Here is a lovely example of an arch that has distorted enough to need a detailed survey. It was almost certainly originally three centred. An ellipse certainly doesn't come near to fitting. I suspect that the pier to the left has settled and rotated a little to the right. The arch has then fleattened to the left, but there is some evidence in the bed joints in the spandrel at the right of the arch having been compressed and so forced up at the right hand haunch.
However, it is certain that nothing short of a proper survey, with perhaps 20 or 30 points in the span, will give any hope of modelling the arch effectively. Even then one should note that the joints in the left half of the span are nearly all open at the bottom giving a very much reduced ability to carry live load.
So here we have a bridge that should not be assessed except as a guide to how it might perform once properly repaired.