Use depth information if dip = 90

[lispandfound]

This PR changes the behaviour of plane.wgs_depth_coordinates_to_fault_coordinates, which converts WGS84 coordinates to coordinates in the fault.

Old Behaviour

The old behaviour used the corner coordinates of the plane to find the fault local coordinates, without using the depth information. Ignoring the depth is desirable for two reasons:

1. If dip != 90, then this information is redundant, and removing the depth improves the numerical stability of the linear algebra used to derive the local strike and dip coordinates.
2. This method is often used for finding converting hypocentre locations to fault local coordinates for simulations. Often the depth is not quite on the plane, or is rounded, which results in errors in the approximation of the fault-local point (see the diagram below).

However, it means that the method fails if dip = 90.

New Behaviour

The new behaviour will now account for the depth of the point in two cases:

1. When dip = 90,
2. If the user provides global coordinates with a depth component.

[AndrewRidden-Harper]

If depth information is not passed, should this message also be shown in the terminal?