Make x0050 default experiment for geos_atmosphere
#501
Conversation
|
@rtodling 3dfgat doesn't have enough resources for C180. Would decreasing iteration be sufficient or should we modify |
Dooruk
changed the title
geos_atmosphere
|
|
@rtodling there may be issues with the |
|
@Dooruk, about the x50_ufo, when I manually OOPS_STATS Run end - Runtime: 242.57 sec, Memory: total: 130.30 Gb, per task: min = 3.59 Gb, max = 3.82 Gb |
With this branch? Weird.. Let me trigger the actions then |
|
I didnot use tier-1 test auto script, I manually
|
|
I realized |
|
@metdyn it fails (I tested both locally and |
Dooruk
merged commit fe3ad63
into
feature/msienkie/aircraft_bias_NC4
|
Thanks. I reproduced your results: I find my previous run get the wrong results, because I source a wrong module that set the $PATH for swell to another branch.
|
…e of nc4 files to use with updated bias correction (#478) * new configuration with NC4 bias file for aircraft bias coefficients - modified aircraft.yaml to use the new file, and modified gsi_bc_to_ioda.py to generate the NC4 coeff. file Working on get_observations.py to get the NC4 aircraft bias coefficient file instead of csv * change indentation of 'obs bias' section (it's not a part of obs space) * some modifications - not yet running * Try 'satbias' type in get_observations.py (?) additions for gsi_bc_to_ioda.py - use an aircraft 'sensor' * remove old 'csv' processing that was left in the missing_obs_test file change aircraftAscentRate to instantaneousAltitudeRate to conform with recently changed nomenclature * Added missing line for appending info on output to yaml Fixed typo: add 'predictors' not 'default predictors' to output * Needed to update the predictor coefficient name to match the new naming convention. Use 'instantaneousAltitudeRate' and 'instantaneousAltitudeRate_order_2' * Remove reference to a non-existent file that was pasted into the bias section * Corrections for python coding norms (indentation, long lines, too many blank lines) * Adjust indentation * Add indentation on two lines for coding standard. (Hope this works) * try increase indent again * tweak indentation again * new aircraft bias format and extension * fixes to provide _cov file separately for aircraft * fix test * reset location of alternative - bkg - files * split aircraft T/W and adjust yamls - including commenting filter not working at this time * needs these too - sorry * add temp and wind for aircraft but intentinally left aircraft alone * eva seems to work now * output files for aircraft varBC * minor fix * updates for handling 20231010 00z case * Make x0050 default experiment for `geos_atmosphere` (#501) * fix tier1 overrides * aircraft is two separate entries now * codestyle * arrange defaults and overrides for tier1 tests * make ufo_testing work for x0050 * deactivate ocean for hofx temporarily * c91 as tier1 test, adapt to geos experiment folder style * hofx has to use x0050 data * add geovals for ufo_testing --------- Co-authored-by: gmao-msienkie <[email protected]> Co-authored-by: dooruk <[email protected]> Co-authored-by: Doruk Ardağ <[email protected]> Co-authored-by: Ricardo Todling <[email protected]>
This PR combined with #478 makes x0050 as default for
3dvar_atmosand3dfgat_atmoswhile increasing the horizontal resolution toc180.Currently, x0050 observations were converted only for October 10th, 2023.
FYI @gmao-wgu @gmao-yzhu