Biased elevation errors early in a trip

[cdimarzio]

Since EveryTrail has started showing elevation tracks, I've noticed an unusual behavior. When I turn on the GPS (Garmin Etrex Legend mounted on a bicycle), the elevation is usually high. This is not an EveryTrail problem, because I see these numbers on the GPS while I am riding. I just never noticed the pattern before.

I started this morning at what I think is 80 feet elevation, with the GPS reporting 150. After a few minutes of flat riding, the elevation came down where I expected it, and my morning commute ended at an elevation of about 30 feet, which I believe is correct for that end. In the afternoon, when I turned it on, the elevation was about 115 feet, and then as I rode, it settled down to what I think is the right number, and when I got home it was reading 80 feet, which I presume is close to correct.

Has anyone else noticed this? Can anyone explain it? One thought is that both ends of the trip are close to buildings, so I'm working with near-zenith satellites. I understand that this means inaccurate elevation, but I'm surprised that it seems to always be high, rather than random.

This is not a big deal, but I'd like to understand the GPS a little better.

--Chuck

[EFletcher]

My Garmin GPSmap 76CS uses a barometric altimeter (presumably in addition to the less-accurate satellite altitude sensing). I noticed that in-car tracks had odd spikes, and determined that the apparent elevation changed whenever the AC came on. Makes sense: the pressure changes! Even opening a window can affect it.

Perhaps your observation is due to something similar.

[prathmann]

The eTrex Legend does not have a pressure sensor so the elevation is based on the GPS data. When first turned on it will start to acquire satellites and needs at least 3 to determine the time and lat/long location (3 satellites needed for the 3 variables). Until it acquires a 4th satellite it has no idea what the actual altitude is and generally uses the value it last stored from a previous use. So the first few points in the tracklog will frequently have elevation data that's way off (and the elevation being off also affects the accuracy of the lat/long position). One way to avoid the problem is to turn it on when you first get to the start of a trip and let it get a good satellite lock while you're getting other things ready. Check to see that it has a good satellite lock and then clear the tracklog so those first few points with the wrong elevation are erased before you start the trip you want to record.

[cdimarzio]

Looking at the records a bit more, I think you are right. The first few points are based on previous history, and there is apparently some "smoothing" that prevents an immediate recovery when the fourth satellite is used.

--Chuck