fishmethods (version 1.10-4)

vbfr: Francis' re-parameterization of the von Bertalanffy growth equation for length-age data

Description

Fits the re-parameterized von Bertalanffy growth equation of Francis (1988) by using nonlinear least-squares

Usage

vbfr(age = NULL, L = NULL, agephi = NULL, agepsi = NULL, graph = TRUE, 
gestimate = TRUE, Lphiparms = c(NA, NA, NA), Lchiparms = c(NA, NA, NA), 
Lpsiparms = c(NA, NA, NA),control = list(maxiter = 10000))

Arguments

age

Vector of ages of individual fish.

L

Vector of lengths of individual fish.

agephi

Arbitrary reference age phi

agepsi

Arbitrary reference age psi. agepsi>agephi.

graph

Logical specifiying whether observed versus predicted, and residual plots should be drawn. Default=TRUE.

gestimate

Logical specifying whether automatic generation of starting values of lphi, lchi and lpsi should be used. Default=TRUE. If gestimate=FALSE, user-specified starting, lower and upper limits of parameters must be entered.

Lphiparms

If gestimate=FALSE, starting value, lower limit and upper limit of lphi used in nls.

Lchiparms

If gestimate=FALSE, starting value, lower limit and upper limit of lchi used in nls.

Lpsiparms

If gestimate=FALSE, starting value, lower limit and upper limit of lpsi used in nls.

control

see control under function nls.

Value

nls object of model results. Use summary to extract results.

Details

Francis (1988) re-parameterized the von Bertalanffy growth equation for age-length in order to make equivalent comparison of parameters to parameters of a common model used to estimate growth from tagging data. Three parameters, lphi, lchi and lpsi, are estimated. The re-parameterization also has better statistical properties than the original equation.

The formulae to get the conventional von Bertalanffy parameters are:

Linf = lphi + (lpsi-lphi)/(1-r^2) where r = (lpsi-lchi)/(lchi-lphi)

K = -(2*log(r))/(agepsi-agephi)

t0 = agephi + (1/K)*log((Linf-lphi)/Linf)

If gestimate=TRUE, unconstrained nonlinear least-squares (function nls) is used to fit the model. If gestimate=FALSE, constrained nonlinear least-squares is used (algorithm "port" in nls).

References

Francis, R. I. C. C. 1988. Are growth parameters estimated from tagging and age-length data comparable? Can. J. Fish. Aquat. Sci. 45: 936-942.

Examples

# NOT RUN {
data(pinfish)
with(pinfish,vbfr(age=age,L=sl,agephi=3,agepsi=6))
# }