Clear-sky masking in COSP-RTTOV

[jshaw35]

Existing clear-sky outputs in COSP-RTTOV follow standard model procedures of repeating radiative transfer calculations without cloud water and ice. This creates a definitional mismatch with clear-sky observations where clouds are actually absent. Due to this mismatch, spectral regions with vapor water absorption will likely inherit a high variance bias and low clear-sky radiance bias due to the inclusion of cloudy scenes with saturated levels (e.g. Figure 1, bottom row here).

This PR brings additional functionality into COSP-RTTOV. Specifically, it allows users to produce an additional "clear-sky" field conditioned on the absence of clouds (all model levels in any column must have <10% cloud fraction). An additional output variable containing the number of counts these new variables accumulate (for tests of statistical significance) is also included. New outputs are triggered using the namelist option "Lrttov_maskclear". New outputs do not replace the standard clear-sky fields and have an additional "m" in the variable name to indicate clear-sky masking.

I am opening this PR now, but also need to ensure that these new outputs are included in CI/CD and clean up changes I have made to namelists. I may need a refresher on that since I am struggling to remember how COSP-RTTOV is currently triggered in CI/CD 🫤

[RobertPincus]

@jshaw35 Maybe you want to make this as a draft PR?