Add PlanetiQ YAM8 SAID=768 RO data in GSI v16#930
Conversation
|
@RussTreadon-NOAA Could you please add @ADCollard and @emilyhcliu as the reviewers? Thank you! |
RussTreadon-NOAA
requested review from
ADCollard,
RussTreadon-NOAA and
emilyhcliu
|
@XuanliLi-NOAA: It's my understanding that this PR does not need to be merged into |
|
@XuanliLi-NOAA : What is the status of this draft PR? |
|
@RussTreadon-NOAA We have completed two 2-week parallel experiments — nopiq and allpiq. The results confirm that temperature and height biases against radiosonde data increase with the number of RO observations assimilated. Therefore, @danholdaway suggested turning on the PlanetiQ YAM-8 data later (together with the MSG data), and @ADCollard confirmed this will likely happen sometime early next year. In the meantime, we are running sensitivity experiments with different k1 coefficients of the obs operator. Preliminary results show a reduction in the biases, and we hope this approach may help address the forecast bias issue. |
|
@XuanliLi-NOAA : Thank you for the update. Given what you shared does it make sense to close this draft PR and open a new PR at a later date? |
|
Yes, that would be fine. Thank you! |
|
Close this PR as development continue. A new PR will be opened at a later date. |
|
@RussTreadon-NOAA and @ADCollard The ctest for develop-v16 on WCOSS2 failed for global_enkf: The Do you have any suggestions on how to solve this? Thank you! |
|
Your Your in the Your run fails because it's trying to read in data for these 2m fields from the surface file but the surface file is not present. This is the error message your quote GSI develop-v16 has fix @ 3c98350. The I can't find your branch |
|
@RussTreadon-NOAA Thank you so much for finding that! I'll use the correct gsi-fix and see how it goes. |
|
Thank you @RussTreadon-NOAA! I changed the branch of fix from develop to 3c98350. Now all 6 tests passed. I'll create a new PR. |
2 participants
Description
This PR addresses issue #928 by adding satellite ID 768 commercial PlanetiQ YAM8 RO data to GSI v16. setupbend.f90 has been updated to apply the commercial RO QC filters for this data, and the print statement for the number of RO obs considered and accepted has been turned off. Note that the fix submodule in this fork is the develop branch, and it works fine for the GSI sample data test runs.
Resolves #928.
Type of change
Please delete options that are not relevant.
How Has This Been Tested?
A GSI test was conducted on WCOSS2 using sample PlanetiQ YAM8 files with SAID = 768 provided by UCAR. Additional testing was performed with files containing PlanetiQ data from both SAID 267 (GNOMES satellites) and 768. Both tests successfully assimilated the YAM8 data, and the statistics in fort.212 look good for PlanetiQ.
A GSI regression test against the develop-v16 branch (262ce16) was run on WCOSS2. All tests passed.
Test project /lfs/h2/emc/da/noscrub/xuanli.li/git/v16_piqyam8_768/GSI/buildStart 1: global_4denvarStart 2: rtmaStart 3: rrfs_3denvar_rdasensStart 4: hafs_4denvar_glbensStart 5: hafs_3denvar_hybensStart 6: global_enkf1/6 Test #3: rrfs_3denvar_rdasens ............. Passed 736.16 sec2/6 Test #2: rtma ............................. Passed 1157.18 sec3/6 Test #5: hafs_3denvar_hybens .............. Passed 1285.03 sec4/6 Test #4: hafs_4denvar_glbens .............. Passed 1534.61 sec5/6 Test #6: global_enkf ...................... Passed 1536.60 sec6/6 Test #1: global_4denvar ................... Passed 1985.11 sec100% tests passed, 0 tests failed out of 6Total Test time (real) = 1985.12 secChecklist