[Cdt-l] accurate GPX track based on the new mapbook data?

Brett blisterfree at yahoo.com
Sun Mar 27 22:11:48 CDT 2011

The shotgun blast of Postholer waypoints is a great resource, Scott... 
but I'm especially needy here. :) The only data I've seen covering the 
official CDTA route in the Gila is the Trailsource GPX file and Bear 
Creek Survey's new waypoint data.

Jerry, thanks for the explanation and I'll look forward to the updated 
tracks when they become available. I'll just add to what Scott noted, 
that GPS Exchange Format (GPX) is indeed a universal format for both GPS 
units and mapping software, and is not inherently limited to 10k points 
the way a GPS unit would be. You could take whatever waypoint data 
you've collected, increase the density considerably beyond the 1/2 mile 
increment (take your full list and strip it down to, say, one minute 
timestamp intervals), and then convert it to a GPX track or KML file, or 
even just provide the waypoint list in txt or whatever and we could go 
from there. The conversion process, turning the waypoints into a 
functional track that we can use in mapping software to plot and analyze 
an actual trail line, is straightforward.

The Trailsource GPX tracks are inaccurate and out of date, at least in 
the Gila / Black Range which is all I've looked at. It appears that it 
was either created from a short list of waypoints, connect-the-dots 
fashion, or was plotted in software based on rough guidance of some 
sort. The errors are beyond what one would expect from a 
recreational-grade handheld GPS, ergo I don't think it's a 
field-recorded track (at least not all of it).

- Brett

> ------------------------------
> Message: 4
> Date: Sun, 27 Mar 2011 17:14:34 -0700
> From: Scott Parks<public at postholer.com>
> Subject: [Cdt-l]  accurate GPX track based on the new mapbook data?
> To: cdt-l at backcountry.net
> Message-ID:<4D8FD2EA.1000107 at postholer.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> First, here are 3000+ way points viewable in google maps. These include
> Dan's, journalists, SPOT users, and others who have contributed their
> data. This is a shot-gun blast of way points. The trail trace is a
> hybrid of various sources and does not necessarily reflect the official
> route. Warning, this CDT google map showing waypoints may kill your
> browser, be patient, works great in FireFox 4 and IE9:
> http://postholer.com/gmap/gmap.php?trail_id=2&zoom=5&dist=0&markertype=14&vw=0
> Thoughts on trail traces and maps. Imagine looking at a 7.5 minute map
> (1:24K). Imagine the trail trace drawn on the map. The colored line
> (trail trace) is going to be a least 20 feet wide, and more like 30.
> Assuming accurate data, having a lat/lon point more frequently than 20
> feet gains nothing in map quality as it wouldn't be displayed. If I had
> hourly accurate tracks with 10K data points I'd grab only points that
> were 20 feet apart and toss the rest. That's 264 points per mile. Much
> more manageable without sacrificing quality. From a 1:24K map
> perspective, 50 points per mile is quite fine.
> The GPX data format is just that, a format. It speaks nothing of data
> quality. The GPX format will hold accurate and inaccurate data equally
> well. It will also hold professional and recreational data equally well! ;)
> I've been dragging my feet about making some 1:100K and 1:250K overview
> maps available for the CDT. If anyone has a need for these, let me know.
> I could use the motivation.
> -postholer
> Brett - It's not surprising to me that the gpx files you have don't match up
> to the maps very well.  The gpx format is something that only exists in the
> recreational  gps world so those tracks were in all probability collected
> with a Garmin or Magellan using the internal antenna while attached to a
> pack or waistband. That method may be fine for navigating a trail but is
> seriously flawed should you desire to collect usable data. The antenna of a
> gps has to be elevated above head high or the data stream is compromised by
> reflections and obstructions. A gps signal bouncing off of your head can
> introduce hundreds of feet of error.
> Our project data is collected using professional grade sub-meter rated gps
> gear with external antennas elevated above head height. These units address
> the data quality issues above but they do not support the gpx format.
> Converting it to gpx requires a lengthy process to dilute the database since
> the most robust track files are limited to 10,000 points. We collect more
> positions than that every hour we are on the trail, so a great deal of
> editing is necessary. I don't anticipate tracks being available from us
> until after the initial project is complete.  The waypoints, however, are on
> a 1/2 mile average spacing and the mileages between them are accurate. Get
> the waypoints and maps and you will get there.
> We will begin in The Great Basin headed north in June.  Hope to see some of
> you out there.
> best wishes,

Jerry Brown
GPS Mapping Specialist
The Continental Divide Trail Alliance
bearcreek at wic.net

More information about the Cdt-l mailing list