There's some interesting complications to the idea of an underlying Cartesian ECEF coordinate system as physical reality. For one thing, gravitational distorti

There's some interesting complications to the idea of an underlying Cartesian ECEF coordinate system as physical reality. For one thing, gravitational distortion (Earth tide) changes the shape of the solid part of the Earth by ~50cm twice a day. This would be a huge unwanted effect when doing high precision surveys, but it's removed when producing coordinates in the international terrestrial reference frame. If you have a nearby reference station which shares the common periodic motion this is relatively easy. If not, the software needs to be aware of the periodic motion and remove it explicitly. I'm constantly amazed that the whole system can be made to work at such high precision - sub 10cm from a moving aerial survey platform with decent gear. There's rather a lot of correction factors to get just right when pinning a consistent coordinate system on a squishy rotating blob like the Earth.