Hello, I have a suite (u-dv923) where I’d like to customize the timesteps when the output is written. I’d like to keep the standard hourly output, but also add a few customized times (down to the second) for output, is this possible? I’d like to compare the output to EarthCARE cloud profiling radar, and the satellite tracks are at very specific times and not on the hour.
Best,
Michelle
Hi Michelle,
The physical timestep is 20 mins and the chemical timestep is an hour by default so I’m not sure output every second would work.
The VISION interpolator (GMD - Peer review - Virtual Integration of Satellite and In-situ Observation Networks (VISION) v1.0: In-Situ Observations Simulator (ISO_simulator)) may be of use.
Cheers,
James
Hi Michelle, James,
To add on VISION, version 2 of the package that has been re-written using cf-python is available here
A satellite simulator is under development, but has not quite been finalised yet. This is currently being finalised but works well and there are examples for flight campaigns here:
As James said, the model timestep is 20 minutes for most climate configurations, so outputting less than this isn’t possible. Also, you would need global output to be able to cover the swath at the neessary local times (the model uses UTC) - we generally output instantaneous hourly global output for interpolation onto the satellite points.
Best wishes,
Luke
Hello, an update - my colleague @siddharth.gumber gave me an example where he wrote output at custom times to compare with satellite observations. Following his example, I went to um > namelist > Model Input and Output > STASH Requests and Profiles > Time Profiles, cloned ‘T1HR’ and created a new set of timesteps for output, with the times of interest in seconds since the model start time. I also cloned some of the output variables and changed their time index to the new timesteps of interest.
However, when I try to run now, I get this error:
???
???!!!???!!!???!!!???!!!???!!! ERROR ???!!!???!!!???!!!???!!!???!!!
? Error code: 40
? Error from routine: READ_NML_ITEMS
? Error message: Duplicate Stashcode entries for Stashcode 413 in items namelists.
? Please check entries.
? Error from processor: 0
? Error number: 1
???
I can’t find a variable with the stashcode 413 that I am requesting for output, nor have I located this code in any of the time/domain profiles. I was wondering if anyone has come across this before?
Hi Michelle,
Can you commit your suite so we can take a look on the repository?
James
Sure thing James! I created a new suite yesterday and replicated this ^ issue but with fewer changes made, so I think that might be the best one to examine. The suite is u-dw064 and I just committed the changes.
I can’t see anything in app/um/rose-app.conf but I wonder if one or more of the conf files in the opt/ subdirectories are being called and overwriting what is in app/um/rose-app.conf . Have you looked at any of those?
413 appears to refer to sea ice categories. Do you know if you are prescribing this in your suite?
1| 1 | 0 | 413 |Sea ice concentration by categories |
2| 2 | 0 | 1 | 1 | 5 | -1 | -1 | 10 | 1 | 10 | 0 |
3| 000000000000700000000000000000 | 00000000000000000001 | 3 |
4| 1 | 0 | -99 -99 -99 -99 -99 -99 -99 -99 -99 -99 |
5| 0 | 1411 | 0 | 129 | 0 | 0 | 0 | 9999 | 0 |
James
Hi James, thank you for checking - I went into the suite and under glm_um > namelist < Reconfiguration and Ancillary control > Configure ancils and initialize dump fields, I was able to locate the duplicate sea ice entry. I am using the default sea ice forcing, and it looked like the two sea ice entries were identical, so I tried removing one and rerunning. Glm_um_recon succeeded in the run this time, we’ll see how the rest of the run goes. Thanks for flagging this as sea ice!
Great, well done, fingers crossed for the rest of the run.
James
Hi, an update on this - I was able to get the suite to run. However, I can’t seem to find the output for the cloned variables with the custom timestep (for both u-dv923, which has the full set of variables I’d like to output at custom times, and u-dw064, which has one variable “theta after timestep” cloned to keep it simpler). I’m wondering if there are any other output or time settings I need to adjust in order for cloned variables with custom timesteps to appear in the output?
Best,
Michelle
Hi Michelle,
Can you run tidy stash transform (rose macro stash_indices.TidyStashTransform in terminal in suite/app/um/) as I can see you have some stash entries in u-dw064 without a proper code, for example:
[namelist:umstash_streq(4)]
dom_name=‘TH_1_120’
isec=0
item=4
package=‘’
tim_name=‘TEC’
use_name=‘th’
Following your use_name = ‘th’ through, I would expect output to be in a file starting de064a_th based on the setup below.
[namelist:umstash_use(th_5f650b55)]
file_id=‘nc2’
locn=3
!!macrotag=0
use_name=‘th’
[namelist:nlstcall_nc(nc2)]
file_id=‘nc2’
filename_base=‘$DATAM/${RUNID}a_th%N.nc’
l_compress=.false.
!!l_shuffle=.true.
!!nccomp_level=1
!!packing=0
reinit_end=-1
reinit_start=0
reinit_step=36
reinit_unit=1
Has the model completed its atmos_main (or equivalent) and failed on postproc?
Cheers,
James