# dots on maps

[Jim Croft]

> I cannot agree with this point. Any map, even very detailed, is just a
> plan of a distributional phenomenon. Thus, ANY map is fibbing to some
> degree. As less thoughtfull it is done as more fibbing it is. Thus,
> constructing the map, the idea is not to dot any known record (though
> on practize for the most distrib. maps it is so, and it is good), but
> to illustrate (or to reveal, or to let the others to discover)some
> pattern (or at least what is thought to be so).
...
> I don't agree that the idea of putting any record on the map just because the
> listed record should be converted into the drawn dot is good idea. Moreover, in
> some cases the original "traditional" list of records, with various label data might
> be better than dots (loss of some information!). I am shure that map, containing
> thousands of dots of records, but somehow (bad design) poorely showing
> its general idea, is of little use. Is there any sense to invest time
> for producing such map from the database of records? Even if it takes few
> seconds for pushing some keys on the keyboard, what is the aim of
> such action?

This is an interesting issue and involves notions of primary and
interpreted data.  We have agonized over this a lot and have decided to
go with point localities rather than gridded or polygonal occurrences.
For visualization and interpretation purposes, you can always create the
second from the first, but not vice versa.

Dots on maps are mutifaceted personalities that may be intended one way
by the mapmaker and interpreted another way by the map reader.  For
ecxample, the dot might be a visual aid to enable the reader to see the
point at the dot's Centre, representing exacltly where the record if from.
On the other hand it might be a deliberately ambiguous blob, imprecisely
placed to give a general impression where organisms might be found.  The
might look the same but the first is a representation of primary data
and in an electronic form is quite scalable, while the second is
interpreted data and attempts at zooming in will lead to errors.  Extreme
cases of interpreted dots include dots within grids - they look very
neat but are totally unscalable.  In the ultra extreme case you dispense
with dots altogether and draw freehand blobs around where the dots would
have occured.

This last approach is more art than science, but nevertheless fun, and
before too long you find yourself drawing biogeographic tracks all over
the place.  In this regard I particularly like the approach of
predictive programs like BIOCLIM and GARP (on-line examples available on
ERIN's http://www.ea.gov.au/search/mapper.html) that take environmental
profiles from known point localities and match them against the
environmental profile for the country to come up with an expected range.
Of course this has to be ground-truthed as things just may not occur in
areas that are quite suitable for them, but it takes the exercise from
the realm of the subjective to the realm of the objective.  Perhaps not
a nice to look at but probably better science.

jim