Impacts of outliers and mis-specification of priors on Bayesian fisheries-stock assessment

2000 ◽  
Vol 57 (11) ◽  
pp. 2293-2305 ◽  
Author(s):  
Y Chen ◽  
P A Breen ◽  
N L Andrew
Keyword(s):  
Normal Distribution ◽  
Stock Assessment ◽  
Management Strategies ◽  
Distribution Functions ◽  
Assessment Model ◽  
Fish Stock ◽  
Model Parameters ◽  
Prior Distributions ◽  
Likelihood Functions ◽  
Robust Likelihood

Bayesian inference is increasingly used in estimating model parameters for fish-stock assessment, because of its ability to incorporate uncertainty and prior knowledge and to provide information for risk analyses in evaluating alternative management strategies. Normal distributions are commonly used in formulating likelihood functions and informative prior distributions; these are sensitive to data outliers and mis-specification of prior distributions, both common problems in fisheries-stock assessment. Using a length-structured stock-assessment model for a New Zealand abalone fishery as an example, we evaluate the robustness of three likelihood functions and two prior-distribution functions, with respect to outliers and mis-specification of priors, in 48 different combinations. The two robust likelihood estimators performed slightly less well when no data outliers were present and much better when data outliers were present. Similarly, the Cauchy distribution was less sensitive to prior mis-specification than the normal distribution. We conclude that replacing the normal distribution with "fat-tailed" distributions for likelihoods and priors can improve Bayesian assessments when there are data outliers and mis-specification of priors, with relatively minor costs when there are none.

scholarly journals A two-stage biomass random effects model for stock assessment without catches: What can be estimated using only biomass survey indices?

2008 ◽  
Vol 65 (6) ◽  
pp. 1024-1035 ◽  
Author(s):  
Verena M. Trenkel
Keyword(s):  
Random Effects ◽  
Stock Assessment ◽  
Random Effect ◽  
Simulated Data ◽  
Fish Stock ◽  
Model Parameters ◽  
Biomass Growth ◽  
Two Stage ◽  
Commercial Catch ◽  
Nested Models

A simple two-stage biomass random effects population dynamics model is presented for carrying out fish stock assessments based on survey indices using no commercial catch information. Recruitment and biomass growth are modelled as random effects, reducing the number of model parameters while maintaining model flexibility. No assumptions regarding natural mortality rates are required. The performance of the method was evaluated using simulated data with emphasis on identifying parameter redundancy, which showed that the variance of the biomass growth random effect might only be estimable if large (>0.2). The full and two nested models were fitted to European anchovy ( Engraulis encrasicolus ) in the Bay of Biscay using two survey series. The best-fitting model had fixed biomass growth and random recruitment following a lognormal distribution.

scholarly journals Impacts of prior mis-specification on Bayesian fisheries stock assessment

2008 ◽  
Vol 59 (2) ◽  
pp. 145 ◽  
Author(s):  
Yong Chen ◽  
Chi-Lu Sun ◽  
Minoru Kanaiwa
Keyword(s):  
Distribution Function ◽  
Stock Assessment ◽  
Cauchy Distribution ◽  
Distribution Functions ◽  
Model Parameters ◽  
Large Uncertainty ◽  
Future Studies ◽  
Accuracy And Precision ◽  
Key Features ◽  
Sea Urchin Fishery

One of the key features of a Bayesian stock assessment is that the modeller needs to provide knowledge on model parameters. Priors summarise modellers’ understanding of model parameters and are often defined by a probability distribution function. Priors are often mis-specified with arbitrary and unrealistic accuracy and precision in perceiving the state of nature for the parameters as a result of our limited understanding of fisheries ecosystems. Commonly used probability functions such as normal distribution functions tend to be sensitive to prior mis-specification, resulting in large uncertainty and/or errors in Bayesian stock assessment. Fat-tailed functions such as the Cauchy distribution function have been found to be robust to prior mis-specification. Using the Maine sea urchin fishery as an example, we evaluated the impacts of mis-specification in defining the prior distributions on Bayesian stock assessment. The present study suggests that the quantification of priors with a Cauchy distribution tends to be robust to the prior mis-specification. Given our limited understanding of fisheries a function such as the Cauchy distribution function that is robust to prior mis-specification tends to be more desirable. Future studies should explore the use of other fat-tailed distribution functions for quantifying priors in fisheries stock assessment.

scholarly journals Model averaging to streamline the stock assessment process

2014 ◽  
Vol 72 (1) ◽  
pp. 93-98 ◽  
Author(s):  
Colin P. Millar ◽  
Ernesto Jardim ◽  
Finlay Scott ◽  
Giacomo Chato Osio ◽  
Iago Mosqueira ◽  
...  
Keyword(s):  
Large Scale ◽  
Stock Assessment ◽  
Model Averaging ◽  
Assessment Model ◽  
Assessment Process ◽  
Fish Stock ◽  
Assessment Procedure ◽  
Observation Error ◽  
Fishery Policy ◽  
Fishery Science

Abstract The current fish stock assessment process in Europe can be very resource- and time-intensive. The scientists involved require a very particular set of skills, acquired over their career, drawing from biology, ecology, statistics, mathematical modelling, oceanography, fishery policy, and computing. There is a particular focus on producing a single “best” stock assessment model, but as fishery science advances, there are clear needs to address a range of hypotheses and uncertainties, from large-scale issues such as climate change to specific ones, such as high observation error on young hake. Key to our discussion is the use of the assessment for all frameworks to translate hypotheses into models. We propose a change to the current stock assessment procedure, driven by the use of model averaging to address a range of plausible hypotheses, where increased collaboration between the varied disciplines within fishery science will result in more robust advice.

Comparison of the frequentist properties of Bayes and the maximum likelihood estimators in an age-structured fish stock assessment model

2002 ◽  
Vol 59 (1) ◽  
pp. 136-143 ◽  
Author(s):  
Anders Nielsen ◽  
Peter Lewy
Keyword(s):  
Maximum Likelihood ◽  
Stock Assessment ◽  
Process Variations ◽  
Bayes Estimator ◽  
Assessment Model ◽  
Fish Stock ◽  
The North ◽  
The North Sea ◽  
Age Structured ◽  
Fish Stock Assessment

A simulation study was carried out for a separable fish stock assessment model including commercial and survey catch-at-age and effort data. All catches are considered stochastic variables subject to sampling and process variations. The results showed that the Bayes estimator of spawning biomass is a useful but slightly biased estimator for which the frequentist variance can be estimated by the posterior variance. Comparisons further show that the Bayes estimator is better than the maximum likelihood in the sense that it is less biased and, surprisingly, has a much lesser variance. The catch simulations were based on the North Sea plaice (Pleuronectes platessa) stock and fishery data.

A recipe for Bayesian network driven stock assessment

2004 ◽  
Vol 61 (9) ◽  
pp. 1647-1657 ◽  
Author(s):  
T R Hammond
Keyword(s):  
Bayesian Network ◽  
Stock Assessment ◽  
Assessment Model ◽  
Fish Stock ◽  
Posterior Density ◽  
Discrete Random Variables ◽  
Analytic Integration ◽  
Network Propagation ◽  
A New Technique ◽  
Fish Stock Assessment

Markov Chain Monte Carlo (MCMC), the most widely used algorithm in Bayesian statistics, can fail to converge. Although convergence is tested by various diagnostics, these can only reveal failure, never success. To avoid these difficulties, this paper suggests a recipe for using Bayesian network propagation (BNP) to compute posterior results for fish stock assessment. Bayesian networks employ discrete random variables and specify relationships between them with conditional probability tables. Therefore, the recipe uses a new technique called "fuzzy discretization" to convert a continuous Bayesian model into a discrete Bayesian network. The technique is illustrated on a Schaefer assessment model by showing how model equations can be converted to probability tables. Posterior density estimates for carrying capacity (K) from both MCMC and BNP were compared with exact results (obtained by analytic integration and grid search) under three scenarios. BNP outperformed MCMC (as implemented in WinBUGS) in all scenarios, though MCMC diagnostics previously deemed sufficient reported no problems. Tightening the grid resolution of discrete numeric variables over regions of high posterior probability greatly improved BNP performance, so a grid selection heuristic is included in the recipe. In summary, this recipe may provide an effective alternative to MCMC for similar Bayesian problems.

Fleksibest—an age–length structured fish stock assessment model

2002 ◽  
Vol 55 (1-3) ◽  
pp. 87-101 ◽  
Author(s):  
Kristin Guldbrandsen Frøysa ◽  
Bjarte Bogstad ◽  
Dankert W. Skagen
Keyword(s):  
Stock Assessment ◽  
Assessment Model ◽  
Fish Stock ◽  
Fish Stock Assessment

scholarly journals A New Extended Mixture Skew Normal Distribution, With Applications

2019 ◽  
Vol 42 (2) ◽  
pp. 167-183
Author(s):  
Haroon M. Barakat ◽  
Abdallh W. Aboutahoun ◽  
Naeema El-kadar
Keyword(s):  
Normal Distribution ◽  
Location Parameter ◽  
Distribution Functions ◽  
Maximum Likelihood Estimates ◽  
Computational Techniques ◽  
Model Parameters ◽  
Skew Normal Distribution ◽  
Data Set ◽  
Wide Range ◽  
Skew Normal

One of the most important property of the mixture normal distributions-model is its flexibility to accommodate various types of distribution functions (df's). We show that the mixture of the skew normal distribution and its reverse, after adding a location parameter to the skew normal distribution, and adding the same location parameter with different sign to its reverse is a family of df's that contains all the possible types of df's. Besides, it has a very remarkable wide range of the indices of skewness and kurtosis. Computational techniques using EM-type algorithms are employed for iteratively computing maximum likelihood estimates of the model parameters. Moreover, an application with a body mass index real data set is presented.

Detecting and correcting underreported catches in fish stock assessment: trial of a new method

2010 ◽  
Vol 67 (8) ◽  
pp. 1247-1261 ◽  
Author(s):  
Nicolas Bousquet ◽  
Noel Cadigan ◽  
Thierry Duchesne ◽  
Louis-Paul Rivest
Keyword(s):  
Gadus Morhua ◽  
Stock Assessment ◽  
Population Analysis ◽  
Assessment Model ◽  
Fish Stock ◽  
Stock Biomass ◽  
Stock Assessments ◽  
Spawning Stock ◽  
Age Structured ◽  
Fish Stock Assessment

Landings from fisheries are often underreported, that is, the true landings are greater than those reported. Despite this bias, reported landings are widely used in fish stock assessments, and this might lead to overoptimistic exploitation strategies. We construct a statistical stock assessment model that accounts for underreported landings using the theory of censoring with sequential population analysis (SPA). The new model is developed and implemented specifically for the cod stock ( Gadus morhua ) from the southern Gulf of St. Lawrence (Canada). This stock is known to have unreported overfishing during 1985–1992. We show in simulations that for this stock, the new censored model can correctly detect the problematic landings. These corrections are nearly insensitive to subjective boundaries placed on real catches and robust to modifications imposed in the simulation of landings. However, when surveys are too noisy, the new SPA for censored catches can result in increased uncertainty in parameters used for management such as spawning stock biomass and age-structured stock size.

Sign in / Sign up

Export Citation Format

Share Document