Garmin edge 305 -Height climbed discrepancy

I doubt any of them are correct, whatever that may be. So my advice would be to ignore the absolute figure, but instead pick one of the programmes and use it to compare the acent figure for your rides.

careful wrote:

I think Garmin use a barometer (air pressure measurement) to measure elevation. If I'm right then they are bound to vary a bit with time of day & weather.

The 305 does have a Barometer but this is to smooth the altitude measurement provided by the WGS84 geoid which is used by the GPS system to provide your position and altitude. In order to have a GPS position to a high enough resolution to give altitude, you need to be locked onto at least 4 satelites. On the Garmin this is shown by the 3-D satelite symbol rather than the blacked out one which means you have 2-D position and the outline one which means you have no position. The WGS84 geoid is a mathematical model and there are areas where it is way out especially in the height. From a previous job I seem to remember a place where the altitude was wrong by 217 feet which when we were looking at flying down to 100' above the sea was abviously a problem.

As I understand it with the 305 and other Garmin devices with a barometer, the theory is when the unit acquires a 3-D GPS fix the unit takes this height and calibrates the local air pressure against the GPS height. Then as you ride it is the change in air pressure that is being used to calculate ascent/descent as this is more accurate. Through the day, as Careful said, weather fronts will move across and local air pressure will change. However as the 305 is constantly callibrating itself against what the unit thinks the barometric height is and the GPS height everything should be smoothed out.

However, theory is not like practice and the Garmin tends to over-read both ascent and descent and as has been posted already, a loop ride always starts and ends at slightly different altitudes. The problem is further complicated by the software that you use on your computer (Training Centre, Garmin Connect and in my case, on a Mac, Ascent). The programmers know about these inaccuracies and try and filter them out even further. Each programme has a different filter to remove what it see as erroneous data (sudden spikes in altitude change) and that results in what you are seeing. Ascent even has a feature where the user can define the sensitivity of the altitude filtering, but how the user can make use of this with any accuracy is beyond me.