CMIP7: Collect ancillaries for UKCM2a-HH

[matthew-mizielinski]

Ancillaries used in previous models include

File	variables
diaptr_basin_masks.nc	atlantic_arctic_ocean_1D_[TV], atlantic_arctic_ocean_2D_[TV], global_ocean_1D_[TV], global_ocean_2D_[TV], indian_pacific_ocean_1D_[TV], indian_pacific_ocean_2D_[TV]
ocean_basin.nc	basin
ocean_byte_masks.nc	mask_2D_[TUV], mask_3D[TUV]
ocean_constants.nc	areacello, difmxlo2d, diftrelo2d, difvmbo difvtrbo, hfgeou, sftof
ocean_zostoga.nc	so_0, zfullo_0, rho_0_mean
qrparm.landfrac.pp	m01s00i505
qrparm.orog.pp	m01s00i033
qrparm.soil.pp	m01s00i207, m01s00i008, m01s00i220, m01s00i223, m01s00i044, m01s00i043, m01s00i048, m01s00i046, m01s00i047, m01s00i041, m01s00i040

Need to confirm which variables are used in mappings and collect appropriate files ready for inclusion in CDDS ancil set

[mo-laurenboon]

Upon initial inspection:

File	Variable	Required? [Y/N]	Where is it used?
diaptr_basin_masks.nc	atlantic_arctic_ocean_1D_[TV]	Y	9 instances in UKCM2_ocean_mappings.cfg
	atlantic_arctic_ocean_2D_[TV]	Y	1 instance in UKCM2_ocean_mappings.cfg
	global_ocean_1D_[TV]	Y	9 instances in UKCM2_ocean_mappings.cfg
	global_ocean_2D_[TV]	Y	1 instance in UKCM2_ocean_mappings.cfg
	indian_pacific_ocean_1D_[TV]	Y	9 instances in UKCM2_ocean_mappings.cfg
	indian_pacific_ocean_2D_[TV]	Y	1 instance in UKCM2_ocean_mappings.cfg
ocean_basin.nc	basin	Y	At least 1 instance in UKCM2_ocean_mappings.cfg
ocean_byte_masks.nc	mask_2D_[TUV]	Y	3 instances in UKCM2_ocean_mappings.cfg
	mask_3D[TUV]	N	-
ocean_constants.nc	areacello	Y	14 instances in UKCM2_ocean_mappings.cfg
	difmxlo2d	N	-
	diftrelo2d	N	-
	difvmbo	N	-
	difvtrbo	N	-
	hfgeou	Y	1 instance in UKCM2_ocean_mappings.cfg
	sftof	Y	1 instance in UKCM2_ocean_mappings.cfg
ocean_zostoga.nc	so_0	Y	1 instance in UKCM2_ocean_mappings.cfg
	zfullo_0	Y	1 instance in UKCM2_ocean_mappings.cfg
	rho_0_mean	Y	1 instance in UKCM2_ocean_mappings.cfg
qrparm.landfrac.pp	m01s00i505	Y	2 instances in UKCM2_aerosol_mappings.cfg, 2 instances in UKCM2_atmos_mappings.cfg, 5 instances in UKCM2_land_mappings.cfg, 1 instance in UKCM2_ocnBgchem_mappings.cfg
qrparm.orog.pp	m01s00i033	Y	1 instance in UKCM2_atmos_mappings.cfg, 1 instance in UKCM2_land_mappings.cfg
qrparm.soil.pp	m01s00i207	N	-
	m01s00i008	N	-
	m01s00i220	N	-
	m01s00i223	N	-
	m01s00i044	N	-
	m01s00i043	N	-
	m01s00i048	N	-
	m01s00i046	N	-
	m01s00i047	N	-
	m01s00i041	Y	1 instance in UKCM2_land_mappings.cfg
	m01s00i040	N	-

Hence:

File	Required? [Y/N]
diaptr_basin_masks.nc	Y
ocean_basin.nc	Y
ocean_byte_masks.nc	Y
ocean_constants.nc	Y
ocean_zostoga.nc	Y
qrparm.landfrac.pp	Y
qrparm.orog.pp	Y
qrparm.soil.pp	Y

[mo-gill]

When testing this variables requiring basin ancils diaptr_basin_masks.nc and ocean_basin.nc haven't been tested, as running them doesn't actually produce anything. Only producing warnings like this when prepare_generate is run:
Unrecognised variable: ocean/hfbasinpadv_tavg-u-hyb-sea['Constraint "basin" is not supported']
And this message from the cdds_extract step:
hfbasinpadv_tavg-u-hyb-sea - not active Constraint "basin" is not supported,

There is an issue for handling these "basin" variables which is currently on hold here:
#963

[mo-gill]

These variables which use these ancils have been tested by running UKCM2a-0-HH requests

ocean_byte_masks.nc

ocean/msftbarot_tavg-u-hxy-sea@mon:onm/grid-U ❌
Can't produce, high resolution ancil required. Current error when using UKCM2 ancil:
iris.exceptions.CannotAddError: Unequal lengths. Cube dimension 1 => 3604; metadata 'latitude' dimension #0 => 330.

ocean/tauuo_tavg-u-hxy-sea@mon:onm/grid-U ✅
Produces (using ocean_byte_masks_HH.nc)

ocean_constants.nc

ocean/so_tavg-ol-hm-sea@mon:onm/grid-T
ValueError: operands could not be broadcast together with shapes (2,75,3604,4320) (330,360)
ocean/thetao_tavg-ol-hm-sea@mon:onm/grid-T
ValueError: operands could not be broadcast together with shapes (2,75,3604,4320) (330,360)

qrparm.orog.pp

atmos/ztp_tavg-u-hxy-u@mon:ap4 ✅
Produces successfully using same qrparm.orog.pp as UKCM2LL

qrparm.soil.pp ✅

land/mrsofc_ti-u-hxy-lnd@fx:afx
Fx file already produced successfully during BCV testing. ✅

ocean_zostoga.nc

ocean/zostoga_tavg-u-hm-sea@mon:onm/grid-T ❌
Requires new ancil. Current error using UKCM2 ancil:

 failed. Error:
    raise ValueError(message.format(a=thetao, b=so, c=zfullo, d=thkcello))
 ValueError: Input arguments have inconsistent shapes:
	thetao = (75, 3604, 4320)
	so = (75, 330, 360)
	zfullo = (75, 330, 360)
	thkcello = (75, 3604, 4320)
	"so" & zfullo come from ocean_zostoga.nc

ocnBgchem/fgdms_tavg-u-hxy-sea@mon:ap5 ❌
Likely can't be produced even with ancil due to a bug in the processor (probably shouldnt be in UKCM2a mappings?). See: #960 for info about processor bug.

qrparm.landfrac.pp

land/evspsblsoi_tavg-u-hxy-u@day:ap6 ❌
land/evspsblveg_tavg-u-hxy-u@day:ap6❌
Can't be produced, probably due to resolution difference. Current error when running using UKCM2 ancils:

ValueError: Cannot resolve cubes, as a suitable transpose of the RHS cube 'land_area_fraction' will not broadcast
with the LHS cube 'Evaporation from canopy'.
ValueError: operands could not be broadcast together with shapes (90, 960, 1280) (1, 144, 192)

aerosol/emibvoc_tavg-u-hxy-u@mon:ap4 ❌
aerosol/emiisop_tavg-u-hxy-u@mon:ap4 ❌
Both emibvoc and emiisop need to be registered as do not produce due to creating erroneous data?

[mo-gill]

We'll need to request new UKCM2a HH ancils for the ones added in this commit
("nemo_eORCA1_grid-U.nc", "nemo_eORCA1_grid-V.nc")

Also need a confirmed replacement for ocean_byte_masks_HH.nc the current one in the ancil dir works, but it's just symlinked to the same file in ancil_testing - which looks like it was just copied from HadGEM3-GC31-HH.