Models
------
All these 3D radial anisotropic models have been obtained using the same dataset of fundamental and overtones data, the same 3D radial anisotropic starting model and NACT calculations for the forward model. Different regularizations have though been used in the inversion.

1) all.crustn.SX.mixed_4_5_doe2.f1h2.0: standard norm damping.
2) all.crustn.SX.mixed_4_5_doe2.f1h2.0.edge100: as 1) but with more damping on the knots at the edge of the region.
3) all.crustn.SX.mixed_4_5_doe2.f1h2.0.edge100.sm0.1: as 2) but with some horizontal smoothing.
4) all.crustn.SX.mixed_4_5_doe2.f1h2.0.edge100.sm0.2: as 3) but with more horizontal smoothing.
5) all.crustn.SX.mixed_4_5_doe2.f1h2.0.edge1000: as 2) but with more damping on the edge knots.
6) all.crustn.SX.mixed_4_5_doe2.f1h2.1.edge100.sm0.1: this model has been obtained in a second iteration starting from model 3).

Script to extract and plot the model: anis_plot_spline_doe.s
------------------------------------------------------------
Inputs on the command line:
1) Parameter: S for isotropic S-velocity, X for radial anisotropy
2) Model name
3) Identifier for the file with the horizontal grid. In your case 4_5_doe2 for the isotropic velocity structure and 4 for radial anisotropy.

The script is calling few executable, which you need to compile on your machine together with two libraries. The source codes (grid_mA3d_anis_doe.c and rmean.f) are in src, the libraries in lib. I used the makefiles grid_mA3d_anis_doe.ms, iolib.ms and mathlib.ms to compile. I have tried to make sure that there is no hardwired path in the source codes, but if you find out that you miss anything, please let me know.