Atmospheric JCB prototype file need unified obs list

[DavidNew-NOAA]

No objections here, but this might be a good time to point out to you and @CoryMartin-NOAA @RussTreadon-NOAA that we have 7 different jcb-prototype files with observations specified in them. Thus, the deterministic analysis may use different observations than the ensemble. We can probably make an obs_list.yaml.j2 file that can one list of obs for each prototype file and then just do an {% include obs_list %} in each jcb-prototype file. I'll make an issue for this.

Originally posted by @DavidNew-NOAA in #1883 (comment)