One of the FE models is that of a disk centered around the North Pole, with "Antarctica" circling the circumference as the Ice Wall.
The FE explanation for sunsets is that the sun gets far enough away such that it appears to sink into the horizon and/or it's light can no longer reach us. Since sunrise and sunsets happen at
very predictable times then this distance needs to be constant. That is, sunsets are hardly dependent on atmospheric conditions if they're dependent on that at all.
However, if you actually map out sunrise and sunsets on this single-poled model, you find that this distance changes dramatically throughout a year. For me it changes by a whole 13.8% of the distance between the north pole and the equator!
Below you can see the to-scale diagram of this. The red is during the summer solstice, the blue during an equinox, and the green during the winter solstice. The dotted lines represent the distance from the sun to me during that time. Notice how they're all different sizes?
Thus, because this model is inconsistent with FE arguments (mainly sunset is caused by distance), it can be safely thrown into the garbage.