[Taxacom] Accuracy of GPS Receivers, was...
mesibov at southcom.com.au
Thu May 9 21:28:21 CDT 2013
GPS manufacturers aren't often clear on this point.
>From that dataset of 50 readings I linked to, you can calculate a mean location - a 'best estimate' by the GPS unit. You can then summarise the distances from this mean location as a measure of the *precision* of the locating process. On average, the readings were 3 m from the mean; the standard deviation was 2 m, the maximum 11 m.
How close was the mean location to the true location? That's the *accuracy* of the averaged GPS readings, and Garmin claimed for that eTrex that 95% of readings would be within 15 m of the true location under best field conditions.
This is a good thing. It means that if I'm getting quick and stable satellite fixes and low 'declared accuracy', I can be reasonably confident that a single GPS reading - taken after patiently waiting for the reading to settle down - is within the '95% circle'.
If I'm deep in a canyon under a tall, dense tree canopy, and even after patiently waiting for 10 minutes my GPS position still isn't stable and the declared accuracy is up around 80 m and I suspect satellite signals are reaching the GPS after bouncing off the canyon walls.... Well, maybe a paper map, or some field notes followed by a look at Google Earth would be a better option than a GPS.
Last year (http://www.pensoft.net/journals/zookeys/article/4195/) I mentioned in a paper the results of checking the location of a permanent survey mark in my neighbourhood. Google Earth put it 2-3 m from the surveyed location, which is pretty impressive georegistration of a satellite image.
I've been using Google Earth more and more in recent years, first to check that a GPS reading is for the site I thought it was, second to get an approximate elevation above sea level for the site, and third to 'displace' the GPS reading when required - like when I couldn't get a reading deep in a forested gully, and had to move 50 m to a nearby clear spot to get a reading. In the latter cases, again, I'm finding the Google Earth position is usually well within that 95%/15m circle advised by the GPS manufacturer.
The biggest variable in my sampling is me. A single collecting site is almost always a patch of bush, in which I take specimens from several scattered spots found while wandering around the patch. The point+radius circle defining that site normally has a radius of 25 m, and often 50 m. My GPS and Google Earth 'errors' are well within that.
Dr Robert Mesibov
Honorary Research Associate
Queen Victoria Museum and Art Gallery, and
School of Agricultural Science, University of Tasmania
Home contact: PO Box 101, Penguin, Tasmania, Australia 7316
Ph: (03) 64371195; 61 3 64371195
More information about the Taxacom