi shery closure
and stock collapse were concerned. These
indicators thus clearly refl ected the effect of
management and can be used to monitor management performance and adapt regulations (Fig. 4 ).
Indicators for Ecosystem-Based Management: Methods and Applications
and impact metrics (FPMs and FIMs) were
calculated from 13 temperate shelf sea communities from the Western and Eastern North Atlantic
and the Mediterranean. Catch statistics were used
to calculate FPM. Bottom trawl survey data were
used to estimate FIMs. Time series of FPMs were
examined to identify ~10-year time periods with
consistent levels, or at least consistent trends, in
fi shing pressure. FPMs were averaged over these
periods; FIMs were averaged across the subsequent 10-year period, allowing for a 10-year lag
between pressure and impact. The relationship
between pressure and impact metrics was examined by a canonical correlation analysis with 27
data points (number of ecological units × number
of time periods). Although stark contrast was
found between FPMs across places and/or time
periods, only a weak link (22 % of total variance)
was found between fi shing selectivity and the
community biodiversity 10 years later: communities from which a more diverse catch was taken
had higher biodiversity, while communities from
which more predators were extracted had a higher
total biomass. Although we examined fi shing
impacts with a reasonable time lag after fi shing
pressure metrics, there is still a suspicion that
these results mostly reveal that fi sheries extract
from a community what is available – if there are
more species in a community, then more species
may have some value and be targeted and caught.
6
Evaluating Management
Performance Indicators
In this example, we used a simulation approach
to study the properties of fi sheries indicators in
view of selecting robust and relevant indicators
for the evaluation of the performance of management measures. A simulation approach allows
disentangling the effects of several factors and
identifying the sensitivity to a particular pressure
(Lehuta et al. 2013b ). Our methodology has two
steps: step 1, review and select metrics for evaluating the long-term performance of management
measures using management strategy evaluation,
and step 2, simulate management strategies
using selected management performance metrics
to assess their capacity of reaching objectives
and their robustness to uncertainties. To carry
out the second step, we used the ISIS-Fish model
(Pelletier et al. 2009 ), a list of management strategies to be assessed with explicit management
objectives and a list of uncertainties and a simulation design to run effi ciently the operating
model (Mahévas and Iooss 2013 ; Mahévas and
Lehuta 2013 ), to compute the selected performance metrics and fi nally to assess their properties. Sensitivity analysis was used to derive
sensitivity indices of model outputs to model
inputs based on variance decomposition.
The approach was applied to the pelagic fi shery in the Bay of Biscay. The population model
included a spatially explicit description of
anchovy dynamics (Lehuta et al. 2010 ) and
global surplus production models for four other
stocks also targeted by the fi shery. The fi shing
behaviour model provided spatial predictions of
fi shing time allocation (Vermard et al. 2008 ).
The overall model was validated for the period
2000–2004 (Lehuta et al. 2013a ). We then simulated seven management measures as defi ned in
the long-term management plan proposed for the
anchovy fi shery and considered six management
performance metrics: (1) P1, the number of years
when biomass drops below Blim; (2) P2, the
number of years of fi shery closure (should not
excess one in 10 years); (3) P3, the number of
years with catch of anchovy higher than 7,000 t
which is the threshold of economical profi tability expressed by the sector; (4) P4, the linear
trend in SSB (should show no degradation to
guaranty reproductive capacity); (5) P5, the
interannual variability in landings (which should
be as low as possible as a proxy of fi shery stability, to guaranty continuous supply to the industry); and (6) P6, the interannual variability in the
proportion of recruits in the population (as low
as possible).
Management scenarios signifi cantly infl uenced the fi shery’s dynamics as far as landings
values and stability, prevention of fi shery closure
and stock collapse were concerned. These
indicators thus clearly refl ected the effect of
management and can be used to monitor management performance and adapt regulations (Fig. 4 ).
Indicators for Ecosystem-Based Management: Methods and Applications