Accuracy of Montana 700 Topo graphics

The Garmin GPSr uses a combination of mapping data, atmospheric pressure, and GPS coordinates to determine current elevation based on many factors - which is why you may need to calibrate it from time to time. Best method is to find a known (published) location elevation, and calibrate the device to that value at those coordinates - which is done from the setup menu on the device.

Makes sense as this is on an island reclaimed from the sea. Should I set it to 0 feet for calibration. 

If you are holding the GPSr directly in  front of you while standing, and the ground is slightly above sea level, I would consider calibrating between 2 and 3 meters.

After calibrating, it has raised the elevation of N23 22.037, E53 38.104 by 2 ft to 490ft.

It does not match with NASA's SRTM3 v3 dataset  elevation value of 126 meters (415 feet) for the coordinates  (N23 22.037, E53 38.104)

Also I am still seeking help on why

"The Garmin shows me Elevation lines at 429 in this area of dunes and sand. When I put a pin on this line and press for more detailed information, it shows me anything BUT 429 ft. I have tested the feature in other nearby deserts and with a 625 ft elevation line. I met with the same problem."

After calibrating it 2 meters above sea level and standing up, it has raised the elevation of N23 22.037, E53 38.104 by 12 ft to 499ft. This is moving away from the 415ft from the Nasa Data

Sorry11 ft

I was afraid you will say that. "Calibrate the device to that value at those coordinates" - Am I right in presuming you are saying being on the spot  to know its elevation.

Surely this defeats the purpose as I am driving to a location with an elevation of which is unknown. How do I plan ahead for the best route if I don't know the height at the coordinates without being in the location yet.

I determined the elevation, 0 ft. AMSL, of my house with this  source: elevation.maplogs.com/.../huntington_beach_ca_usa.50350.html

WRT using a GPS to determing elevations at other locations, I will answer in the context of "precision" in contrast to "accuracy".  I will offer two methods, one each at both ends of the time spectrum as opposed to an intermediate position.

A decade or so ago on another forum CHAM01 asked for advice on measuring the height of a walkable embankment roughly 50 feet from top to bottom.  I advised using the barometer of the GPS instead of the GPS calculated elevations at both top and bottom and subtract the bottom value from the top after walking down.  Doing this during a calm period, with essentially zero wind speed, the precision of the barometric readings  will be a factor of 6 less than those GPS derived.

When I am going for a period of several days across the Mojave Desert and up into the Sierra Nevada Mountains, I will use the GPS derived elevation values.  The precision of these GPS calculated values is a factor of 13 less than those baromnetrically derived due to weather fluctuations over longer periods of time.

Note: atlas_cached can provide a link to a characterization of precision and accuracy variations.

When I am planning a route for a drive through sand dunes, I can live with accuracy of +/- 30 ft that is +/- 3 storey building sheer drop. But when a sheer drop is 100 ft or 10 storey, I may want to avoid it and take a different route. Topo graphic maps are meant to show elevation and depression contours to help one figure out the most accessible route to a location. However if these contour lines are incorrect by ten stories it can be dangerous driving. Certainly my American brethren would have their lawyers on speed dial if this caused  them to have an accident or injury or lose their vehicle down the bottom of a steep sand dune.

Can you suggest a more accurate device? As it aint Garmin 700 even after calibration.