Replace CRTM static file with something consistent w/ JEDI's use

[rtodling]

This is aimed at minimizing the differences between GSI and JEDI. I already have these "newer" files available (20-surf types). This is now a matter of teaching GSI to read them and have a little comparison experiment where these newer fields are using instead of the 13-types we've been using.

[rtodling]

Just to illustrate to effect of this. Here is the Jo(o-b) between a run w/ the thus-far used 13 surface types (right) and another using the 20 sfc types used in, say, JEDI. Both runs here are obviously GSI runs.

and here is the final Jo(o-a) comparing the same two analyses:

The differences from the perspective of the cost function are somewhat small - but not negligible.

[rtodling]

In an initial analysis (before cycling) the only field that shows any noticeable difference in the increment, as compared to an original experiment (nveg = 13) is Tskin - even in this, the differences are very minor. The figure show the T-skin increment in the control exp (nveg = 13; top) and the difference of the experimenting using 20 types w/ that control (bottom).

If I were to show similar figures for Tv and other fields, no differences would show in the bottom plot - at the scales used in the plot - the temperature differences due to this change are on the order of tenth of a degree Kelvin.

[rtodling]

@mjkagnes123 I consider these differences insignificant. However, I do want to make this change in the GSI analysis since it reduces further any possible differences w/ JEDI.