GPS Watches Demystified
By Coaches Stephane Moulin, Graham Hand and Brendan Davies
There are a lot of running “smart” devices on the market and it is hard to make a choice sometimes…. One key attribute that some runners may be looking for is elevation gain to capture those trail adventures.
Most of us use a GPS (or multiple) watch to capture our running which is then exported to some third party/proprietary application for review, sharing, analysis.
Global Positioning System – GPS
The global positioning system is a satellite-based navigation system consisting of a network of 32 or so orbiting satellites that are 20,200km above the Earth. The satellites are constantly moving, making two complete orbits around the Earth in just under 24 hours. If you do the math, that’s about 2.6 kilometers per second. Get that Strava!
GPS watches vary in sizes mainly as a result of the CPU and GPS antenna in use, which in turn, affects the speed of lock /acquisition and strength.
GLONASS is available on the higher end of the Garmin watches. It is Russia’s version, and it stands for Globalnaya Navigazionnaya Sputnikovaya Sistema, aka Global Navigation Satellite System for those of us who don’t speak Russian. With several fewer satellites than GPS, GLONASS is comprised of 24 and they are at an orbital height higher than GPS at 21150 Km. They also differ in orbital plane inclination, tilted at 64.8 degrees. GLONASS provides your position with 5–10 m of accuracy. If you turn on GLONASS combined with the GPS you get over 55 Satellites that will give you more accuracy but at a cost of Battery Life.
If your device receives information from both GPS and GLONASS, you’re in for a more accurate tracking experience. So if you happen to be in a spot where GPS isn’t picking up your signal, GLONASS can relay its information to you so there is no gap in coverage, and vice versa. Another benefit of having both GPS and GLONASS working together comes when you’re running on a course with tight turns, as in a lot of trail running. The heightened accuracy will help track fine movements more precisely.
What types of GPS Watches are there?
There are two types of GPS watches on the market:
- With barometer
- Without barometer
Why does that matter you may ask…? Well it does if the focus is to capture accurate measurement of ascent/descent.
GPS satellites accuracy is approximately 5m in altitude per position locked. So one can imagine that on a 30km run with plenty of terrain variations the “errors” in measurements compounds resulting in an approximation.
Ascent can be defined as an instance of rising up through the air, for the higher you go the less pressure the air exerts. Incidentally, the lower you go, the higher the air pressure is up to sea level where it is stable at approx. 101 kPA/1013 mbar.
One of the most accurate method for measuring gain/loss in altitude is to measure the atmospheric/barometric variations in pressure. This is where some of the GPS watches come into their own as they rely on the physics to measure gain. Those watches require being calibrated/set-up whereby a reference is input (sea level pressure as altitude would be 0m).
Some of the running watches with Barometers are Suunto Peak, Suunto Vertical, Fenix 3.
Most other watches use GPS as a gain/loss measurements and it has been seen a difference up to 20% in meters between the 2 type of watches.
Is that an issue? Well not really, as long as you understand the “restrictions” or unless you are Killian Jornet or Emily Forsberg …
What affects measurements?
Your GPS watch measurements will be affected by the terrain (visibility to the sky), terrain (along cliffs some GPS watches struggle to gain lock due to the “noise” created by the cliffs and signal bounces off), polling frequency to some extent although your ascent/descent speed will not be fast enough to make huge differences unless you are gaining/dropping more than 10m per sec J (but this will affect distance).
GPS watches measure straight lines so running in a circle/oval will result in some anomalies as the watch will compile straight lines instead of curves. This can result in different distances measured on watch vs the number of laps done on a track – the later being accurately known.
To add to the confusion, each watch exports data to its own app and also allows sharing it with 3rd party applications such as Strava, TrainingPeaks etc.
Some of those applications use their own algorithm to measure ascent/descent which could result in showing different results between your Strava record and your GarminConnect/Suunto Movescount files. Strava calculates gain by summing up all gains in an activity, where a gain is defined as the difference between a low point and a high point. The “issue” is that Stava detect peaks to determine high points and a gain must exceed a threshold in order to be counted and some of the data is smoothed – this typically results in some loss in gain due to the law of averages.
Additionally, Strava ignore elevation data from devices without a barometric altimeter and instead determine elevation by comparing data to a set of elevation databases.
TrainingPeaks uses data supplied by your watch and performs some averaging and carries some inacurracies outside of the US (as their reference database is mainly US based).
Most applications do have the option to perform Elevation Correction for your data which would allow remediating some of the discrepancies.
A Classic Example on how it can all go wrong!
Below is a screenshot of some Training Peaks data from an UP Coaching Athlete who lives in Blackheath and regularly trains in and around the Grose Valley and uses a Garmin device.
At first glance it looks like at around 15km the athlete decided to take a side trip down to the valley floor, perhaps via Govettes Leap or Parry’s Lookdown stairs. This would be further confirmed by the elevation data (the last two columns of the ‘Laps and Splits’ table). The 2nd last column is Elevation Gain and the last column is Elevation Loss. You can see that Training Peaks, from where this data has been taken, has given for Lap (km in this case) 16 as -109m loss and +158 gain. Please note that this data has been originally taken from Garmin Connect via transfer.
However, in this case looks are deceiving. Upon talking to the athlete, it was confirmed that indeed there was no trip to the valley floor and the athlete stayed well and truly on the top section of the cliff. So how could Garmin get this so wrong?
Just as has been explained. Most Garmin devices don’t use barometric readings to ascertain altitude, rather relying on map contour data which is then ‘overlayed’ on top of your GPS running route to acquire a ‘guesstimate of elevation gain/loss’. On flatter terrain this will be mostly accurate. But on terrain with sharp drops, cliffs and the like, it only takes one bad GPS recording (in this case GPS taking a recording of a position ‘over’ the side of the cliff) for it to become suddenly very inaccurate. So in actuality the recording has ‘overlayed’ the GPS Point as being on the other side of the cliff face and so ‘assumed’ the athlete did go over the cliff face…with a device that relies on air pressure readings this simply wouldn’t have happened.
Why is this important?
In the above example, the issue in the 16th km has added approximately 150m of loss and gain to the athletes final total of 560m. This makes a big difference, particularly in a longer session where the coach may have set an elevation target to reach and if the inaccuracy occurs multiple times. The athlete may not be working as hard even for all intents and purposed their device may be telling them that they are reaching their targets.
Hopefully this clarifies. Happy elevation gain hunting
Coach Stephane, Graham and Brendan