(cherry picked from commit 863d628c)

Philippe 08/07/2020: bugfix: budgets: add missing multiplication by prhodj for scalar variables budgets

Philippe 30/06/2020: budgets: bugfix for HENU source in concentration of activated nuclei C2R2/KHKO budget

Philippe + Benoit Vié 18/06/2020: improve removal of negative scalar variables: use ZEXN instead of PEXNREF

Philippe + Benoit Vié 11/06/2020: improve removal of negative scalar variables + adapt the corresponding budgets

Philippe 04/06/2020: set_conc_lima: correct array start for microphysics concentrations + add kmi dummy argument
(this subroutine is also called for other models)

Marc M. 28/04/2020: minor correction of ACPRT accumulated precipitation conversion to mm taken into account twice in DIAG program in written output

 Quentin 21/04/2020: major correction for GFS grib use : the u and v wind component is now written in the correct vertical order at prep_real

This reverts commit bbae8eb8.

This reverts commit d7ac06a0.