Nicolas Bousquet - Profile on Academia.edu (original) (raw)

Papers by Nicolas Bousquet

Research paper thumbnail of Diagnostics of prior-data agreement in applied Bayesian analysis

Research paper thumbnail of Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Risk analysis : an official publication of the Society for Risk Analysis, 2015

We consider the problem of estimating the probability of detection (POD) of flaws in an industria... more We consider the problem of estimating the probability of detection (POD) of flaws in an industrial steel component. Modeled as an increasing function of the flaw height, the POD characterizes the detection process; it is also involved in the estimation of the flaw size distribution, a key input parameter of physical models describing the behavior of the steel component when submitted to extreme thermodynamic loads. Such models are used to assess the resistance of highly reliable systems whose failures are seldom observed in practice. We develop a Bayesian method to estimate the flaw size distribution and the POD function, using flaw height measures from periodic in-service inspections conducted with an ultrasonic detection device, together with measures from destructive lab experiments. Our approach, based on approximate Bayesian computation (ABC) techniques, is applied to a real data set and compared to maximum likelihood estimation (MLE) and a more classical approach based on Mark...

Research paper thumbnail of Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Risk analysis : an official publication of the Society for Risk Analysis, 2015

We consider the problem of estimating the probability of detection (POD) of flaws in an industria... more We consider the problem of estimating the probability of detection (POD) of flaws in an industrial steel component. Modeled as an increasing function of the flaw height, the POD characterizes the detection process; it is also involved in the estimation of the flaw size distribution, a key input parameter of physical models describing the behavior of the steel component when submitted to extreme thermodynamic loads. Such models are used to assess the resistance of highly reliable systems whose failures are seldom observed in practice. We develop a Bayesian method to estimate the flaw size distribution and the POD function, using flaw height measures from periodic in-service inspections conducted with an ultrasonic detection device, together with measures from destructive lab experiments. Our approach, based on approximate Bayesian computation (ABC) techniques, is applied to a real data set and compared to maximum likelihood estimation (MLE) and a more classical approach based on Mark...

Research paper thumbnail of Inversion probabiliste en analyse d'incertitude

Inversion probabiliste en analyse d'incertitude

Research paper thumbnail of A protocol for integrating FED and expert data in a study of durability using the Weibull distribution

The main issues raised by the estimation of lifetime parametric models used in industrial modelli... more The main issues raised by the estimation of lifetime parametric models used in industrial modelling of reliability are censoring and FED (Feedback Experience Data) sample size. Many studies are facing homogeneous, small-sized, censored failure times which have to be integrated into Bayesian procedures with informative prior parameter. This way of dealing with statistical inference has been especially followed by EDF for predicting failures on nuclear material. The example of the Weibull distribution will be here thoroughfully treated. Firstly experts have to be asked about the durability of a material with precise and simple questions. According to the choice of the considered model, prior point estimations and confidence intervals about parameters must be given, directly or indirectly, by experts. Secondly efficient modelling has to be chosen for informative prior distributions. Once, it must produce posterior distributions easily estimated by classical methods. But computation complexity is often a limiting factor of Bayesian inference. The impact of prior choices on posterior results must be simple to derive. Then, hyperparameters of these prior distributions must be evaluated linking the intrinsic properties of the prior densities (mean, mode, variance, etc.) with expert information on parameters.

Research paper thumbnail of Chassot2009

Research paper thumbnail of Estimation de modèles markoviens discrets dans un cadre industriel fiabiliste à données manquantes

Les modèles markoviens sont particulièrement utiles pour décrire des systèmes qui, au long de leu... more Les modèles markoviens sont particulièrement utiles pour décrire des systèmes qui, au long de leur vie, passent à travers différents états. Les paramètres de ces modèles sont les probabilités de transition entre les états. Normalement, les données disponibles pour l'inférence statistique sont des séquences temporelles d'états pour un nombre donné d'individus. Quand les séquences sont incomplètes, l'estimation de la matrice de transition n'est pas triviale et demande l'utilisation de techniques plus avancées. Dans cette communication, nous nous focalisons sur l'estimation bayésienne des probabilités de transition. Premièrement, nous présentons différentes méthodes MCMC, en fonction de la structure des données manquantes. Ensuite, nous proposons une manière d'accélérer les calculs MCMC en tenant compte de la dépendance entre les lignes de la matrice de transition. Finalement, nous montrons les résultats d'essais simulés menés sur des matrices typiqu...

Research paper thumbnail of On the Practical Interest of Discrete Inverse Pólya and Weibull-1 Models in Industrial Reliability Studies

Quality and Reliability Engineering International, 2015

Engineers often cope with the problem of assessing the lifetime of industrial components, under t... more Engineers often cope with the problem of assessing the lifetime of industrial components, under the basis of observed industrial feedback data. Usually, lifetime is modelled as a continuous random variable, for instance exponentially or Weibull distributed. However, in some cases, the features of the piece of equipment under investigation rather suggest the use of discrete probabilistic models. This happens for an equipment which only operates on cycles or on demand. In these cases, the lifetime is rather measured in number of cycles or number of demands before failure, therefore, in theory, discrete models should be more appropriate. This article aims at bringing some light to the practical interest for the reliability engineer in using two discrete models among the most popular: the Inverse Pólya distribution (IPD), based on a Pólya urn scheme, and the so-called Weibull-1 (W1) model. It is showed that, for different reasons, the practical use of both models should be restricted to specific industrial situations. In particular, when nothing is a priori known over the nature of ageing and/or data are heavily right-censored, they can remain of limited interest with respect to more flexible continuous lifetime models such as the usual Weibull distribution. Nonetheless, the intuitive meaning given to the IPD distribution favors its use by engineers in low (decelerated) ageing situations.

Research paper thumbnail of Une modelisation de duree de vie a risques de defaillance concurrents

Une modelisation de duree de vie a risques de defaillance concurrents

A simple competing risk distribution as a possible alternative to the Weibull distri- bution in l... more A simple competing risk distribution as a possible alternative to the Weibull distri- bution in lifetimes analysis is proposed. This distribution is the minimum between an exponential and a Weibull distributions. First, its main characteristics are pre- sented. Then the estimation of its parameters are considered through maximum likelihood and Bayesian inference. Statistical tests to choose between a Weibull distribution

Research paper thumbnail of Bayesian inference for inverse problems occurring in uncertainty analysis

International Journal for Uncertainty Quantification, 2014

Pr♦•❡❝t✲❚❡❛♠s ❙❊▲❊❈❚ ■♥r✐❛ ❛♥❞ ❊❉❋ |✫❉ |❡s❡❛r❝❤ |❡♣♦rt ♥➦ ✼✾✾✺ ✖ ❏✉✐♥ ✷✵✶✷ ✖ ✷✼ ♣❛❣❡s ❆❜str❛❝t✿ ❚... more Pr♦•❡❝t✲❚❡❛♠s ❙❊▲❊❈❚ ■♥r✐❛ ❛♥❞ ❊❉❋ |✫❉ |❡s❡❛r❝❤ |❡♣♦rt ♥➦ ✼✾✾✺ ✖ ❏✉✐♥ ✷✵✶✷ ✖ ✷✼ ♣❛❣❡s ❆❜str❛❝t✿ ❚❤❡ ✐♥✈❡rs❡ ♣r♦❜•❡♠ ❝♦♥s✐❞❡r❡❞ ❤❡r❡ ✐s t♦ ❡st✐♠❛t❡ t❤❡ ❞✐str✐❜✉t✐♦♥ ♦❢ ❛ ♥♦♥✲♦❜s❡r✈❡❞ r❛♥❞♦♠ ✈❛r✐❛❜•❡ X ❢r♦♠ s♦♠❡ ♥♦✐s② ♦❜s❡r✈❡❞ ❞❛t❛ Y •✐♥❦❡❞ t♦ X t❤r♦✉❣❤ ❛ t✐♠❡✲❝♦♥s✉♠✐♥❣ ♣❤②s✐❝❛• ♠♦❞❡• H✳ ❇❛②❡s✐❛♥ ✐♥❢❡r❡♥❝❡ ✐s ❝♦♥s✐❞❡r❡❞ t♦ t❛❦❡ ✐♥t♦ ❛❝❝♦✉♥t ♣r✐♦r ❡①♣❡rt ❦♥♦✇•❡❞❣❡ ♦♥ X ✐♥ ❛ s♠❛•• s❛♠♣•❡ s✐③❡ s❡tt✐♥❣✳ ❆ ▼❡tr♦♣♦•✐s✲❍❛st✐♥❣s ✇✐t❤✐♥ •✐❜❜s ❛•❣♦r✐t❤♠ ✐s ♣r♦♣♦s❡❞ t♦ ❝♦♠♣✉t❡ t❤❡ ♣♦st❡r✐♦r ❞✐str✐❜✉t✐♦♥ ♦❢ t❤❡ ♣❛r❛♠❡t❡rs ♦❢ X t❤r♦✉❣❤ ❛ ❞❛t❛ ❛✉❣♠❡♥t❛t✐♦♥ ♣r♦❝❡ss✳ ❙✐♥❝❡ ❝❛••s t♦ H ❛r❡ q✉✐t❡ ❡①♣❡♥s✐✈❡✱ t❤✐s ✐♥❢❡r❡♥❝❡ ✐s ❛❝❤✐❡✈❡❞ ❜② r❡♣•❛❝✐♥❣ H ✇✐t❤ ❛ ❦r✐❣✐♥❣ ❡♠✉•❛t♦r ✐♥t❡r♣♦•❛t✐♥❣ H ❢r♦♠ ❛ ♥✉♠❡r✐❝❛• ❞❡s✐❣♥ ♦❢ ❡①♣❡r✐♠❡♥ts✳ ❚❤✐s ❛♣♣r♦❛❝❤ ✐♥✈♦•✈❡s s❡✈❡r❛• ❡rr♦rs ♦❢ ❞✐✛❡r❡♥t ♥❛t✉r❡ ❛♥❞✱ ✐♥ t❤✐s ♣❛♣❡r✱ ✇❡ ♣❛② ❡✛♦rt t♦ ♠❡❛s✉r❡ ❛♥❞ r❡❞✉❝❡ t❤❡ ♣♦ss✐❜•❡ ✐♠♣❛❝t ♦❢ t❤♦s❡ ❡rr♦rs✳ ■♥ ♣❛rt✐❝✉•❛r✱ ✇❡ ♣r♦♣♦s❡ t♦ ✉s❡ t❤❡ s♦✲❝❛••❡❞ ❉❆❈ ❝r✐t❡r✐♦♥ t♦ ❛ss❡ss ✐♥ t❤❡ s❛♠❡ ❡①❡r❝✐s❡ t❤❡ r❡•❡✈❛♥❝❡ ♦❢ t❤❡ ♥✉♠❡r✐❝❛• ❞❡s✐❣♥ ❛♥❞ t❤❡ ♣r✐♦r ❞✐str✐❜✉t✐♦♥s✳ ❆❢t❡r ❞❡s❝r✐❜✐♥❣ ❤♦✇ ❝♦♠♣✉t✐♥❣ t❤✐s ❝r✐t❡r✐♦♥ ❢♦r t❤❡ ❡♠✉•❛t♦r ❛t ❤❛♥❞✱ ✐ts ❜❡❤❛✈✐♦r ✐s ✐••✉str❛t❡❞ ♦♥ ♥✉♠❡r✐❝❛• ❡①♣❡r✐♠❡♥ts✳ ❑❡②✲✇♦r❞s✿ ■♥✈❡rs❡ ♣r♦❜•❡♠s✱ ❇❛②❡s✐❛♥ ❛♥❛•②s✐s✱ ❑r✐❣✐♥❣✱ ❉❡s✐❣♥ ♦❢ ❊①♣❡r✐♠❡♥ts✱ ❆ss❡ss♠❡♥t ❊rr♦r✳ ❊✲♠❛✐• ❛❞❞r❡ss❡s✿ s❤✉❛✐✳❢✉❅❡❞❢✳❢r✱❣✐••❡s✳❝❡•❡✉①❅♠❛t❤✳✉✲♣s✉❞✳❢r✱♥✐❝♦•❛s✳❜♦✉sq✉❡t❅❡❞❢✳❢r✱♠❛t❤✐❡✉✳❝♦✉♣•❡t❅❡❞❢✳❢r * ❯♥✐✈❡rs✐t② ♦❢ P❛r✐s✲❙✉❞ ✶✶✱ ▼❛t❤❡♠❛t✐❝s ❉❡♣t✳✱ ❇❛t✳ ✹✷✺✱ ✾✶✹✵✺ ❖rs❛② ✭❋r❛♥❝❡✮ † ❊❉❋ |✫❉✱ ■♥❞✉str✐❛• |✐s❦ ▼❛♥❛❣❡♠❡♥t ❉❡♣t✳✱ ✻ q✉❛✐ ❲❛t✐❡r✱ ✼✽✹✵✶ ❈❤❛t♦✉ ✭❋r❛♥❝❡✮ hal-00708814, version 1 -15 Jun 2012

Research paper thumbnail of Density modification-based reliability sensitivity analysis

Sensitivity analysis of a numerical model, for instance simulating physical phenomena, is useful ... more Sensitivity analysis of a numerical model, for instance simulating physical phenomena, is useful to quantify the influence of the inputs on the model responses. This paper proposes a new sensitivity index, based upon the modification of the probability density function (pdf) of the random inputs, when the quantity of interest is a failure probability (probability that a model output exceeds a given threshold). An input is considered influential if the input pdf modification leads to a broad change in the failure probability. These sensitivity indices can be computed using the sole set of simulations that has already been used to estimate the failure probability, thus limiting the number of calls to the numerical model. In the case of a Monte Carlo sample, asymptotical properties of the indices are derived. Based on Kullback-Leibler divergence, several types of input perturbations are introduced. The relevance of this new sensitivity analysis method is analysed through three case studies.

Research paper thumbnail of Accounting for Age Uncertainty in Growth Modeling, the Case Study of Yellowfin Tuna (Thunnus albacares) of the Indian Ocean

PLoS ONE, 2013

Age estimates, typically determined by counting periodic growth increments in calcified structure... more Age estimates, typically determined by counting periodic growth increments in calcified structures of vertebrates, are the basis of population dynamics models used for managing exploited or threatened species. In fisheries research, the use of otolith growth rings as an indicator of fish age has increased considerably in recent decades. However, otolith readings include various sources of uncertainty. Current ageing methods, which converts an average count of rings into age, only provide periodic age estimates in which the range of uncertainty is fully ignored. In this study, we describe a hierarchical model for estimating individual ages from repeated otolith readings. The model was developed within a Bayesian framework to explicitly represent the sources of uncertainty associated with age estimation, to allow for individual variations and to include knowledge on parameters from expertise. The performance of the proposed model was examined through simulations, and then it was coupled to a two-stanza somatic growth model to evaluate the impact of the age estimation method on the age composition of commercial fisheries catches. We illustrate our approach using the saggital otoliths of yellowfin tuna of the Indian Ocean collected through large-scale mark-recapture experiments. The simulation performance suggested that the ageing error model was able to estimate the ageing biases and provide accurate age estimates, regardless of the age of the fish. Coupled with the growth model, this approach appeared suitable for modeling the growth of Indian Ocean yellowfin and is consistent with findings of previous studies. The simulations showed that the choice of the ageing method can strongly affect growth estimates with subsequent implications for age-structured data used as inputs for population models. Finally, our modeling approach revealed particularly useful to reflect uncertainty around age estimates into the process of growth estimation and it can be applied to any study relying on age estimation. Citation: Dortel E, Massiot-Granier F, Rivot E, Million J, Hallier J-P, et al. (2013) Accounting for Age Uncertainty in Growth Modeling, the Case Study of Yellowfin Tuna (Thunnus albacares) of the Indian Ocean. PLoS ONE 8(4): e60886.

Research paper thumbnail of Forecasting the Major Influences of Predation and Environment on Cod Recovery in the Northern Gulf of St. Lawrence

PLoS ONE, 2014

The northern Gulf of St. Lawrence (NGSL) stock of Atlantic cod (Gadus morhua), historically the s... more The northern Gulf of St. Lawrence (NGSL) stock of Atlantic cod (Gadus morhua), historically the second largest cod population in the Western Atlantic, has known a severe collapse during the early 1990 s and is currently considered as endangered by the Committee on the Status of Endangered Wildlife in Canada. As for many fish populations over the world which are currently being heavily exploited or overfished, urgent management actions in the form of recovery plans are needed for restoring this stock to sustainable levels. Stochastic projections based on a statistical population model incorporating predation were conducted over a period of 30 years (2010-2040) to assess the expected outcomes of alternative fishing strategies on the stock recovery under different scenarios of harp seal (Pagophilus groenlandicus) abundance and environmental conditions. This sensitivity study shows that water temperature is key in the rebuilding of the NGSL cod stock. Model projections suggest that maintaining the current management practice under cooler water temperatures is likely to maintain the species in an endangered status. Under current or warmer conditions in the Gulf of St. Lawrence, partial recovery might only be achieved by significant reductions in both fishing and predation pressure. In the medium-term, a management strategy that reduces catch could be favoured over a complete moratorium so as to minimize socio-economic impacts on the industry.

Research paper thumbnail of Role of predation by harp seals Pagophilus groenlandicus in the collapse and non-recovery of northern Gulf of St. Lawrence cod Gadus morhua

Marine Ecology Progress Series, 2009

A statistical catch-at-age model was developed to assess the effects of predation by the northwes... more A statistical catch-at-age model was developed to assess the effects of predation by the northwest Atlantic harp seal population on northern Gulf of St. Lawrence cod by estimating the relative importance of different sources of mortality that affected the stock during a period of collapse and non-recovery. Cod recruitment at age 1 is modeled via a non-linear stock-recruitment relationship based on total egg production and accounts for changes in female length-at-maturity and cod condition. Natural mortality other than seal predation also depends on cod condition used as an integrative index of changes in environmental conditions. The linkage between seals and cod is modeled through a multi-age functional response that was derived from the reconstruction of the seal diet using morphometric relationships and stomach contents of more than 200 seals collected between 1998 and 2001. The model was fitted following a maximum likelihood estimation approach to a scientific survey abundance index (1984 to 2006). Model results show that the collapse of the northern Gulf of St. Lawrence cod stock was mainly due to the combination of high fishing mortality rates and poor environmental conditions in the early to mid-1990s contributing to the current state of recruitment overfishing. The increase in harp seal abundance during 1984 to 2006 was reflected by an increase in predation mortality for the young cod age-groups targeted by seals. Although current levels of predation mortality affect cod spawning biomass, the lack of recovery of the NGSL cod stock seems mainly due to the very poor recruitment.

Research paper thumbnail of An alternative competing risk model to the Weibull distribution for modelling aging in lifetime data analysis

Lifetime Data Analysis, 2006

A simple competing risk distribution as a possible alternative to the Weibull distribution in lif... more A simple competing risk distribution as a possible alternative to the Weibull distribution in lifetime analysis is proposed. This distribution corresponds to the minimum between exponential and Weibull distributions. Our motivation is to take account of both accidental and aging failures in lifetime data analysis. First, the main characteristics of this distribution are presented. Then the estimation of its parameters are considered through maximum likelihood and Bayesian inference. In particular the existence of a unique consistent root of the likelihood equations is proved. Decision tests to choose between an exponential, Weibull and this competing risk distribution are presented. And this alternative model is compared to the Weibull model from numerical experiments on both real and simulated data sets, especially in an industrial context.

Research paper thumbnail of Redefining the maximum sustainable yield for the Schaefer population model including multiplicative environmental noise

Journal of theoretical biology, Jan 7, 2008

The focus of this article is to investigate the biological reference points, such as the maximum ... more The focus of this article is to investigate the biological reference points, such as the maximum sustainable yield (MSY), in a common Schaefer (logistic) surplus production model in the presence of a multiplicative environmental noise. This type of model is used in fisheries stock assessment as a first-hand tool for biomass modelling. Under the assumption that catches are proportional to the biomass, we derive new conditions on the environmental noise distribution such that stationarity exists and extinction is avoided. We then get new explicit results about the stationary behavior of the biomass distribution for a particular specification of the noise, namely the biomass distribution itself and a redefinition of the MSY and related quantities that now depend on the value of the variance of the noise. Consequently, we obtain a more precise vision of how less optimistic the stochastic version of the MSY can be than the traditionally used (deterministic) MSY. In addition, we give empi...

Research paper thumbnail of Density-dependence can be revealed by modelling the variance in the stock-recruitment process: an application to flatfish

ICES Journal of Marine Science, 2014

Recruitment success in marine species is mostly driven by the high and variable mortality of firs... more Recruitment success in marine species is mostly driven by the high and variable mortality of first life stages, and the relationships between stock and recruitment are then largely dominated by residual variability. We show that analysing the residual variability may provide insights on the density-dependence process occurring during the recruitment. Following the seminal formulation of Minto et al. (Survival variabilityand population density in fish populations. Nature, 2008), we show that when recruitment is considered as a sequence of a pelagic stage with stochastic density-independent mortality followed by a second stage with stochastic density-dependent mortality, then the variability of the recruitment rate per spawning biomass (RPSB) should be a decreasing function of the spawning biomass. Using stock-recruit data of 148 stocks from the RAM legacy database, we provide a test of this hypothesis by showing that the variability of RPSB is lower for fish species with the higher concentration during juvenile stages. Second, a hierarchical Bayesian model (HBM) is built to derive a meta-analysis of stock-recruit data for 39 flatfish stocks, characterized by a high concentration of juveniles in coastal nursery habitats. Results of the HBM show that the variance of the RPSB decreases with the spawning biomass for almost all stocks, thus providing strong evidence of density-dependence during the recruitment process. Finally, we attempt to relate patterns in recruitment variance to relevant life-history traits of flatfish species.

Research paper thumbnail of Estimating discrete Markov models from various incomplete data schemes

Computational Statistics & Data Analysis, 2012

The parameters of a discrete stationary Markov model are transition probabilities between states.... more The parameters of a discrete stationary Markov model are transition probabilities between states. Traditionally, data consists in sequences of observed states for a given number of individuals over the whole observation period. In such a case, the estimation of transition probabilities is straightforwardly made by counting one-step moves from a given state to another. In many real-life problems, however, the inference is much more difficult as state sequences are not fully observed, namely the state of each individual is known only for some given values of the time variable t. In this paper we give a review of this field, focusing on Monte Carlo Markov Chain (MCMC) algorithms to perform Bayesian inference and evaluate posterior distributions of the transition probabilities in this missing-data framework. We also propose a way to accelerate the classical Metropolis-Hastings technique for typical reliability problems, taking advantage of the dependence between the matrix rows to build an adaptive MCMC.

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

Canadian Journal of Fisheries and Aquatic Sciences, 2010

Landings from fisheries are often underreported, that is, the true landings are greater than thos... more 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.

Research paper thumbnail of An integrated Bayesian modeling approach for the growth of Indian Ocean yellowfin tuna

Fisheries Research, 2015

The Indian Ocean Tuna Tagging Program provided a unique opportunity to collect demographic data o... more The Indian Ocean Tuna Tagging Program provided a unique opportunity to collect demographic data on the key commercially targeted tropical tuna species in the Indian Ocean. In this paper, we focused on estimating growth rates for one of these species, yellowfin (Thunnus albacares). Whilst most growth studies only draw on one data source, in this study we use a range of data sources: individual growth rates derived from yellowfin that were tagged and recaptured, direct age estimates obtained through otolith readings, and length-frequency data collected from the purse seine fishery between 2000 and 2010. To combine these data sources, we used an integrated Bayesian model that allowed us to account for the process and measurement errors associated with each data set. Our results indicate that the gradual addition of each data type improved the model's parameter estimations. The Bayesian framework was useful, as it allowed us to account for uncertainties associated with age estimates and to provide additional information on some parameters (e.g., asymptotic length). Our results support the existence of a complex growth pattern for Indian Ocean yellowfin, with two distinct growth phases between the immature and mature life stages. Such complex growth patterns, however, require additional information on absolute age of fish and transition rates between growth stanzas. This type of information is not available from the data. We suggest that bioenergetic models may address this current data gap. This modeling approach explicitly considers the allocation of metabolic energy in tuna and may offer a way to understand the underlying mechanisms that drive the observed growth patterns.

Research paper thumbnail of Diagnostics of prior-data agreement in applied Bayesian analysis

Research paper thumbnail of Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Risk analysis : an official publication of the Society for Risk Analysis, 2015

We consider the problem of estimating the probability of detection (POD) of flaws in an industria... more We consider the problem of estimating the probability of detection (POD) of flaws in an industrial steel component. Modeled as an increasing function of the flaw height, the POD characterizes the detection process; it is also involved in the estimation of the flaw size distribution, a key input parameter of physical models describing the behavior of the steel component when submitted to extreme thermodynamic loads. Such models are used to assess the resistance of highly reliable systems whose failures are seldom observed in practice. We develop a Bayesian method to estimate the flaw size distribution and the POD function, using flaw height measures from periodic in-service inspections conducted with an ultrasonic detection device, together with measures from destructive lab experiments. Our approach, based on approximate Bayesian computation (ABC) techniques, is applied to a real data set and compared to maximum likelihood estimation (MLE) and a more classical approach based on Mark...

Research paper thumbnail of Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Nonparametric Estimation of the Probability of Detection of Flaws in an Industrial Component, from Destructive and Nondestructive Testing Data, Using Approximate Bayesian Computation

Risk analysis : an official publication of the Society for Risk Analysis, 2015

We consider the problem of estimating the probability of detection (POD) of flaws in an industria... more We consider the problem of estimating the probability of detection (POD) of flaws in an industrial steel component. Modeled as an increasing function of the flaw height, the POD characterizes the detection process; it is also involved in the estimation of the flaw size distribution, a key input parameter of physical models describing the behavior of the steel component when submitted to extreme thermodynamic loads. Such models are used to assess the resistance of highly reliable systems whose failures are seldom observed in practice. We develop a Bayesian method to estimate the flaw size distribution and the POD function, using flaw height measures from periodic in-service inspections conducted with an ultrasonic detection device, together with measures from destructive lab experiments. Our approach, based on approximate Bayesian computation (ABC) techniques, is applied to a real data set and compared to maximum likelihood estimation (MLE) and a more classical approach based on Mark...

Research paper thumbnail of Inversion probabiliste en analyse d'incertitude

Inversion probabiliste en analyse d'incertitude

Research paper thumbnail of A protocol for integrating FED and expert data in a study of durability using the Weibull distribution

The main issues raised by the estimation of lifetime parametric models used in industrial modelli... more The main issues raised by the estimation of lifetime parametric models used in industrial modelling of reliability are censoring and FED (Feedback Experience Data) sample size. Many studies are facing homogeneous, small-sized, censored failure times which have to be integrated into Bayesian procedures with informative prior parameter. This way of dealing with statistical inference has been especially followed by EDF for predicting failures on nuclear material. The example of the Weibull distribution will be here thoroughfully treated. Firstly experts have to be asked about the durability of a material with precise and simple questions. According to the choice of the considered model, prior point estimations and confidence intervals about parameters must be given, directly or indirectly, by experts. Secondly efficient modelling has to be chosen for informative prior distributions. Once, it must produce posterior distributions easily estimated by classical methods. But computation complexity is often a limiting factor of Bayesian inference. The impact of prior choices on posterior results must be simple to derive. Then, hyperparameters of these prior distributions must be evaluated linking the intrinsic properties of the prior densities (mean, mode, variance, etc.) with expert information on parameters.

Research paper thumbnail of Chassot2009

Research paper thumbnail of Estimation de modèles markoviens discrets dans un cadre industriel fiabiliste à données manquantes

Les modèles markoviens sont particulièrement utiles pour décrire des systèmes qui, au long de leu... more Les modèles markoviens sont particulièrement utiles pour décrire des systèmes qui, au long de leur vie, passent à travers différents états. Les paramètres de ces modèles sont les probabilités de transition entre les états. Normalement, les données disponibles pour l'inférence statistique sont des séquences temporelles d'états pour un nombre donné d'individus. Quand les séquences sont incomplètes, l'estimation de la matrice de transition n'est pas triviale et demande l'utilisation de techniques plus avancées. Dans cette communication, nous nous focalisons sur l'estimation bayésienne des probabilités de transition. Premièrement, nous présentons différentes méthodes MCMC, en fonction de la structure des données manquantes. Ensuite, nous proposons une manière d'accélérer les calculs MCMC en tenant compte de la dépendance entre les lignes de la matrice de transition. Finalement, nous montrons les résultats d'essais simulés menés sur des matrices typiqu...

Research paper thumbnail of On the Practical Interest of Discrete Inverse Pólya and Weibull-1 Models in Industrial Reliability Studies

Quality and Reliability Engineering International, 2015

Engineers often cope with the problem of assessing the lifetime of industrial components, under t... more Engineers often cope with the problem of assessing the lifetime of industrial components, under the basis of observed industrial feedback data. Usually, lifetime is modelled as a continuous random variable, for instance exponentially or Weibull distributed. However, in some cases, the features of the piece of equipment under investigation rather suggest the use of discrete probabilistic models. This happens for an equipment which only operates on cycles or on demand. In these cases, the lifetime is rather measured in number of cycles or number of demands before failure, therefore, in theory, discrete models should be more appropriate. This article aims at bringing some light to the practical interest for the reliability engineer in using two discrete models among the most popular: the Inverse Pólya distribution (IPD), based on a Pólya urn scheme, and the so-called Weibull-1 (W1) model. It is showed that, for different reasons, the practical use of both models should be restricted to specific industrial situations. In particular, when nothing is a priori known over the nature of ageing and/or data are heavily right-censored, they can remain of limited interest with respect to more flexible continuous lifetime models such as the usual Weibull distribution. Nonetheless, the intuitive meaning given to the IPD distribution favors its use by engineers in low (decelerated) ageing situations.

Research paper thumbnail of Une modelisation de duree de vie a risques de defaillance concurrents

Une modelisation de duree de vie a risques de defaillance concurrents

A simple competing risk distribution as a possible alternative to the Weibull distri- bution in l... more A simple competing risk distribution as a possible alternative to the Weibull distri- bution in lifetimes analysis is proposed. This distribution is the minimum between an exponential and a Weibull distributions. First, its main characteristics are pre- sented. Then the estimation of its parameters are considered through maximum likelihood and Bayesian inference. Statistical tests to choose between a Weibull distribution

Research paper thumbnail of Bayesian inference for inverse problems occurring in uncertainty analysis

International Journal for Uncertainty Quantification, 2014

Pr♦•❡❝t✲❚❡❛♠s ❙❊▲❊❈❚ ■♥r✐❛ ❛♥❞ ❊❉❋ |✫❉ |❡s❡❛r❝❤ |❡♣♦rt ♥➦ ✼✾✾✺ ✖ ❏✉✐♥ ✷✵✶✷ ✖ ✷✼ ♣❛❣❡s ❆❜str❛❝t✿ ❚... more Pr♦•❡❝t✲❚❡❛♠s ❙❊▲❊❈❚ ■♥r✐❛ ❛♥❞ ❊❉❋ |✫❉ |❡s❡❛r❝❤ |❡♣♦rt ♥➦ ✼✾✾✺ ✖ ❏✉✐♥ ✷✵✶✷ ✖ ✷✼ ♣❛❣❡s ❆❜str❛❝t✿ ❚❤❡ ✐♥✈❡rs❡ ♣r♦❜•❡♠ ❝♦♥s✐❞❡r❡❞ ❤❡r❡ ✐s t♦ ❡st✐♠❛t❡ t❤❡ ❞✐str✐❜✉t✐♦♥ ♦❢ ❛ ♥♦♥✲♦❜s❡r✈❡❞ r❛♥❞♦♠ ✈❛r✐❛❜•❡ X ❢r♦♠ s♦♠❡ ♥♦✐s② ♦❜s❡r✈❡❞ ❞❛t❛ Y •✐♥❦❡❞ t♦ X t❤r♦✉❣❤ ❛ t✐♠❡✲❝♦♥s✉♠✐♥❣ ♣❤②s✐❝❛• ♠♦❞❡• H✳ ❇❛②❡s✐❛♥ ✐♥❢❡r❡♥❝❡ ✐s ❝♦♥s✐❞❡r❡❞ t♦ t❛❦❡ ✐♥t♦ ❛❝❝♦✉♥t ♣r✐♦r ❡①♣❡rt ❦♥♦✇•❡❞❣❡ ♦♥ X ✐♥ ❛ s♠❛•• s❛♠♣•❡ s✐③❡ s❡tt✐♥❣✳ ❆ ▼❡tr♦♣♦•✐s✲❍❛st✐♥❣s ✇✐t❤✐♥ •✐❜❜s ❛•❣♦r✐t❤♠ ✐s ♣r♦♣♦s❡❞ t♦ ❝♦♠♣✉t❡ t❤❡ ♣♦st❡r✐♦r ❞✐str✐❜✉t✐♦♥ ♦❢ t❤❡ ♣❛r❛♠❡t❡rs ♦❢ X t❤r♦✉❣❤ ❛ ❞❛t❛ ❛✉❣♠❡♥t❛t✐♦♥ ♣r♦❝❡ss✳ ❙✐♥❝❡ ❝❛••s t♦ H ❛r❡ q✉✐t❡ ❡①♣❡♥s✐✈❡✱ t❤✐s ✐♥❢❡r❡♥❝❡ ✐s ❛❝❤✐❡✈❡❞ ❜② r❡♣•❛❝✐♥❣ H ✇✐t❤ ❛ ❦r✐❣✐♥❣ ❡♠✉•❛t♦r ✐♥t❡r♣♦•❛t✐♥❣ H ❢r♦♠ ❛ ♥✉♠❡r✐❝❛• ❞❡s✐❣♥ ♦❢ ❡①♣❡r✐♠❡♥ts✳ ❚❤✐s ❛♣♣r♦❛❝❤ ✐♥✈♦•✈❡s s❡✈❡r❛• ❡rr♦rs ♦❢ ❞✐✛❡r❡♥t ♥❛t✉r❡ ❛♥❞✱ ✐♥ t❤✐s ♣❛♣❡r✱ ✇❡ ♣❛② ❡✛♦rt t♦ ♠❡❛s✉r❡ ❛♥❞ r❡❞✉❝❡ t❤❡ ♣♦ss✐❜•❡ ✐♠♣❛❝t ♦❢ t❤♦s❡ ❡rr♦rs✳ ■♥ ♣❛rt✐❝✉•❛r✱ ✇❡ ♣r♦♣♦s❡ t♦ ✉s❡ t❤❡ s♦✲❝❛••❡❞ ❉❆❈ ❝r✐t❡r✐♦♥ t♦ ❛ss❡ss ✐♥ t❤❡ s❛♠❡ ❡①❡r❝✐s❡ t❤❡ r❡•❡✈❛♥❝❡ ♦❢ t❤❡ ♥✉♠❡r✐❝❛• ❞❡s✐❣♥ ❛♥❞ t❤❡ ♣r✐♦r ❞✐str✐❜✉t✐♦♥s✳ ❆❢t❡r ❞❡s❝r✐❜✐♥❣ ❤♦✇ ❝♦♠♣✉t✐♥❣ t❤✐s ❝r✐t❡r✐♦♥ ❢♦r t❤❡ ❡♠✉•❛t♦r ❛t ❤❛♥❞✱ ✐ts ❜❡❤❛✈✐♦r ✐s ✐••✉str❛t❡❞ ♦♥ ♥✉♠❡r✐❝❛• ❡①♣❡r✐♠❡♥ts✳ ❑❡②✲✇♦r❞s✿ ■♥✈❡rs❡ ♣r♦❜•❡♠s✱ ❇❛②❡s✐❛♥ ❛♥❛•②s✐s✱ ❑r✐❣✐♥❣✱ ❉❡s✐❣♥ ♦❢ ❊①♣❡r✐♠❡♥ts✱ ❆ss❡ss♠❡♥t ❊rr♦r✳ ❊✲♠❛✐• ❛❞❞r❡ss❡s✿ s❤✉❛✐✳❢✉❅❡❞❢✳❢r✱❣✐••❡s✳❝❡•❡✉①❅♠❛t❤✳✉✲♣s✉❞✳❢r✱♥✐❝♦•❛s✳❜♦✉sq✉❡t❅❡❞❢✳❢r✱♠❛t❤✐❡✉✳❝♦✉♣•❡t❅❡❞❢✳❢r * ❯♥✐✈❡rs✐t② ♦❢ P❛r✐s✲❙✉❞ ✶✶✱ ▼❛t❤❡♠❛t✐❝s ❉❡♣t✳✱ ❇❛t✳ ✹✷✺✱ ✾✶✹✵✺ ❖rs❛② ✭❋r❛♥❝❡✮ † ❊❉❋ |✫❉✱ ■♥❞✉str✐❛• |✐s❦ ▼❛♥❛❣❡♠❡♥t ❉❡♣t✳✱ ✻ q✉❛✐ ❲❛t✐❡r✱ ✼✽✹✵✶ ❈❤❛t♦✉ ✭❋r❛♥❝❡✮ hal-00708814, version 1 -15 Jun 2012

Research paper thumbnail of Density modification-based reliability sensitivity analysis

Sensitivity analysis of a numerical model, for instance simulating physical phenomena, is useful ... more Sensitivity analysis of a numerical model, for instance simulating physical phenomena, is useful to quantify the influence of the inputs on the model responses. This paper proposes a new sensitivity index, based upon the modification of the probability density function (pdf) of the random inputs, when the quantity of interest is a failure probability (probability that a model output exceeds a given threshold). An input is considered influential if the input pdf modification leads to a broad change in the failure probability. These sensitivity indices can be computed using the sole set of simulations that has already been used to estimate the failure probability, thus limiting the number of calls to the numerical model. In the case of a Monte Carlo sample, asymptotical properties of the indices are derived. Based on Kullback-Leibler divergence, several types of input perturbations are introduced. The relevance of this new sensitivity analysis method is analysed through three case studies.

Research paper thumbnail of Accounting for Age Uncertainty in Growth Modeling, the Case Study of Yellowfin Tuna (Thunnus albacares) of the Indian Ocean

PLoS ONE, 2013

Age estimates, typically determined by counting periodic growth increments in calcified structure... more Age estimates, typically determined by counting periodic growth increments in calcified structures of vertebrates, are the basis of population dynamics models used for managing exploited or threatened species. In fisheries research, the use of otolith growth rings as an indicator of fish age has increased considerably in recent decades. However, otolith readings include various sources of uncertainty. Current ageing methods, which converts an average count of rings into age, only provide periodic age estimates in which the range of uncertainty is fully ignored. In this study, we describe a hierarchical model for estimating individual ages from repeated otolith readings. The model was developed within a Bayesian framework to explicitly represent the sources of uncertainty associated with age estimation, to allow for individual variations and to include knowledge on parameters from expertise. The performance of the proposed model was examined through simulations, and then it was coupled to a two-stanza somatic growth model to evaluate the impact of the age estimation method on the age composition of commercial fisheries catches. We illustrate our approach using the saggital otoliths of yellowfin tuna of the Indian Ocean collected through large-scale mark-recapture experiments. The simulation performance suggested that the ageing error model was able to estimate the ageing biases and provide accurate age estimates, regardless of the age of the fish. Coupled with the growth model, this approach appeared suitable for modeling the growth of Indian Ocean yellowfin and is consistent with findings of previous studies. The simulations showed that the choice of the ageing method can strongly affect growth estimates with subsequent implications for age-structured data used as inputs for population models. Finally, our modeling approach revealed particularly useful to reflect uncertainty around age estimates into the process of growth estimation and it can be applied to any study relying on age estimation. Citation: Dortel E, Massiot-Granier F, Rivot E, Million J, Hallier J-P, et al. (2013) Accounting for Age Uncertainty in Growth Modeling, the Case Study of Yellowfin Tuna (Thunnus albacares) of the Indian Ocean. PLoS ONE 8(4): e60886.

Research paper thumbnail of Forecasting the Major Influences of Predation and Environment on Cod Recovery in the Northern Gulf of St. Lawrence

PLoS ONE, 2014

The northern Gulf of St. Lawrence (NGSL) stock of Atlantic cod (Gadus morhua), historically the s... more The northern Gulf of St. Lawrence (NGSL) stock of Atlantic cod (Gadus morhua), historically the second largest cod population in the Western Atlantic, has known a severe collapse during the early 1990 s and is currently considered as endangered by the Committee on the Status of Endangered Wildlife in Canada. As for many fish populations over the world which are currently being heavily exploited or overfished, urgent management actions in the form of recovery plans are needed for restoring this stock to sustainable levels. Stochastic projections based on a statistical population model incorporating predation were conducted over a period of 30 years (2010-2040) to assess the expected outcomes of alternative fishing strategies on the stock recovery under different scenarios of harp seal (Pagophilus groenlandicus) abundance and environmental conditions. This sensitivity study shows that water temperature is key in the rebuilding of the NGSL cod stock. Model projections suggest that maintaining the current management practice under cooler water temperatures is likely to maintain the species in an endangered status. Under current or warmer conditions in the Gulf of St. Lawrence, partial recovery might only be achieved by significant reductions in both fishing and predation pressure. In the medium-term, a management strategy that reduces catch could be favoured over a complete moratorium so as to minimize socio-economic impacts on the industry.

Research paper thumbnail of Role of predation by harp seals Pagophilus groenlandicus in the collapse and non-recovery of northern Gulf of St. Lawrence cod Gadus morhua

Marine Ecology Progress Series, 2009

A statistical catch-at-age model was developed to assess the effects of predation by the northwes... more A statistical catch-at-age model was developed to assess the effects of predation by the northwest Atlantic harp seal population on northern Gulf of St. Lawrence cod by estimating the relative importance of different sources of mortality that affected the stock during a period of collapse and non-recovery. Cod recruitment at age 1 is modeled via a non-linear stock-recruitment relationship based on total egg production and accounts for changes in female length-at-maturity and cod condition. Natural mortality other than seal predation also depends on cod condition used as an integrative index of changes in environmental conditions. The linkage between seals and cod is modeled through a multi-age functional response that was derived from the reconstruction of the seal diet using morphometric relationships and stomach contents of more than 200 seals collected between 1998 and 2001. The model was fitted following a maximum likelihood estimation approach to a scientific survey abundance index (1984 to 2006). Model results show that the collapse of the northern Gulf of St. Lawrence cod stock was mainly due to the combination of high fishing mortality rates and poor environmental conditions in the early to mid-1990s contributing to the current state of recruitment overfishing. The increase in harp seal abundance during 1984 to 2006 was reflected by an increase in predation mortality for the young cod age-groups targeted by seals. Although current levels of predation mortality affect cod spawning biomass, the lack of recovery of the NGSL cod stock seems mainly due to the very poor recruitment.

Research paper thumbnail of An alternative competing risk model to the Weibull distribution for modelling aging in lifetime data analysis

Lifetime Data Analysis, 2006

A simple competing risk distribution as a possible alternative to the Weibull distribution in lif... more A simple competing risk distribution as a possible alternative to the Weibull distribution in lifetime analysis is proposed. This distribution corresponds to the minimum between exponential and Weibull distributions. Our motivation is to take account of both accidental and aging failures in lifetime data analysis. First, the main characteristics of this distribution are presented. Then the estimation of its parameters are considered through maximum likelihood and Bayesian inference. In particular the existence of a unique consistent root of the likelihood equations is proved. Decision tests to choose between an exponential, Weibull and this competing risk distribution are presented. And this alternative model is compared to the Weibull model from numerical experiments on both real and simulated data sets, especially in an industrial context.

Research paper thumbnail of Redefining the maximum sustainable yield for the Schaefer population model including multiplicative environmental noise

Journal of theoretical biology, Jan 7, 2008

The focus of this article is to investigate the biological reference points, such as the maximum ... more The focus of this article is to investigate the biological reference points, such as the maximum sustainable yield (MSY), in a common Schaefer (logistic) surplus production model in the presence of a multiplicative environmental noise. This type of model is used in fisheries stock assessment as a first-hand tool for biomass modelling. Under the assumption that catches are proportional to the biomass, we derive new conditions on the environmental noise distribution such that stationarity exists and extinction is avoided. We then get new explicit results about the stationary behavior of the biomass distribution for a particular specification of the noise, namely the biomass distribution itself and a redefinition of the MSY and related quantities that now depend on the value of the variance of the noise. Consequently, we obtain a more precise vision of how less optimistic the stochastic version of the MSY can be than the traditionally used (deterministic) MSY. In addition, we give empi...

Research paper thumbnail of Density-dependence can be revealed by modelling the variance in the stock-recruitment process: an application to flatfish

ICES Journal of Marine Science, 2014

Recruitment success in marine species is mostly driven by the high and variable mortality of firs... more Recruitment success in marine species is mostly driven by the high and variable mortality of first life stages, and the relationships between stock and recruitment are then largely dominated by residual variability. We show that analysing the residual variability may provide insights on the density-dependence process occurring during the recruitment. Following the seminal formulation of Minto et al. (Survival variabilityand population density in fish populations. Nature, 2008), we show that when recruitment is considered as a sequence of a pelagic stage with stochastic density-independent mortality followed by a second stage with stochastic density-dependent mortality, then the variability of the recruitment rate per spawning biomass (RPSB) should be a decreasing function of the spawning biomass. Using stock-recruit data of 148 stocks from the RAM legacy database, we provide a test of this hypothesis by showing that the variability of RPSB is lower for fish species with the higher concentration during juvenile stages. Second, a hierarchical Bayesian model (HBM) is built to derive a meta-analysis of stock-recruit data for 39 flatfish stocks, characterized by a high concentration of juveniles in coastal nursery habitats. Results of the HBM show that the variance of the RPSB decreases with the spawning biomass for almost all stocks, thus providing strong evidence of density-dependence during the recruitment process. Finally, we attempt to relate patterns in recruitment variance to relevant life-history traits of flatfish species.

Research paper thumbnail of Estimating discrete Markov models from various incomplete data schemes

Computational Statistics & Data Analysis, 2012

The parameters of a discrete stationary Markov model are transition probabilities between states.... more The parameters of a discrete stationary Markov model are transition probabilities between states. Traditionally, data consists in sequences of observed states for a given number of individuals over the whole observation period. In such a case, the estimation of transition probabilities is straightforwardly made by counting one-step moves from a given state to another. In many real-life problems, however, the inference is much more difficult as state sequences are not fully observed, namely the state of each individual is known only for some given values of the time variable t. In this paper we give a review of this field, focusing on Monte Carlo Markov Chain (MCMC) algorithms to perform Bayesian inference and evaluate posterior distributions of the transition probabilities in this missing-data framework. We also propose a way to accelerate the classical Metropolis-Hastings technique for typical reliability problems, taking advantage of the dependence between the matrix rows to build an adaptive MCMC.

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

Canadian Journal of Fisheries and Aquatic Sciences, 2010

Landings from fisheries are often underreported, that is, the true landings are greater than thos... more 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.

Research paper thumbnail of An integrated Bayesian modeling approach for the growth of Indian Ocean yellowfin tuna

Fisheries Research, 2015

The Indian Ocean Tuna Tagging Program provided a unique opportunity to collect demographic data o... more The Indian Ocean Tuna Tagging Program provided a unique opportunity to collect demographic data on the key commercially targeted tropical tuna species in the Indian Ocean. In this paper, we focused on estimating growth rates for one of these species, yellowfin (Thunnus albacares). Whilst most growth studies only draw on one data source, in this study we use a range of data sources: individual growth rates derived from yellowfin that were tagged and recaptured, direct age estimates obtained through otolith readings, and length-frequency data collected from the purse seine fishery between 2000 and 2010. To combine these data sources, we used an integrated Bayesian model that allowed us to account for the process and measurement errors associated with each data set. Our results indicate that the gradual addition of each data type improved the model's parameter estimations. The Bayesian framework was useful, as it allowed us to account for uncertainties associated with age estimates and to provide additional information on some parameters (e.g., asymptotic length). Our results support the existence of a complex growth pattern for Indian Ocean yellowfin, with two distinct growth phases between the immature and mature life stages. Such complex growth patterns, however, require additional information on absolute age of fish and transition rates between growth stanzas. This type of information is not available from the data. We suggest that bioenergetic models may address this current data gap. This modeling approach explicitly considers the allocation of metabolic energy in tuna and may offer a way to understand the underlying mechanisms that drive the observed growth patterns.