Juan 01/07/2021: configure, for Irene/TGCC upgrade to intel/19.0.5.281+mpi/openmpi/4.0.2 & add module load cmake/3.18.1 for Intel/AMD arch

Quentin 30/06/2021: bugfix OCEAN move constants change in ini_modeln to ensure change in every sub-model n

F.Couvreux 30/06/2021: add LRELAX_UVMEAN_FRC forcing. Instead of LRELAX_UV_FRC which applies the u,v relaxation read in namelist on each atmospheric column,
LRELAX_UVMEAN_FRC applies the relaxation on the horizontal mean of u and v at each level. To be used in LES mode.

P.Tulet 30/06/2021: bugfix diag with chemistry variables : no time splitted needed at DIAG so SVT must be computed with XSVT and GSPLIT=F

P.Tulet 29/06/2021: SURFEX bugfix, PK%NSIZE_P can be zero leading to missing initialization of dust variables + missing logical LCH_SURF_EMIS for chemical emission activation

23/07/2018  Correction from Arsene Druel for maj average
23/07/2018 Correction to read the 10-days periods in the PGD file (Rafife Nheili)
23/07/2018 Correction for PGD and NAM_DATA (Arsene Druel)
22/08/2018 Bug for parallel PREP & snow (array sizes) (Q. Rodier)
23/08/2018 Correction for LSPLIT_PATCH and LRESET_BUDGET=F (P. Samuelsson)
23/08/2018 Correction to have the same diagnostic output fields names with LSPLIT_PATCH=F
05/09/2018  bug fix from Aaron for MEB - initialisation
14/09/2018 Bugfix ISBA MEB (Aaron Boone)
25/09/2018 Error of dimension in allocation
10/10/2018 Bugfix ISBA MEB (Aaron Boone)
23/10/2018 Correction for upward compatibility for snow albedo
14/05/2019 BugFix: correction of H_TREE calculation by patch (impact only if H_TREE given in namelist, and if NPATCH different of 1 or NVEGTYPE)
21/06/2019 BugFix: correction of ZMASK used to initialise TEB variables when GTEB=FALSE in the extern file read for initialisation
23/08/2019 Bugfix NVEGTYPE_ECOSG / NVEGTYPE in read_nam_pgd_isba_meb (CEN)
17/11/2020 Bugfix: wrong initialisation of LIGNIN_STRUC and SOILCARB (when CRESPSL=CNT)
17/11/2020 Bugfix: correction for the writing of the variable ROOTFRAC in PREP file
29/03/2021 BugFix: YPHFILETYPE initialized twice, and missing initialization for YFERTFILETYPE
Updated to last commit (e826177a)

Quentin 29/06/2021: bugfix SURFEX avoid abort for interpolation of a field which is fully XUNDEF at spawning with ECOCLIMAP-SG. With ECOSG, a field can be fully XUNDEF after PGD.

Philippe 28/06/2021: small optimisation in the call to ice4_rainfr_vert (create arrays with zeros instead of multiplications by zero)

Philippe 28/06/2021: netCDF: use correct integer datatypes for netCDF integers + respect Doctor norm

Philippe 28/06/2021: IO: better management of pre-5.5 files (default values for missing fields instead of crash)

Philippe 28/06/2021: allocate to zero-size arrays in ini_model_n to prevent use of unallocated arrays later
+ allocate some of them to non zero-size only if necessary