MAJOR BUGFIX ------------ New wind implicitation in all schemes (Arpege crash) : Optional key in NAM_SURF_ATM : CIMPLICIT_WIND = 'OLD' => direct old implementation 'NEW' => Taylor serie, order 1 (default) XQS_TILE now initialized (calculator crash) Initialization of some local variables in Hydro_soildif.F90 (calculator crash) Default value for Jackson coefficient (cumulative root fraction for DIF) over non vegetated area (calculator crash in some user cases) Supress numerical problems with topographic index pdf calculation (calculator crash with some special grid) Time splitting for soil ice with DIF (calculator crash for large time step) Bug in friction calculation with z04D (calculator crash) and friction due to subscale orography to momentum fluxes added only over continental area Bug in local variables declaration in PROVAR_TO_DIAG subroutine (calculator crash) Missing interface (use modi) in prep_inland_water.F90 (compilation crash with gmkpack) Wrong interface in write_header_mnh.F90 (compilation crash with gmkpack) LINIT_PRECIP now in MODD_SURF_ATM_n instead of MODD_SURF_ATM (parallelization problems with atmosperic model) Suppress test on NWG_LAYER to write/read isa_dif water prognostic reservoirs (parallelization problems) Bug in specific humidity implicitation in Crocus Bug in Mass conservation in Snow3l Bug in total albedo calculation if ES or crocus are used: Save snow albedo values at beginning of the time step in isba.F90 Bug in total Ri (Richardson number) diagnostic if ES or crocus are used: snow Ri not taken into account. Now Ri = (1-Pn) * Ri_isba + Pn * Ri_snow Bug in total sublimation flux calculation if ES or crocus are used (isba_snow_agr.F90) Bug in soil exponential profile calculation with DIF Bug in Ri calculation (PVMOD instead of ZVMOD because limitation of wind modulus done in surface_ri.F90 subroutine) for coare30, ecume and isba_flood_properties. Bug in hydro_glacier snowfall input argument and wrong calculation when ES or Crocus are used (hydro_glacier.F90 and prep_perm_snow.F90). Bug in initialization of an uniform snow pack (read_prep_isba_snow.F90, prep_hor_isba_field.F90) Limitation of Ri by Rimax now in surface_ri.F90 to prevent potential bugs (as it was the case with 3L and Crocus) Latent heat flux of sublimation required by atmsopheric model, so added in get_fluxn.F90 Latent heat flux of sublimation for TEB because required by atmsopheric model. Must be improved in next surfex version! MINOR BUGFIX ------------ Good ponderation coefficient for cumulative root fraction of evergreen forest for DIF Added flood fraction in COTWORES Wrong calculation of unsaturated hydraulic conductivity function over frozen soil with DIF Add some useful comments in Hydro_soildif.F90 and Ice_soildif.F90 Computation efficiency for HRAIN=='SGH' in hydro_veg.F90 Bug in wg2 ice energy budget calculation with ISBA-FR Bug in restore flux calculation in isba_fluxes.F90 (only used for diag but important to close energy budget) If sea or water imposed to 1 in a grid cell: no extrapolation in pgd_cover.F90 (too slow and not required) Optimization of subscale_z0eff_1d.F90 (complex subroutine must not be called inside a loop) Loop optimization in unpack_diag_patchn.F90 Loop optimization in tridiag_ground.F90 (very important because frequently used with ES and Crocus) Erroneous description in diag comments of LETR (transpiration and not evapotranspiration in w/m2) and LER (canopy direct evaporation in w/m2) Supress RETURN (in write_surf_atmn.F90) if LPROVAR_TO_DIAG in order to write prognostic fields in diag files if user want.