Update calculation of time-averaged radiation variables

[LarissaReames-NOAA]

Description

This PR fixes incorrect diagnostic output for fhzero-averaged radiation output variables. Presently, there is a significant "sawtooth" appearance to these fields (such as dswrf_ave and ulwrf_avetoa) when we expect smoother lines with jumps only at fhzero diagnostic reset times. Here is an example of the present (OLD) result and the result from the current PR (FIX) from a 24-hour S2S regression test forecast.

Issue(s) addressed

• fixes a problem found in investigating ufs-weather-model/#1767

Testing

Ran full regression test suite on Hera. See log files here: /scratch1/NAGAPE/hpc-wof1/lreames/ufs-weather-model/tests/logs/log_hera

Forecast output can be found here: /scratch1/NCEPDEV/stmp2/Larissa.Reames/FV3_RT/rt_4109245

All regression tests fail as sfcf*.nc files, as well as post-processed grib2 files as a result, are not identical. New baselines will be necessary.

Dependencies

• waiting on noaa-emc/ccpp-physics/pull/241

Requirements before merging

• All new code in this PR is tested by at least one unit test
• All new code in this PR includes Doxygen documentation
• All new code in this PR does not add new compilation warnings (check CI output)