Review and evaluation of likelihood functions for composition data in stock-assessment models: Estimating the effective sample size

Maunder, Mark N.

doi:10.1016/j.fishres.2011.02.018

Cited by 55 publications

(21 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These surveys provided an excellent opportunity to evaluate sample sizes for age composition because of the longevity of the surveys and the number of different species sampled for ages. Data (AFSC bottom trawl surveys, website) were used from trawl surveys conducted in the GOA from 1984 to 2011 (triennially from 1984 to 1999 then biennially from 2001 to 2011, e.g., ), in the AI from 1980to 2010(triennially from 1980to 1986and from 1991to 1997, then biennially from 2000to 2006and 2010on, e.g., von Szalay et al, 2011, and in the BS from 1982 to 2011 (annually; e.g., Lauth and Conner, 2014). In the interest of brevity we refer to von for a detailed description of how the survey index and age-length compositions are estimated from the surveys.…”

Section: Methodsmentioning

confidence: 99%

“…Intrahaul correlation in sampling for ages has recently been the subject of several studies (Pennington et al, 2002;Hulson et al, 2011), as well as investigations of how to account for intrahaul correlations in SCAA models (Francis, 2011;Maunder, 2011;Hulson et al, 2012). Although intrahaul correlation has received attention in the literature in terms of integration into stock assessment models, it is not clear at this point in time how intrahaul correlation could be incorporated in estimating optimal distribution of age-composition sample sizes across species before a stock assessment.…”

Section: Figurementioning

confidence: 99%

“…Age-composition data sets provide the fundamental information necessary for application of statistical catch-at-age assessment (SCAA) models used to estimate population dynamics and to manage commercial fisheries (Quinn and Deriso, 1999;Maunder and Punt, 2013). Age composition data allow the tracking of year classes through time, providing an improved understanding of population dynamics compared with that gained from analyzing fishery catch per unit of effort or survey biomass indices alone.…”

mentioning

confidence: 99%

“…This is, however, a different investigation than the application of sample size within an SCAA model, or, the idea of effective sample size. The topic of effective sample size and the use of age or length composition data within stock assessments has received extensive attention in recent years (Francis, 2011;Maunder, 2011). This study is also a different investigation from that of optimizing the number of hauls from which age-composition samples are obtained (Pennington et al, 2002;Hulson et al, 2011).…”

mentioning

confidence: 99%

See 3 more Smart Citations

Investigations into the distribution of sample sizes for determining age composition of multiple species

Hulson¹,

Hanselman²,

Shotwell³

2017

View full text Add to dashboard Cite

Abstract-Collecting age-composition data is a critical aspect of stock assessment; however, there are no biological or statistical investigations that support optimization of the distribution of sample size across species. Sample sizes for both collection and age-reading are often set by using ad hoc or historical values. Investigations into quantifying the trade-offs when allocating sample sizes across species are needed because resources for age determination are always limited. In this study we performed analyses to investigate the distribution of sample sizes to determine ages across multiple species by using methods derived from sampling theory and simulation testing of stock assessment models. We found that, in terms of methods based on sampling theory, distribution of sample size under 2-stage sampling could be significantly related to the life-history characteristics of the species. Results from simulation analysis illustrated that the influence of sample sizes required to determine age composition of fish on uncertainty in stock assessment models was related to uncertainty in a survey index and recruitment variability of the species being assessed. The simulation analysis highlighted cases in which larger age-composition sample size did not appreciably decrease uncertainty in the stock assessment model, in particular, for species with lower recruitment variability and larger survey index uncertainty.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Figurementioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Investigations into the distribution of sample sizes for determining age composition of multiple species

Hulson¹,

Hanselman²,

Shotwell³

2017

View full text Add to dashboard Cite

show abstract

“…Index data are almost always assumed to be log-normally distributed about the model predictions of vulnerable biomass. However, there are several choices for the likelihood function for compositional data (Francis, 2011;Maunder, 2011). Maunder (2011) evaluated several potential likelihood functions and recommended the "Punt-Kennedy" approach, which is widely used in assessments of South African fish stocks, but is not included in any of the major stock assessment packages.…”

Section: Data Weightingmentioning

confidence: 99%

Assessing billfish stocks: A review of current methods and some future directions

Punt

Sun

2015

Fisheries Research

View full text Add to dashboard Cite

Data Quality, Data Quantity, and Its Effect on an Applied Stock Assessment of Cisco in Thunder Bay, Ontario

Fisch

Bence

2020

N American J Fish Manag

View full text Add to dashboard Cite

Stock assessments, or population models developed to support fishery management decisions, require informative data to produce reliable estimates. However, resources available to collect these data are limited. Thus, information relating the effects of different data collection schema on stock assessment performance should be of interest to fishery managers. We used an existing data set on a stock of Cisco Coregonus artedi in Thunder Bay to simulate various degrees of reduction in available data. We considered both cluster subsampling of biological data from the commercial fishery harvest (which determine the observed harvest age composition) and reductions in the frequency of hydroacoustic surveys in order to examine their effect on fits of an age-structured stock assessment model for the stock. Our results indicate that reductions in the frequency of hydroacoustic surveys would have a greater effect on applied stock assessment performance for Thunder Bay Cisco than would reductions in biological sampling to randomly selected temporal clusters of the fishery harvest. Reduction in the frequency of the hydroacoustic survey resulted in different point estimates and larger estimated uncertainty for spawning biomass and natural mortality rate compared with the original assessment model. This was likely largely driven by increases in lag between the final year of the survey and the current year of the assessment. The lower influence of reduced biological sampling may be due to the highly variable nature of Cisco recruitment, where large or "boom" year-classes were still evident in the reduced biological samples, combined with information from survey age compositions. We suggest a priority be placed on performing hydroacoustic surveys with some regularity, such that when they are performed, they are done extensively to minimize uncertainty (measurement error). The data subsampling approach used here could be used in many assessments to determine if a reduction in sampling of various types could be implemented without materially changing assessment results.

show abstract

Review and evaluation of likelihood functions for composition data in stock-assessment models: Estimating the effective sample size

Cited by 55 publications

References 33 publications

Investigations into the distribution of sample sizes for determining age composition of multiple species

Investigations into the distribution of sample sizes for determining age composition of multiple species

Assessing billfish stocks: A review of current methods and some future directions

Data Quality, Data Quantity, and Its Effect on an Applied Stock Assessment of Cisco in Thunder Bay, Ontario

Contact Info

Product

Resources

About