[Geysers] GIS Work and Accuracy

[Janet White]

My husband's a cartographer who has done everything in GIS work from field
work to final map.  He says the equipment the Park Service uses makes a
difference.  For example, if they use Ashtech Z-Surveyors  - depending on
many factors (the user,  the unit, the location and accuracy of the base
station, and how long they sit on a point, atmospherics, number of
satellites, post processing, etc.) - they can get it down to cm and even mm
accuracy.  If they're using a Tremble Sub-meter unit (again, depending on
many factors), then the accuracy can be gotten down to about 1.5 foot
accuracy when done well.  

This page has the specs on the Z-surveyor:
http://www.eclipse.net/~essco/gps/gps%20only.pdf 

As for a slight jiggle in the unit when sitting on a point...Ok, this is an
explanation in layman's term based on my husband explaining it to me, me
trying to translate it to layman's terms and him reading through this to
verify.  Just to try and get it understandable and yet accurate - lots of
info left out here, in other words.  

Dashboard GPS units in cars are usually accurate to about 50 feet.  Mapping
is often done to decimeter accuracy (~ 3").  But the same units used for
mapping can also be used for surveying accuracy, down to cm or even mm
depending on how it's used and who's using it.

Sitting on a point, (holding the pole with the big white round thing on top
as steady as possible), it's getting a stream of data from the satellites
out there - and you can set it to take a point reading from the satellite
every 1 second or 5 seconds or whatever - most folks set it to every 1
second.  That means it's getting 60 readings every minute.  The longer you
sit on a point, the more data you get.  Then the unit will average all of
those readings - some have panels that tell you when you're within a
specified accuracy (decimeter, 3 mm, or whatever the desired accuracy).  So,
if you move the unit for a moment an inch over - yeah, if it takes the data
point from the satellite when moved, then it will affect the accuracy.  But
what accuracy you're looking for is the key.  

If you need it for survey type accuracy, then they usually use a tripod and
sit on the point for an hour to four hours to really tie down that point and
avoid movement.  This type of accuracy is used to create control points or
base stations.  A base station is a unit that is always on and always in one
point.  A reason for using a base station, which the Park Service has in the
Old Faithful area (and I'm sure in other areas as well), is to have a second
unit getting the same satellite info.  Usually you want the 'roving' unit
within 50 miles of the base station if you want mapping type accuracy.  For
surveying type accuracy, you need to be within 5 miles.  Then, using
mathematics to triangulate everything, they compare the data from the two,
and can check and see how accurate the remote unit was at a point (and
retake points if the accuracy isn't good enough for the project).  

My husband's been out in stiff (and cold) winds that make it difficult to
hold the pole still, and has had to retake points when they don't prove to
be as accurate as needed.  So, if you move the pole while taking the point
in an area around an inch, you may only have accuracy down to an inch.  If
you're waving it around a foot or two, then that's probably about as
accurate as it'll get.  And that might work great if the requirements for
the project only needs that.  Accuracy in GIS work depends on numerous
factors, but the reality is that if you have knowledgeable people on the
project, and good equipment then you can trust the data.  Obviously, if
they're claiming that type of accuracy, they trust the equipment, people,
data, and post processing to make that claim.   

The National Geodetic Survey uses RTK to establish  GPS control points down
to the 100th of a meter, which would be mm accuracy which the Park Service
likely used to get control points for their work. 

As for the lat/lon listed, it's a number - my husband says his work normally
lists in degrees, minutes, and seconds with it taken to 4  decimal points
after the second.  Trying to equate that to the area encompassed is
possible, but difficult simply because, as stated,  longitude curves and
arches across the Earth and the distance changes - the distance between
longitude at the equator is different than at 45 degrees.  So figuring the
distance based on lat/lon numbers depends on what your latitude is and would
need to be recalculated for each change in latitude.   

But what's the point?  Does it really matter whether it's mm accuracy or
meter accuracy to use the information?  The way they list it on the RCN may
simply work as a way for people to use the lat/lon as a way point on their
handheld GPS units.   It should get you to the feature, I would imagine - at
least to the general area depending on the accuracy of the handheld unit.
And to do research, that would be the point, I would think.  The photos, to
me, would add confirmation that everyone is talking about the same feature -
and may not exist for comparison for studies in changes of the feature.
Between accurate GIS info and photos - it's doubtful scientists or
researchers could mix up the features which I see as a good thing.

Janet (and Mike) White

-----Original Message-----
From: geysers-bounces at lists.wallawalla.edu
[mailto:geysers-bounces at lists.wallawalla.edu] On Behalf Of SCOTT BRYAN
Sent: Sunday, December 02, 2007 8:58 PM
To: geyser observation reports
Subject: Re: [Geysers] People on Minute Man Geyser

Those people were involved with the production of the Geothermal Inventory
that has gotten recent attention in these missives. As I've noted, the NPS
people seem to believe that their GPS positions are accurate to the
1/10,000,000th of a degree -- for example here, they cite Minute Man Geyser
as being at latitude 44.3547549 [deg N] longitude 110.7984134 [deg W]. As
pointed out previously, it is [my opinion again] rather dumb to presume that
accuracy: 1 degree of latitude is a bit more than 69 miles (1 degree of
longitude is, too, but only at the equator). So 1/10,000,000 of 69 miles (1
degree) = approximately 0.437 inch of latitude and considerably less than
that of longitude. So what happens if the equipment jiggles just a little
bit....?]