Latitude & Longitude data

[Robert Raven]

Cute problem. Thankfully this one is way behind us. Our approach was to
have a person calculate each point from maps. Very tedious but now we
have a GIS too and things can be easier.

Try this.

Your database (R:Base, at last checking) should be a ble to do
select distinct locale state county etc. So work only on the unique rows.

Second, normalize the locality data. So that 13km NW of Chiapas and 8
miles NW of Chiapas look identical (if they are): Chiapas, 13 km NW

Three, try geocoding your register against the GIS in which I gather you
have a gazeteer. Fine in principle but typically taxonomists don't
collect in places found in gazeteers but in more interesting areas so
manual handling is necessary.

Four, make sure your database (in hwich I hope your locality/expedition
data are stored as a separate table) has a locality error field, i.e. how
you determined the value. From long experience with museum data and a lot
of academic input from GIS users, i have learnt that assigning
quantitative values to the locality error field is an unreal expectation.
Just take Long Island, New York. If that is the only data given then you
have to have 2 different error fields, one for longitude and one for
latitude and even then a wedge shaped island may confound those values.
My WORKABLE approach is to assign a letter code to the locality error:
COLL the coordinates were given by the collector who is thorough in such
matters
QGAZ taken from Queensland Gazeteer (on autolookup in the database)
AGAZ taken from Australian Gazeteer (book)
CALC direct line according to the description (13km NW of Chiapas)
UNIT The centerpoint of the unit (eg Long Island, Chiapas, etc)
HIST Historical research revealed very closely where the collection was made

This will not of course satisfy the GIS puritans but they don't live in
the real world of museum data that were collected before GIS was even
conceived.

Having done this you should then print your maps (we did it to an A1
plotter) at a sufficient scale to have all points separate so that you
can determine that three points around Chiapas should not all be to the
north when one is south of Chiapas. Enjoy the proofing.

The final result can be fun (see Ingram & Raven, 1991 (eds), An Atlas of
Queensland's frogs, reptiles, birds and mammals. Qld Museum Board of
Trustees. Which are the 1200 plus maps of the museum's records. The
volume is almost out of print.)

Next comes the awakening when the government basically demands that
environmental impact statements do a search on the Museum's records.
Things are quite busy here with such searches.

Enjoy!

Rob


Dr Robert J. Raven
Museum Scientist (Arachnology)
Queensland Museum, Grey St, South Brisbane, 4101, Q.
Queensland Museum, PO Box 3300, South Brisbane, 4101, Q.
Australia

Phone: 61-7-38407698 or 61-18-748 467 (24 hour)
Fax: 61-7-3846 1918