Accounting for variable recruitment and fishing mortality in length-based stock assessments for data-limited fisheries

Abstract

In fisheries with limited capacity for monitoring, it is often easier to collect length measurements from fishery catch than quantify total catch. Conventional stock assessment tools that rely on length measurements without total catch do not directly account for variable fishing mortality and recruitment over time. However, this equilibrium assumption is likely violated in almost every fishery, degrading estimation performance. We developed an extension of length-only approaches to account for time-varying recruitment and fishing mortality. This Length-based Integrated Mixed Effects (LIME) method at a minimum requires a single year of length data and basic biological information but can fit to multiple years of length data, catch, and an abundance index if available. We use simulation testing to demonstrate that LIME can estimate how much fishing has reduced spawning output in the most recent year across a variety of scenarios for recruitment and fishing mortality. LIME improves data-limited fisheries stock assessments by its flexibility to incorporate additional years or types of data if available and obviates the need for equilibrium assumptions.

Predicting recruitment density dependence and intrinsic growth rate for all fishes worldwide using a data‐integrated life‐history model

“…I believe that multivariate predictions for multiple population‐dynamics parameters will be useful for stock assessment, decision theory, ensemble modelling, and strategic decision‐making, and discuss each of these briefly: Stock assessment : These multivariate predictions can be used to define a joint prior distribution (or penalty) for parameters used in a Bayesian (or maximum likelihood) stock assessment model. For example, a recent data‐poor assessment model using length‐composition data to estimate stock status requires values for maturity, mortality, size, steepness and recruitment variability (Rudd & Thorson, ), and a joint prior for all of these parameters can be generated using the results. Using a joint prior distribution prevents assessments from estimating a combination of life‐history parameters that are highly unlikely (Brandon, Breiwick, Punt, & Wade, ; Kindsvater et al, ), for example, by estimating high mortality M in combination with large maximum body size .…”

Global reconstruction of life‐history strategies: A case study using tunas

“…The described approach for imputing missing life‐history data could potentially be applied to any data‐limited group of species. Within a fisheries context, the reconstructed life‐history data can be employed in data‐limited assessments, such as ecological risk assessments and length‐based models (Hordyk, Ono, Prince, & Walters, ; Rudd & Thorson, ). In addition, the approach has immediate application to the world‐wide management of fisheries for tuna in which the steepness of the Stock Recruitment Relationship (SRR) is used.…”

Evaluating the Performances of Size‐Frequency‐Based Methods for Estimating Fishing Mortality of Pholis fangi

“…(Substantially large measurement errors may result in unreasonable estimates for both methods, shown in Table S7.) The length data was generated by simulating underlying age structure (Table ; Rudd and Thorson ). First, the probability of being in a length bin for individuals of each age was calculated (Table , equation 3).…”

“…) offers a size‐based version to estimate spawning potential ratio. Likewise, a length‐based, integrated, mixed‐effects model (LIME; Rudd and Thorson ) provides a comprehensive framework to account for variable fishing mortality and recruitment when only length data are available. Also, a length‐based Bayesian biomass estimation method (Froese et al.…”