Here's the warp vector illustration of deflection for case 1 (Very coarse mesh, 1st Order).
The resulting deflections for the different mesh fineness is shown below.
Interestingly, the use of second order mesh significantly improved the accuracy of the coarse mesh to near 100%, and had little benefit for the other finer meshes. I am suspicious of this result and would like to know why this was the case and why I was unable to carry out simulations on the fine 2nd order meshes due to memory allocation errors - even when increasing the number of cores. I suspect that I got something wrong with this.