3). Changes in the output fields

1). As explained here, all output fields are now 1D (there is no longer the dimension PATCH in the output files).

It leads that:

  • Names and shapes of fields previously defined with the dimension PATCH (field names like FIELD_P and prognostic fields) are changed:
    • for 1 field FIELD_P defined in V8_0 for all ISBA patches separately, there are now NPATCH fields: their names are FIELDP1, FIELDP2, etc.
    • for 1 prognostic fields TG1 for example, there are now NPATCH fields: their names are TG1P1, TG1P2, etc.
  • To keep field names with at most 12 characters, it’s possible that some fields names changed:
    • for example, DV_NAT_[SPEC], DV_TWN_[SPEC], DV_WAT_[SPEC], DV_SEA_[SPEC] respectively become DVNT[SPEC], DVTN[SPEC], DVWT[SPEC], DVSE[SPEC].
    • if you are looking for a field in an output file, check the number of characters of its name. If it’s more than 9 characters, the name is probably changed in V8_1.

2). We changed the names of the MEB diagnostics to harmonize them with the names of the associated variables and with them one another:

V8_0 V8_1 V8_0 V8_1
LEVCV LEV_CV LETRGV LETR_GV
LESC LES_CV LERGV LER_GV
LETRCV LETR_CV LE_G_C LE_GV
LERCV LER_CV H_G_C H_GV
LE_V_C LE_CV LE_C_A LE_CA
H_V_C H_CV H_C_A H_CA
MELTCV MELT_CV H_N_C H_GN
FRZCV FRZ_CV LE_N_C LE_GN

3) For the OUT.nc output files (CTIMESERIES_FILETYPE = "OFFLIN"):

  • The suppression of the init_outfn_.. definition files leads that all fields able to be written (ie for which write_surf is called) are written in the OUT.nc files (on the contrary to before when only fields defined in the init_outfn_.. files were written).
  • When LSNOWDIMNC = .TRUE.:
    • the dimension snow_layer is added to files ISBA_PROGNOSTIC.OUT.nc, ISBA_DIAGNOSTICS.OUT.nc, ISBA_DIAG_CUMUL.OUT.nc, TEB_PROGNOSTIC.OUT.nc, GARDEN_PRONOGSTIC.OUT.nc, GREENROOF_PROGNOSTIC.OUT.nc.
    • GARDEN_PROGNOSTIC.OUT.nc and GREENROOF_PROGNOSTIC.OUT.nc are created to simplify the management of dimensions : only one snow_layer dimension for each OUT.nc file.

4) PGD fields "L_..." for ISBA (presence or not of data in NAM_DATA_ISBA):

* These L_... fields are now dimensioned to NVEGTYPE and NTIME*NVEGTYPE for the time dimensioned fields.
* They are written and read with this new dimensions.

5) For SODA-EKF, several diagnostics are added:

* Jacobian values: HO[NUM_OBS]_[NUM_VAR]_[TIME]P[NUM_PATCH]
* innovations: INNOV[NUM_OBS]_[TIME]
* residuals: RESID[NUM_OBS]_[TIME]

Also in this section