Initial modifications for standalone Icepack bgc capability

[apcraig]

These are some initial changes to support standalone Icepack with the updated bgc.

• Add some checks to avoid divides by zeros
• Update icepack driver bgc calls
• Turn back on a "z_tracers" test in icepack_zbgc.F90, otherwise get an array bounds error.

The code still has an array bounds issue with skl bgc that I will continue to work on.

[njeffery]

@apcraig : I believe I've sorted through the debug compile errors. Essentially, I didn't put all the bgc parameter array definitions in init_zbgc. It's corrected here in updates to my branch: https://github.com/E3SM-Project/Icepack/tree/f459b0c5de3d39b40e181ea2641ce4ac3039d329

I also added your columnphysics changes above into my branch with a couple of variations. Instead of querying in the ice_algae routine, I put some bounds on the parameters when they're initialized. They shouldn't be zero or negative. Let me know if you'd prefer that in another routine. I'm testing the updates to icepack in E3SM now.

[apcraig]

@njeffery, I'll give your modifications a test. Could you also merge in the changes in icedrv_init_column. Those allow your branch to be tested in icepack standalone mode.

It might be more efficient if you want "most" of the changes I PR to accept the PR then to create a new commit that updates that PR. Pulling bits and pieces here and there feels like we end up with more commits and more iterations and also fixing the same thing multiple times. Thanks.

Let me know when you've updated icedrv_init_column and I'll check out the latest and retest. Thanks.

[njeffery]

Merged!