I'm looking at this SPICE kernel, which contains orientation data for planets and other bodies. Model description and examples in "Orientation models" section look pretty simple. But actual numbers make me confused.
Firstly, it says that Earth rotation along its axis (prime meridian angle) is 190.147 degrees at J2000 epoch, and I don't get it. From my understanding this angle must be between the equinox vector and prime meridian which is right in front of the sun at J2000. On 21 of December the north pole is turned directly away from the sun, so at J2000 it is about the same since J2000 is the 1st of January, the difference is about 10 degrees. So if the north pole looks away from the sun, the equinox vector should be about parallel to the Earth orbit, which is about perpendicular to the sun direction. So the prime meridian angle should be not 190 degrees, but more like 280 degrees. Why is it like so?
I even checked it in Cosmographia, a tool downloaded from SPICE website, and this visualisation just confirms my thoughts.
Secondly, the numbers say that the declination of the Earth rotation axis starts from 90 degrees and decreases by 0.557 degrees every century. Which means that in about 16000 years Earth rotation axis will lay in ecliptic plane and it doesn't make any sense. Is it because this model should be applied on short time scales only?