Statistical Catch at Age Model
(STATCAM)

STATCAM is a statistical catch-at-age model.  It is a likelihood-based assessment model for joint analyses of age-specific fishery and survey data. Age-structured population stock dynamics are modeled using standard forward-projection methods for statistical catch-at-age analyses. The population dynamics model is fit to observed fishery and survey data using an iterative maximum likelihood estimation approach. 

The calculation engine was developed by Dr. Jon Brodziak, NMFS Pacific Islands Fishery Science Center, using AD Model Builder.

Back to Top 

STATCAM Version 1.4.1 (February 2007)

Replaces 1.3.1 and includes significant updates to the calculation engine with associated GUI changes.

• One can select fishery selectivity and survey selectivity functions to be either parametric or nonparametric. If fishery selectivity is parametric then either a 3-parameter Thompson or a 2-parameter logistic curve is estimated Otherwise, a nonparametric selectivity value will be estimated for each age group from 1 up to the number of estimated ages. In either case, the selectivity vector is normalized to have a maximum value of 1.

• One can specify whether a smoothing function should be applied to fishery selectivity estimates.

• One can specify a natural mortality at age multiplier. This multiplier times the estimate of the fully-vulnerable natural mortality rate determines the natural mortality at age vector. One can also choose to fix or estimate the fully-vulnerable natural mortality rate. If the fully-vulnerable natural mortality rate is estimated, one needs to specify the coefficient of variation of the lognormal prior distribution for fully-vulnerable natural mortality.

• One can apply a penalty function to shrink recruitment deviations to be near the  mean recruitment estimate starting in the terminal year of the assessment time horizon and proceeding backwards for a fixed number of  years, nshrinkage; this is the number of recent years for which incomplete observations of cohort strength may lead one to shrink estimates towards their mean value.

• One can apply the bootstrap to construct bootstrap replicates of the population numbers at age vector in the terminal year plus one, analogous to the Toolbox VPA program. These bootstrap replicates are written to file and can be input to the Toolbox AGEPRO module.

• One can choose to add a small nonnegative constant to the argument of natural log functions to ensure that there is no numeric overflow if the argument becomes close to zero during the parameter optimization.

• One can set mean recruitment or mean fishing mortality to be fixed constants instead of estimated parameters. One can set annual recruitment deviations or fishing mortality deviations to be fixed constant vectors instead of estimated parameter vectors.

• One can use the fit_survey_index flag to determine whether to fit each survey index. This simplifies the choice of including or excluding a given survey index.

• One can change the optimization phases for estimating parameters to fit recruitment, fishing mortality, survey, fishery, and natural mortality processes.

• Program outputs incude time series of predicted catch, population, spawning, and exploitable biomass at age, similar to VPA outputs, along with the fraction of unfished spawning biomass realized in each year.

STATCAM Version 1.3.1 (September 2006)

Replaces Beta Version 1.3

• Recruitment year is now set to Start Year - Number of Ages +1 (Formerly +2).

• Added the ability to capture error messages and output to the command window to a log file and have the log file viewed in the graphical interface.

• Upgraded Plotting Package.

Back to Top 

Older Version Changes 

Printer friendly view