bhnoneq: Length-based Beverton-Holt Nonequilibrium Z Estimator

In fishmethods: Fishery Science Methods and Models

Length-based Beverton-Holt Nonequilibrium Z Estimator

Description

A nonequilibrium Beverton-Holt estimator of instantaneous total mortality (Z) from length data.

Usage

bhnoneq(year=NULL,mlen=NULL, ss=NULL, K = NULL, Linf = NULL, 
Lc = NULL, nbreaks = NULL, styrs = NULL, stZ = NULL, 
graph = TRUE)

Arguments

year	the vector of year values associated with mean length data. The number of year values must correspond to the number of length records. Include year value even if mean length and numbers (see below) are missing.
mlen	the vector of mean lengths for lengths >=Lc. One record for each year.
ss	the vector of numbers of observations associated with the mean length.
K	the growth coefficient from a von Bertalanffy growth model.
Linf	the L-infinity coefficient from a von Bertalanffy growth model.
Lc	the length at first capture.
nbreaks	the number of times (breaks) mortality is thought to change over the time series. Can be 0 or greater.
styrs	the starting guess(es) of the year(s) during which mortality is thought to change. The number of starting guesses must match the number of mortality breaks, should be separated by commas within the concatentation function and should be within the range of years present in the data.
stZ	the starting guesses of Z values enclosed within the concatentation function. There should be nbreaks+1 values provided.
graph	logical indicating whether the observed vs predicted and residual plots should be drawn. Default=TRUE.

Details

The mean lengths for each year for lengths>=Lc. Following Gedamke and Hoening(2006), the model estimates nbreaks+1 Z values, the year(s) in which the changes in mortality began, the standard deviation of lengths>=Lc, and standard errors of all parameters. An AIC value is produced for model comparison. The estimated parameters for the number of nbreaks is equal to 2*nbreaks+2. Problematic parameter estimates may have extremely large t-values or extremely small standard error. Quang C. Huynh of Virginia Institute of Marine Science revised the function to make estimation more stable. Specifically, the derivative method BFGS is used in optim which allows more reliable convergence to the global minimum from a given set of starting values, a function is included to estimate Z assuming equilibrium, sigma is estimated analytically and convergence results . Use 0 nbreaks to get Z equilibrium.

Value

summary	list element containing table of parameters with estimates, standard errors, and t-values.
convergence	list element specifying if convergence was reached.
hessian	list element specifying if hessian is positive definite
results	list element containing, observed value, predicted values, and residuals from the model fit.

Note

Todd Gedamke provided the predicted mean length code in C++.

Author(s)

Gary A. Nelson, Massachusetts Division of Marine Fisheries [e-mail]

Quang C. Huynh of Virginia Institute of Marine Science

References

Gedamke, T. and J. M. Hoenig. 2006. Estimating mortality from mean length data in nonequilibrium situations, with application to the assessment of goosefish. Trans. Am. Fish. Soc. 135:476-487

See Also

bheq

Examples

data(goosefish)
bhnoneq(year=goosefish$year,mlen=goosefish$mlen, ss=goosefish$ss,
K=0.108,Linf=126,Lc=30,nbreaks=1,styrs=c(1982),stZ=c(0.1,0.3))

fishmethods documentation built on April 27, 2023, 9:10 a.m.