Simulation using time-varying SST boundary condition? #257

AndrewILWilliams · 2023-08-09T20:17:40Z

AndrewILWilliams
Aug 9, 2023

Hi there, I'm trying to set up a simulation where the sea-ice is set to a climatology, but the SSTs are time-varying and taken from the AMIP database. I've tried to copy the style of sst_clim_amip.nc as much as possible (see screenshotted xarray output of my file, below), but when I run the realistic_continents test case using this SST input file I get the following error:

2023-08-09 21:08:36,856 - isca - DEBUG - ATMOS MODEL DOMAIN DECOMPOSITION
2023-08-09 21:08:36,856 - isca - DEBUG - X-AXIS =  128
2023-08-09 21:08:36,856 - isca - DEBUG - Y-AXIS =    2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2   2
2023-08-09 21:08:36,856 - isca - DEBUG - 2   2   2   2   2   2   2   2
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.00000000E-07  fraction_smoothed=  3.32996833E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  2.00000000E-07  fraction_smoothed=  4.27244378E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  5.95506981E-07  fraction_smoothed=  6.24173077E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  9.48633651E-07  fraction_smoothed=  7.08755612E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.32957313E-06  fraction_smoothed=  7.64227308E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.57523410E-06  fraction_smoothed=  7.89588033E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.67609138E-06  fraction_smoothed=  7.98416297E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.69418418E-06  fraction_smoothed=  7.99918309E-01
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine regularize: lambda=  1.69418418E-06  fraction_smoothed=  7.99918309E-01
2023-08-09 21:08:36,856 - isca - DEBUG -
2023-08-09 21:08:36,856 - isca - DEBUG - Message from subroutine get_topography:
2023-08-09 21:08:36,856 - isca - DEBUG - lambda=  1.69418418E-06  fraction_smoothed=  7.99918309E-01
2023-08-09 21:08:36,857 - isca - DEBUG -
2023-08-09 21:08:36,857 - isca - DEBUG - mean surface pressure= 973.2448 mb
2023-08-09 21:08:36,857 - isca - DEBUG - NOTE from PE    0: idealized_moist_phys: Using Betts-Miller convection scheme.
2023-08-09 21:08:36,857 - isca - DEBUG - NOTE from PE    0: Interpolator: Using input file with no time dimension - sst_timeVarying_amip.nc
2023-08-09 21:08:36,857 - isca - DEBUG - Initializing src field : sst_timeVarying_amip
2023-08-09 21:08:36,857 - isca - DEBUG -
2023-08-09 21:08:36,857 - isca - DEBUG - Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
2023-08-09 21:08:36,857 - isca - DEBUG -
2023-08-09 21:08:36,857 - isca - DEBUG - Backtrace for this error:
2023-08-09 21:08:36,857 - isca - DEBUG -
2023-08-09 21:08:36,857 - isca - DEBUG - Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
...
...
2023-08-09 21:08:36,864 - isca - DEBUG - Backtrace for this error:
2023-08-09 21:08:36,864 - isca - DEBUG -
2023-08-09 21:08:36,864 - isca - DEBUG - Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
2023-08-09 21:08:36,864 - isca - DEBUG -
2023-08-09 21:08:36,864 - isca - DEBUG - Backtrace for this error:
2023-08-09 21:08:36,864 - isca - DEBUG - malloc(): invalid size (unsorted)
2023-08-09 21:08:36,864 - isca - DEBUG -
2023-08-09 21:08:36,865 - isca - DEBUG - Program received signal SIGABRT: Process abort signal.

Is there something obvious I'm doing wrong here? I couldn't see much on the docs about setting up time-varying SST runs.

Answered by sit23

Aug 11, 2023

Just as a quick guess, it can sometimes object if there are capital letters in the field name (so you'd have to change both the file and field name to remove the capitals). But if you're able to supply Neil with the experiment script then we can test it properly.

View full answer

AndrewILWilliams · 2023-08-10T15:44:23Z

AndrewILWilliams
Aug 10, 2023
Author

Hi there, I've managed to get passed my initial error by using the scripts in src/extra/python/scripts, but now I am having an issue with the Interpolator:

2023-08-10 16:39:26,342 - isca - DEBUG - mean surface pressure= 973.2448 mb
2023-08-10 16:39:26,375 - isca - DEBUG - NOTE from PE    0: idealized_moist_phys: Using Betts-Miller convection scheme.
2023-08-10 16:39:26,406 - isca - DEBUG - NOTE from PE    0: interpolator_mod: Model and file calendars are the same for file sst_timeVarying_amip_subset.nc; no calendar conversion  needed
2023-08-10 16:39:26,406 - isca - DEBUG - NOTE from PE    0: interpolator_mod :sst_timeVarying_amip_subset.nc is a timeseries file
2023-08-10 16:39:26,408 - isca - DEBUG - Initializing src field : sst_timeVarying_amip
2023-08-10 16:39:26,417 - isca - DEBUG -
2023-08-10 16:39:26,417 - isca - DEBUG - FATAL from PE   23: Interpolator: the field name is not contained in this intepolate_type: sst_timeVarying_amip_subset
2023-08-10 16:39:26,417 - isca - DEBUG -
2023-08-10 16:39:26,417 - isca - DEBUG - MPICH ERROR [Rank 23] [job id 4191488.0] [Thu Aug 10 16:39:26 2023] [nid005734] - Abort(1) (rank 23 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 23
2023-08-10 16:39:26,417 - isca - DEBUG -
2023-08-10 16:39:26,417 - isca - DEBUG - aborting job:

I can do some debugging, but does anyone have any tips for how the 2D interpolator is working here? It looks like the offending section is here (L2359 of interpolator.F90):

do i= 1,size(clim_type%field_name(:))
!++lwh
  if ( field_name == clim_type%field_name(i) ) then
!--lwh

    found_field=.true.

    if(present(clim_units)) then
      call mpp_get_atts(clim_type%field_type(i),units=clim_units)
      clim_units = chomp(clim_units)
    endif

I'm not really sure what clim_type or field_name are though...it seems to be something to do with interpolator_init?

0 replies

sit23 · 2023-08-10T15:54:57Z

sit23
Aug 10, 2023
Maintainer

Hi @AndrewILWilliams - the variable within the SST input file will need to have the same name as the input file itself. It looks like your file is called sst_timeVarying_amip_subset.nc, and so it's looking for the variable called sst_timeVarying_amip_subset. If you make sure that's the name of the variable in your file then that should work.

10 replies

sit23 Aug 10, 2023
Maintainer

It's a silly feature - sometimes when we use input files we've tried to provide namelist options for the file name and the field name, but that's not currently the case for SST. The field name is what's passed to the call interpolator function, which in this case is just sst_file:
https://github.com/execlim/Isca/blob/master/src/atmos_spectral/driver/solo/mixed_layer.F90#L334

AndrewILWilliams Aug 10, 2023
Author

Hmmm, I still get the same error. I've uploaded my SST input file to dropbox, would you be able to have a quick look at it? https://www.dropbox.com/scl/fi/0gyyf956m9be4gxgtf8mx/sst_timeVarying_amip_subset.nc?rlkey=yek900jpr44u940oih491vo47&dl=0

(base) ailw3142@ln04:/work/n02/n02/ailw3142/Isca/exp/test_cases/realistic_continents/input> ncdump -h sst_timeVarying_amip_subset.nc 
netcdf sst_timeVarying_amip_subset {
dimensions:
	lat = 180 ;
	lon = 360 ;
	latb = 181 ;
	lonb = 361 ;
	time = UNLIMITED ; // (61 currently)
variables:
	double latb(latb) ;
		latb:units = "degrees_N" ;
		latb:cartesian_axis = "Y" ;
		latb:long_name = "latitude edges" ;
	double lonb(lonb) ;
		lonb:units = "degrees_E" ;
		lonb:cartesian_axis = "X" ;
		lonb:long_name = "longitude edges" ;
	double lat(lat) ;
		lat:units = "degrees_N" ;
		lat:cartesian_axis = "Y" ;
		lat:long_name = "latitude" ;
		lat:edges = "latb" ;
	double lon(lon) ;
		lon:units = "degrees_E" ;
		lon:cartesian_axis = "X" ;
		lon:long_name = "longitude" ;
		lon:edges = "lonb" ;
	double time(time) ;
		time:units = "days since 0001-01-01 00:00:00.0" ;
		time:calendar = "THIRTY_DAY_MONTHS" ;
		time:calendar_type = "THIRTY_DAY_MONTHS" ;
		time:cartesian_axis = "T" ;
	float sst_timeVarying_amip_subset(time, lat, lon) ;
}

ntlewis Aug 11, 2023
Maintainer

@AndrewILWilliams do you have a copy of the python script you're using to run the model? I can try to run it with your input file and see what happens.

sit23 Aug 11, 2023
Maintainer

Just as a quick guess, it can sometimes object if there are capital letters in the field name (so you'd have to change both the file and field name to remove the capitals). But if you're able to supply Neil with the experiment script then we can test it properly.

Answer selected by AndrewILWilliams

ntlewis Aug 11, 2023
Maintainer

+1 to the capital letter thing, perhaps try that before sending me the script :)

AndrewILWilliams Aug 11, 2023
Author

Woohoo! Looks like capital letters were the crux of the problem - got to love GCMs eh ;)

Thanks both! I'll bug you again if I (when) fall into further disarray...

sit23 Aug 11, 2023
Maintainer

Great news - thanks for letting us know @AndrewILWilliams.

AndrewILWilliams Aug 11, 2023
Author

@sit23 maybe one more quick question, is there a list somewhere of the output variables? Or is that just found in the code (register_diag_field etc)?

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Simulation using time-varying SST boundary condition? #257

{{title}}

Replies: 2 comments 10 replies

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

Select a reply

Simulation using time-varying SST boundary condition? #257

AndrewILWilliams Aug 9, 2023

Replies: 2 comments · 10 replies

AndrewILWilliams Aug 10, 2023 Author

sit23 Aug 10, 2023 Maintainer

sit23 Aug 10, 2023 Maintainer

AndrewILWilliams Aug 10, 2023 Author

ntlewis Aug 11, 2023 Maintainer

sit23 Aug 11, 2023 Maintainer

ntlewis Aug 11, 2023 Maintainer

AndrewILWilliams Aug 11, 2023 Author

sit23 Aug 11, 2023 Maintainer

AndrewILWilliams Aug 11, 2023 Author

AndrewILWilliams
Aug 9, 2023

Replies: 2 comments 10 replies

AndrewILWilliams
Aug 10, 2023
Author

sit23
Aug 10, 2023
Maintainer

sit23 Aug 10, 2023
Maintainer

AndrewILWilliams Aug 10, 2023
Author

ntlewis Aug 11, 2023
Maintainer

sit23 Aug 11, 2023
Maintainer

ntlewis Aug 11, 2023
Maintainer

AndrewILWilliams Aug 11, 2023
Author

sit23 Aug 11, 2023
Maintainer

AndrewILWilliams Aug 11, 2023
Author