Registration by invitation. Write to invitations at this website.
RoadSkater.Net skating & cycling photos!

Barometric and Digital Elevation Model (DEM) Correction with Cycling Running Skating Global Positioning System (GPS) Data

roadskater's picture

Congrats to all who are spending plenty of time exercising! Kudos if you do that AND spend lots of time thinking about this while exercising or not. Thinking about exercising is part of actually going to exercise. And any mind game that gets me out exercising is a mind game I like. Also, thinking about scientific and mathematical issues keeps brains exercised along with the rest of the body. If you feel elevation data is irrelevant to your training, you won't want to waste your time reading this article. If reviewing your data informs you and inspires you, read on to see if you agree.

I am a big fan of Garmin devices in general (great hardware, great software), and I love my Forerunner 305 wrist GPS/HRM. It is the combination of data on how hard you worked (not power, but heart rate over time at least) with the automatic logging of your exercise diary that make it a real incentive for me. I don't like keeping records of workouts by writing them down or typing them in, so I really feel I benefit by knowing what I did or didn't do when, including average HRMs as they improve over the season. The distance data is also important to me, and to some extent, the elevation data (but usually after I upload the GPS data and use my position to determine elevation (see below).

However, for some discussion of the weak points (elevation, calorie estimates) of this wonderful device (the Forerunner 305 GPS/HRM) and of similar systems, see the main articles and comments here...

  • http://roadskater.net/garmin-forerunner-305-gps-heart-rate-monitor-has-useless-calorie-counting-feature*
  • http://roadskater.net/inline-speedometer-recommendations-replies-actual-users-discuss-and-review-forerunner-305-hrm-gps
  • http://roadskater.net/garmin-forerunner-305-training-center-gps-unit-updates-pdf-manual-edge-305-705-forerunner-405

GPS is more accurate in the XY (lat long) than the Z (elev). The Z error is in general 1.5x (150%) of the XY. So if it's 10-foot accuracy for position (about as good as it gets for standard GPS), it's about 15 feet off for elevation. Trouble is, 10 feet across is not much, but 15 feet up and down feels like a lot usually. Winter readings for all dimensions can be better on the same course (e.g. Country Park, or climbing at Mt. Mitchell) due to more tree cover in the summer in many cases, but in some cases the augmenting satellites can be too low for your horizon (see WAAS below). 

While in the actual roadway, usually tree cover is less of a concern than for trail users, but it is still a variable affecting results. When there is a clear southern exposure, units with WAAS (wide area augmentation service) capability improve accuracy. (WAAS provides extra satellites to enhance the accuracy of standard GPS...WAAS is more helpful on flat land or at least with a clear southern exposure, and as you get closer to the equator, since the equatorial sats are higher in your sky view).

A 12-channel receiver helps, too. The Garmin 300 generation has improved signal strength over the old 200 series with the use of the SiRFstar III receiver (12 channel receivers like the SiRFstar III can track the maximum likely available 12 satellites in view at a time simultaneously rather than switching among satellites to find the strongest signals...producing better data more quickly). 

One good solution that improves consistency of data over many variable circumstances is to use the XY data from the GPS device and to look up the elevation based on government-produced digital elevation model (DEM) data. This would explain why the information in your GPS unit can be different than what you see once you upload your lat long data to a web service, for example. I use "best available" DEM data when I have the choice, but DEM data can be USGS NED 30m resolution data, or NASA's SRTM1 30m resolution (these two for the USA) or NASA SRTM3 for many latitudes outside the USA (there are also newer models from more recent USA/Japan cooperation using Space Shuttle instruments).

If I understand http://ned.usgs.gov/Ned/faq.asp, National Elevation Data (NED) is roughly accurate in the z (elev) at about 8 feet. About the best one can hope for in GPS is 15 feet optimally, and usually worse. But NED is taken every 30m or 98 feet, roughly, the rest being extrapolated algorithmically. See

  • http://ned.usgs.gov/downloads/documents/NED_Accuracy.pdf. Also...
  • http://www.gpsvisualizer.com/map_input provides the opportunity to use DEM data in its General Map Parameters by clicking on Show Advanced Options. 

I trust the elevation data from USGS and NASA more, in general, even if it has lower granularity. This provides some consistently measured data based on my GPS data, which is subject to numerous variances of accuracy when riding, skating or walking, but it avoids the MULTIPLICATION of my error by a factor of 1.5 as would be the case typically with GPS elevation data.

For brief sessions where weather is not changing rapidly, barometric elevation can be a good alternative or aid to GPS. Some devices (especially those that do not have GPS capability, notably the old Polar 700 series) use barometric elevation, which is good if: 

  • the user sets it to a known standard at the start of exercise, and for long events, resets the elevation at known points (for example, when hiking); 
  • the user has worn the device long enough to get the device up to a steady temperature (so that it doesn't get MUCH warmer or cooler during the exercise period...it helps if you do not wear in on the wrist or elsewhere next to your body);
  • if the ambient temperature remains reasonably constant, and 
  • if a weather front does not come through during exercise (meaning the ambient barometric pressure doesn't change much during exercise). 

Note that a barometric device does not know if a change in pressure was due to a change in your elevation or a high or low pressure front passing through. A GPS, likewise, measures your position with error, so that you could sit in one place with either device and read false elevation changes. However, GPS errors are less important as you move faster, and even a GPS at rest will average to a very good position estimate, and DEM look up can give you a good estimate of the elevation at that point. It is good to blend common sense with either barometric or satellite data. 

Some devices, like the older (not sure if in production now) handheld Garmin Etrex Vista use the GPS elevation as a basic correction standard to keep the barometric numbers more reasonable in differing weather conditions without requiring the intervention of the user at known elevations. However, it is still better for the user to correct as they hike, skate, bike or whatever. 

More recently, accelerometers have been employed in situations where GPS or barometric (or hybrids of these) have not had success alone. However, my mostly uninformed impression is that accelerometric data is not as accurate as barometric-corrected GPS data or DEM-corrected GPS data. But that is based on a few conversations with users, not anything resembling research. 

So if your goal is to COMPARE workout courses with some accuracy in the xy (lat long) and consistency in the z (elevation) my take is to trust the xy lat long of your GPS (in general) and upload that data to a service using the data to provide you with DEM elevation results. The shorter the workout, the more reliably barometric data can be used alone or as augmentation for GPS. Another way to think of it is if 30 meters (98 feet) x 30 meters is a large area compared with your route, lean toward barometric measurements. If 30x30 is small compared to your route, DEM is probably a good method, especially for comparing various routes of many differing weather conditions.

Right or wrong, those are my thoughts at this point.


eebee's picture

Very informative!

Thanks for writing this. I will probably refer back to it often. 

roadskater's picture

Glad to Hear it Might Help

I am sure there's more that we can learn about all of this, but there was a recent discussion on a local online group about whether to believe your "on-board" device or website-provided info based on your GPS tracks. I never got a reply from anyone there, but I trust people will be reading the info here along the way as time rolls on. We'll see! Today's top articles were:

Those are all your contributions I think, eebee! Thanks for sharing!

If you look at the "Active Content" block, which helps find great popular content for the day, week, month, quarter, half-year and year, the CLEAR and CONSISTENT winner, for some reason, is...

How to Reboot a Garmin Forerunner 305 That Locked Up When Recharging on the Cradle and Won't Power On

Yes I write SEO titles I admit and that one is a clunker, but I wanted to include words from the kinds of searches I'd make to find information on that topic, then mashed them into a title only the writer could love. I wonder if that article has been linked to somewhere...liked, digged, tweeted or something. 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Search Ebay for Skates and More
Search Ebay for Inline Skates
Search Ebay for Skate Wheels
Search Ebay for 100mm Skate Wheels