FIX: require zarr>=3 for DataTree.to_zarr compatibility

[syedhamidali]

• Closes Zarr compatibility issue: TypeError on to_zarr() with zarr_format argument #398
• Changes are documented in history.md

Summary

The multiple-sweeps-into-volume-scan notebook fails on DataTree.to_zarr(). There are actually two problems, both surfaced by current xarray:

1. zarr<3's open_group() rejects the zarr_format argument that recent xarray forwards → TypeError: open_group() got an unexpected keyword argument 'zarr_format'.
2. With zarr>=3, its codecs can't serialize numpy MaskedArray — and xradar's IRIS reader returns masked moments (e.g. VRADH) → ValueError: cannot reshape array of size N into shape (4N,).

Fix

1. Bump the zarr pin to >=3 (requirements_dev.txt, environment.yml, ci/notebooktests.yml).
2. Fill masked moments in the notebook before to_zarr (small fill_masked helper via map_over_datasets).

Verified

With zarr 3.3.0 + xarray 2026.7.0 (the exact combo that fails in CI), on a real IRIS/Sigmet file (which yields a masked VRADH): the to_zarr + open_datatree roundtrip succeeds and VRADH values are preserved (masked → NaN).

Note: I couldn't run the full notebook locally (its data is on the IDEAM s3-radaresideam bucket), but reproduced and fixed the exact failing operation with representative IRIS data.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.23%. Comparing base (7b81ba0) to head (a9051c2).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #399   +/-   ##
=======================================
  Coverage   94.23%   94.23%           
=======================================
  Files          29       29           
  Lines        6452     6452           
=======================================
  Hits         6080     6080           
  Misses        372      372           

Flag	Coverage Δ
notebooktests	0.00% <ø> (?)
unittests	94.23% <ø> (ø)

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

☔ View full report in Codecov by Harness.
📢 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.

[kmuehlbauer]

Thanks @syedhamidali, for trying. There is an issue in the writer. Maybe @aladinor has a clue?

[kmuehlbauer]

Thanks @syedhamidali! This will unlock other PR, which got stuck because of zarr < 3.