Converting Township Range Section data to lat/lon

[Mary Barkworth]

My primary interest is in broad scale data, but I also think that we can
so tied up in the need for great precision that we get to the point where
we can do nothing.  If you see a dot on a map that is accurate within a
mile and you think the record worth investigating - for any reason - then,
because the data are associated with a specimen, you can ask the curator
concerned for more information on that specimen.  To my mind, this puts us
ahead of not having the dot in the first place.  Yes, we can do things
much more accurately now, but are you disputing the value of the dot maps
that have been published over the years in which the dots were put on by
reading the label, looking at an atlas, and then putting the dot as on the
baseman for the species, probably by guestimating where it should go?

There are times when plant people need accuracy to the nearest centimeter.
I would not portray the maps developed from our herbarium specimens as
having anything like that kind of accuracy - within a mile would be good.
Nevertheless, I see the maps as useful.

I apologize if this sounds a bit curt.


-----Original Message-----
From: Doug Yanega [mailto:dyanega at POP.UCR.EDU]
Sent: Monday, August 21, 2000 8:42 PM
To: TAXACOM at USOBI.ORG
Subject: Re: Converting Township Range Section data to lat/lon


Mary Barkworth wrote:

>for many specimens
>collected in the U.S., the most accurate locality data on a herbarium
>label is the TRS data.  It provides location within a mile.

Given that, how do you deal with the uncertainty involved when converting?
Do you give a list of the four lat/longs that define the corners of the
box, or do you list the lat/long of the exact center of the box? I can
imagine that giving the corners, while more useful in one respect, would
not be very good for exporting data into mapping programs. Are there
mapping programs that will happily map a locality as a box covering, say,
one full degree of lat/long, or are all mapping programs limited to single
points in space?

Perhaps more to the point, do we as a community have anything like a
consensus as to how precise a coordinate must be before we will consider
it
*viable* for use in mapping? After all, if you have records as in my
example above, where only the nearest *degree* is available, then it's
theoretically possible for the actual collecting site to be something
around 40 miles away from the "centroid" (X'30", Y'30"). Is that kind of
error acceptable? If not, and people just go ahead and enter all records
into their databases and have them converted to single lat/long values,
how
do users of such data sets know whether or not such unwelcome points are
IN
a data set they may be accessing remotely, for example? Does every
collection-record database in existence actually have a "linear error +/-"
field in it? What's the cutoff, anyway? 10 miles? 5? 1?

Perpetually wrestling with this sort of thing,


Doug Yanega        Dept. of Entomology         Entomology Research Museum
Univ. of California - Riverside, Riverside, CA 92521
phone: (909) 787-4315 (standard disclaimer: opinions are mine, not UCR's)
           http://insects.ucr.edu/staff/yanega.html
  "There are some enterprises in which a careful disorderliness
        is the true method" - Herman Melville, Moby Dick, Chap. 82