Support combining distributed output for FieldTimeSeries with Reduced directions

[simone-silvestri]

Needed for CliMA/ClimaSeaIce.jl#109 to pass

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 72.30%. Comparing base (2781181) to head (d00df50).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5188      +/-   ##
==========================================
- Coverage   72.93%   72.30%   -0.64%     
==========================================
  Files         393      393              
  Lines       21872    22135     +263     
==========================================
+ Hits        15953    16005      +52     
- Misses       5919     6130     +211     

Flag	Coverage Δ
buildkite	68.23% <0.00%> (-0.03%)	⬇️
julia	68.23% <0.00%> (-0.03%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[glwagner]

I don't understand what the PR does from the title -- are you saying that sometimes we get nothing instead of FieldTimeSeries? How does this happen?

[glwagner]

Reading the code I think the issue is when some locations of the field in question are Nothing?

[navidcy]

Hm.... CliMA/ClimaSeaIce.jl#109 passed.

[simone-silvestri]

Ah yeah, it was the combining of distributed fields that does not pass because it assumed halos in z-direction, so I was trying with single ranks
https://github.com/CliMA/ClimaSeaIce.jl/blob/61451d83921321b0ac38577c58c77b29e5495fae/test/test_distributed_sea_ice.jl#L134-L138

basically if one direction is a Nothing we need to avoid computing the range with the halos