If you have a line angled above the horizontal, pointing upwards, it can't end up below the horizontal.
You need to define your horizontals, though. What IS the horizontal to which you refer?
One hour after sunset, the photographer is 360/24 = 15 degrees beyond the Earth terminator line, at an unspecified latitude, with the camera inclined upward by 45 degrees. What does the horizontal of the camera frame then represent, other than itself?
It's certainly not comparable to the line of illumination connecting the Moon and Sun. That's a different 'horizontal'. The camera frame is inclined on one axis by some 15 degrees relative to this, and by unspecified angles in the other axes.
Bottom line is that we could model this in 3D, but we need to know where the photo was taken from, i.e. the latitude of the photographer. We know his 'longitude', that's 15 degrees beyond the terminator, but we don't know his latitude. It would also help to know the date of the photo, to determine how far off the sun/earth plane the moon was at the time.
You're regarding the line connecting moon and sun as being 'above the horizontal', but all that horizontal represents is a line in the camera frame, non-parallel to any of the lines connecting sun, moon and earth.