1

Elevation accuracy on RA logs (Read 141 times)


delicate flower

    Maybe this has been answered but I couldn't find it...

     

    How accurate are the elevation figures on the RA workout maps?  My run yesterday indicates a 952' elevation gain according to the RA estimates.  Garmin Connect indicates 375' elevation gain.  Kind of a big difference.  I notice most of the time that the RA numbers are quite higher.  I use a Garmin 610, so it doesn't have an altimiter.

     

    The hill profiles are similar, but the numbers are quite different.  It certainly feels more like 375 than 950.  Just wondering.

    proud sherpa


    A Dance with Monkeys

      Depends.


      When you upload, do you have the "Replace GPS elevation measurements" checkbox checked or not?


      delicate flower

        I'll need to check tonight after my run.  I leave it at whatever the default is.  Do I want to check it or uncheck it?

        proud sherpa


        A Dance with Monkeys

          The intrinsic elevation data your GPS device collects tends to be highly imprecise and inaccurate, with loads of scatter. This tends to lead to an overestimate of total elevation change.

           

          If you have the box checked, RA replaces this data with reference data from the US Geologic Services, which is generally better.


          an amazing likeness

            You want that box checked -- that causes RA to use the reference data (USGS ?) source for the elevation of each track point, rather than the elevation reported by the GPS for each track point.

             

            You can also force RA to resample the elevation via the 'two mountain' icon, upper right on the menu tool bar when view the map of your GPS data.

             

            I'll need to check tonight after my run.  I leave it at whatever the default is.  Do I want to check it or uncheck it?

            Choosing my words carefully has never been my strength I've been known to be vague and often pointless

              Looking at your run, the highest point in your run was ~375' above your lowest point.

               

              Perhaps the garmin is quoting the range in elevation (MAX-MIN) while RA is adding up all of the elevation change with a positive slope to get the +XXX.X' and adding up the elevation change with the negative slope to get the -XXX.X' values which are added together to get the net elevation change between the starting and ending points of your run.

               

              To me it seems like there is close agreement between the two methods but they are just presenting the data to you in different ways.


              delicate flower

                I loaded today's run with that box checked.  RA shows 799' elevation gain.

                http://www.runningahead.com/logs/f9d0a30c6bbd4510baabdc79caca8ebe/workouts/bc4f917544414f7b9e179508117ee714/map

                 

                Here is the same route from last week, loaded with the box unchecked.  This one shows 731' elevation gain.

                http://www.runningahead.com/logs/f9d0a30c6bbd4510baabdc79caca8ebe/workouts/d369e6748d0d45e091cb29c898139320/map

                 

                Garmin shows 215' elevation gain.

                http://connect.garmin.com/activity/370695267

                 

                Maybe delucj is onto something and I'm just misunderstanding the data.  Not a big deal either way...I'm just trying to understand what I'm looking at.

                proud sherpa


                delicate flower

                  I just clicked that resample button and it did bring the elevation gain down.  583' for today's run.  Still more than twice what Garmin Connect shows.

                  proud sherpa

                    Depends.


                    When you upload, do you have the "Replace GPS elevation measurements" checkbox checked or not?

                     

                    Where is this option?

                    And you can quote me as saying I was mis-quoted. Groucho Marx

                     

                    Rob


                    delicate flower

                       

                      Where is this option?

                       

                      When you click "Upload GPS data" and RA finds your file, a box appears with four check boxes.  That box is one of them.

                      proud sherpa

                        Elevation is not a precise measurement, even if you are standing on the exact spot with a yard stick.  It's hard to say which website (RA, GC, or any other site) has the more accurate elevation because it depends on many factors:

                         

                        Data source

                        RA uses multiple data sources to maximize global coverage.  Each data source has a different measurement method, which would affect the accurate of each data point.

                         

                        Data source version

                        Like everything else, the elevation measurements are updated regularly.

                         

                        Data resolution

                        A data source can have the data at various resolutions.  Typical resolutions include 90m, 30m and 10m.  That is, the elevations are sampled at every 90m, 30m or 10m.  The higher the resolution, the more data points there are, and the more disk space is needed.  The sampling frequency can have dramatic effects on hilly routes.

                         

                        Sampling algorithm

                        Even if you have the 10m resolution data, you still have to interpolate the elevation at a given point.  There are various methods of coming up with that number.  Depending on the formula used, you can have differing elevation values for the same location using the same data.

                         

                        Data Smoothing

                        After the elevation points are sampled, the data might be smoothed to provide a less jagged view.

                         

                        I'm sure I missed couple of other factors that can contribute to the difference.  My point is that you should always take the elevation with a grain of salt.

                          To add to Eric's comments, take a look at some of this in TopoFusion

                          http://www.topofusion.com/climb.php to see some of the variation based on methodology.

                           

                          Take a look at the topo map version to see much of the topography. Among other things, my hunch is you may be picking up "hills" near relatively steep "embankments" or whatever that are next to your course. If your data point is off a few feet, it may think you're up on the hillside, then back down, while you were really traversing a relatively flat area. I have that happen on an abandoned railroad that's on a bench - steep up on one side to the road above, steep down to the river below. It's funny to see the hills that the gps (without altimeter) generates.

                           

                          Also, note the jaggedness of the RA profile when the terrain doesn't look like that based on topo. (TopoFusion used to refer to that as "hairiness", iirc, but I can't find that page now.)  That said, sometimes a lot happens between the contours on those USGS maps. Our trails that go up and down old glacial moraines are notoriously hilly (and why we love them) look like they're on flat terrain because the old topos had 100ft contours (and probably not that accurate) and most of the hills are < 100ft.

                           

                          I'm just looking at your data and maps, but you were there so you'd know more about what you actually ran through. Something like "Terry's Plain" doesn't sound particularly hilly. There's a climb coming out of that, but then the map looks like it's following a contour, which it may or may not be. There are a couple hills in there though - crossing the river as well as a couple noses of terrain sticking out. I was looking mostly at miles 3-6, and other areas looked flatter. There's a couple benchmarks in there just after miles 4 (on Terry's Plain, iirc).

                           

                          I would agree with you that the garmin connect numbers may be more realistic - somewhere in the 300-500ft range rather than 800ft range. My FR305 could multiply the hills by 3 - so 500ft of up became 1500ft. It was decent on the mountains - like 3000ft up, but it was bad on the rolling hills.

                          "So many people get stuck in the routine of life that their dreams waste away. This is about living the dream." - Cave Dog