Update ufs-s2s-model to use the most recent FV3 and NEMS components

[binli2337]

Use fv3atm committed on 3/18/2020.
Use NEMS committed on 3/19/2020.

close #59

[DeniseWorthen]

I ran the 2012 benchmark case for 35d and compared to a previous saved version. Everything appeared reasonable. However, the compset run did not complete correctly. I received the following after the job completed. At that point, cd'ing back into the directory where I had launched the job showed it had been cleaned out. Only subdirectory NEMS and log were present. I will attempt to recreate and if it reproduces I will create an issue on NEMS.

[junwang-noaa]

I had this error a coupled of times, usually because I forgot to run: git submodule update --init --recursive after I checked out the branch, so there are no conf files under NEMS. Running the submodule update command fixed the error for me.

…

On Sun, Mar 22, 2020 at 2:25 PM Denise Worthen ***@***.***> wrote: ***@***.**** approved this pull request. I ran the 2012 benchmark case for 35d and compared to a previous saved version. Everything appeared reasonable. However, the compset run did not complete correctly. I received the following after the job completed. At that point, cd'ing back into the directory where I had launched the job showed it had been cleaned out. Only subdirectory NEMS and log were present. I will attempt to recreate and if it reproduces I will create an issue on NEMS. [image: Screen Shot 2020-03-22 at 2 22 42 PM] <https://user-images.githubusercontent.com/40498404/77257109-a03cb680-6c48-11ea-8232-9a154443c552.png> — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#60 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AI7D6TIWTYALZ2G6DWCFHDTRIZJYDANCNFSM4LRANKTA> .

[JessicaMeixner-NOAA]

@junwang-noaa I thought you wanted the WW3 regression test before this? I can update to PR #57 that adds the WW3 regression test now (just need to merge the changes to that branch). However, both the WW3 regression test and the benchmark regression test will have changed answers after this PR with the update of fv3atm. Since I saw this PR yesterday, I also prepared another branch that I can do a PR from that will have an updated baseline using this fv3atm, depending on the desired order of commits. Please let me know if I should update PR #57 or close that PR and make a new one that will go in after this PR with the added regression tests.