Altitude

[milanuk]

Hello,

 

I've seen a number of posts on various boards about how either the GPS derived altitude is considered very inaccurate, or the sensors in the GPSr's w/ the compass/barometer/altimeter functions aren't that accurate, etc. Some places (usually backpacking/hiking oriented folks focused on weight savings) almost consider a GPS to be an 'extra' compared to a map, compass, and altimeter to locate where they are on a map.

 

I'm a little confused by some of this. I don't yet have a GPSr (still shopping), but I've been reading thru 'GPS Made Easy', a book that was recommended to me along the way. In there it cites the horizontal accuracy of normal/WAAS GPS as being 15m/3m, respectively, and the vertical accuracy as being 19m/6m respectively. I realize that's not *as* accurate, but it seems pretty dang close to me unless I'm missing something here. The region where I go traipsing about is kind of rugged (eastern Cascades, central Washington) and I've heard that getting enough satellites in view to get a 3-D lock can be a problem in some locations due to steep canyon walls, dense foliage over head, etc. Assuming that you *can* get a 3-D lock... how is 6m vertical not accurate enough for most uses?

 

Recently I purchased a dedicated pocket weather station (Kestrel 4000) as I wanted it for another hobby of mine. It reports wind speed, temperature, humidity... as well as altitude and barometric pressure. Seems fairly accurate... once it's calibrated, as in set to a reference, which require me to either be at a known altitude marker or somewhere where I can call or otherwise check and get an accurate station pressure, which is somewhat of a PITA.

 

How often does a GPS derived altitude not coincide w/ a known point on a map, or a marker of some sort? How much are they usually off? Is it w/i the 19/6m listed above for regular/WAAS units?

 

Barring weather changes, how closely does the altitude computed by a GPSr track w/ that from either the sensors onboard the unit, or that of a dedicated tool like the Kestrel that I have? Is the computed altitude from a GPS accurate enough to use for setting the reference for a barometer sensor?

 

Do the GPSr's that have mapping, as in detailed 24k or even 50k maps, have any sort of provision for determining altitude based on where you are on the horizontally?

 

Sorry for all the questions; I'm just envisioning myself standing someplace looking at a map, a GPS, and my Kestrel, all giving me different information and me trying to figure out which one to trust.

 

TIA,

 

Monte

Edited December 2, 2005 by milanuk

[peter]

My fairly long-term experience has been that when comparing my GPS altitude at known spots I get an accuracy within 35' at least 95% of the time without WAAS and WAAS improves that to within about 20'. But I look at the satellite page and make sure that I actually have a 3D lock and also make sure that if I only have 4 satellite signals the geometry is reasonably good (i.e. not all clustered together or in a straight line. Yes, this isn't quite as accurate as the horizontal position due to the inherent geometry consideration of never being able to receive satellites from below us; but I find it to be quite usable in conjunction with topographic maps and to be the most reliable means of determining my altitude without surveying equipment. Barometric-based sensors can do better for short time periods but can vary considerably over longer times when weather conditions are changing.

 

The Garmin GPS models that include barometric sensors also have a provision where they use the two methods together to give a better result than either method alone. The pressure method is more stable and sensitive to small changes but is subject to long-term variations due to weather. OTOH, the GPS altitude is subject to more random variation, drops out sometimes due to bad reception, but is unaffected by weather and is very accurate when averaged over longer time periods.

So to get the best of both, the unit uses a time-weighted average of the valid GPS readings to continuously recalibrate the barometric sensor. The time constant appears to be about 30 minutes which is long enough to average over the GPS fluctuations but is fairly short relative to most weather changes.

[stevesisti]

I don't see anywhere near those accuracies....I spent a fair amount of time using the altimeter, and checking it with the GPS derived altitude....seems theres always a fairly big difference.

 

When used in autocalibrate mode, the altimeter relies on the GPS altitude to correct it. Keeping that in mind, it's not unusual for me to get GPS derived altitudes that vary by 50 ft on a 200 ft altitude. And I'm not talking about a 1/2 hour time gap between readings....I'm talking about 10 sec. This is also not in a "slot canyon" I was in, or deep in a ravine...this is on top of a mountain under heavy tree foliage. I've done it several times and had the same result.

 

Maybe the figures you quoted are what they advertise, but I highly doubt you will see that accuracy.

 

Maybe more important is do you really need that much accuracy in an altitude measurement. I was initially annoyed because the readings seemed to fall far short of what was published, but after thinking about it, just how accurate a reading do you really need. Personally I find the contour lines on my "homebrew" topo to be more accurate without the calibration fuss. My experience has been at "relatively" low altitudes, and it's entirely possible the units are more accurate (% wise) at higher elevations.

[peter]

Maybe the figures you quoted are what they advertise, but I highly doubt you will see that accuracy.

...

My experience has been at "relatively" low altitudes, and it's entirely possible the units are more accurate (% wise) at higher elevations.

I have no idea what is "advertised" - what I reported on are my observations based on using the altitude readings of my GPS receivers on an almost daily basis over the last 6 years and comparing them to known values where that was possible - primarily when at surveyed summits, at other USGS survey markers, along lakeshores that are dam-regulated, and when at sea level. But, as stated before, I excluded measurements when the satellite page showed a particularly poor satellite geometry. I'd note that your variations of up to 50' in altitude are not inconsistent with a variation of plus or minus 25' from the real value so they don't seem out of line with the accuracy that I reported.

 

Since GPS measurements are based on determining distances to satellites that are about 12000 miles away and using a coordinate system with an origin at the center of the earth, there's no reason why altitudes near sea level should be any more accurate than altitudes on the top of Everest. An error of 25' would be equally likely in either situation.

 

BTW, an additional factor to consider is the accuracy of the geoid-ellipsoid table used by the GPS manufacturer to translate altitudes based on the WGS-84 ellipsoid shape of the earth to the altitude above mean sea level. When I first started monitoring altitudes with GPS I noticed a systematic error of about 16' in addition to the random fluctuations. Checking the USGS site for the geoid-ellipsoid correction for my area showed that Garmin's value was off by 5 m thus accounting for the discrepancy. You can check the value used by your GPS receiver by looking at the NMEA GGA sentence.

[+EScout]

I have not paid much attention to this until I got an eXplorist 500 earlier this year with the Topo 3d mapping software loaded into it. I have done tests at many sites, benchmarks, etc. I compare the altitude shown by the GPS to that shown on the map page, and I bring the waypoint home and put it on National Geo Topo on my PC. For benchmarks, I use the listed altitudes, when they are accurate adjusted ones.

 

I have been amazed with the accuracy of the GPS altitude (derived from satelites.) It is always within 25 feet, very often within 15 feet of the other sources, often within 10 feet and many times within 5 feet. My location is in So Calif, near the coast at altitudes of 300 to 1200 feet, with a clear view of the sky with typically 7 to 9 sats received (and often 11 to 14 received, including both WAAS.)

 

The key is the location, and how many satellites are received. The quality of the explorist receiver and the topo 3d software also helps.

[stevesisti]

I have no idea what is "advertised" - what I reported on are my observations based on using the altitude readings of my GPS receivers on an almost daily basis over the last 6 years and comparing them to known values where that was possible

I wasn't referring to your comment regarding quoted accuracies, but rather the original poster...... If I was referring to your comment, I would have replied to your post.

 

Also note that my comment regarding possible different accuracies was not referring to the GPS derived altitude...but rather the Pressure sensor contribution to the final altitude. All sensors, pressure, temp etc...do not respond in a straight linear fashion...especially not near the top or bottom of their recommended range....not only is it likely, but it's extremely probable that the Pressure sensor is more accurate in the middle of it's intended range and not near the top or bottom.

 

Lastly...If one has to , as you say, start discerning whether or not they have a "good lock" a "not so good lock", or a "bad lock" , and\or factor in geoid elipsoid tables for the particular area they are in to get accurate results,Geeeez..... why bother with the thing in the first place. Bottom line is the altimeter is a rough estimate of what your altitude is......being off by 25-30% is "rough" to me........to you it might be hair splitting accuracy.

[+EraSeek]

Not a horribly educated responds here but I carry a barometric altimeter and of course my GPS. I rely on my GPS reading more than my altimeter because of the rapid barometric changes around here (western Washington). I find the GPS more reliable and just as accurate. The times when I have compared my GPS reading to a benchmark it has been within around 20' of the stated hieght.

 

It appears from the following article that newer GPS's will give a more accurate altitude and has loads of interesting info (not all of which I yet understand):

"However, most users expect accurate elevation readings that are related to MSL. Consequently, newer GPS devices output orthometric (geoid) height measurements as a product of "behind the scenes" calculations based on a combination of formulas, tables, and matrices that use geographic coordinates as inputs. "

 

Mean Sea level, GPS's , and the Geoid

Edited December 3, 2005 by EraSeek

[peter]

Consequently, newer GPS devices output orthometric (geoid) height measurements

It's not so much the age of the unit as the manufacturer's design decisions. For example, even the oldest eTrex reports geoid heights and this is also true of Magellans of that vintage. AIRC, the old Lowrance GM100 did not correct for the geoid-ellipsoid, but they have corrected that since then. OTOH, some current GPS devices, primarily ones for PDAs and notebooks, still report the uncorrected ellipsoid height.