Total Column water vapor calculation for microwave radiance data QC in all-sky assimilations

[jianjunj]

The total column water vapor (tcwv) is incorrectly calculated in crtm_interface.f90.
It is an integrated values of atmosphere(1)%absorber(k,1) at
https://github.com/NOAA-EMC/GSI/blob/develop/src/gsi/crtm_interface.f90#L2154
if (present(tcwv)) tcwv = tcwv + (atmosphere(1)%absorber(k,1)*0.001_r_kind)*c6(k)

where atmosphere(1)%absorber(k,1) is a mixing ratio of vapor to dry air mass
atmosphere(1)%absorber(k,1) = q(kk2)*c3(kk2)

q(k) is specific humidity and c3(k) is a converter to convert specific humidity to mixing ratio.
c3(k)=r1000/(one-q(k))

The total column water vapor should be a integration of specific q(k), and https://github.com/NOAA-EMC/GSI/blob/develop/src/gsi/crtm_interface.f90#L2154 should be replaced by
if (present(tcwv)) tcwv = tcwv + q(kk2)*c6(k)

In UFO, tcwv is calculated correctly by transforming mixing water vapor mixing ratios:
https://github.com/JCSDA-internal/ufo/blob/develop/src/ufo/filters/obsfunctions/TotalColumnVaporGuess.cc#L53

This issue only impacts the cold-air-outbreak QC for microwave radiance data assimilated in all-sky conditions.