Logging Tour Vert Geek-ery Questions

So here's a question about device accuracy when measuring vert gained/lost during a tour and the elevation accuracy of tour planning tools in general (i.e. Nat Geo, etc.)

So did this nice tour from the top of Chair 8 at Baker (started elevation log here) out the Arm beyond the drop in for the Hourglass (lots of up and down travel), skied south into the "East Fork" of Swift Creek (for lack of a better name). Skied three more runs in the Ann vicinity (great snow btw), climbed back up to the Arm and skied the right side of the Safety Line back to Chair 8 (stopped log back at WS parking lot).

Pretty simple.

My Sunnto Vector logged the tour at 5140' of gain.

BC Nav app on Samsung Galaxy 4 logged the same tour at 4344' of gain.

Exporting the gpx track from BC Nav and importing into Nat Geo (older version) - the elevation profile of the track in Nat Geo shows 4630' of elevation gain.

Pretty big difference I think. Which one is right-er?

This isn't the first time I've had some elevation gain discrepancies of this size between watch and phone gps/mapping app, but those tours had way more elevation gain/loss and much larger variations in elevation (like the Watson Trav).

This tour is much smaller, is in a fairly narrow elevation band (less error in watch altimeter?), and has a shorter duration - 6:29:21 to be precise versus the +12 hour day for the Watson (also less time for changes in air pressure due to atmospheric change).

Please help, my fragile ego is hanging in the balance . . .

Under open skies I would probably trust the GPS over the barometric altimeter unless you were calibrating it regularly.

I have the Suunto Ambit2 which uses both GPS and barometric altimeter to determine altitude. I imagine it calibrates using the GPS when the signal is good and falls back to the barometric altimeter when it goes poor. In my experience it is quite accurate.

I've wondered about recalibrating, but then thought that recalibrating in middle of recording an elevation log would artificially add gain or loss to the log since you are manually changing the elevation.

Another thing is the Sunnto always records greater elevation gain. If it were an issue of elevation I would think it would be off either direction (either more or less gain recorded).

The fact the watch always records greater gain might also just be a function of how the watch always 'decalibrates' over the course of a tour.

Finally, I've noticed the watch doesn't necessarily decalibrate but loses accuracy in elevation with a gain in elevation but then regains accuracy as I return to the starting point.

For instance - on a summit ski of Baker, if I set the watch at 3700' at the Glacier Creek TH it will be hundreds of feet off at the summit, but then be right back at 3700' upon return to the TH. I wish I had kept track of which direction the watch is off on the summit . . . .

Would be stoked to hear about how phone GPS elevation compares with handheld GPS elevation. Meanwhile I'll just ask the oracle and see what comes up . . .

GPS elevation is much less accurate than lat/long position.

However, some phone systems will correct for elevation after the fact, since if you know lat/long accurately, and have a good elevation database, you can figure out what your elevation was.

I use RideWithGPS (designed for cycling), and upload the days tracks. It corrects for elevation after the fact, and I consider it to be a pretty accurate record of the vertical feet. It doesn't count runs or other fancy stuff but works for my purposes.

Not that snow depth is an issue right now but later in the season there could be 40' of snow in specific places. I don't think the GPS data can correct for that since it is using historical measurements presumably based on the conditions in which the measurement was taken unless the device is trying to triangulate in 3 dimension in which case I've heard the same thing about GPS having poor accuracy for altitude.

I think the issue that you're having is probably related more to the measurement resolution (frequency and precision) rather than accuracy of the devices. I recently bought an InReach and don't have much experience with it yet but it only reports position every 10 minutes which basically cut the reported mileage of a recent trail run in half since it missed a lot of switch backs in the trail.

OP: why don't you post the .gpx so we can see what's in there? If you pm me I can post a link here to it. How many points are in the file? Are elevations listed in the .gpx? If you overlay the .gpx on the map, does the drawn path actually go where you went? As others have said if the gps is recording your position infrequently you might miss some elevation if you travelled both up and down between recorded points.

Micah - that's a really good idea. I'll break open that gpx file later today and then post it up when I get a chance.

Jason - I hear you on the frequency and position vs. accuracy.  From what you are saying regarding snow heights, a modern gps with map installed is cross referencing it's position with the listed height of that position on a map?

On Baker tour yesterday, with  minimal up and down, my friend's phone GPS came in at ~6700 feet of gain when there was just no way that seemed possible. The Sunnto logged it at 6130' which seemed way more inline with TH and turnaround elevations (3100'to 9100' with a few short up and down's thrown in).

On a final note, when I worked as a land survey tech we would use GPS for recording physical location (like on a property boundary survey) but never used GPS for elevations (as in setting cut/fill depths for an excavation contractor). The tolerances for that kind of work were super tight - error allowances of 0.05' or five-hundreths of a foot so to speak - GPS was not that accurate in elevation. I never considered that relevant for this topic but maybe has more relevance than once thought.

Thanks for the input to all who replied

I have the same problem and I believe it will not be solved soon. I assume that my vertical data is off by 10% either way.

I am not familiar with the software you are using, I use Google Earth, but it is very likely the elevation data comes from the same source: Space Shuttle SRTM data, which offers 30m vertical accuracy at 90m horizontal grid spacing. Elevation data is quite expensive to collect and the SRTM data is the one that is most accessible so it is safe to assume most of the mapping apps and programs are using this (they sometimes mention this in the About section or in the user manual).

There are more precise elevation models, it is widely accepted the military has much better data for at substantial parts of the world (where the "bad" guys live). There are claims that Google Earth is using this in some parts of the world but I have not seen it.

Anyway the 30m over 90m accuracy is pretty good in flatlands but can and does produce substantial errors in the mountains, it will easily miss 600 ft up or down along a jagged ridges that abound around here.

For backcountry use I'd argue that in therms of vertical accuracy altimeter is still the best (especially the ones with recording interval of 1 m), followed by readings obtained form a GPS track in programs like Google Earth (i.e using the SRTM data), followed by smartphones using signal triangulation and GPS, followed by GPS/smartphones calculating the vertical from just the GPS signal.

Anyway if you want to go super geeky this is a good place to start:
eros.usgs.gov/elevation-products

Also this is what Garmin says about GPS vertical accuracy (applies to GPS chipsets in smartphones as well):
support.garmin.com/support/searchSupport...dc-4733-000000000000}

Have fun!