Richard Jensen rjensen at SAINTMARYS.EDU
Tue Sep 3 10:05:25 CDT 2002

One thing that seems to have been overlooked is that, if the models are
fundamentally sound,  the distributions they predict might suggest areas where
we should be looking for certain taxa.  If we then discover the taxa are there,
that reinforces the model; if the taxa are not there, that tells us the model
needs refining.  Either way, Tom is right - we must rely on field work itself to
verify the distributions predicted by the models.


Thomas DiBenedetto wrote:

> -----Original Message-----
> From: Robert Mesibov
> ....
> While I respect the efforts and intentions of range modellers, I can't help
> thinking that modelling is no substitute for field work,...
> I wonder about the goals of predictive range modellers. I understand that if
> one were to have a well-working model, capable of predicting areas where
> particular species might be found, that that might be an interesting thing.
> However I dont see that as being really very useful. My instinct would be to
> agree wholeheartedly with Robert about the primacy of observations in the
> field when it comes to defining the actual range of species.
> This is not to say however, that I dont see great use for predictive models,
> but it is not for the predictions per se. As I see it, the model serves as a
> vehicle for organizing our knowledge of the factors that influence
> distributions. The models will incorporate some factors, may choose not to
> incorporate others, and will necessarily define relationships between the
> factors. All of this equals a complex hypothesis regarding the causes for
> particular distributions. Armed with the output of such models, we can then
> examine real-world distributions as a test of the models. And in such a
> manner, we can refine the models until we arrive at a model that best
> indicates that which we find in nature. We will then have reason to claim to
> having increased our knowledge of why species are distributed as they are.
> Yet I constantly hear discussions of these models that seem to indicate that
> their proponents are not so much focussed on seeing the models as formal,
> general hypotheses, but as useful predictors of actual occurences. Even if
> they were, I dont see that as anywhere near their most valuable use.
> Tom diBenedetto

Richard J. Jensen              TEL: 574-284-4674
Department of Biology      FAX: 574-284-4716
Saint Mary's College         E-mail: rjensen at
Notre Dame, IN  46556

More information about the Taxacom mailing list