Fix incorrect double reversal of topological line sub-segment normals.

The topo line's sub-segment geometries are already getting reversed
(when necessary) which already reverses its great circle arc normals.

gplately/ptt/subduction_convergence.py

@@ -422,24 +422,21 @@ def subduction_convergence(
                     )
 
                     sub_sub_segment_geometry = sub_sub_segment.get_resolved_geometry()
-                    sub_sub_segment_trench_normal_reversal = trench_normal_reversal
                     # If sub-sub-segment was reversed when it contributed to the topological line shared sub-segment then
                     # we need to use that reversed geometry so that it has the same order of points as the topological line.
                     if sub_sub_segment.was_geometry_reversed_in_topology():
                         # Create a new sub-sub-segment polyline with points in reverse order.
                         sub_sub_segment_geometry = pygplates.PolylineOnSphere(
                             sub_sub_segment_geometry[::-1]
                         )
-                        #  The trench normal direction is also reversed.
-                        sub_sub_segment_trench_normal_reversal = -trench_normal_reversal
 
                     _sub_segment_subduction_convergence(
                         output_data,
                         time,
                         sub_sub_segment_geometry,
                         trench_plate_id,
                         subducting_plate_id,
-                        sub_sub_segment_trench_normal_reversal,
+                        trench_normal_reversal,
                         trench_length_radians,
                         distance_along_trench_radians,
                         threshold_sampling_distance_radians,