Good day,
We are comparing size-at-age curves for polar bears in Southern Hudson Bay,
Ontario, Canada. We hypothesized that polar bears grew more slowly and
possibly had smaller maximum sizes in recent years compared to previously.
Preliminary examination of the data supports this hypothesis, and furthermore
indicates that the von Bertalanffy growth model fits polar bear size-at-age
data quite well. The data consist of size-at-age data for live-captured polar
bears. Many individuals were captured only once, but many were also captured 2
- 4 times over many years.
It would appear, therefore, that a non-linear mixed effects model, where at
least two of the parameters of the von Bertalanffy growth function (growth
rate, K, and maximum size Linf) could be modeled as functions of covariates,
and bear.id could be included as a random effect, would provide a good
framework for analysis and inference.
I have found a few articles presenting results from similar models. For
example:
Kimura (2008. Extending the von Bertalanffy growth model using explanatory
variables. Can.J.Fish.Aquat Sci.) developed models that allowed for explanatory
covariates, but not random effects, and with no implementation in R.
Baudron et al (2011. Implications of a warming North Sea for the growth of
Haddock. Journal of Fish Biology) implemented a similar model in R.
I was hoping to find, but have not as yet, an R package or code that implements
the von Bertalanffy growth model as a non-linear mixed effects model in R (e.g.
relying on nlme or lme4). I have also been working on coding one myself in
nlme.
I haven't posted this question to r-sig-mixed-models.
Thanks in advance,
Eric
[[alternative HTML version deleted]]
_______________________________________________
R-sig-ecology mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
