EAMxx: Incorrect Time Step Values in Sub-Stepped Processes Using timestamp().get_num_steps()

[singhbalwinder]

I’m attempting to print the model time step in the run_impl method of an atmosphere process using auto ts = timestamp().get_num_steps();. This approach works correctly when I use it in a process that isn’t sub-stepped, such as rrtmgp, where I get the expected time step value. However, when I use the same method in processes that are sub-stepped (e.g., P3 or SHOC), the returned time step values are incorrect—they differ from the time step values logged in the atm.log.* file.

[bartgol]

Note to self. two issues to fix related to the timestamp handling in atm proc:

• store two timestamps in the atm proc class: the start-of-step stamp and the end-of-stamp one.
• update the start/end-of-step stamps inside the subcycling loop, right before/after calling run_impl.