Using a genetic algorithm to optimize a data-limited catch rule

Fischer, Simon; Oliveira, José Antônio Apparecido de; Mumford, John; Kell, Laurence T.

doi:10.1093/icesjms/fsab018

Cited by 12 publications

(11 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2017). But several methods have been recently adopted to improve the efficiency of computational optimization including genetic algorithms (Fischer et al . 2021), partially observable Markov decision process (Memarzadeh & Boettiger 2018), bootstrapping (ICES 2020a), and Bayesian statistics (ICES 2020a).…”

Section: Discussionmentioning

confidence: 99%

“…Another caveat of our approach is computational intensity (requiring extensive parallel computing on a high-performance computer cluster), which may pose challenges in its application especially for more complex management objectives (more control parameters) Hilborn 1978, Chadès et al 2017). Methods have been recently adopted to improve the efficiency of computational optimization including genetic algorithms (Fischer et al 2021), partially observable Markov decision process (Memarzadeh & Boettiger 2018), stochastic 20 process (Wiedenmann et al 2015), bootstrapping (ICES 2020a), and Bayesian statistics (ICES 2020a). Future research would benefit from applying these techniques to expand this feedbackbased approach to tackling estimation bias in assessment.…”

Section: Managing Risks Under Rising Uncertaintymentioning

confidence: 99%

See 1 more Smart Citation

Shaping sustainable harvest boundaries for marine populations despite estimation bias

Goto¹,

Devine²,

Umar³

et al. 2020

Preprint

View full text Add to dashboard Cite

Imperfect knowledge of social–ecological systems can obscure predictability of resource fluctuation and, in turn, lead to erroneous risk assessments and delayed management actions. Systematic error in population status such as persistent overestimation of abundance is a pervasive conservation problem and has plagued assessments of commercial exploitation of marine species, threatening its sustainability.Using North Sea saithe (Pollachius virens)–a demersal (bottom-water) predatory fish–as a real-world case study, we illustrate a precautionary approach to diagnose robustness of harvest rules to persistent estimation bias (overestimated stock abundance and underestimated fishing mortality rate) and to develop alternative protective measures that minimize population depletion (quasi-extinction) risk by propagating known sources of uncertainty (process, observation, and implementation) through closed-loop simulation of resource–management feedback systems (management strategy evaluation, MSE).Analyses showed that the harvest rules set for saithe are robust to a moderate amount (10– 30%) of estimation bias. More severe bias sets overly optimistic catch limits and promotes overexploitation only in the short term; unacceptably high quasi-extinction risks, however, result primarily from progressively amplified amplitudes of catch fluctuation. Although these undesirable outcomes were, to some extent, mitigated by applying a policy tool to suppress catch fluctuation, this tool falls short of being an effective measure to achieve management goals.More consistent performance of management measures was achieved by developing and applying more precautionary harvest rules through MSE by explicitly accounting for bias. When bias became more severe, raising threshold abundance (by 8–24%) that triggers management actions and lowering target exploitation rate (by 6–29%) would not only safeguard against overexploitation and depletion but also provide catch stability (less disruption in fishing operations).We show that the precautionary approach to risk management through MSE offers a powerful tool to set safe harvest boundaries when assessments are persistently biased. Given challenges in identifying the sources, we suggest bias be routinely evaluated through MSE, and alternative measures be developed to set catch limits when needed. By explicitly accounting for key sources of uncertainty in managing commercial exploitation, our proposed approach ensures effective conservation and sustainable exploitation of living marine resources even under profound uncertainty.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Managing Risks Under Rising Uncertaintymentioning

confidence: 99%

Shaping sustainable harvest boundaries for marine populations despite estimation bias

Goto¹,

Devine²,

Umar³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Another caveat of our approach is computational intensity (requiring extensive parallel computing on a high-performance computer cluster), which may pose challenges in its application especially for more complex management objectives (more control parameters) (Chadès et al, 2017;Walters & Hilborn, 1978). Methods have been recently adopted to improve the efficiency of computational optimization including genetic algorithms (Fischer et al, 2021), partially observable Markov decision process (Memarzadeh & Boettiger, 2018), stochastic process (Wiedenmann et al, 2015), bootstrapping (ICES, 2020a), and Bayesian statistics (ICES, 2020a). Future research would benefit from applying these techniques to expand this feedback-based approach to tackling estimation bias in assessment.…”

Section: Managing Risks Under Rising Uncertaintymentioning

confidence: 99%

Shaping sustainable harvest boundaries for marine populations despite estimation bias

et al. 2022

View full text Add to dashboard Cite

Biased estimates of population status are a pervasive conservation problem. This problem has plagued assessments of commercial exploitation of marine species and can threaten the sustainability of both populations and fisheries.We develop a computer-intensive approach to minimize adverse effects of persistent estimation bias in assessments by optimizing operational harvest measures (harvest control rules) with closed-loop simulation of resourcemanagement feedback systems: management strategy evaluation. Using saithe (Pollachius virens), a bottom water, apex predator in the North Sea, as a realworld case study, we illustrate the approach by first diagnosing robustness of the existing harvest control rule and then optimizing it through propagation of biases (overestimated stock abundance and underestimated fishing pressure) along with select process and observation uncertainties. Analyses showed that severe biases lead to overly optimistic catch limits and then progressively magnify the amplitude of catch fluctuation, thereby posing unacceptably high overharvest risks. Consistent performance of management strategies to conserve the resource can be achieved by developing more robust control rules. These rules explicitly account for estimation bias through a computational grid search for a set of control parameters (threshold abundance that triggers management action, B trigger , and target exploitation rate, F target ) that maximize yield while keeping stock abundance above a precautionary level. When the biases become too severe, optimized control parameters-for saithe, raising B trigger and lowering F target -would safeguard against a overharvest risk (<3.5% probability of stock depletion) and provide short-term stability in catch limit (<20% year-to-year variation), thereby minimizing disruption to fishing communities. The precautionary approach to fine-tuning adaptive risk management through management strategy evaluation offers a powerful tool to better shape sustainable harvest boundaries for exploited resource populations when estimation bias persists. By explicitly accounting for emergent sources of

show abstract

“…The tested MPs are detailed in Table 4 and included the data-rich (ICES category 1) MSY rule (ICES, 2021f), the data-limited (ICES category 3) empirical 2 over 3 rule (ICES, 2012a), the rfb rule (Fischer et al, 2020(Fischer et al, , 2021a(Fischer et al, , 2021b and the hr rule (Fischer et al, 2022). The ICES MSY rule MP mimicked the process conducted by ICES working groups, including a SAM assessment and a short-term forecast.…”

Section: Management Proceduresmentioning

confidence: 99%

“…The genetic algorithm was set up with a population size of 1000 individuals. Variability was introduced through two genetic operators, crossover with p = .8 and mutation with p = .1, as well as elitism with p = .05 (Fischer et al, 2021a). Convergence of the optimisation was achieved when either a maximum of 100 generations was reached or no further improvement was achieved within 10 generations.…”

Section: Ices 2019b 2020b)mentioning

confidence: 99%

Risk equivalence in data‐limited and data‐rich fisheries management: An example based on the ICES advice framework

et al. 2022

Self Cite

View full text Add to dashboard Cite

Fisheries management needs to ensure that resources are exploited sustainably, and the risk of depletion is at an acceptable level. However, often uncertainty about resource dynamics exists, and data availability may differ substantially between fish stocks. This situation can be addressed through tiered systems, where tiers represent different data limitations, and tier-specific stock assessment methods are defined, aiming for risk equivalence across tiers. As case studies, we selected stocks of European plaice, Atlantic cod and Atlantic herring, where advice is provided by the International Council for the Exploration of the Sea (ICES). We conducted a closedloop simulation to compare risk equivalence between the data-rich ICES MSY rule, based on a quantitative stock assessment, and the revised data-limited empirical management procedures of the ICES advice framework. The simulations indicated that the data-limited approaches were precautionary and did not lead to a higher risk of depletion than the data-rich approach. Although the catch based on generic datalimited approaches was lower, stock-specific optimisation improved management performance with catch levels comparable with the data-rich approach. Furthermore, the simulation indicated the ICES MSY rule can fail to meet management objectives due to increased depletion risk when management reference points are set suboptimally. We conclude that the recent revisions of the ICES system explicitly account for risk equivalence for data-limited fisheries management and are a major step forward.Finally, we advocate further consideration of simple empirical management procedures irrespective of data limitations due to their ability to meet fisheries management objectives with greater simplicity.

show abstract

Using a genetic algorithm to optimize a data-limited catch rule

Cited by 12 publications

References 27 publications

Shaping sustainable harvest boundaries for marine populations despite estimation bias

Shaping sustainable harvest boundaries for marine populations despite estimation bias

Shaping sustainable harvest boundaries for marine populations despite estimation bias

Risk equivalence in data‐limited and data‐rich fisheries management: An example based on the ICES advice framework

Contact Info

Product

Resources

About