Correct tobs in read_gsb to indicate that the obs are sensible.

[ClaraDraper-NOAA]

Description

Minor edits to correctly identify windborne and saildrone T obs as sensible, not virtual.

Resolves #993
Resolves #996

Type of change

Please delete options that are not relevant.

• [x ] Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Testing underway

Checklist

• [ x] My code follows the style guidelines of this project
• [x ] I have performed a self-review of my own code
• [ x] I have commented my code, particularly in hard-to-understand areas
• [ x] New and existing tests pass with my changes
• [x ] Any dependent changes have been merged and published

[ClaraDraper-NOAA]

Draft while I run the tests.

[ClaraDraper-NOAA]

Passes the global c-tests - which do not assimilate the windborne or saildrone obs (files aren't included in the test directories). Converting to ready for review.

Let me know if you'd like me to do more testing.

[ClaraDraper-NOAA]

The figure shows obs (x-axis) v mod (y_axis) for the original code (in which the obs are understood to be Tv) and this PR (sensible T). C384 example, based on Cory's JEDI g_w test case. For the 285 obs (small sample!) above 275K the stdev of the O-F increases from 0.85 K to 0.91 K. @ADCollard - any chance that the obs are Tv, and have the wrong descriptor?

[RussTreadon-NOAA]

@CatherineThomas-NOAA : Do we need the changes from this PR in GFS v17?

[ADCollard]

@ClaraDraper-NOAA The increase in standard deviation is concerning (and it looks like the bias is also increased). Let me take the time to double check this.

Looking at the stats from a previous run of Windborne there is a negative O-B bias in the lower troposphere. If B changes from sensible to virtual, it will increase and make things worse!

Also, we have a small bug in Saildrone processing (#993) is it worth bundling the fix in with this?

[CatherineThomas-NOAA]

@RussTreadon-NOAA - Yes, any changes that will be eventually going into ops also need to get into our dev/gfs.v17 branch and the real time. It sounds like #993 is a candidate as well. That said, I'd prefer the additional scrutiny that @ADCollard mentioned before incorporating any new changes.

[ClaraDraper-NOAA]

@ClaraDraper-NOAA The increase in standard deviation is concerning (and it looks like the bias is also increased). Let me take the time to double check this.

Looking at the stats from a previous run of Windborne there is a negative O-B bias in the lower troposphere. If B changes from sensible to virtual, it will increase and make things worse!

Also, we have a small bug in Saildrone processing (#993) is it worth bundling the fix in with this?

The caveat was that my example was just from the g_w test case I was running, so may not representative.

[ADCollard]

My logic was wrong in the above comment, if we make this change B goes from Tv to Ts, so O-B will increase. This should improve the O-Bs at low altitudes. I ran 31 GSI cycles and the summary stats are below. "Clara" indeed improves over the control (called "newobs" here).

[ADCollard]

Thanks for finding this bug!

[RussTreadon-NOAA]

Approve

[ADCollard]

Is it too late to push the changes needed for #993 ?

[RussTreadon-NOAA]

@ADCollard : @CatherineThomas-NOAA also mentioned folding #993 into this PR. We can add #993, rerun ctests, approve, and merge.

[ADCollard]

@ADCollard : @CatherineThomas-NOAA also mentioned folding #993 into this PR. We can add #993, rerun ctests, approve, and merge.

As I cannot push to Clara's fork, please see ClaraDraper-NOAA#1

[RussTreadon-NOAA]

WCOSS2 ctests

Install fix/gsb_tvfg at 407bba6 as updat and develop at 005343a as contrl. Run GSI ctests with the following results

Test project /lfs/h2/emc/da/noscrub/russ.treadon/git/gsi/pr997/build
    Start 1: global_4denvar
    Start 2: rtma
    Start 3: rrfs_3denvar_rdasens
    Start 4: hafs_4denvar_glbens
    Start 5: hafs_3denvar_hybens
    Start 6: global_enkf
1/6 Test #3: rrfs_3denvar_rdasens .............   Passed  1590.22 sec
2/6 Test #5: hafs_3denvar_hybens ..............   Passed  2201.75 sec
3/6 Test #4: hafs_4denvar_glbens ..............   Passed  2446.36 sec
4/6 Test #6: global_enkf ......................   Passed  2456.57 sec
5/6 Test #2: rtma .............................   Passed  3039.18 sec
6/6 Test #1: global_4denvar ...................   Passed  3202.78 sec

100% tests passed, 0 tests failed out of 6

Total Test time (real) = 3202.89 sec

[RussTreadon-NOAA]

WCOSS2 ctests pass. Approve.

[RussTreadon-NOAA]

@ADCollard and @CatherineThomas-NOAA : We can move this PR forward with your reviews and approvals.

[RussTreadon-NOAA]

Will work with GSI review team to schedule this PR for merger today. I assume we want these changes in GFS v17.