shepherd: Shepherd-Nicholson Sigma

In hafroAssmt: Hafro Assessment Toolbox

Description

Estimate Shepherd-Nicholson sigma for a given catch-at-age matrix.

Usage

shepherd(x, plot=FALSE)

Arguments

x	Catch at age, as a data.frame (year, age, frequency in columns 1–3), matrix (years in rows, ages in columns), or table (years in rows, ages in columns).
plot	Whether to plot model fit to data.

Value

List containing

model	lm model.
fit	data.frame containing Year, Age, Series, and Freq.
res	matrix of residuals.
age.sigma	standard deviation of log residuals within each age.
year.sigma	standard deviation of log residuals within each year.
sigma	standard deviation of all log residuals.

Note

The log-linear model (Shepherd and Nicholson 1991, Eq. 5), lm(log(Catch)~factor(Year)+factor(Age)+factor(Cohort)), is exact only if the fishing mortality is separable and the overall fishing mortality is constant over the period considered.

A common purpose of fitting the Shepherd-Nicholson model is to evaluate a simple diagnostic of commercial and survey catch-at-age, describing how well the data conform to a set of basic assumptions. The scale-free sigmas can be compared between ages, years, datasets, and species.

The sigmas reported are maximum likelihood estimates, with the usual negative bias. They combine measurement error, process variability, and model misspecification.

Author(s)

Sigurdur Thor Jonsson and Arni Magnusson.

References

Shepherd, J.G. and M.D. Nicholson. 1991. Multiplicative modelling of catch-at-age data, and its application to catch forecasts. Journal du Conseil international pour l'Exploration de la Mer 47:284–294.

See Also

pope is a simple assessment model to estimate stock size.

hafroAssmt-package gives an overview of the package.

Examples

shepherd(catage)
shepherd(catage, plot=TRUE)

shepherd(catage)$sigma                 # MLE
summary(shepherd(catage)$model)$sigma  # unbiased estimate

hafroAssmt documentation built on May 2, 2019, 5:47 p.m.