Results 1 to 10 of 10
  1. #1
    mtbr member
    Reputation:
    Join Date
    May 2006
    Posts
    311

    Elevation - Garmin connect vs Google Earth

    Hi Guys (and gals)
    Just a question my buddy and I were pondering. He has a Garmin Dakota with no barometric altimeter. After the last ride he loaded the ride data into both Google Earth and Garmin Connect. Over the distance of 35KM or so, google earth indicated approx 250 meters more in elevation gain than what was shown in the Garmin Connect application.

    Just curious as to what people figure would be more accurate? My money is on the Garmin connect site being more accurate especially considering the altitude correction algorithms etc. but I thought I'd see what others thought?

  2. #2
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,197
    there's no really good way to tell which is better unless you know the absolute highest and lowest elevations for a given hill and can compare to the elevation profile given by your favorite trail. and even then, that's just going to be an estimate.

    all those elevation gain calculations done by different programs and websites will be roughly based on the same data. Most of them use 30m digital elevation models, that means each 30m pixel on the DEM is given the same elevation. different algorithms for line smoothing and whatnot of your elevation profile or interpolation will change that. further, 10m DEM data is becoming increasingly available for free, so some sources may be using it already. it's hard to say because they typically don't publish that information.

  3. #3
    mtbr dismember
    Reputation: Wherewolf's Avatar
    Join Date
    Jan 2004
    Posts
    6,333

    Statistically insignificant

    Quote Originally Posted by bikeriderguy View Post
    Over the distance of 35KM or so, google earth indicated approx 250 meters more in elevation
    250 m over 35 km is statistically insignificant.

  4. #4
    mtbr member
    Reputation:
    Join Date
    May 2006
    Posts
    311
    Quote Originally Posted by Wherewolf View Post
    250 m over 35 km is statistically insignificant.
    That's not really the point. I was looking for feedback on what people thought was more accurate.

  5. #5
    mtbr dismember
    Reputation: Wherewolf's Avatar
    Join Date
    Jan 2004
    Posts
    6,333

    No way to tell

    Quote Originally Posted by bikeriderguy View Post
    That's not really the point. I was looking for feedback on what people thought was more accurate.
    No way to tell. Every Garmin GPS unit is a little different too. My 705 and 800 get vastly different numbers. I'd trust Topofusion over both Garmin and Google Earth
    Last edited by Wherewolf; 06-06-2012 at 04:10 PM.

  6. #6
    trail rat
    Reputation: slocaus's Avatar
    Join Date
    Jul 2005
    Posts
    7,647
    To add to what Nate and Wherewolf said, the best way to tell is this.

    CCCMB web - Facebook
    SLO trail maps - conditions

    "The physician heals, Nature makes well" - real fortune cookie

  7. #7
    Rep Power: Infinity
    Reputation: NateHawk's Avatar
    Join Date
    Dec 2003
    Posts
    11,197
    Quote Originally Posted by bikeriderguy View Post
    That's not really the point. I was looking for feedback on what people thought was more accurate.
    it's precisely the point.

  8. #8
    mtbr member
    Reputation:
    Join Date
    May 2006
    Posts
    311
    Quote Originally Posted by slocaus View Post
    To add to what Nate and Wherewolf said, the best way to tell is this.

    Awesome!
    thx guys

  9. #9
    mtbr member
    Reputation:
    Join Date
    Feb 2004
    Posts
    545
    Yep no real way to tell, you have to realize that exact accurate elevations are a real issue even at a single point unless you are using survey grade equipment much less a whole route.

    If you use an elevation datum such as google earth, or the GPS does elevation based on your position (if it's not barometric) there is an error factor because of how the datum is calculated, plus the error factor in your exact position. Barometric elevations are only accurate if you start with a known elevation such as a USGS survey marker, calibrate it to that elevation, and then only if the weather/pressure does not change significantly while you are riding.

    Typically I've generally seen a gps position based elevation within 10 meters of what survey markers said, but I've seen them off as much as 100 meters in places that are remote and have high relief. The barometric ones do better if it's over a short time frame, it's calibrated to a known elevation at the start, and the weather does not change. However, I've gone on hikes in the mountains doing just that and had the barometric indicated elevation be off 20-30 meters at the same spot between the start/end of the hike.

    InSAR (Interferometric synthetic aperture radar) work can see elevation changes in the range of millimeters, with good data and processing, but again it's only really good for a single point or small scene.

    If you could put together a high quality InSAR image in an area where you had a known USGS marker to calibrate the InSAR image (an InSAR image gives elevation change not altitude so it needs to be calibrated to a known point), and you had a GPS route to overlay on that data I'd say that's your best bet for the closest real elevation change measurements. Even then you'd have the problems of the GPS recorded route not being perfect and being segmented.

    LIDAR would be another option, that typically can get accurate elevations to the sub foot range but there's very little coverage aside urban areas. Plus you still need a on ground calibration point, or the plane is using GPS for it's elevation as a reference when flying the data, which has it's own problems.

    All that said as a general rule I don't bother to use the gps barometric altitude functions, a barometric graph is fine for predicting weather trends, but I prefer to use the GPS position based elevation, I've found them more consistent especially since most of the time you won't have a USGS survey marker to calibrate to at the start of your ride/hike.

    Even topo maps have this issue, they are still an average at some level, if you put an GPS track into ArcGIS and get an elevation profile for it, and then change the datum, you can get significant elevation changes for the whole route, because all the datums used as a reference are calculated slightly differently.

    Last I checked google earth was using WGS84 for it's datum (typically the best option for worldwide data), so you might also verify that your GPS is set to the same, because if they are not, some are compatible and some are not, so that will increase the error factor in both position and elevation when you try to compare/import them. WGS84 should be the default but ya never know.

  10. #10
    mtbr member
    Reputation:
    Join Date
    May 2006
    Posts
    311
    great info - thanks for taking the time to write these responses up guys!

    Personally I've had pretty good luck with my garmin 500. I have preset elevation points in a few local areas (one being my garage where I start most of my rides) If the weather stays consistent is seems pretty good. I work for the City where I reside and have access to the official survey maps with elevation so I was lucky in that regard.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •