So how does the automatic recalibrating work? How frequently did Garmin decide to have it recalibrate itself? Do you know when and how much it is recalibrating? Doesn't this make calculations of total ascent/descent worthless? Does the instrument know that a 20 feet positive recalibration (while you are climbing) is different from your climbing 20 feet? If the GPS altitude is always assumed to be the primary data, then what was the logic for including a barometric altimeter in the first place?
I don't mean to be argumentative, I just don't understand the logic of Garmin's design. If the GPS altitude values are so good now (and almost never lose satellite coverage with the new chipset), what good is the barometric altimeter?
The GPS measured position of a stationary GPS will wander around on a fairly short time scale, (eg
http://edu-observatory.org/gps/images/single.gif) but the long term average will be significantly more accurate. (The figure is for the horizontal position, but the same happens for the GPS altitude measurement.) Thus, any particular GPS altitude measurement may have a fairly large error.
The nominal horizontal error of a GPS is about 95% probability of being less than 10 meters. The vertical error of a GPS is 2-3 times the horizontal, so the vertical error would be 95% probability of being within 20-30 meters. These numbers assume the GPS is receiving good signals--if it is receiving poor signals or some satellites are blocked, the errors could be [much] larger.
In contrast, the error of a barometric altimeter varies slowly with changes in sea-level pressure caused by changes in the weather, but is good at following short-term changes caused by a hiker changing altitude. If the calibration was poor or there has been a significant sea-level atmospheric pressure change since the last calibration the error can be large. Also, barometric altimeters assume a standard temperature and humidity profile in the air and if the actual profiles are significantly different from the standard profile additional errors will be introduced.
The auto-calibration uses the long-term average of the GPS to calibrate the barometric altimeter. The calibration is performed as a continuous slow process so it follows (sea-level) atmospheric pressure changes. The GPS also follows ones changes in altitude and thus the calibration process is compensated for any changes in actual altitude.
Mathematically:
* Basic barometric altimeter calibration:
est_baro_alt = raw_baro_alt + baro_cal
baro_error = est_baro_alt - actual_alt
When you calibrate a barometric altimeter, you adjust the baro_cal to make the est_baro_alt the same as your known altitude from an outside source (such as a topo map). ie
baro_cal = baro_cal - baro_error
* GPS auto-calibration of barometric altimeter:
perform once per second:
est_baro_alt = raw_baro_alt + baro_cal
est_baro_error = est_baro_alt - gps_alt
baro_cal = baro_cal - cal_speed*est_baro_error
cal_speed (typ = .001 to .00001) is the calibration speed: larger makes the calibration faster but ultimately less accurate (because it averages over a shorter time) and smaller makes the calibration slower but ultimately more accurate (because it averages over a longer time). The speed, of course, has to be fast enough to follow the expected changes in the sea-level pressure.
The above process automatically compensates for any changes of altitude during the auto-calibration process.
Garmin GPSes have been observed to auto-calibrate faster for the first half hour or so (for faster initial calibration) and then slower for better ultimate accuracy.
It is a 2-3 hr drive to the Whites for me. I simply turn the GPS on during the drive and it is adequately auto-calibrated by the time I reach the trail head. When I check the barometric altitude at the summit, it is usually with 10 ft of the altitude given on my topo map.
I intentionally don't reset my Suunto barometric altimeter during the day to maintain the accuracy of the ascent/descent data, and I know at the end of the hike how much the pressure/altitude change was (and can see why the total ascent/descent are different, if they are, because of pressure changes).
Calibrating to a known altitude at the beginning of the day and checking against a (potentially different) known altitude at the end of the day is a manual method of obtaining the est_baro_error accumulated over the day*. You would then reset your altimeter to the new known altitude and repeat the next day. The GPS auto-calibration of the barometric altimeter does this continuously throughout the day. The final result is that your indicated altitude is more accurate than you would obtain using either the GPS altitude or the barometric altitude alone.
* This is used for weather prediction: if the est_baro_error is >0, the sea-level pressure is decreasing and the weather is likely to get worse or stay bad and if the est_baro_error is <0, the sea-level pressure is increasing and the weather is likely to improve or stay good. Climbers often check for pressure changes while sleeping overnight since they are at a fixed (even if unknown) altitude.
Doug
