Add write_restart_at_endofrun functionality to rtm same as in clm

Fix hXi (instantaneous) file metadata

Match code to what was already working in the clm

Put instantaneous and non-inst. fields on separate history files

This is the RTM counterpart to ESCOMP/MOSART#117 and ESCOMP/CTSM#2445.

Resolves #32 (my branch name should have referenced issue #32 instead of 52)

Fix izumi_nag compile

Fix so can compile and run on izumi_nag. Also add a single test to the rtm testsuite for izumi_nag

Standardize time metadata

Standardizes a dimension name of output variable time_bounds, as well as attributes for that plus mcdate, mcsec, mdcur, and mscur.

The RTM rpointer file now has the simulation timestamp in the filename

Have the restart pointer filenames for RTM with the simulation time-stamp in them. Make it backwards compatible to using filenames without the simulation time-stamp as well. This changes time-manager setup from reading in the restart time from the restart file -- to using the current time sent from the coupler. Previously you would EITHER call timemgr_init OR timemgr_restart. Now, you always call timemgr_init -- and only call timemgr_restart for a restart case, with most of what it does is to make sure the time on the restart file agrees with the time sent from the coupler.

Also remove use of shr_file_mod throughout the code.

Change the ChangeLog file into a markdown file and start using some markdown syntax.

Resolves: #53

Resolves: #46

PR
#57

Eliminate the 0th time step from RTM simulations

time in hist now equals the middle of the time_bounds

RTM equivalent to CTSM work done in ESCOMP/CTSM#2838
Answers change only for the time variable.

Update testlist_rtm

fix area scaling to correct RTM river flux to MOM6

fix how area is applied to input runoff for rtm