Fix type mismatch compiler error when gfortran 10 is used without '-fallow-argument-mismatch' flag

[DusanJovic-NOAA]

Starting with version 10 of GNU Fortran compiler, subroutine argument mismatches are flagged as errors, and must be silenced by using '-fallow-argument-mismatch' flag, which can potentially hide actual bugs, which is undesirable.

This PR fixes all such mismatches by switching to use mpi_f08 MPI module.

[pjpegion]

@DusanJovic-NOAA does anything need to be changed on the compile side because of the addition "use mpi_f08"?
I ask because the stochastic physics code is also being used in branches of CESM and other models.

[DusanJovic-NOAA]

@DusanJovic-NOAA does anything need to be changed on the compile side because of the addition "use mpi_f08"? I ask because the stochastic physics code is also being used in branches of CESM and other models.

Nothing needs to be changed in the build, but the mpi library used must provide mpi_f08 module, which most of the recent mpi libraries do provide. But just to be sure it would be great if other models (CESM) can test this change, because the type of the mpi communicator is changed from integer to MPI_Comm, which means other models will probably need to make that change. If that is not possible for whatever reason we'll need to find a way to support both versions either by some kind of compile time macro or something else.

[pjpegion]

I will see if we can get this tested in CESM, but since it is currently only in a one branch, and they are using a much older version of stochastic physics. I think it would be difficult for them to test this version. At this point I just want to be aware that there may be an issue if they update to the newest code.

[zach1221]

Hello, @pjpegion . This PR is ready to merge. Testing is complete on ufs-wm PR #1147.