|
Shareef,
In answer to your:
> Your second point does not make sense to me. Let me write that again:
>
> "The above functional relation indicates that natural mortality is largely governed by size
> and temperature.
>
> On that too, I must disagree most strongly."
>
> Yes, Pauly's equation is a mechanistic model
No, Pauly's model is _not_ mechanistic. It is purely descriptive. There
is no evidence, drawn across a wide variety of fish (as was Pauly's
regression), that water temperature and/or growth patterns are factors
that control the rate of natural mortality.
> Your answer raises another question
> whether one could conclude cause and effect based on good fit alone. This is one of the key
> issues in model fitting.
It is not an issue at all, let alone a "key" one. Quite simply,
correlation can never be evidence of causality, no matter how good the fit.
> Yes, one may fit growth as a function of mortality. But it does not help to achieve what he
> or she wants in this particular case; so, we don't need to talk about it.
Of course it does not help in practical applications, since we commonly
have estimates of growth parameters but do not have a way of estimating
the natural mortality rate. However, you would find it rewarding to
reconsider the possibility of regressing growth on mortality. That it
could be done with Dan's dataset should show you that you cannot assume
causality from his regressions: If a regression of mortality on growth
showed that growth controls mortality, then a regression of growth on
mortality would show that mortality controls growth. Both regressions
are entirely possible with the same dataset, yet growth cannot control
mortality if mortality controls growth. The solution to the conundrum is
that, while both regressions are real, they do not demonstrate causality
in either direction.
> What we have
> discussed is for fish having continuous growth, not for invertebrates where length-based
> model is most needed.
Most (not all) fish have continuous growth but so do some invertebrates.
Others, such as the clawed lobsters, have growth patterns that
approximate to von Bertalanffy curves, allbeit overlain by fluctuations
corresponding to moults. (Many teleosts have similar fluctuations driven
by the seasonal cycle, though they are usually ignored when drawing
growth curves.)
However, I would dispute your supposition that length-based models are
most needed for invertebrates. There are a great many, valuable finfish
resources for which the collection of adequate age data is not
practical. Conversely, there are important shellfish resources that can
be easily and cheaply aged.
Trevor Kenchington
--
Trevor J. Kenchington PhD [log in to unmask]
Gadus Associates, Office(902) 889-9250
R.R.#1, Musquodoboit Harbour, Fax (902) 889-9251
Nova Scotia B0J 2L0, CANADA Home (902) 889-3555
Science Serving the Fisheries
http://home.istar.ca/~gadus
><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><>
To leave the Fish-Sci list, Send blank message to:
mailto:[log in to unmask]
Web page at http://segate.sunet.se/archives/fish-sci.html
><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><> ><>
|
