Differential GPS being one example.
I thought that was switched off (or at least people stopped buying into it) when the Ammericans stopped scrabling the GPS signals?
History lesson as to why GPS is now so much more popular than it was 10 years ago, the USA used to add a deliberate error to the signals so that your position could only be known withing +/- 50m or so. Which was fine for shipping in the middle of the ocean, but less usefull for armies.
There were 2 ways arroud it, a) make friends with Uncle Sam and get the code's to decifer the GPS signals accurately. b) buy a dGPS (differential) set. These worked by a signal on the coast (as most users were on boats) which knew it's position, figures out the error in the signal and broadcaast this to dGPS sets.
About 10 years ago the European Space Agency was going to launch a competitor to GPS without scrambling, the Americans didn't like this as they'd lose control of it so swithed off the scrabling, removing the need for an ESA version, but maintaining their ability to turn it off to various parts of the world. Since they turned off the scrambling there's very little if any difference in accuracy between them.
GPS can work out altitude, but because it's referenced to a mathematical model of the earth it can't, by definition, be accurate anywhere.
Actualy it can, burried in the settings you can pick a map projection (even without a map installed), this is what local maps are drawn in and gives the closest aproximation of a sphere for wherever you are in the world. So as long as you've got the right one selected then (within the accuracy of the signals/calculations) the altitude should be accurate.
IME the GPS claculated total climbing on my edge 305 was bang on when you count contours on a OS map. The barrometric one on my 800 seems to get confused by the weather (gaining/loseing a few hundred feet over a ride) and doesn't pick up small hills. And STRAVA's topo maps are even worse.