N arrow accuracy in GE?

[crass]

Hello all. I have 3 or 4 examples where I have imported a Google Earth aerial and surface to Sketchup and also overlayed a scan of the site's actual survey and there is considerable discrepancy in the angle of the North Arrow (around 10-15 degrees) when I line up the images. My form of georeferencing is crude (ie using SU transform tools) so I would accept a minor discrepancy, but this is way too much. This also happens to me in AutoCAD Civil 3D so I feel that it is a Google Earth issue, or less likely an issue with the surveys. I assume both a survey and GE would use true north so declination would not be a factor here. Before I post examples, I was wondering if anyone else had this issue. Obviously accuracy in the orientation of one's model is critical for solar design. I just don't know what to trust. Any insight would be greatly appreciated. Thanks

[pyroluna]

er... I don't suppose your model is so close to the north pole that it might be a difference between the magnetic pole and the geographic north pole?
See http://en.wikipedia.org/wiki/File:IGRF_2000_magnetic_declination.gif

[Gaieus]

When I served in the army as an artillery scout, we indeed had to take magnetic deviation (declination) into account (well, it is obviously not indifferent where you shoot with a 152 mm cannon over a 25 km distance of course).

I really doubt that such a small deviation would have any serious affect on how a model is placed on GE. Also, it changes from place to place and not even only up to the geographic location relative to the geographical and magnetic North but a bunch of other things can affect it.

Is it really worth the hassle? What project do you have where it would be so important?