lat/long designation and collection data

[Robin Panza]

>Message-ID: <9508298124.AA812426117 at casmail.calacademy.org>
>From: Library <Library at CASMAIL.CALACADEMY.ORG>
>
>          The California Academy of Sciences has recently acquired
>          GIS software.  It is of great interest to us to use this
>          software in conjunction with our existing collection's data.
>          However, much of our existing data is without latitudinal
>          and longitudinal designations.  Has anyone already
>          undertaken the huge job of retrospectively assigning
>          latitude and longitude to their collections data.  If so,
>          what was done to maximize the speed and accuracy of the job?
>
>          Karen Cebra
>          Collections Manager
>          Department of Ornithology and Mammalogy
>          California Academy of Sciences
>          kcebra at calacademy.org


I can't say we did much to maximize speed and accuracy, but we have worked on
lat/long data for our specimens (ca. 200,000 birds).  As part of our
computerization process, I spent a year (half- to two-thirds time) verifying,
correcting, and finding localities.  I must admit that lat/long are not in the
specimen record nor on the specimen.  At this point, I have a computerized
gazetteer of our localities, which I hope someday to publish.

First, we printed out all our specimens' localities (a massive
printout, but well worth it to us), sorted by country, state, county, and
specific locality, in that order.  The printout made it easy to look for
inconsistencies in the data.  These included typos or too few/many spaces,
resulting in specimens not sorting properly; differences of opinion on the
exact formatting of a locality; and outright misinformation (such as British
Columbia and Costa Rica being entered as states in the US, or Algeria as a
state in the country of Africa--I was not the original data entry
person).

The next stage was to check atlases & gazetteers for things like what state
it's in, what's the proper/modern spelling, and coordinates for it.  My primary
reference was the Times Atlas (1989 edition) for outside South America (for
which I used the Ornithological Gazetteers by Paynter et al.).  I also used
older Times Atlases (1965 index and 1922) and National Geographic (several
editions).  We selected the 1989 Times Atlas as our primary reference because
it is both quite thorough and readily available to anyone wanting to find the
places on maps of their own (e.g., for mapping distributions).

Next priority for outside the US was our extensive map collection,
many going back to the era when the specimens were collected, many with
annotations of the localities in question.  You might wonder why these were not
first priority.  First, older maps are often unreliable--they were created by
people wandering the wilderness and trying to keep track of where they were.
Second, they often used someplace other than Greenwich for the prime meridian
and required finding out where the PM is and calculating to Greenwich
longitude.  Third, boundaries change and, as maps become more accurate, I
generally had to compare the old map with newer ones to place the locality
(found on the old map) into context on a new map (to find modern determination
of lat/long).  For example, you can tell the age of a Cameroon map by how far
north Bafia is placed--this city has been migrating northward, at least on the
4 or 5 maps we have.

For US localities, I must tell you we haven't dealt with lat/long.  I used the
1980 (because that's what we have) Rand McNally Commercial Atlas and Marketing
Guide.  If a village, however small, is not in there, it doesn't exist.  This
was useful for townships (often used in the past) and for tiny towns not found
on other maps.  Once found, a tiny town can be listed with reference to a
larger town that is on more maps--e.g., Millvale (1.5 mi N; Evergreen).  The
Commercial Atlas is pretty much useless for geographic localities (mountain,
river, etc.).  Because of time constraints (I did US last), I just "found" the
localities and will one day go back and get lat/long.

After that, there were many sources:  gazetteers or route descriptions
published in major ornithological works; non-ornithological route description
and gazetteers; DeLorme atlases of US states (very detailed and very
inclusive); modern maps bought for this purpose or borrowed from other
departments; and USBGN gazetteers.  These last might seem like they'd have a
higher priority, but I found them more useful for "last resort" work--there
are so many duplicate names and so little information about the localities
that I found it hard to decide which one was the locality I wanted.

The actual changing in the computer was done mostly using a global change
facility.  When there were only 1 or 2 to be changed in a particular way, the
record was called up (the massive printout had specimen numbers listed with
each locality record, and corrections and additions were made onto the printout
and later done to the computer database).

I'm sorry that this came out so long--the only "maximization" we did had to do
with choices of sources of information.  I hope this helps.  By all means,
contact me if you have questions.

Robin Panza
Section of Birds                      panzar at clp2.clpgh.org
Carnegie Museum of Natural History
4400 Forbes Ave.
Pittsburgh  PA  15213                 (412) 622-3255