Avoid kc becomes zero when NDVI == 1.1 * NDVI_min

[LINAMARIAOSORIO]

Checklist

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same update/change?
• Added tests for changed code.
• Updated documentation for changed code.

Description

In dynamic() the crop coefficient used two strict masks (NDVI < 1.1 * NDVI_min and NDVI > 1.1 * NDVI_min), so a cell whose NDVI equals 1.1 * NDVI_min got kc = 0 and no vegetated-area evapotranspiration. The branch follows the documented rule with pcr.ifthenelse(current_ndvi <= 1.1 * self.ndvi_min, min_crop_coef, partial_crop_coef): kc = kc_min whenever NDVI <= 1.1 * NDVI_min.

Follow-up commits on top of the fix:

• A behaviour test in tests/unit/core/test_dynamic_model.py: three synthetic runs with the first-step NDVI exactly on the threshold, at NDVI_min and above it; the first-step eta of the threshold run must equal the NDVI_min run and differ from the interpolated run (fails on main, where the threshold run has eta = 0).
• Merge of fix_319 and a changelog entry under Unreleased › Fixed.

Stacked on #323 (base branch fix_319): the diff of this pull request is rubem/_dynamic_model.py, the test and the changelog line.

Related Issue

• Resolve kc becomes zero when NDVI == 1.1 * NDVI_min #320

Motivation and context

doc/source/overview.rst (equation kc) and the supplementary document (equation SE11) state kc = kc_min if NDVI <= 1.1 * NDVI_min; the equality case was left without a branch. The regression dataset does not contain the equality (the exact job reports byte-identical outputs with and without this change), so the behaviour is covered by the new test.

How has this been tested

• Local suite (pytest --ignore=tests/integration/doc): 998 passed, 1 skipped.
• test_ndvi_at_the_crop_coefficient_threshold_takes_the_minimum_branch fails with rubem/_dynamic_model.py from main and passes here. The threshold value is PCRaster's own float32 product of 1.1 and the ndvi_min raster, so the model's comparison sees exact equality.
• Golden fixtures are not affected by this change: byte-exact reproduction verified in the golden environment container with this branch.
• CI on this pull request: ci-success green.

Screenshots

• N/A

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.47%. Comparing base (1613d18) to head (ed3be1c).

Additional details and impacted files

@@             Coverage Diff             @@
##           fix_319     #324      +/-   ##
===========================================
- Coverage    92.47%   92.47%   -0.01%     
===========================================
  Files           60       60              
  Lines         3908     3906       -2     
  Branches       487      487              
===========================================
- Hits          3614     3612       -2     
  Misses         238      238              
  Partials        56       56              

☔ 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.