Search

Annual run size and timing of Atlantic salmon smolt migration was estimated using Bayesian model framework and data from six years of a video monitoring survey. The model has a modular structure. It separates sub-processes of departing, traveling and observing, of which the first two together define the arrival distribution. The sub-processes utilize biological background and expert knowledge about the migratory behavior of smolts and about the probability to observe them from the video footage under varying environmental conditions. Daily mean temperature and discharge were used as environmental covariates. The model framework does not require assuming a simple distributional shape for the arrival dynamics and thus also allows for multimodal arrival distributions. Results indicate that 20% - 43% of smolts passed the Utsjoki monitoring site unobserved during the years of study. Predictive studies were made to estimate daily run size in cases with missing counts either at the beginning or in the middle of the run, indicating good predictive performance.

Many studies underscore the importance of incorporating the effect of environmental data within a life-history-stage–specific framework for determining the recruitment and survival of small pelagic fish. The recruitment of anchovy (Engraulis encrasicolus) in the Gulf of Cádiz (NE Atlantic) is sensitive to the effect of intense easterlies, stratification of the water column, and discharges from the Guadalquivir River on early life stages. As a proof of concept, we have developed the basis for a new Bayesian model with a dual time step resolution: monthly for juveniles and adults, and weekly for earlier life stages. This dual time step resolution resolves environmental effects on prerecruits while simulating the effect of fishing on recruits. Our estimates for juvenile abundances are validated with field data. The Bayesian framework accounts for the uncertainty, thus providing consistent length-frequency estimates and a plausible environmentally driven stock-recruitment relationship., The research leading to these results has received funding from the European Union’s Seventh Framework Programme (FP7/ 2007-2013) under grant agreement 244706/ ECOKNOWS project and M. Rincón was funded by P09-RNM-5358 of the Junta de Andalucía.

The paper presents an effort to build a biologically realistic, age structured Bayesian model for the stock assessment of data poor fisheries where only aggregated catch data is available. The model is built using prior information from other areas and ecologically or taxonomically similar species. The modeling approach is tested with data poor fisheries on the Cyclades islands in Greek archipelago. The two most important species in the area are selected: bogue (Boops boops) and picarel (Spicara smaris). Both are hermaphroditic. The only data available is the total catch from 1950 to 2010. Information was gathered about natural mortality, recruitment, growth, body size, fecundity, and sex ratio. There were significant problems in finding reliable prior information and a uniform prior was used for fishing mortality. The models at their present stage are not used to give management advice. The biological characteristics of the species in that area should be further studied. However, the posteriors of biological parameters reflect the best available knowledge on these species and they could be used in future studies or in simpler biomass dynamics models as priors.

Despite their name, hierarchical stock–recruit meta-analyses are often parameterized in terms of steepness, which depends not only on the assumed stock–recruitment relationship but also on the recruit–spawner relationship. This parameterization requires assumptions about the reproductive potential of the recruit that are not desirable if the focus of the study is limited to the spawning–recruitment phase instead of the full life cycle. Thus, usage of steepness should be avoided in studies that aim to produce informative priors for the stock–recruit relationship for use in studies of other salmon stocks. An alternative key parameter for stock–recruit models is the maximum survival of eggs, which is the slope at the origin of the stock–recruitment curve when spawning stock size is defined in terms of the number of eggs. Furthermore, the current widely used practices in stock–recruit modeling could be improved by taking into account the stock-specific model uncertainty. We use the method of Bayesian model averaging to build a hierarchical stock–recruit model that allows stock-specific model structures with Beverton–Holt, Ricker, and hockey stick models as alternatives, all of which can be parameterized with the maximum survival of eggs. We illustrate our approach by analyzing nine previously published datasets for Atlantic salmon (Salmo salar).

Mäntyniemi, S., Romakkaniemi, A., Dannewitz, J., Palm, S., Pakarinen, T., Pulkkinen, H., Gårdmark, A., and Karlsson, O. 2012. Both predation and feeding opportunities may explain changes in survival of Baltic salmon post-smolts. – ICES Journal of Marine Science, 69: 1574–1579. The survival of wild and hatchery-reared post-smolts of salmon (Salmo salar) in the Baltic Sea has declined since the 1990s. Direct observations of the processes affecting survival are, however, lacking. Here, the importance of food availability and predation in regulating post-smolt survival is analysed. Based on previous studies, the following explanatory variables were selected: (i) availability of herring (Clupea harengus membras) recruits in the Gulf of Bothnia (Bothnian Sea, Bothnian Bay) in the northern Baltic Sea; (ii) sprat (Sprattus sprattus balticus) and herring abundance in the southern Baltic Sea; and (iii) abundance of grey seal (Halichoerus grypus) along the post-smolt migration route. Bayesian analysis was used to estimate the relative probability of each of the 32 combinations of these variables and revealed that the model including grey seal abundance and herring recruits per post-smolt had the highest posterior probability and a high coefficient of determination. The results suggest that the declining trend in post-smolt survival is explained by the increased number of grey seals, whereas the annual variation in survival coincides with variation in the recruitment of Bothnian Sea herring. However, it remains uncertain whether the observed correlations arise from direct causalities or other mechanisms.

Juntunen, T., Vanhatalo, J., Peltonen, H., and Mäntyniemi, S. 2012. Bayesian spatial multispecies modelling to assess pelagic fish stocks from acoustic- and trawl-survey data. – ICES Journal of Marine Science, 69: 95–104. A Bayesian spatial model was constructed to estimate the abundance of multiple fish species in a pelagic environment. Acoustic- and trawl-survey data were combined with environmental data to predict the spatial distribution of (i) the acoustic backscattering of fish, (ii) the relative proportion of each species, and (iii) their mean length in the Gulf of Finland in the northeastern Baltic Sea. By combining the three spatial model layers, the spatial distribution of the biomass of each species was estimated. The model consists of a linear predictor on environmental variables and a spatial random effect given by a Gaussian process. A Bayesian approach is a natural choice for the task because it provides a theoretically justified means of summarizing the uncertainties from various model layers. In the study area, three species dominate pelagic waters: sprat (Sprattus sprattus), herring (Clupea harengus), and three-spined stickleback (Gasterosteus aculeatus). Results are presented for each model layer and for estimated total biomass for each species in 2 × 2 km lattices. The posterior mean and central 95% credible intervals of total biomass were sprat 45.7 kt (27.7–71.6), herring 24.6 kt (9.7–41.3), and three-spined stickleback 1.9 kt (0.9–3.2).

In fisheries stock assessment, the amount of information available from a specific stock is often limited, and the resources to gather new data can be insufficient. This is especially the case when management advice is required for by-catch species which are not always well monitored. However, information is often available from other stocks of the same or closely related species. Also, potential correlations between the life history parameters may contain information which is not usually taken into account in stock assessments. Utilizing all available information can be a cost-efficient way to diminish the amount of uncertainty about key parameters for a case with limited data or when constructing an informative prior for a new case study. In hierarchical modeling, different hierarchical levels can be used to connect closely related and more distant stocks. For example, a lower level of hierarchy can link the stocks within the same species, and a higher level, stocks of related species. We use length-weight and length-fecundity datasets from the FishBase database. Using these datasets, we demonstrate how Bayesian hierarchical correla- tion analysis can improve understanding of fecundity and formalize available knowledge about length-weight and length-fecundity relationships in terms of informative priors for future analysis.

Mäntyniemi, S., Kuikka, S., Rahikainen, M., Kell, L. T., and Kaitala, V. 2009. The value of information in fisheries management: North Sea herring as an example. – ICES Journal of Marine Science, 66: 2278–2283. We take a decision theoretical approach to fisheries management, using a Bayesian approach to integrate the uncertainty about stock dynamics and current stock status, and express management objectives in the form of a utility function. The value of new information, potentially resulting in new control measures, is high if the information is expected to help in differentiating between the expected consequences of alternative management actions. Conversely, the value of new information is low if there is already great certainty about the state and dynamics of the stock and/or if there is only a small difference between the utility attached to different potential outcomes of the alternative management action. The approach can, therefore, help when deciding on the allocation of resources between obtaining new information and improving management actions. In our example, we evaluate the value of obtaining hypothetically perfect knowledge of the type of stock–recruitment function of the North Sea herring (Clupea harengus) population.

This paper presents a sequential Bayesian framework for quantitative fisheries stock assessment that relies on a wide range of fisheries-dependent and -independent data and information. The presented methodology combines information from multiple Bayesian data analyses through the incorporation of the joint posterior probability density functions (pdfs) in subsequent analyses, either as informative prior pdfs or as additional likelihood contributions. Different practical strategies are presented for minimising any loss of information between analyses. Using this methodology, the final stock assessment model used for the provision of the management advice can be kept relatively simple, despite the dependence on a large variety of data and other information. This methodology is illustrated for the assessment of the mixed-stock fishery for four wild Atlantic salmon (Salmo salar) stocks in the northern Baltic Sea. The incorporation of different data and information results in a considerable update of previously available smolt abundance and smolt production capacity estimates by substantially reducing the associated uncertainty. The methodology also allows, for the first time, the estimation of stock–recruit functions for the different salmon stocks.

This study presents a state-space modelling framework for the purposes of stock assessment. The stochastic population dynamics build on the notion of correlated survival and capture events among individuals. The correlation is thought to arise as a combination of schooling behaviour, a spatially patchy environment, and common but unobserved environmental factors affecting all the individuals. The population dynamics model isolates the key biological processes, so that they are not condensed into one parameter but are kept separate. This approach is chosen to aid the inclusion of biological knowledge from sources other than the assessment data at hand. The model can be tailored to each case by choosing appropriate models for the biological processes. Uncertainty about the model parameters and about the appropriate model structures is then described using prior distributions. Different combinations of, for example, age, size, phenotype, life stage, species, and spatial location can be used to structure the population. To update the prior knowledge, the model can be fitted to data by defining appropriate observation models. Much like the biological parameters, the observation models must also be tailored to fit each individual case., The research leading to these results has received funding from the European Union's Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 244706/ECOKNOWS project and from the Academy of Finland (through a grant to AK).

In this paper, we demonstrate how information from broodstocks can be combined with lab information on alevins to obtain annual stock-specific mortality estimates from early mortality syndromes (EMS) using a probabilistic approach, how a hierarchical model structure can be used to predict these mortality rates for related, partly sampled, or unsampled stocks, and why these estimates should be used to remove the effect of this mortality on stock–recruit estimates. The approach has been illustrated for Atlantic salmon (Salmo salar) stocks in the Baltic Sea affected by the M74 syndrome. Results indicate that data on the proportion of M74-affected females, commonly used to approximate M74 mortality, overestimate actual M74-related mortality because of a declining trend in mortality among offspring of these females. The stock-specific M74 mortality estimates are used to account for nonstationarity in the stock–recruitment relationship caused by this fluctuating mortality. Because hierarchical meta-analyses assume exchangeability, the effect of M74 mortality is removed before including these stocks within hierarchical stock–recruit analyses of Atlantic salmon stocks, which are commonly unaffected by M74 mortality. Failure to remove the effect of M74 mortality on the stock–recruit data results in underestimation of the stock's productivity and resilience to exploitation, especially in the case of stocks with steep stock–recruit curves.

A Bayesian state–space mark–recapture model is developed to estimate the exploitation rates of fish stocks caught in mixed-stock fisheries. Expert knowledge and published results on biological parameters, reporting rates of tags and other key parameters, are incorporated into the mark–recapture analysis through elaborations in model structure and the use of informative prior probability distributions for model parameters. Information on related stocks is incorporated through the use of hierarchical structures and parameters that represent differences between the stock in question and related stocks. Fishing mortality rates are modelled using fishing effort data as covariates. A state–space formulation is adopted to account for uncertainties in system dynamics and the observation process. The methodology is applied to wild Atlantic salmon (Salmo salar) stocks from rivers located in the northeastern Baltic Sea that are exploited by a sequence of mixed- and single-stock fisheries. Estimated fishing mortality rates for wild salmon are influenced by prior knowledge about tag reporting rates and salmon biology and, to a limited extent, by prior assumptions about exploitation rates.

We introduce a Bayesian probability model for the estimation of the size of an animal population from re- moval data. The model is based on the assumption that in the removal sampling, catchability may vary between indi- viduals, which appears to be necessary for a realistic description of many biological populations. Heterogeneous catchability among individuals leads to a situation where the mean catchability in the population gradually decreases as the number of removals increases. Under this assumption, the model can be fitted to any removal data, i.e., there are no limitations regarding the total catch, the number of removals, or the decline of the catch. Using a published data set from removal experiments of a known population size, the model is shown to be able to estimate the population size appropriately in all cases considered. It is also shown that regardless of the statistical approach, a model that assumes equal catchability of individuals generally leads to an underestimation of the population. The example indicates that if there is only vague prior information about the variation of catchability among individuals, a very high number of suc- cessive removals may be needed to correctly estimate the population size. Resume : Nous utilisons un modele de probabilite bayesien pour estimer la taille d'une population animale a partir de donnees de retrait. Le modele se base sur la presupposition que, dans un echantillonnage par retraits, la capturabilite peut varier d'un individu a un autre, ce qui semble necessaire pour decrire de facon realiste de nombreuses populations biologiques. Une capturabilite individuelle heterogene engendre une situation dans laquelle la capturabilite moyenne de la population decroit graduellement a mesure que le nombre de retraits augmente. Avec cette presupposition, le modele s'ajuste a n'importe quelle serie de donnees de retrait; il n'y a donc pas de limites en ce qui a trait a la capture totale, au nombre de retraits ou au declin de la recolte. L'utilisation de donnees de retrait publiees obtenues sur une popula- tion de taille connue montre que le modele est capable d'estimer la taille de la population de facon adequate dans tous les cas etudies. De plus, independamment de l'approche statistique utilisee, un modele qui presuppose une capturabilite individuelle uniforme sous-estime la population. L'exemple montre que, s'il n'existe que des informations prealables vagues sur la capturabilite individuelle, il faudra peut-etre un nombre tres eleve de retraits successifs pour pouvoir estimer correctement la taille de la population. (Traduit par la Redaction) Mantyniemi et al. 300

We developed a Bayesian probability model for mark–recapture data. Three alternative versions of the model were applied to two sets of data on the abundance of migrating Atlantic salmon (Salmo salar) smolt populations, and the results were then compared with those of two widely used maximum likelihood models (Petersen method and a model using stratified data). Our model follows the basic principles of stochastic models presented for stratified data. In contrast to the earlier models, our model can deal with sparse data. Moreover, even weak dependencies between the studied parameters and the possible factors affecting them can be used to improve the plausibility of the estimates. The assumptions behind our approach are more realistic than those of earlier models, taking into account such factors as overdispersion, which is expected to be present in the mark–recapture data of salmon smolts because of their schooling behavior. Our examples also show that assumptions about the model structure can have a substantial impact on the resulting inferences on the size of the smolt run, especially in terms of the precision of the estimate.

We developed a hierarchical Bayesian integrated life cycle model for Atlantic salmon that improves on the stock assessment approach currently used by ICES and provides some interesting insights about the population dynamics of a stock assemblage. The model is applied to the salmon stocks in eastern Scotland. It assimilates a 40-year (1971–2010) time-series of data compiled by ICES, including the catches in the distant water fisheries at Faroes and West Greenland and estimates of returning fish abundance. Our model offers major improvements in terms of statistical methodology for A. salmon stock assessment. Uncertainty about inferences is readily quantified in the form of Bayesian posterior distributions for parameters and abundance at all life stages, and the model could be adapted to provide projections based on the uncertainty derived from the estimation phase. The approach offers flexibility to improve the ecological realism of the model. It allows the introduction of density dependence in the egg-to-smolt transition, which is not considered in the current ICES assessment method. The results show that this modifies the inferences on the temporal dynamics of the post-smolt marine survival. In particular, the overall decrease in the marine survival between 1971 and 2010 and the sharp decline around 1988–1990 are dampened when density dependence is considered. The return rates of smolts as two-sea-winter (2SW) fish has declined in a higher proportion than return rates as one-sea-winter (1SW) fish. Our results indicate that this can be explained either by an increase in the proportion maturing as 1SW fish or by an increase in the mortality rate at sea of 2SW fish, but the data used in our analyses do not allow the likelihood of these two hypotheses to be gauged.

We present a method by which the knowledge of stakeholders can be taken into account in stock assessment. The approach consists of a structured interview process followed by quantitative modelling of the answers. The outcome is a set of probability models, each describing the views of different stakeholders. Individual models are then merged to a large model by applying the techniques of Bayesian model averaging, and this model is conditioned on stock assessment data. As a result, the views of interviewed stakeholders have been taken into account and weighed based on how well their views are supported by the observed data. We applied this method to the Baltic Sea herring (Clupea harengus) stock assessment by interviewing six stakeholders and conditioning the resulting models on stock assessment data provided by the International Council for the Exploration of the Sea.

How can uncertain fisheries science be linked with good governance processes, thereby increasing fisheries management legitimacy and effectiveness? Reducing the uncertainties around scientific models has long been perceived as the cure of the fisheries management problem. There is however increasing recognition that uncertainty in the numbers will remain. A lack of transparency with respect to these uncertainties can damage the credibility of science. The EU Commission’s proposal for a reformed Common Fisheries Policy calls for more self-management for the fishing industry by increasing fishers’ involvement in the planning and execution of policies and boosting the role of fishers’ organisations. One way of higher transparency and improved participation is to include stakeholders in the modelling process itself. The JAKFISH project (Judgment And Knowledge in Fisheries Involving StakeHolders) invited fisheries stakeholders to participate in the process of framing the management problem, and to give input and evaluate the scientific models that are used to provide fisheries management advice. JAKFISH investigated various tools to assess and communicate uncer- tainty around fish stock assessments and fisheries management. Here, a synthesis is presented of the participatory work carried out in four European fishery case studies (Western Baltic herring, North Sea Nephrops, Central Baltic Herring and Mediterranean swordfish), focussing on the uncertainty tools used, the stakeholders’ responses to these, and the lessons learnt. It is concluded that participatory modelling has the potential to facilitate and structure discussions between scientists and stakeholders about uncertainties and the quality of the knowledge base. It can also contribute to collective learning, increase legitimacy, and advance scientific understanding. However, when approaching real-life situations, modelling should not be seen as the priority objective. Rather, the crucial step in a science–stakeholder collaboration is the joint problem framing in an open, transparent way.