Moist-Dry conversions

[jatkinson1000]

This addresses #15

The variable conversion routines now also include a moist - dry conversion and back.

The moist mixing ratios from CAM are now converted to dry mixing ratios as expected by SAM for use in that code, and then back for calculation of tendencies and return values.

[MarionBWeinzierl]

This looks sensible to me. Some more comments might help, in particular what the variables are and what happens, e.g., in the iteration in lines 687-716. Otherwise fine.

[jatkinson1000]

I have added some comments and references to explain the variable conversion code in f4e7198

[paogorman]

I'm a bit worried about the introduction of the r variables for mixing ratios. We have ended up with lines like:
qn = max(0., r_temp-qsat)
Here qn, qsat, and r_temp are all mixing ratios but only one is denoted r.
An alternative approach would be to keep everything as q, and have distinct blocks of code that do the moist to dry mixing ratio conversions.

[MarionBWeinzierl]

I am happy with this now, but will refrain from approving it until @paogorman is happy, too

[jatkinson1000]

I have renamed the variables used by 'SAM' to be r instead of q, and renamed the functions extracted from SAM to use r in their names as well.
If you can review and check you are happy before I merge it that would be appreciated.

[paogorman]

Change to rv etc. generally looks good and does make things clearer.
For CAM_var_conversion:

• might be cleaner to introduce rc,ri explicitly (but what you have looks right)
• are you now missing a declaration for incoming variables qc,qi?

[jatkinson1000]

This should now be ready to go.
@paogorman the changes we discussed are summarised in 15e21cc if you could have a quick glance to check you are happy with that. I also used it to simplify the calculation of r(i,k) as well, as this also uses ri and rc so avoids a double calculation.

[paogorman]

Yes this looks good - thanks Jack