Search

Scientific bottom-trawl surveys are ecological observation programs conducted along continental shelves and slopes of seas and oceans that sample marine communities associated with the seafloor. These surveys report taxa occurrence, abundance and/or weight in space and time, and contribute to fisheries management as well as population and biodiversity research. Bottom-trawl surveys are conducted all over the world and represent a unique opportunity to understand ocean biogeography, macroecology, and global change. However, combining these data together for cross-ecosystem analyses remains challenging. Here, we present an integrated dataset of 29 publicly available bottom-trawl surveys conducted in national waters of 18 countries that are standardized and pre-processed, covering a total of 2,170 sampled fish taxa and 216,548 hauls collected from 1963 to 2021. We describe the processing steps to create the dataset, flags, and standardization methods that we developed to assist users in conducting spatio-temporal analyses with stable regional survey footprints. The aim of this dataset is to support research, marine conservation, and management in the context of global change.

Marine heatwaves have been linked to negative ecological effects in recent decades1,2. If marine heatwaves regularly induce community reorganization and biomass collapses in fishes, the consequences could be catastrophic for ecosystems, fisheries and human communities3,4. However, the extent to which marine heatwaves have negative impacts on fish biomass or community composition, or even whether their effects can be distinguished from natural and sampling variability, remains unclear. We investigated the effects of 248 sea-bottom heatwaves from 1993 to 2019 on marine fishes by analysing 82,322 hauls (samples) from long-term scientific surveys of continental shelf ecosystems in North America and Europe spanning the subtropics to the Arctic. Here we show that the effects of marine heatwaves on fish biomass were often minimal and could not be distinguished from natural and sampling variability. Furthermore, marine heatwaves were not consistently associated with tropicalization (gain of warm-affiliated species) or deborealization (loss of cold-affiliated species) in these ecosystems. Although steep declines in biomass occasionally occurred after marine heatwaves, these were the exception, not the rule. Against the highly variable backdrop of ocean ecosystems, marine heatwaves have not driven biomass change or community turnover in fish communities that support many of the world's largest and most productive fisheries.

Generating accurate data for stock assessments is resource-demanding, necessitating periodic evaluation of survey sampling designs and potential impacts on stock assessments. We developed a framework for bootstrapped resampling of survey age data and calculation of input sample sizes as a function of among-bootstrap variance in age compositions. Data from this bootstrap estimator were then used to evaluate the influence of alternative sampling rates and methods on uncertainty in estimates of the overfishing limit (OFL) calculated using stock assessment models. For dusky rockfish (Sebastes variabilis) and Pacific ocean perch (Sebastes alutus), a 10% decrease in the number of tows sampled upon led to a predicted 5%-6% increase in the CV of OFL (log-log slope = -0.576 to -0.486), which was greater than the 0%-2% increase from a 10% decrease in otolithsper-tow (log-log slope = -0.238 to -0.029). Application of this approach across all stocks monitored in the survey of interest is required to identify which stocks (i) benefit the most from increased sampling of ageing structures or (ii) cost the least in terms of OFL uncertainty owing to reduced sampling. Key words: age composition, data-weighting, stock assessment, sample size, otoliths, Introduction Sustainable fisheries management is contingent upon sound advice informed by the results of stock assessments, which describe the population dynamics of the resource and provide recommendations for future catch [...]

Marine biota are redistributing at a rapid pace in response to climate change and shifting seascapes. While changes in fish populations and community structure threaten the sustainability of fisheries, our capacity to adapt by tracking and projecting marine species remains a challenge due to data discontinuities in biological observations, lack of data availability, and mismatch between data and real species distributions. To assess the extent of this challenge, we review the global status and accessibility of ongoing scientific bottom trawl surveys. In total, we gathered metadata for 283,925 samples from 95 surveys conducted regularly from 2001 to 2019. We identified that 59% of the metadata collected are not publicly available, highlighting that the availability of data is the most important challenge to assess species redistributions under global climate change. Given that the primary purpose of surveys is to provide independent data to inform stock assessment of commercially important populations, we further highlight that single surveys do not cover the full range of the main commercial demersal fish species. An average of 18 surveys is needed to cover at least 50% of species ranges, demonstrating the importance of combining multiple surveys to evaluate species range shifts. We assess the potential for combining surveys to track transboundary species redistributions and show that differences in sampling schemes and inconsistency in sampling can be overcome with spatio-temporal modeling to follow species density redistributions. In light of our global assessment, we establish a framework for improving the management and conservation of transboundary and migrating marine demersal species. We provide directions to improve data availability and encourage countries to share survey data, to assess species vulnerabilities, and to support management adaptation in a time of climate-driven ocean changes.

The spatial distributions of marine fish populations are influenced by environmental conditions, intrinsic properties of the populations, and prior distribution. The influence of these factors may not be consistent across age classes. For this study, age composition estimates for walleye pollock (Gadus chalcogrammus) collected on bottom-trawl surveys in the Bering Sea were used to estimate range correlation indices, population centers of gravity, and effective area occupied. Age- specific density maps suggested a circular ontogenic migration during the summer feeding season, with the youngest and oldest groups most broadly distributed. Range correlation analysis among age groups and year classes provided clear evidence of a population cohort effect in the spatial distribution of the population. Variance decomposition analysis indicated that the variance in the spatial distribution of age groups during summer was influenced by the initial distribution of that cohort as recruits. Model-based analyses showed that extrinsic temperature variables affected the youngest and oldest age classes the most, but provided no indication of age-related effects for intrinsic population factors. This study showed that both cohort and age-specific factors are important drivers of spatial distribution. Key words: demographics, age composition, distribution, walleye pollock, modeling, VAST La repartition spatiale de populations de poissons marins est influencee par les conditions ambiantes, des proprietes intrinseques des populations et la repartition prealable. L'influence de ces facteurs pourrait ne pas etre uniforme d'une classe d'age a l'autre. Nous avons utilise des estimations de la composition par age pour des goberges de l'Alaska (Gadus chalcogrammus) prelevees dans le cadre d'evaluations au chalut de fond dans la mer de Behring pour estimer des indices de correlation des aires de repartition, les centres de gravite de la population et la superficie occupee effective. Des cartes de densite selon l'age indiqueraient une migration ontogenetique circulaire durant la saison d'alimentation estivale, les groupes les plus jeunes et les plus vieux presentant les repartitions les plus vastes. L'analyse de correlation des aires de repartition de differents groupes d'age et classes d'annee fournit des preuves claires d'un effet populationnel de la cohorte dans la repartition spatiale de la population. L'analyse de decomposition de la variance indique que la variance de la repartition spatiale des groupes d'age durant l'ete est influencee par la repartition initiale de la cohorte au moment du recrutement. Des analyses reposant sur des modeles montrent que des variables extrinseques associees a la temperature exercent la plus grande influence sur les classes d'age les plus jeunes et les plus vieilles, mais elles ne revelent aucun effet relie a l'age de facteurs intrinseques a la population. L'etude demontre que des facteurs associes tant a la cohorte qu'a l'age exercent une importante influence sur la repartition spatiale. [Traduit par la Redaction] Mots-cles : demographie, composition par age, repartition, goberge de l'Alaska, modelisation, VAST, Introduction The spatial distributions of fish populations are influenced by many factors. Environmental variables that define habitat suitability are clearly important. For example, physiological tolerances and resource availability directly affect [...]

Szuwalski argues that varying age structure can affect surplus production and that recruitment is a better metric of productivity. We explain how our null model controlled for age structure and other processes as explanations for the temperature-production relationship. Surplus production includes growth, recruitment, and other processes and provides a more complete description of food production impacts than does recruitment alone.

Climate change is altering habitats for marine fishes and invertebrates, but the net effect of these changes on potential food production is unknown. We used temperature-dependent population models to measure the influence of warming on the productivity of 235 populations of 124 species in 38 ecoregions. Some populations responded significantly positively (n = 9 populations) and others responded significantly negatively (n = 19 populations) to warming, with the direction and magnitude of the response explained by ecoregion, taxonomy, life history, and exploitation history. Hindcasts indicate that the maximum sustainable yield of the evaluated populations decreased by 4.1% from 1930 to 2010, with five ecoregions experiencing losses of 15 to 35%. Outcomes of fisheries management-including long-term food provisioning-will be improved by accounting for changing productivity in a warmer ocean.

Changing distribution and abundance of small pelagic fishes may drive changes in predator distributions, affecting predator availability to fisheries and surveys. However, small pelagics are difficult to survey directly, so we developed a novel method of assessing the aggregate abundance of 21 small pelagic forage taxa via predator stomach contents. We used stomach contents collected from 22 piscivore species captured by multiple bottom trawl surveys within a vector autoregressive spatio-temporal model to assess trends of small pelagics on the Northeast US shelf. The goal was to develop a spatial "forage index" to inform survey and (or) fishery availability in the western North Atlantic bluefish (Pomatomus saltatrix) stock assessment. This spatially resolved index compared favorably with more traditional design-based survey biomass indices for forage species well sampled by surveys. However, our stomach content-based index better represented smaller unmanaged forage species that surveys are not designed to capture. The stomach-based forage index helped explain bluefish availability to the recreational fishery for stock assessment and provided insight into pelagic forage trends throughout the regional ecosystem. [ABSTRACT FROM AUTHOR]

In the face of biodiversity loss worldwide, it is paramount to quantify species' extinction risk to guide conservation efforts. The International Union for the Conservation of Nature (IUCN)'s Red List is considered the global standard for evaluating extinction risks. IUCN criteria also inform national extinction risk assessments. Bayesian models, including the state‐of‐the‐art JARA ('Just Another Red List Assessment') tool, deliver probabilistic statements about species falling into extinction risk categories, thereby enabling characterisation and communication of uncertainty in extinction risk assessments.We coupled the state‐of‐the‐art VAST ('Vector Autoregressive Spatio‐Temporal') modelling tool and JARA, for better informed Red List assessments of marine fishes. In this framework, VAST is fitted to scientific survey catch rate data to provide indices to JARA whose uncertainty is propagated to JARA outcomes suggesting extinction risk categories (under the population reduction criterion). In addition, VAST delivers a valuable habitat assessment to better understand what may be driving extinction risk in the study region. Here, we demonstrate the coupled VAST‐JARA modelling framework by applying it to five contrasting North Sea species, with or without a quantitative stock assessment and with different conservation statuses according to the latest global Red List assessments.The North Sea application coupled with previous assessments and studies suggest that, among the three elasmobranchs, starry ray is in most need of urgent research (and conservation actions where appropriate), followed by spurdog, while lesser‐spotted dogfish is increasing in biomass. Moreover, both the VAST‐JARA modelling framework and previous research indicate that, while European plaice is not of conservation concern, cod has likely met the IUCN criteria for being listed as Endangered recently.Synthesis and applications. The predictions of the VAST‐JARA modelling framework for North Sea species, including JARA output and VAST habitat assessment, constitute valuable supporting information to make interpretations based on Red List guidelines, which will help decision‐makers in their next North Sea Red List assessment. We foresee applications of the modelling framework to assist Red List assessments of numerous marine fishes worldwide. Our modelling framework has many potential advantageous uses, including informing resource management about climate change impacts on species' extinction risks. [ABSTRACT FROM AUTHOR]

Spatiotemporal predictions of bycatch (i.e., catch of nontargeted species) have shown promise as dynamic ocean management tools for reducing bycatch. However, which spatiotemporal model framework to use for generating these predictions is unclear. We evaluated a relatively new method, Gaussian Markov random fields (GMRFs), with two other frameworks, generalized additive models (GAMs) and random forests. We fit geostatistical delta-models to fisheries observer bycatch data for six species with a broad range of movement patterns (e.g., highly migratory sea turtles versus sedentary rockfish) and bycatch rates (percentage of observations with nonzero catch, 0.3%-96.2%). Random forests had better interpolation performance than the GMRF and GAM models for all six species, but random forests performance was more sensitive when predicting data at the edge of the fishery (i.e., spatial extrapolation). Using random forests to identify and remove the 5% highest bycatch risk fishing events reduced the bycatch-to-target species catch ratio by 34% on average. All models considerably reduced the bycatch-to-target ratio, demonstrating the clear potential of species distribution models to support spatial fishery management. Les predictions spatiotemporelles de prises accessoires (c.-a-d. prises d'especes non visees) sont prometteuses comme outils de gestion dynamique des oceans pour reduire ces prises. Cependant, la nature du cadre de modelisation spatiotemporelle a utiliser pour produire ces predictions n'est pas bien etablie. Nous avons evalue une methode relativement nouvelle, les champs aleatoires de Markov gaussiens (CAMG), et deux autres cadres, les modeles additifs generalises (MAG) et les forets aleatoires. Nous avons cale des modeles delta geostatistiques a des donnees d'observateurs des peches sur les prises accessoires representatives d'un vaste eventail de motifs de deplacement (p. ex. de tortues de mer tres migratrices a des sebastes sedentaires) et de frequences de prises accessoires (pourcentages d'observations de prise non nulle de 0,3 % a 96,2 %). Les forets aleatoires donnent de meilleures interpolations que les modeles CAMG et MAG pour les six especes, mais la performance des forets aleatoires est plus sensible quand il s'agit de predire des donnees a la limite de la peche (c.-a-d. pour l'extrapolation spatiale). L'utilisation de forets aleatoires pour cerner et retirer les 5 % des evenements de peche qui presentent le risque le plus eleve de prises accessoires reduit le rapport des prises accessoires et des prises de l'espece visee de 34 % en moyenne. Tous les modeles reduisent considerablement ce rapport, ce qui demontre clairement le potentiel des modeles de distribution d'especes comme outil pour appuyer la gestion spatiale des peches. [Traduit par la Redaction], Introduction Bycatch--catch of nontargeted species--occurs in nearly every commercial and recreational fishery and in many cases is a serious environmental and economic problem (Alverson et al. 1994; Davies et al. [...]

Accounting for marine stocks spatiotemporal complexity has become one of the most pressing improvements that should be added to the new generation of stock assessment. Disentangling persistent and dynamic population subcomponents and understanding their main drivers of variation are still stock-specific challenges. Here, we hypothesized that the spatiotemporal variability of density in two adjacent fish stocks is associated with spatially structured environmental processes across multiple spatiotemporal scales. To test this, we applied a generalized empirical orthogonal function and dynamic factor analysis to fishery-independent and -dependent data of red mullet, a highly commercial species, in the Western Mediterranean Sea. Areas with persistent and dynamic high aggregations were detected for both stock units. A large-scale climatic index and local open-ocean convection were associated with both stocks, while other variables exhibited stock-specific effects. We also revealed spatially structured density dynamics within the examined management units. This suggests a metapopulation structure and supports the future implementation of a spatial stock assessment. Considering the common—generally unrealistic—assumptions of panmictic structure and absence of connectivity with neighbouring stock units, our methodology can be applied to other species and systems with putative spatial complexity to inform more accurate population dynamics and structure. [ABSTRACT FROM AUTHOR]

Designing effective spatial management strategies is challenging because marine ecosystems are highly dynamic and opaque, and extractive entities such as fishing fleets respond endogenously to ecosystem changes in ways that depend on ecological and policy context. We present a modelling framework, marlin, that can be used to efficiently simulate the bio‐economic dynamics of marine systems in support of both management and research. We demonstrate marlin's capabilities by focusing on two case studies on the conservation and food production impacts of marine protected areas (MPAs): a coastal coral reef and a pelagic tuna fishery. In the coastal coral reef example, we explore how heterogeneity in species distributions and fleet preferences can affect distributional outcomes of MPAs. In the pelagic case study, we show how our model can be used to assess the climate resilience of different MPA design strategies, as well as the climate sensitivity of different fishing fleets. This paper demonstrates how intermediate complexity simulation of coupled bio‐economic dynamics can help communities predict and potentially manage trade‐offs among conservation, fisheries yields and distributional outcomes of management policies affected by spatial bio‐economic dynamics. [ABSTRACT FROM AUTHOR]

Species distribution models (SDMs) are an important tool for conservation and resource management. However, managers are often interested in derived quantities such as range or area occupied, and how these are calculated can have a large impact.Ecosystem‐based management typically requires spatial information about species distributions, which is increasingly generated from SDMs that are then processed to identify occupied habitat. Many types of SDMs exist, but there is little research regarding how this model‐choice affects outcomes when defining occupied habitat, in part because these models generate different types of output.We fit a suite of five SDMs to data for 208 species/life stage combinations in three marine ecosystems while ensuring that they all estimate a 'common currency' of numerical abundance. We then calculate out‐of‐sample predictive performance to weight these constituents in an ensemble SDM.Results show that this approach can reduce bias arising from a priori specification of individual SDMs resulting in a better fit to survey data (constituent SDMs had a median of 7% higher RMSE). The SDMs had a range of responses relative to the ensemble, with MaxEnt typically predicting a median 1.3% higher area occupied, and negative‐binomial GAMs predicting 21.4% lower area occupied.Two potential methods of identifying the area of occupied habitat from SDM outputs are compared—probability‐based and cumulative density‐based methods. We find that cumulative densities result in smaller estimates of area occupied, and we recommend careful consideration of how model‐choice affects occupied‐habitat estimates in spatial management.Policy implications: Finally, we discuss how the patterns identified during the 5‐year Review of Essential Fish Habitat for Alaska should be carefully considered by managers using SDMs to identify habitat that may be impacted by anthropogenic activities. [ABSTRACT FROM AUTHOR]

Fisheries scientists compare processes among species to estimate species productivity, management reference points, and climate sensitivities. Ecologists have developed "phylogenetic comparative methods" (PCMs) to address these questions, but there is surprisingly little application of PCM within fisheries science. Here, I bridge this gap by introducing PCM (including Brownian motion, Ornstein–Uhlenbeck, and Pagel's kappa and lambda models for species covariance), thereby showing that PCM generalizes the nested taxonomic random effects that are commonly used in fisheries science. I next summarize phylogenetic structural equation models (PSEMs), which extend the linear models that are commonly used in fisheries. Finally, I re‐analyse a high‐quality database used to predict mortality rates from longevity and/or growth parameters. I specifically propose a PSEM that reverts to a longevity‐based prediction when longevity information is available but uses phylogenetic corrected growth parameters otherwise. Using this single PSEM replaces the common practice of fitting and predicting using separate linear models depending upon what data are available for a given species. Cross‐validation suggests that the relationship between log‐mortality rate and longevity does not vary based on phylogeny, and therefore, linear models and PSEM both explain 82% of variance when longevity is available. When longevity is unavailable, by contrast, the linear model explains only 37% of variance while the PSEM explains 52% of variance, where this gain occurs from conditioning predictions on phylogenetic similarities. I therefore conclude that PCM and PSEM provide a general and user‐friendly replacement for linear models and can improve performance for fisheries meta‐analyses that are used for fisheries management applications. [ABSTRACT FROM AUTHOR]

Climate can impact fish population dynamics through changes in productivity and shifts in distribution, and both responses have been observed for many fish species. However, few studies have incorporated climate into population dynamics or stock assessment models. This study aimed to uncover how past variations in population vital rates and fishing pressure account for observed abundance variation in summer flounder (Paralichthys dentatus). The influences of the Gulf Stream Index, an index of climate variability in the Northwest Atlantic, on abundance were explored through natural mortality and stock-recruitment relationships in age-structured hierarchical Bayesian models. Posterior predictive loss and deviance information criterion indicated that out of tested models, the best estimates of summer flounder abundances resulted from the climate-dependent natural mortality model that included log-quadratic responses to the Gulf Stream Index. This climate-linked population model demonstrates the role of climate responses in observed abundance patterns and emphasizes the complexities of environmental effects on populations beyond simple correlations. This approach highlights the importance of modeling the combined effect of fishing and climate simultaneously to understand population dynamics. Le climat peut avoir un impact sur la dynamique des populations de poissons par le biais de modifications de la productivity et de la repartition, ces deux types de reactions ayant ete observes pour de nombreuses especes de poissons. Peu d'etudes ont toutefois integre le climat a des modeles de dynamique des populations ou d'evaluation des stocks. La presente etude avait pour objectif de determiner l'incidence de variations passees d'indices vitaux de populations et de la pression de la peche sur les variations observees de l'abondance de cardeaux d'ete (Paralichthys dentatus). Les influences de l'indice du Gulf Stream, un indice de variabilite du climat dans le nord-ouest de l'ocean Atlantique, sur cette abondance ont ete examinees par l'entremise de relations entre la mortalite naturelle et le recrutement au stock dans des modeles bayesiens hierarchiques structures par age. La perte predite a posteriori et le critere d'information associe a la somme des carres des ecarts indiquent que, des modeles mis a l'essai, les meilleures estimations des abondances des cardeaux d'ete sont obtenues du modele de mortalite naturelle dependant du climat qui integre des reactions log-quadratiques a l'indice du Gulf Stream. Ce modele de population relie au climat demontre le role des reactions au climat dans les motifs d'abondance observes et souligne la complexite des effets environnementaux sur les populations au-dela de correlations simples. L'approche fait ressortir l'importance de modeliser simultanement l'effet combine de la peche et du climat pour comprendre la dynamique des populations. [Traduit par la Redaction], Introduction Water temperatures are increasing globally (Lozier et al. 2008; Belkin 2009; Hansen et al. 2010; GISTEMP Team 2018) and global climate models indicate that this warming is likely to [...]

Stock assessment models are fitted to abundance-index, fishery catch, and age-length-sex composition data that are estimated from survey and fishery records. Research has developed spatiotemporal methods to estimate abundance indices, but there is little research regarding model-based methods to generate age-length-sex composition data. We demonstrate a spatiotemporal approach to generate composition data and a multinomial sample size that approximates the estimated imprecision. A simulation experiment comparing spatiotemporal and design-based methods demonstrates a 32% increase in input sample size for the spatiotemporal estimator. A Stock Synthesis assessment used to manage lingcod (Ophiodon elongatus) in the California Current also shows a 17% increase in sample size and better model fit using the spatiotemporal estimator, resulting in smaller standard errors when estimating spawning biomass. We conclude that spatiotemporal approaches are feasible for estimating both abundance-index and compositional data, thereby providing a unified approach for generating inputs for stock assessments. We hypothesize that spatiotemporal methods will improve statistical efficiency for composition data in many stock assessments and recommend that future research explore the impact of including additional habitat or sampling covariates. Les modeles d'evaluation des stocks sont cales sur des donnees d'indice d'abondance, de prises des peches et de composition selon l'age, la longueur ou le sexe qui sont estimees a partir de registres de releves et de peche. Des travaux de recherche ont mene a l'etablissement de methodes spatiotemporelles pour estimer les indices d'abondance, mais peu de travaux ont touche aux methodes basees sur les modeles pour produire des donnees de compositions selon l'age, la longueur ou le sexe. Nous faisons la demonstration d'une approche spatiotemporelle pour produire des donnees de composition et une taille d'echantillon multinomiale qui approxime l'imprecision estimee. Une experience de simulation qui compare des methodes spatiotemporelles et basees sur le schema d'evaluation fait ressortir une augmentation de 32 % de la taille de l'echantillon intrant pour l'estimateur spatiotemporel. Une evaluation avec Stock Synthesis utilisee pour la gestion de la morue-lingue (Ophiodon elongatus) dans le courant de Californie montre egalement une augmentation de 17 % de la taille de l'echantillon et un meilleur calage du modele en utilisant l'estimateur spatiotemporel, ce qui se traduit par des ecarts quadratiques moyens plus faibles pour les estimations de la biomasse de geniteurs. Nous concluons que les approches spatiotemporelles sont viables pour estimer des donnees d'indice d'abondance et de composition, fournissant ainsi une approche unifiee pour generer des intrants pour les evaluations de stocks. Nous postulons que les methodes spatiotemporelles amelioreront la puissance statistique pour les donnees de composition dans de nombreuses evaluations de stocks et recommandons que des travaux futurs se penchent sur l'incidence de l'inclusion d'autres covariables associees a l'habitat ou a l'echantillonnage. [Traduit par la Redaction], Introduction Fish production is highly variable over space and time, and variation is driven by both human and natural causes. Scientific advice regarding fish status and productivity is often based [...]

Selectivity is a key parameter in stock assessments that describes how fisheries interact with different ages and sizes of fish. It is usually confounded with other processes (e.g., natural mortality and recruitment) in stock assessments and the assumption of selectivity can strongly affect stock assessment outcome. Here, we introduce a new semi-parametric selectivity method, which we implement and test in Stock Synthesis. This selectivity method includes a parametric component and an autocorrelated nonparametric component consisting of deviations from the parametric component. We explore the new selectivity method using two simulation experiments, which show that the two autocorrelation parameters for selectivity deviations of data-rich fisheries are estimable using either mixed-effect or simpler sample-based algorithms. When selectivity deviations of a data-rich fishery are highly autocorrelated, using the new method to estimate the two autocorrelation parameters leads to more precise estimations of spawning biomass and fully selected fishing mortality. However, this new method fails to improve model performance in low data quality cases where measurement error in the data overwhelms the pattern caused by the autocorrelated process. Finally, we use a case study involving North Sea herring (Clupea harengus) to show that our new method substantially reduces autocorrelations in the Pearson residuals in fit to age composition data.La selectivity est un parametre cle des evaluations de stocks qui decrit comment les peches interagissent avec differents ages et differentes tailles de poissons. Elle est habituellement confondue avec d'autres processus (p. ex. mortalite naturelle et recrutement) dans les evaluations de stocks, et l'hypothese de la selectivity peut avoir une importante incidence sur le resultat de revaluation de stocks. Nous presentons une nouvelle methode de selectivite semi-parametrique que nous appliquons et validons dans Stock Synthesis. Cette methode de selectivite comprend une composante parametrique et une composante non parametrique autocorrelee qui est constituee d'ecarts par rapport a la composante parametrique. Nous explorons la nouvelle methode de selectivite en utilisant deux experiences de simulation qui montrent que les deux parametres d'autocorrelation pour les ecarts de selectivite de peches pour lesquelles les donnees sont abondantes peuvent etre estimes en utilisant des algorithmes a effets mixtes ou des algorithmes plus simples bases sur les echantillons. Quand les ecarts de selectivite d'une peche aux donnees abondantes sont fortement autocorreles, l'utilisation de la nouvelle methode pour estimer les deux parametres d'autocorrelation produit des estimations plus precises de la biomasse reproductrice et de la mortalite par peche entierement selectionnee. Cette nouvelle methode n'ameliore toutefois pas la performance des modeles dans les cas de faible qualite des donnees pour lesquels l'erreur de mesurage dans les donnees masque le motif resultant du processus autocorrele. Enfin, nous utilisons une etude de cas du hareng (Clupea harengus) de la mer du Nord pour demontrer que notre nouvelle methode reduit sensiblement les autocorrelations dans les residus de Pearson dans des donnees ajustees a la composition par age. [Traduit par la Redaction], IntroductionIn stock assessment models, selectivity is usually referred to as the combination of contact selectivity and population selectivity. It describes how relative fishing effectiveness varies with the age or size [...]