Issue634 pen sw

[einola]

This issue links back to the neXtSIM_DG issue #320:

There are errors in the neXtSIM Lagrangian implementation of the Winton (2000) and Semtner (1976) thermodynamics models.

• In the Winton model, I use "I = I_0*Qsw" without taking into account that I_0 is 0 with (thick enough) snow cover
• In the Semtner model, I don't apply the gamma fudge factor, and only the beta fudge factor through the albedo scheme - but not the one we usually use.

Both errors stem from improper treatment of the penetration of short-wave radiation into the ice, so a fix to both fits together.

These are being addressed in neXtSIM Lagrangian by adding I (the penetrating short-wave radiation) as an output to IABulkFluxes and using this in the Winton and Semtner models. A similar approach should be taken in neXtSIM_DG.

This pull request is the fix in neXtSIM Lagrangian. I've tested the neXtSIM_DG version against Winton's results, and @timspainNERSC has compared neXtSIM_DG and neXtSIM Lagrangian outputs.

[einola]

Asking @timspainNERSC to review, as this is basically a re-implementation of what I did in neXtSIM_DG ... and to be sure that his comparison of the two models gave ok results.