It seems to me I remembered as a kid that the topographic maps near where I lived were something like 3 degrees W declination, so I figured I'd check the current status of Indiana since I know the north magnetic pole kinda moves around. I got two totally different readings; 1 was -5 degrees which would be five degrees West, and the other was positive 10 degrees east...! Both appeared to be current data.
Then I realized that somehow when I looked for 'Indiana', one of the searches was giving me results for a town named 'Indiana' in Pennsylvania...
What struck me was I had figured the 'west' declination for Indiana meant that the magnetic north was somewhere west (relative to Indiana) of the rotational axis, sort of northern Alaska or something. Yet Pennsylvania data makes it look like more towards Greenland. Then if you go out to Wyoming it's also east, which at least sort of 'intersects' with the Indiana declination.
I guess the deal is that big mountain ranges affect it because of all the material there, But even that seems funny if you think about the fact that a mountain range is massive and maybe averages a mile higher than the surrounding terrain, but it's on a Planet with a diameter of 8000 miles or so, and a 'crust' that likely isn't smooth on the inside boundary either.
I saw one map that showed a bunch of wiggly lines representing declination. It was just a map of the central United States but the lines were certainly very wavy, probably relating to geologic features. I couldn't help but wonder if you got above the surface of the Earth a few dozen or 100 miles if the magnetic field then becomes much more uniform and geometric.
Supposedly the poles will soon (in geologic time) and rather abruptly (dunno how 'geologic' they mean by 'abrupt') switch north-to-south. THAT should be kinda interesting. Humans (if still around) will figure it out no doubt (and Al Gore's descendants will no doubt make money fear-mongering and offering 'solutions'...