Geolocation Decimal Degree

424 views
Skip to first unread message

Mike Giddens

unread,
Sep 1, 2008, 11:27:34 PM9/1/08
to Biodiversity Collections Index
I was looking at the example of Edinburgh Herbairum and see that it
shows:
Longitude 55.9669610000
Latitude -3.2062350000

My questions is if a GPS is only accurate to +- 10 Meters then that
would only be 4 decimals places at the equator and more accurate
elsewhere so what is the significance to go past 10 meters when a room
or building would never be smaller then that?

I am just wondering what you guys though about this little detail.

-Mike

rogerhyam

unread,
Sep 3, 2008, 9:39:36 AM9/3/08
to Biodiversity Collections Index
Hi Mike,

You are correct in that there are far too many decimal places in the
longitude and latitude fields. These are actually the default decimal
sizes for MySQL I think. I will raise a bug to clip them down next
time a visit the form interface.

I'll have to disagree with you on the number of places needed though.
We tend to use 6 decimal places as a standard for location data. Here
is my somewhat round about logic for why:

1 degree of arc = 60 minutes
1 minute = 1 nautical mile
1 nautical mile = 1,852 meters
therefore 1 degree = 60 * 1,825 = 111,120 meters.
therefore 6 decimal places gives at least 1 meter *precision*
(actually down to about 11 cm) i.e. define the center of the 10
meters of uncertainty with precision rather than the corner of a 10
meter square. It could be argued that 5 decimals is enough but I think
4 is probably too few to get the most from regular GPS. Ideally one
needs to record some measure of certainty as well as the actual
reading.

Emotionally being <1 meter (3 feet, a yard) from something seems right
for precision because if you are that close you are close enough to
tough whatever it is - give or take 10meter ;) also recording one more
decimal than you need is preferable to having recorded one less than
you need.

As we are talking about buildings this is all pretty academic but it
would be good to get it right in the long run.

Thanks for your contribution,

Roger
Reply all
Reply to author
Forward
0 new messages