All measurements have error, but not all errors are the same. Measurement errors may be:
- Totally random (noise), in which case averaging measurements reduces error.
- Totally systematic (bias); if the error is known, it can be corrected for (for example, through calibration)
- Anything in between: this includes drift or periodic effects, or effects which are correlated on short time scales but random over long time scales (such as regular calibration based on a reference measurement which itself has random error)
Over on Outdoors.SE we were having some discussion about how "accurate" the altitude measurement of a consumer-grade GNSS receiver may be. Sources make imprecise statements such as The general rule of the thumb is that vertical error is three times the horizontal error., the theory and practice that GPS altitude at 10 to 20 meters (35 to 70 feet) error is superior to Pressure Altitude, or that barometric altimeters are shaped correctly but may have the wrong baseline.
Assuming good reception without prior knowledge, between the three sources of error mentioned above, which one is larger for GPS altitude measurements and which one for barometric altitude measurements?
My impression is that GPS altitude measurements have a smaller systematic error but barometric altitude measurements may have a smaller random error.
For the sake of this question, please ignore any altitude corrections based on mapping data or other forms of assistance.