SURFEX

Current bugs: 7_2_1

2012-12-20T14:43:05Z

revisions 734 & 735: XW33 is removed + important correction in extrapol_fields.F90 (B. Decharme)

revision 794: Correction to re-initialize ANDAY and ANFM diagnostics each day: vegetation_evol.F90 (S. Lafont)

revision 804: modification of the regression for W33 with CO84 for high values of clay: mode_soil.F90 (B. Decharme)

revision 805: default value for NTIME in case init_isba_mixpar is used: read_pgd_isba_parn.F90 (B. Decharme)

revision 810 : fields names for CSELECT were limited to 12 characters whereas some of them have until 16 characters. It prevented them to be written: modd_diag_surf_atmn.F90, modn_surf_atmn.F90, default_diag_surf_atm.F90 (D. Carrer)

revision 834 : if tests on HPHOTO, the option NCB is added with NIT, LST and AST: init_isba_mixpar.F90 (C. Delire).

revision 852 : a problem with FA compaction was coming from a typo in a routine from the XRD38 library, because modifications where done not to use the OML_LIB part of the XRD38: mxmn_mf.F90

revision 874 : during the vertical interpolation of input ground data on the soil depths of the user, with CISBA=DIF in PREP step, there was a confusion concerning the use of the root fraction as a limit to delimit the root depth: prep_hor_isba_field.F90, prep_hor_teb_garden_field.F90 (B. Vincendon).

Current bugs: 7_2_0

2012-06-19T16:41:48Z

revision 419: clearer messages printed at screen for user at compilation: conf/profil_surfex.ihm, src/configure

revision 420: add of the name of the seeked record in the error message of abor1_f90: ol_find_file_read.F90

revision 421: correction on the definition of the date for FA format: offline.F90

revision 422: so that 1D binary forcing files can be read in the framework of a 1D experiment (like in ASCII): open_close_bin_asc_forc.F90, read_surf_atm.F90.

revision 424: if a PREP file in defined for TEB, but uniform values are defined in the namelist, the fields not defined in namelist (T_CAN,Q_CAN) are deduced from uniform values and not from the PREP input file: default_prep_teb.F90,prep_teb_unif.F90,read_prep_teb_conf.F90, mode_read_grib.F90.

revision 425 & revision 455 & revision 457: errors in intent of arguments: init_dst.F90, init_slt.F90

revision 456: files beginning by assim_... are moved in OFFLIN directory.

revision 468: declaration of ZHOOK_HANDLE forgotten in OPTIMSOILGRID: soilgrid.F90.

revision 499 : by default, SURFEX was compiled in DEBUG mode so no optimizations were performed (-O0 compilation option). So runs were slower. The default mode for compilation on a cnrm PC linux is set to O2 (-O2 compilation option). To take it into account, the whole installation has to be redone.

revision 500 : error in water balance for DIF (LETR forgotten in FRC calculation): hydro_soildif.F90.

revision 512 : correction inversion between VTANG1 and VTANG2 : modd_snow_metamo.F90

revision 514 : correction for CO84 regressions for WWILT, WFC et W33: mode_soil.F90

revision 515 : problem with an infinite loop in GAUSS case: read_direct.F90, read_direct_gauss.F90

revision 541 : numerous corrections of bugs by B. Decharme. Detail of the modifications here:

README_BUGFIX7.2

(B. Decharme)

revision 564 : with large time step, snow melt with D95 causes instability on the TG1. This is solved by a XTAU_SMELT in namelist NAM_SURF_CSTS and with the use of max(XTAU_SMELT,XTSTEP) in isba_fluxes instead of XTAU_SMELT directly (E. Brun & B. Decharme). Impacted files are : modd_snow_par.F90, ini_surf_csts.F90, isba_fluxes.F90.

revision 597 :
- mode_snow3l.F90: optimization of do-loops (A. Boone)
- snow3L_isba.F90: check suspicious low temperature for ES and CROCUS
- snow3l.F90: correction of a problem with the discretization of snow layers for intense snow precipitation (E. Brun)
- snow3l.F90, e_budget.F90, isba.F90, isba_fluxes.F90, snowcro.F90: problem with the wind implicitation and the calculation of USTAR (B. Decharme).

revision 600 (P. Le Moigne):
- modd_flaken.F90, modd_diag_misc_flaken.F90: errors in GOTO routines
- start_lake_of.F90, modd_data_lake.F90: in case there's a lake but no climatic data associated then fill with neighbour existing data instead of aborting.
- put_zsn.F90: calculation of masks for tiles if not ever done.

revisions 662 & 663 (B. Decharme): errors relative to soil ice: isba_fluxes.F90, hydro_soil.F90.

Reproducibility

2012-02-20T12:09:14Z

CISBA=2-L
WR:
diff max: 3.40951007E-18
diff min: -3.40951007E-18
average: 1.57537911E-21
AV_PGD is called before INI_DATA_SOIL in convert_patch_isba

CISBA=3-L CRAIN=SGH:
TG1:
diff max: 0.17199707
diff min: -1.00708008E-03
average: 2.55551597E-04
In hydro_veg.F90, deletes SGH for very fine precipitation.

CISBA=DIF with otimized grid:
TG1:
diff max: 3.6489868
diff min: -1.3560181
average: 9.08908024E-02
In ice_soildif.F90, formulations for ice melt and freeze have changed.
TG1:
diff max: 3.79943848E-02
diff min: -0.77600098
average: 7.77693540E-02
W33 is used instead of WFC in drag.F90 for DIF.
TG1:
diff max: 5.59997559E-02
diff min: -0.79901123
average: 4.18960899E-02
The number of layers for hydrology, different from this for temperature, is introduced.
TG2:
diff max: 1.00708008E-03
diff min: 0.0000000
average: 4.37100738E-08
In ice_soildif.F90, the threshold to remove very small numbers for ice content has changed (WGI<1.E-10 becomes WGI<1.E-6).

TG12:
diff max: 0.0000000
diff min: -1.00708008E-03
average: 4.37100738E-08

CISBA=DIF CRUNOFF=DT92
TG1
diff max: 2.01416016E-03
diff min:-1.00708008E-03
average: 2.69558695E-05
In compute_isba_parameters.F90, RUNOFFD if calculated differently for DIF.
TG1
diff max: 1.00708008E-03
diff min: -1.00708008E-03
average: 2.55902614E-06

News in v7_2

2012-02-20T11:47:38Z

Bertrand Decharme:

Rewriting of case CISBA=DIF and SGH options:
- the initialization of DIF grid is now done through an optimized grid with 14 layers when ecoclimap is activated. The user has the possibility to define his own grid in namelist NAM_ISBA.
- the hydrologic soil depth or number of layers is calculated or given by the user and possibly different from the temperature soil depth.
- runoff depth is calculated from the hydrologic soil depth for DIF; dunne and horton numbers of layers are also calculated to limit certain calculations.
- CRUNOFF=SGH is implemented for DIF
optimizations through intermediate variables.
Fix fields of PGD are now written only in PGD file.
CAUTION: to make a run, it's necessary to have PGD and PREP file in the run directory.

Champs fixes PGD/PREP

(Bertrand Decharme)

add of new hydrologic diagnostics.
writing of 2D netcdf output files for IGN grid.

Aaron Boone:

Changes in calculations of ice melt and freeze in ice_soildif.F90.

Dominique Carrer:

New radiative transfert for CPHOTO/=NON.
Representation of the nitrogen dilution functions of CO2

Eric Martin:

Canopy fields that are initialized not during PREP step but at first time step of run are no more written in PREP output file.
Land/Sea mask is used to locate missing values when they are not equal to XUNDEF in read file. For snow, WSNOW (amount of water in snow) is used as mask.

Eric Bazile / StÃ©phanie Faroux:

Forcing gradients for CISBA=TSZ0 are now givable in namelist NAM_DATA_TSZ0.
Numbers and times of forcing for flux and temperature are given in namelist NAM_IDEAL_FLUX with CISBA/CTEB/CSEA/CWATER=FLUX

StÃ©phanie Faroux:

netcdf forcing files can be 2D
adaptations to run Meso-nh with external PREP files : compatibility with separated PGD / PREP.
grouping of initializations for ISBA and GARDEN.
Rewriting of parts for DUST and SALT.
With HINIT='PGD', surface parameters are no more calculated (for MÃ©so-nh nesting).
Bugs in the calculation of implicit coefficients:
- when the vertical shift is taken into account, there is an inconsistency in dimensions for TA in ISBA (temperature mixed with theta).
- for water and sea surfaces, implicit temperature coefficient are calculated with PS instead of PA.
Updating of libraries FM, LFI and XRD with versions from cycle 38.
Installation and compilation environments token from recent MÃ©soNH.

Patrick Le Moigne:

update of FLAKE to be consistent with watflux scheme + corrections of bugs

Pierre Tulet:

Error on unit conversion for chemicals.

To run offline 2D with 1D forcing files (very useful for testing).
Simplification of namelist NAM_IGN in case of a regular grid

David Barbary:

Possibility to read recent mercator netcdf files (with a scale factor and an offset on SST).

Trygve Aspelien:

Development of new surface assimilation interface SODA.
Adds and tests the capability to compile Surfex OFFLIN with ifort.

Mireille Tomasini:

Changes in namelists

2012-02-20T11:22:31Z

NAM_SSOn: add of XCOEFBE (def=2.): coefficient for Beljaars calculation of SSO drag.

NAM_ISBA:
- add of XSOILGRID(NGROUND_LAYER) for CISBA=DIF.
- add of LTR_ML, if TRUE, new radiative transfert is calculated (for CPHOTO/=NON).

NAM_DATA_ISBA: for CISBA=DIF, add of
- XUNIF_GROUND_DEPTH/CFNAM_GROUND_DEPTH/CFTYP_GROUND_DEPTH: hydrologic soil depth.
- XUNIF_ROOT_DEPTH/CFNAM_ROOT_DEPTH/CFTYP_ROOT_DEPTH
- XUNIF_ROOT_EXTINCTION/CFNAM_ROOT_EXTINCTION/CFTYP_ROOT_EXTINCTION
- XUNIF_ROOT_LIN/CFNAM_ROOT_LIN/CFTYP_ROOT_LIN

NAM_ISBAn: add of XCDRAG (default=0.15) then used in isba_canopy.F90.

NAM_SURF_CSTS: add of XZ0SN: roughness length of pure snow surface (m), default=0.001, and XZ0HSN: roughness length for heat of pure snow surface (m), default=0.0001.

NAM_DATA_TSZ0/NTIME, XUNIF_DTS, XUNIF_DHUGRD/: new namelist containing gradients for temperature and humidity in case of CISBA=TSZ0. NTIME=1 or NTIME=25 (1 value by hour in a day).

NAM_SEAFLUXn/CSEA_ALB/: new value for CSEA_ALB= MK10 : Modified version taken from SAM6.8 (Oct 2010) Marat Khairoutdinov
25 from CCM3).

NAM_IDEAL_FLUX: add of NFORCF, NFORCT, XTIMEF(NFORCF), XTIMET(NFORCT).

NAM_FLAKEn: add of CSNOW_FLK (DEF), CFLK_FLUX(DEF/FLK) CFLK_ALB(UNIF/TA96).

NAM_IGN: add of XX_LLCORNER, XY_LLCORNER, XCELLSIZE, NCOLS, NROWS. XCELLSIZE can be used instead of XDX and XDY; the 4 other can replace NPOINTS, XX and YY (case of a regular grid).