Incorrect Initial Lithostatic Pressure boundary traction when Y is negative

[ricitron]

While using the Initial Lithostatic Pressure boundary traction condition I notice different results if I set the base of the domain at Y = 0 or at Y = -200km.

For a plate under flexure from a disk load at the surface, if I place the base of the domain at Y = 0, surface at Y = 200km, I get the following:

If I instead set the base of the domain at Y = -200km, surface at Y=0, then the lithostatic pressure condition on the right boundary appears to not work, allowing excessive outflow from the right boundary causing the model to sink as material outflows:

Any ideas what causes the lithostatic pressure to give a different boundary condition based on where I put the surface? I modified the Representative point for the lithostatic pressure condition appropriately (see input files below).

Input file for surface at 200km:
plate_surf200km.txt
Input file for surface at 0km:
plate_surf0km.txt