New GPS buyers are frequently concerned about the accuracy (or lackof it) of the altitude readout on their newly purchased GPS. Many suspecttheir equipment may even be defective when they see the altitude readoutat a fixed point vary by many hundreds of feet. This is NORMAL.
With most low cost GPS receivers, the horizontal error (without SA nowthat it is off) is specified to be within about +/- 15 meters (50 feet)95% of the time. Most users find this is a conservative specification andthat their modern GPS receivers routinely perform better than this worstcase specification. But.. Users should expect that SOMETIMES they may seethe error approach the specification limits. AND.. 5%of the time, the error may be "any value" from zero to whatever".Note: Unless you have a CLEAR AND UNOBSTRUCTED view of the sky (on yourdash or looking out of an airplane window with no externally mounted antenna,or similar obstructed view does not count!) you can count on yourerror excursions to be much greater than the above numbers.Your GPS <depends> on this clear and unobstructed view or it cannotmake accurate range measurements to the satellites.
Generally, Altitude error is specified to be 1.5 x Horizontalerror specification. This means that the user of standard consumerGPS receivers should consider +/-23meters (75ft) with a DOP of 1 for 95%confidence. Altitude error is always considerably worse than thehorizontal (position error). Much of this is a matter of geometry. If we(simplistically) consider just four satellites, the "optimum" configurationfor best overall accuracy is having the four SVs at 40 to 55 degrees abovethe horizon and one (for instance) in each general direction N, E, W, andS. (Note: You will get a very BAD DOP if the SVs are at theexact same elevation. Luckily, this is a rare occurrence.)See: DOPdemonstration site by Norris Weimer> How SV geometry affects GPS accuracy(JavaRequired).. The similar "best" arrangement for vertical positionis with one SV overhead and the others at the horizon and 120 degrees inazimuth apart. Obviously, this arrangement is very poor from a signal standpoint.As a result, of this geometry the calculated solution for altitude is notas accurate as it is for horizontal position. Almost any calibratedaltimeter will be more stable at reading altitude than a GPS.
GPS altitude measures the users' distance from the center of the SVsorbits. These measurements are referenced to geodetic altitude or ellipsoidalaltitude in some GPS equipment. Garmin and most equipment manufacturersutilize a mathematical model in the GPS software which roughly approximatesthe geodetic model of the earth and reference altitude to this model. Aswith any model, there will be errors as the earth is not a simple mathematicalshape to represent. What this means is that if you are walking onthe seashore, and see your altitude as -15 meters, you shouldnot be concerned. First, the geodetic model of the earth canhave much more than this amount of error at any specific point and Second,you have the GPS error itself to add in. As a result of this combinederror, I am not surprised to be at the seashore and see -40 metererrors in some spots.
DGPS operation (where available) will dramatically improve the performanceof even low cost GPS receivers. Horizontal accuracy of +/- 5 meters andaltitude accuracy of +/- 10 meters (relative to the WGS-84 geode) withsuitable DGPS receivers and low cost GPS receivers such as the Garmin GPS-12XLcan be expected.
In any case, it is extremely unwise to overly depend on the altitudereadout of a GPS. Those who use GPS altitude to aid in landing their smallplane should have their insurance policies paid up at all times.
Joe Mehaffey