A Bayesian hierarchical analysis of stock–recruit data: quantifying structural and parameter uncertainties

2004 ◽  
Vol 61 (6) ◽  
pp. 1032-1047 ◽  
Author(s):  
Catherine GJ Michielsens ◽  
Murdoch K McAllister
Keyword(s):  
Alternative Model ◽  
Prior Probability ◽  
Stock Assessment ◽  
Population Variation ◽  
Data Sets ◽  
Model Structure ◽  
Hierarchical Analysis ◽  
Parameter Uncertainties ◽  
Bayesian Hierarchical ◽  
Parameter Values

Stock–recruit functions are important in fisheries stock assessment, but there is often uncertainty surrounding the appropriate stock–recruit model and its parameter values. Combining different stock–recruit data sets of related species through Bayesian hierarchical analysis can decrease these uncertainties and help to characterize appropriate stock–recruit forms and ranges of plausible parameter values. Two different stock–recruit functions (Beverton–Holt and Ricker) have been parameterized in terms of the steepness, which is a parameter that is comparable between populations. In the hierarchical analysis, the prior probability distribution of parameters for the cross-population variation in steepness is determined through a concise model structure. By calculating the Bayes' posteriors for alternative model forms, model uncertainty is accounted for. This methodology has been applied to Atlantic salmon (Salmo salar) stock–recruit data to provide predictions for the steepness of the stock–recruit function for Baltic salmon for which no stock–recruit data exist.

scholarly journals Identification of behavioural model input data sets for WWTP uncertainty analysis

2019 ◽  
Vol 81 (8) ◽  
pp. 1558-1568 ◽  
Author(s):  
E. Lindblom ◽  
U. Jeppsson ◽  
G. Sin
Keyword(s):  
Uncertainty Analysis ◽  
Data Model ◽  
Input Data ◽  
Treatment Plant ◽  
Data Sets ◽  
Model Structure ◽  
Sources Of Uncertainty ◽  
Behavioural Model ◽  
Plant Data ◽  
Parameter Values

Abstract Uncertainty analysis is important for wastewater treatment plant (WWTP) model applications. An important aspect of uncertainty analysis is the identification and proper quantification of sources of uncertainty. In this contribution, a methodology to identify an ensemble of behavioural model representations (combinations of input data, model structure and parameter values) is presented and evaluated. The outcome is a multivariate conditional distribution of input data that is used for generating samples of likely inputs (such as Monte Carlo input samples) to perform WWTP model uncertainty analysis. This article presents an approach to verify uncertainty distributions of input data (otherwise often assumed) by using historical observations and actual plant data.

Analysis of Contradictory Data Sources in Fish Stock Assessment

1993 ◽  
Vol 50 (9) ◽  
pp. 1916-1923 ◽  
Author(s):  
Jon T. Schnute ◽  
Ray Hilborn
Keyword(s):  
Stock Assessment ◽  
Fish Stock ◽  
Parameter Estimates ◽  
Data Sets ◽  
Likelihood Functions ◽  
Northern Cod ◽  
Stock Assessments ◽  
Fish Stock Assessment ◽  
Parameter Values ◽  
Contradictory Data

Fisheries stock assessments sometimes prove, in retrospect, to be wrong. Errors may be due to poor model assumptions or to data that do not reflect the biological process of interest. We develop a method that formally admits the possibility of such errors. Likelihood functions derived from this method indicate greater uncertainty in parameter values than conventional likelihoods, whose derivations presume that models correctly describe the observed data. The problem of uncertainty is particularly acute when more than one data source is available and different data sets provide contradictory parameter estimates. Traditional methods of stock assessment involve weighted averages of the contradictory data, and these generally produce parameter estimates intermediate to those obtained from the data sets individually. We demonstrate that, when model or data errors are considered, the most likely parameter values are not intermediary to conflicting values; instead, they occur at one of the apparent extremes. We provide an example using contradictory trends in catch-per-u nit-effort data for the Canadian northern cod stock (1978–88).

Nonstationarity of Production Relationships in Exploited Populations

1987 ◽  
Vol 44 (S2) ◽  
pp. s156-s165 ◽  
Author(s):  
Carl J. Walters
Keyword(s):  
Environmental Changes ◽  
Historical Data ◽  
Stock Assessment ◽  
Policy Implications ◽  
Data Sets ◽  
Stock Structure ◽  
Stock Size ◽  
Rate Processes ◽  
And Shifts ◽  
Time Invariant

Stock assessment usually proceeds from the assumption that there are time-invariant relationships between stock size and rate processes such as recruitment, although such relationships are difficult to discern due to noise caused by factors other than stock size. There are good biological reasons not to trust this assumption in exploited populations, where persistent environmental changes and shifts in stock structure may cause various parameters to change. Graphical and statistical procedures can be used to detect this nonstationarity in historical data sets for which stock size has varied so as to repeatedly sample a range of sizes. The policy implications of nonstationarity depend on whether the changes are clearly observable as deviations from known, Song-term baseline responses. If the changes are observable, it is usually best to pretend that the current deviation will persist unless strong constraints on policy change make it necessary to plan for changes that may occur far into the future. If the changes are not observable (the usual case), then it is necessary to make a difficult policy choice between passively waiting for informative stock responses versus actively experimenting with harvest rates so as to quickly get information about responses over a range of stock sizes.

scholarly journals Modelling the mass-SFR relation at high redshifts; predicted constraints from JWST

2019 ◽  
Vol 15 (S352) ◽  
pp. 114-114
Author(s):  
Emma Curtis-Lake
Keyword(s):  
Data Sets ◽  
Bayesian Hierarchical ◽  
Hierarchical Modelling ◽  
Bayesian Hierarchical Modelling ◽  
The Galaxy ◽  
Low Mass ◽  
Robust Measurements ◽  
Galaxy Population ◽  
Added Uncertainty ◽  
Intrinsic Scatter

AbstractThe mass-SFR relation of galaxies encodes information of present and historical star formation in the galaxy population. We expect the intrinsic scatter in the relation to increase to low mass where SFR becomes more stochastic. Measurements at z ‰ 4 from the Hubble Frontier fields have hinted at this (Santini et al., 2017), however, with the added uncertainty of lensing magnification we await JWST to provide robust measurements. Even with data-sets provided by JWST, uncertainties on mass and SFR estimates are often large, potentially covariant and dependent on assumptions used. I will present our method of Bayesian hierarchical modelling of the mass-SFR relation that self-consistently propagates uncertainties on mass and SFR estimates to uncertainties on the mass-SFR relation parameters. I will expose the biases imposed by standard SED-modelling practices, and address to what significance we can measure an increase in intrinsic scatter to low masses with JWST.

scholarly journals Understanding the Average Impact of Microcredit Expansions: A Bayesian Hierarchical Analysis of Seven Randomized Experiments

2019 ◽  
Vol 11 (1) ◽  
pp. 57-91 ◽  
Author(s):  
Rachael Meager
Keyword(s):  
External Validity ◽  
Hierarchical Models ◽  
Hierarchical Analysis ◽  
Randomized Experiments ◽  
Average Effect ◽  
Bayesian Hierarchical ◽  
Heterogeneous Effects ◽  
Average Impact ◽  
The Impact ◽  
Better Than

Despite evidence from multiple randomized evaluations of microcredit, questions about external validity have impeded consensus on the results. I jointly estimate the average effect and the heterogeneity in effects across seven studies using Bayesian hierarchical models. I  find the impact on household business and consumption variables is unlikely to be transformative and may be negligible. I find reasonable external validity: true heterogeneity in effects is moderate, and approximately 60 percent of observed heterogeneity is sampling variation. Households with previous business experience have larger but more heterogeneous effects. Economic features of microcredit interventions predict variation in effects better than studies’ evaluation protocols. (JEL D14, G21, I38, O12, O16, P34, P36)

A new way to visualize and report structural and data uncertainty in stock assessments

2020 ◽  
Vol 77 (8) ◽  
pp. 1275-1280
Author(s):  
Jason Cope ◽  
Vladlena Gertseva
Keyword(s):  
Sensitivity Analysis ◽  
Alternative Model ◽  
Reference Model ◽  
Stock Assessment ◽  
Assessment Model ◽  
Data Uncertainty ◽  
Model Comparisons ◽  
Modelling Framework ◽  
Stock Assessments ◽  
General Method

We present a visual and tabular representation of fisheries stock assessment model outputs to rapidly examine and effectively communicate sensitivity analysis results from numerous alternative model comparisons. This approach uses multiple output metrics to identify which alternative stock assessment model configurations relative to the reference model deserve further attention when quantifying intermodel uncertainty. An accompanying table of likelihood components, parameters, and model-derived quantities highlights where major changes exist compared with the reference model. The general method is applicable to any stock assessment and should aid in model behavior diagnosis and communicating uncertainty to managers. Specific examples and code are provided for the Stock Synthesis modelling framework.

Sign in / Sign up

Export Citation Format

Share Document