The highest resolution is 1 meter (but they don't have that data for the entire US).
The crappiness of GPS based total elevation gain is pretty well understood, so they probably wouldn't use that. So as long as they used data (or a topo map) instead of GPS, you can assume that the sampling frequency was not ridiculous (and it probably only includes gain that is apparent over something like 10 or 30 feet).
http://ned.usgs.gov/
The highest resolution is 1 meter (but they don't have that data for the entire US).
The crappiness of GPS based total elevation gain is pretty well understood, so they probably wouldn't use that. So as long as they used data (or a topo map) instead of GPS, you can assume that the sampling frequency was not ridiculous (and it probably only includes gain that is apparent over something like 10 or 30 feet).