[Herpnet] Questions about georeferencing- coordinate precision from GIS layers

[carol spencer]

This message is relevant to those who are currently georeferencing, so I am
forwarding it to the list:
---------- Forwarded message ----------

Dear John,

It's true though that I still have some confusion regarding precision
of my co-ordinate position and extent estimation. It seems to me that
the precision (and I may have been using the term accuracy
incorrectly in my questions below, when I actually meant precision)
given by the GIS is often much greater than I feel is justified. This
is why I was asking the question about how many decimal points I
should report. The thing is, if I move my cursor even imperceptibly
(particularly when measuring an extent of a relatively large feature,
like a park) the extent will change on the GIS. To me, it seems a bit
subjective as to where I finally take my value, yet the 'exact'
measurement or co-ordinate that I get off the GIS suggests that I was
able to measure much more precisely than I was.

So, that was my question in more detail. Thank you for the
clarification of my other ones. I hope that what I did on the
maps/excel spreadsheet that I sent on Friday made sense.

Once again, thank you for your patience and assistance.

Response back from John W:
Ah, OK. Good question. I got caught up in the semantics and missed the
point. I understand what you are saying about unwarranted precision. My
recommendation is to conserve seven digits of precision. The reason isn't
what you'd expect. The precision of the GIS measurement is real, so it
shouldn't be a concern. Your ability to place the point accurately is quite
a different thing. This isn't a judgement of your ability (:-)), it's a
pragmatic view of the lack of certain knowledge about the original
information. This uncertainty is already accounted for in the various
sources of uncertainty the calculator takes into account, and it's the real
measure of the quality of the end product - the precision of the point is
not.

So why seven digits? It's because that is the number required to preserve a
decimal degree coordinate through transformation to and back from
sexagesimal degrees (degrees minutes seconds) at the level of a meter on the
ground. If you don't keep that many digits, changes to a different
coordinate system, then back again will not give you the original value back
again.

-- 
Carol L. Spencer, Ph. D.
Museum of Vertebrate Zoology
3101 Valley Life Sciences Building
University of California, Berkeley, CA, USA
94720-3160
atrox10 at gmail.com
atrox at berkeley.edu
phone 510-643-1620 /fax 510-643-8238
http://www.herpnet.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.nhm.ku.edu/pipermail/herpnet/attachments/20070226/5094ecba/attachment.html