Testing environment for merging POP-TRENDY branch and main

[ccarouge]

For bitwise-comparisons, the testing is relatively easy but we need a good code coverage. Cases:

• POP: use the 1000 random grid cells. Only do spin up once (every time we don't have bitwise comparability anymore)
• biogeophysics
• SLI
• Single site CASA tests (@rml599gh )

For science merging, we want to run single sites tests all the time and TRENDY simulations at some milestones.

Test for restart reproducibility vs continuation run. It's important to fix that.

ACCESS-NRI Whole community input on shared configurations @juergenknauer : pushing TRENDY v13 config.