Originally Posted by
CommuteCommando
Actually what is happening is the barometric pressure was rising. If you look at the elevation track it looks like an angled saw blade. The low point is right after he turns onto Lincoln from Jefferson. If you look to the north of that you will see a creek/drainage channel that confirms that the terrain is sloping down in that direction. The difference between there and the opposite corner, Polk and Adams, appears to be about 15-20 ft. If you look at the graph of speed against elevation, the general shape of the elevation profile is confirmed.
Right, assuming this is an actual difference in barometric pressure, rising pressure would correspond to a lower elevation detection by the altimeter in his Garmin. I'm not so sure that's what's really going on though; check out the "segment" he created for a one mile loop:
http://www.strava.com/activities/93277888#1967799801
The start and end points are the same, yet his data shows they're 12 feet different in elevation. I could figure out how much of a pressure drop that would require, but I doubt that it happened in the 2-3 minutes it took him to ride that loop. The barometric swing though would explain why the elevation trend is consistently down over the whole ride, and not randomly fluctuating.
My conclusion is that the altimeters used in most bike computers just aren't that good at producing consistent repeatable readings... that is I think their users assume they're more precise than they really are.
Now I've got dozens of Strava records up the local mountain from my Android-app, and they vary by 1% or so over 3800 feet of elevation change... it's hard to expect much more precision than that.
So altimeters and GPS-map data are more/less accurate/precise than each other in different situations.
