Set time value for averaged quantities to midpoint of averaging period

[phillips-ad]

This is linked to ESCOMP/CAM#159 and ESCOMP/CTSM#1059, and involves setting the CISM output time value to the center of the averaging period. Different components may end up handling output differently, but the going consensus so far is to separate instantaneous data streams from averaged/min/max data streams.

[Katetc]

Note from the cochairs discussion on July 25:
From this morning's discussion on CESM history output, this is my understanding of the group's modifications to the presented proposal:

• Keep time = end of bounds for instantaneous fields
• Remove time_bounds from instantaneous tapes

I am implementing these changes to the CTSM in the following PR:
ESCOMP/CTSM#2019

Thank you all again for your input.

Sam

Samuel Levis
Terrestrial Sciences Section, NCAR

[Katetc]

LIkely part of the CISM3 release

[whlipscomb]

Thanks, I will work on this with @gunterl