Elevation error in Cyclemeter
 

I rode a time trial this past weekend that was multiple laps around a 10-km course. I was using the Cyclemeter app to track my performance. Cyclemeter will use the barometric sensor on my iPhone, but I was on the course long enough that barometric pressure changed. It's interesting to see how this affected the altitude readings. Multiple laps of the same route make it clear. This screenshot is taken from Cyclometer's web-based "explorer."

I knew beforehand that the course has 97 meters of climbing per lap, and I could see as I was riding it that even over the course of a single lap, it wasn't recording my altitude very accurately. This is not a big surprise.

Ride with GPS can correct the altitude of a track after the fact—I uploaded this route to RwGPS, applied altitude correction, and it looks correct there now. AFAICT, Cyclemeter cannot perform post-hoc altitude corrections.


https://cimg4.ibsrv.net/gimg/bikefor...46cbaa6e02.png
Altitude in Cyclemeter

How much did the barometer fall during your ride? 0.1 inches of Hg is 100 feet of altitude.

Here's the weather almanac for the first day (race went noon on the 5th to noon on the 6th). Not taken from a weather station in the park, but near enough. You can see it went from 1035 mb at noon to about 1025 mb at midnight. 10 millibar is about .29 inches of mercury, which would imply about 300 feet of altitude change. Not too far off.

Of course, my screenshot has distance in miles on the X axis, not time, but I would have been at about 173 miles at midnight.

One of my Garmin GPS units has a pressure sensor, but the unit attempts to self calibrate the elevation by using elevation data from GPS.

I am not familiar with the app you used and my Android phone does not have a pressure sensor. But I am surprised that whomever wrote the software for that app did not try to incorporate a self calibration based on GPS elevation data.

Since your X axis is distance, I am surprised there is no big jump late in the graph when you stopped to rest for several hours, as I would have expected the barometric pressure to continue to change.

I have a couple old Casio watches that are over a couple decades old that have no GPS electronics, thus the pressure sensor in them has nothing to calibrate to. The plot you have is similar to what I would expect from my uncalibrated Casio watches.

The problem there is that GPS-derived elevation is also unreliable. I think the only way to get an honest elevation profile is to reference topographic data.

I was surprised by that too.

I can turn off barometric altitude sensing in the app, and for rides that will go through major weather changes, seems like there's no reason not to.

It use to be on Garmin devices from circa 2010 with barometers, that they'd only show the uncorrected elevation given by the barometric sensor. If an elevation correction waypoint was set, and you happen to be near that waypoint at the start of your ride, they'd adjust everything based on that elevation. However it was a one time correction and only if found during the first few minutes of the ride.

Don't know what they do today. I don't use them for actual elevation. They do still give as decent as any estimates of elevation gain/loss if one isn't a bean counter and can live with the inaccuracies that occur on some rides.

I'd just ignore the elevation data from a ride you know is suspect. If you've done other rides of the same, then you'll have an idea what is more correct.

I calibrate my barometer manually, from a topo or known height like at a saddle or summit. Garmin can auto calibrate them once at the start of an activity from GPS or continually nudge it towards agreement, but 99% of the time it works very well.

When knowing my elevation is important, I have a data screen on my Garmin that shows the altitude according to both sources. And I can work it out from a topographic map.