Current bugs

2011-12-12T16:10:08Z

revision 4: initialization of LTREEDRAG at FALSE was missing: modd_treedrag.F90

revision 5: Beljaars coefficient for the calculation of SSO_DRAG may be put in namelist, so value in the code was removed:
- sso_beljaars04.F90

revision 6: slight corrections to compile with new version of gmkpack :
- sxpost.F90
- zoom_pgd_cover.F90

revision 13 (from branch bugs_valery revision 12) : mode_read_extern.F90, bug of initialization of the water reservoir of garden when using an initial file with garden activated.

revision 15: diag_isba_initn.F90, allocation of XIACAN must be done as soon as CPHOTO/=NON, to be consistent with unpack_diag_isban.F90.

Transpiration fix for ISBA for dry conditions

revision 17 (from branch mesonh_bug revision 10): a coefficient for the calculation of sub scale orography drag in case of Beljaars parameterization is added in namelist NAM_SSO, and initialized at 2. (COEFBE). The files impacted are:
- modd_surf_atm_sson.F90, modn_sson.F90: definition of COEFBE
- default_sso.F90, read_namelists_surfn.F90, init_surf_atmn.F90: set of default value 2. for COEFBE.
- sso_beljaars04.F90: inclusion of COEFBE factor in the calculation of sso drag.

revision 22 (from branch smorin_bug revision 19): maximum number of snow layers allowed to 50 (was 20): modd_prep_snow.F90

revision 23 (from branch smorin_bug revision 20): snowcro.F90
- update snow albedo coefficients for CROCUS (for consistency with Vionnet et al., GMDD 2011)
- test albedo weighting based on SWE 5mm

revision 24 (from branch bugs_valery revision 21):
bugs for some TEB diagnostics when several models are used : modd_diag_misc_tebn.F90

revision 44 : error of index in calculation of XIACAN : unpack_diag_patchn.F90

revision 96 (from branch mesonh_bug revisions 39,40,41) :
to guarantee ascendant compatibility in zoom_pgd_.. part.
Files impacted are zoom_pgd_isba.F90, zoom_pgd_surf_atm.F90, zoom_pgd_cover.F90.

revision 97 (initially from branch mesonh_bug revision 42) :
to be able to write 2M fields in parallelized mode of Surfex, the function SUM_ON_ALL_PROCS is added in a test: write_diag_seb_surf_atmn.F90

revision 47 (from branch bertrand_dev) : to avoid division by zero: hydro_glacier.F90

revision 48 : default value for NYEAR and NMONTH must be NUNDEF and not XUNDEF: read_teb_date.F90, read_watflux_date.F90, read_flake_date.F90, read_isba_date.F90, read_seaflux_date.F90

revisions 50 and 108 : FLOOD parameters must not be computed in case of PREP computation of TSRAD. Files impacted are :
- averaged_albedo_emis_isba.F90: LFLOOD given in argument
- prep_isba.F90: .FALSE. given in argument for LFLOOD
- update_rad_isban.F90, update_esm_isban.F90, compute_isba_parameters.F90, coupling_isban.F90: LFLOOD given in argument

revision 52 : in case of an ASCLLV file for WG and TG in NAM_PREP_ISBA, WGI is initialized at 0., because only the option XUNIF is possible for WGI: read_prep_teb_garden_conf.F90.

revision 58 : problem when nb of layers in extern file grower than current one : prep_snow_extern.F90.

revision 96 : ascendant compatibility for Meso-NH: zoom_pgd_isba.F90, zoom_pgd_surf_atm.F90, zoom_pgd_cover.F90.

revision 97 : Use of SUM_ON_ALL_PROCS for parallelization : write_diag_seb_surf_atmn.F90.

revision 106 : there was a problem in the position of the call of flag_teb_garden_n(1). Files impacted are teb_garden.F90 and garden.F90.

revision 107 : the test on values of vegtypes given in namelist was not relevant in the case of interpolations needed: pgd_isba_par.F90.

revision 136 : one argument forgotten in call to latlon_grid : pgd_surf_atm.F90

revision 165 : when reading of auxiliary file is intercalated with reading of PGD or PREP file, problems occur with the buffer (read_prep_file_date.F90) and with the mask for GARDEN (prep_teb_garden_extern.F90).

revision 166 : the computation of SST values according to the time in the run was wrong : sst_update.F90.

Reproducibility: tests and results

2011-11-25T15:55:53Z

Fery few modifications have a significant impact on the scientific results in Surfex v7.1, as differences between v7 and v7.1 are mainly technical.

2011-11-25T11:26:48Z

New tests cases with specific configurations were also implemented for TEB and GARDEN schemes.

If LTREEDRAG = T in NAM_TREEDRAG, the tree drag is taken into account in the atmospheric model instead of Surfex. In this case, the Z0 forest is reduced to the Z0 grass.
the writting of the covers in PGD and PREP files is simplified. It has a slight impact on writing / reading routines of Surfex.
The PGD is parallelized. This induces the creation of several new routines notably to split the grid and sum on all processors. It also has an impact on the interpolation routines.
update of the writing in the .des file

Some tests were also done in AROME configuration, and lead to some corrections:

the missing USE MODI_... were added; the useless USE MODI_... where removed.
the covers characteristics are read in .bin files and no longer in mode_cover....F90 files. However, it's still possible to read the information without these .bin files activating LREAD_DATA_COVER=T in NAM_READ_DATA_COVER.
lines in F77 were adapted

Surfex v7.1 was compiled on IBM and led also to little corrections.