Altimeter Watch Accuracy

[Bonzo]

For those with an altimeter on their watch: how accurate is it?

I’ve just now gotten around to getting an ABC watch, and although I’ve been able to verify that the B and C are both quite accurate, I have noticed variations of anywhere between 2 and 120 feet on the A portion…and sometimes, the variation itself changes from day to day (or hour to hour).  I looked up a map of geodetic survey points in my immediate area and I’ve done altimeter calibrations while standing as close to those points as I could get – and with a reasonably clear line of sight to whatever GPS birds were flocking overhead – I’m still seeing variation from both the survey data and previously-reported readings from the watch itself.  Is this par for the course, or am I dealing with something abnormal?

[Philip Tschersich]

ABC = altimeter, barometer, compass?

You would expect the altimeter reading to change as the barometric pressure changes. The barometric pressure changes with the weather or by changing your elevation. The watch doesn’t know which of these is happening (weather or elevation changes, or both), so you would need to regularly tell the watch your actual elevation or what the nearby barometric pressure at sea level is to keep it ‘calibrated’.

Barometric pressure / elevation relationship

[Bruce Tolley]

Are you calibrating daily? I have a Suunto Vector  watch with (barometric) altimeter.

I cannot recall the claimed accuracy and usually calibrate against a GPS or map every morning or at the trailhead when doing a trip. I would say within a 12 hour period, it is accurate to perhaps plus or minus 50 to 75 feet.  If I do not calibrate it can drift to plus or minus 100 feet since it is calculating altitude based on barometric pressure.

[Bonzo]

ABC = altimeter, barometer, compass?

Correct.

You would expect the altimeter reading to change as the barometric pressure changes…

I cannot recall the claimed accuracy and usually calibrate against a GPS or map every morning or at the trailhead when doing a trip…

Yes…understood on the barometric variation and calibration against GPS, but does the barometric variation still happen with a combined GPS/baro altimeter?

…so you would need to regularly tell the watch your actual elevation or what the nearby barometric pressure at sea level is to keep it ‘calibrated’.

So, this is what I thought the combined GPS/barometer solved: once the altimeter and GPS are calibrated – and yes, I’ve been calibrating daily as part of my “what all does this thing actually do?” routine for the last week – shouldn’t the integrated GPS correct the barometer so that the altimeter knows what it’s doing?  Do I have that right?

If 50 to 75 is normal, then okay, great: I can live with that.  Supposedly there’s a 1-meter accuracy, but I’ve yet to see it, according to the geodetic data.

[Brad Rogers]

I’ve used an ABC watch since 2005 and find it reasonably accurate in most situations.  I generally calibrate it at the trailhead of a trip to a known elevation and it usually stays within ~50ft or so during a week long trip and that’s accurate enough for me.   If a big storm rolls through it can change the barometric pressure obviously and it’s best to re- calibrate when you get to the next known elevation, but I can only think of a time or two where it went off more than 100ft or so.

The past few of years I’ve been using a Garmin GPS watch but I have to charge it every 3 days or so if using the GPS during the day.  The ABC watch was a change the battery once a year piece of equipment and thus simpler.

[Brad W]

My Garmin Instinct is hit or miss. To be accurate, it needs to be calibrated before the activity and then, the port on the watch is suspectible to sweat interfering with a good reading. I always verify later with GPS Visualizer and or Caltopo. GPS based altimeter seems to be much better IME.

[Todd T]

In my experience, they’re “accurate enough.”

If the thing’s right within a couple hundred feet, I’m happy.  And it usually is.

[Bonzo]

Thanks for all of the responses thus far.  I’ve located a few local spots that have easily-found survey markers and that are also in open areas, so I’m going to give this a few days of calibrating and start keeping some accurate records of what I’m getting.  I’m also wondering if there could be variance due to the software on the watch: it’s brand new and there have already been two updates, so maybe they’re working on stuff behind the scenes.

[Todd T]

I’m also wondering if there could be variance due to the software on the watch

My (now dead) Suunto Core had a nice “auto” system that intuited whether or not you were moving to decide whether to interpret pressure changes as elevation changes or weather changes.  It worked impressively well to reduce the frequency of needed recalibrations.

It was based on the rate of pressure change.  There was a threshold considered too fast to be caused by the weather; when that threshold was exceeded over a three-minute interval (IIRC), it switched into elevation mode and logged all further pressure changes as elevation changes.  When the pressure didn’t vary above a low threshold for about 12 minutes, it assumed you’d quit moving and switched into barometer mode, logging all subsequent pressure changes to the weather.  It was quite good at avoiding the need for calibration in the morning after sitting in camp overnight.

[Bonzo]

Thanks for posting that, Todd; it’s helpful to know a bit more about how the system in the watch actually works.  The watch in question is a Suunto 9 Peak, so I’m assuming that it has either the same system or a derivative thereof.  I deliberately skipped the calibration this morning in order to see if it was drifting overnight, and I found that today’s readings are off by .66 meters from yesterday’s, at the exact same location.  That’s pretty damned good, in my opinion…except that the readings themselves are off by about 6.5 meters from the survey data for this location, as of calibration yesterday morning.

On my way home from work today I’m going to stop at one of the previously-located survey points that is exactly 304.80 meters: 1000 feet, for those of us in the US, Liberia or Myanmar.  The round number doesn’t really matter, but it feels like it’s more accurate.

[Jenny A]

My experience over the past 40 years with using barometric pressure altimeters in the Rockies is that they might or might not require recalibration on a daily basis and sometimes more frequently depending on changing weather conditions.  I just always count on verifying the altimeter reading with a known point on a topo map whenever it is convenient.

As far as GPS-based elevation readings, those can vary a fair amount depending on the orientation of the satellites the unit is accessing.  Sometimes the elevations will change +/- 15 ft while standing in the same spot.  I really like pressure-based altimeters for the information they convey about approaching storms etc.

It is interesting to read that somehow the GPS elevation  can try to interface with and “educate” the pressure altimeter.  Seems like that would assume the information the watch calculated from the satellites is accurate, which might not always be the case.

 

[Paul Wagner]

^^  This.  I’ve used a couple of different barometric pressure watches over the past fifteen years.  They do require recalibration more than once a day—I have always assumed it’s because of temperature, as well as pressure changes.  But I have also never counted on the information from that watch alone to make a navigation decision based on a difference of ten feet, or even fifty feet of elevation.

It’s just one tool, to be combined with maps, my own eyesight, a compass, etc.

[Bonzo]

After a bit of pseudo-scientific investigation and data-tracking, it looks like the normal deviation is settling down a bit.  Right now, the altimeter is about 15′ off from a known elevation, and has been all day; I’m going to keep tabs on that over the next few hours as a rain squall moves through and changes the pressure.  I’ve also done a benchmark calibration against the known 1,000-ft. elevation point, and the auto-adjustment zeroed the watch to 1001 feet: I can certainly live with that.  I’ll drive back to that area in the morning and see what I find; hopefully the reading will be close.

Also: I very much agree that this tool is one among many.  Even if it was perfectly accurate each and every time, I think that it’s prudent to never totally rely on a single instrument or method.

[Rex Sanders]

GPS latitude-longitude accuracy, particularly with WAAS, can be 0.9 meters. Elevation accuracy is a whole different problem. Due to one-sided triangulation distortions, it’s easy to get multi-meter height errors that drift over time, as satellites appear, move, and disappear. In addition to many other GPS problems.

My challenge is often this: GPS says one thing, ABC watch another, map another, and my own lying eyes yet another. Which to believe? They all have problems that vary by time and location.

— Rex

[Bonzo]

Great article, Rex; thanks for posting the link and taking the time to compile it.  I’ve noticed that buildings and other structures are often somewhat confusing for any sort of GPS-based navigation; I had awful problems for an hour or so in Berlin, once, until I just gave up on the navigation app and started looking at street signs and asking for directions.

Still logging data, today; I’m hoping to figure out a good, reliable calibration point for my local area.  If that fails, I’ll just head up into the hills and look for a few known, posted elevation markers and see what I get.  Should be an interesting time, one way or another!

[Bob Kerner]

Hit or miss but definitely calibrate. I have a Coros. Some days it’s spot-on, others it’s not at all.

I recently went for a bike ride with it, a Wahoo bike computer and an Apple Watch. All 3 showed different data! Similarly, I’ve had data variations between the Coros, Apple Watch and Gaia on my smart phone. And yet other times all 3 will  exactly agree. I wouldn’t bet my life on any of them.

[Bonzo]

Interesting effect this morning: last night the altimeter at my location read 1006 ft.  Today, it’s reading 960 ft. with no calibration…but the barometer has been falling for the last four hours.

[Roger Caffin]

Bear in mind that most ‘altimeters’ are just barometers with a different scale. As the weather changes the air pressure goes up and down, and so the altitude seems to go up and down – while you are sitting there.

The only exceptions to this are satellite radar readings, from orbit.

Cheers

[Steve S]

Roger, Bonzo seems to have barometer-based altimeter watch, with a GPS autocorrection factor controlled by an unknown algorithm. In other words, a faith-based navigation device.

[Roger Caffin]

a faith-based navigation device
I like it!
I do use an altimeter in alpine areas: it tells me how many hundreds of metres I have yet to climb …

Cheers

[Bonzo]

…a faith-based navigation device.

I’m not a faith-based person by any stretch of the imagination, but I might make an exception for this instance, because that’s funny.

[Matthew / BPL]

I thought a GPS enabled ABC watch would be more accurate than they appear to be. I have used two different brands and neither has been as accurate as I hoped. They are, however, much more accurate than the non-GPS ABC watch I have used.

My $.02 is that you should recalibrate a couple times a day if you want an accurate read of altitude. I like knowing how far through the climb I am and I think altitude is useful for locating oneself on maps.

Calibrate at the trailhead, then recalibrate at a lake or pass during the day. It doesn’t take long and I’m pretty sure the GPS model is going to be more accurate in between calibrations.

[Bonzo]

Matthew, that’s a good summary of my feelings; maybe not as moment-to-moment accurate as I had hoped, but pretty decent nonetheless.  Now that I’ve had a few days to work with the thing, I’ve begun to see that it’s actually a bit better than I thought.  Being inside with no GPS signal seems to mess with it at times, but calibrating with decent lines of sight to the sky around – i.e. few/no buildings – seems to zero it pretty well: usually within a meter or two.  I took it on three trail runs over the last three days, all along the same path: it recorded elevation gain/loss within 1 foot each time, over each of the three runs…so, six near-identical values in total (note: this was not a pre-recorded path, and Snap-to-Route is disabled anyway).   Coincidentally, I have been calibrating twice a day – morning and sometime in early afternoon – and that seems to be the ticket to keeping the altimeter happy.

[Lowell k]

What happens to your feet gained count when your correct the altimeter of a GPS ABC watch in the field? I just purchased a Solar Instinct and am learning how to use it. When using the altimeter gain feature what happens when it says I gained 2k feet and then I recalibrate it when I’m at a known elevation?

I will try this Saturday, but just curious.

[Bonzo]

Hmm… interesting question.  Theoretically the value should be slightly corrected if there’s any correction made to the altimeter during calibration.  I’ll try it this weekend and see what happens; I have a nearby trail that would be ideal for it.  I can calibrate at the trailhead, start recording an activity, go up, and then stop the activity.  Recalibrate on top of the hill, start a second activity, go down, and then see what I get on the exact same path in the opposite direction.  It won’t be a scientific result, but it will at least start to answer the question…and I’ll be at known elevations, so I can track the accuracy throughout.

[Author]

Posts