FWIW this looks steeper than 22 degrees.
I can't help but wonder about contributions from both sides feeding the rear end. Could that support a higher angle?
I also think about how much less steep is the ridge of a hip roof compared to the sides. Point being that flow vectors with some transverse component might be considered effectively shallower with their azimuth taken into account. When working in any field (electric, magnetic, etc) cosine projections into a principle plane usually offer great advantage but maybe there are cases where a fresh or even naive look reveals something previously missed. (shrug)
Cheers
KB