Something that happened today (see Part 2) reminded me when I first got a GPS. This was 2008 and it was a hand-held Garmin which cost a wee bit ($280 USD, inc $50 USD shipping). Arguably, most of the function I used then is now available in a $4.99 phone app.
So the first time I used it for navigation testing and route logging was in the Dandenongs. I had a paper map with a grid on it. My planned route (pale red, below) was roughly N-E; Edgar Track, Bills Tk, Camelia Tk.
Yet the GPS was telling me I was hundreds of metres away from where I should be. I could expect it to be 1 or 2 meters, but hundreds? If it wasn’t for the signs, I would have missed the turn offs.
Later, when I loaded the GPS log into their map program on the PC, I saw the log was ‘correct’ but displaced, in both directions. The image below is a very rough recreation, showing the walk I really did in pale red and the GPS log in red.
My first reaction was “my new GPS from America is faulty!” But a bit of reading taught me something: the paper map had the same co-ordinate system, but a different datum than the GPS.
A datum is “a system which allows the location of latitudes and longitudes (and heights) to be identified onto the surface of the Earth – i.e. onto the surface of a ’round’ object.” Source.
There are different ways mathematically of doing this, so lots of different datums. I discovered that my map had an old Australian one (AGD66) but the GPS was set up to use the new Australian one; GDA94
The difference is about 200 metres to the north-east. Which was exactly what I saw. I’ve never been tripped up like this again. In fact Google Earth uses an international datum virtually identical to GDA94, so that makes things nice and consistent. As I said earlier, this is sort of related to what happened today, as covered in Part 2.