Distribute - Albedo - set vis and ir albedo to None when sun is down

[jmichellehu]

• Zero-fill of vis and ir albedo from PR #27 (lines 193-194) overrides broadband albedo branch in SolarHRRR.calculate_net_solar
• Results in zero albedo and net_solar radiation being made equivalent to incoming solar values
• Explicit None definition ensures broadband albedo branch kicks in correctly
• Expand tests to check for albedo values after sun is down period and assert net_solar < hrrr_solar everywhere

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused fix addresses the branch-selection bug and includes appropriate regression coverage.

Review effort: Balanced
Findings: None

What changed in this PR

Fixes HRRR net-solar calculation after nighttime steps by preserving broadband albedo selection.

Changes:

• Resets visible and infrared albedo to None at night.
• Adds regression tests for broadband albedo after nighttime.

File	Description
smrf/​distribute/​albedo.py	Clears spectral albedo values when sunlight is absent.
smrf/​tests/​distribute/​test_albedo.py	Tests nighttime-to-daytime broadband albedo behavior.
smrf/​tests/​distribute/​test_solar_hrrr.py	Verifies broadband albedo reduces HRRR net solar correctly.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[jomey]

Can we please make an effort and explain changes in PRs in a human way?

These auto-generated LLM descriptions make it hard to understand things. This is also a copy of the commit message and more context should be given in a PR

[jomey]

There is also a need to further explain why this is needed with net solar only being calculated when the sun is up

[jmichellehu]

There is also a need to further explain why this is needed with net solar only being calculated when the sun is up

Net solar is only calculated when the sun is up, but the albedo distribution is called at every timestep, day or night. So when illumination angles are None at night, the else branch kicks in when the sun is down and the albedo vis and ir grids are set to zero. Nothing resets them to None when the sun comes back up on that run day, so the grids are not None, and the broadband albedo branch gets skipped.

When using a broadband albedo source like SPIReS, this leads to net_solar == hrrr_solar for every sun-up timestep after the nighttime grids are set to zero. Even awsm resetting grids to None the following run day doesn't help the rest of the sun-up timesteps after the nighttime zero-setting timestep.

Theoretically (I didn't test this), this issue doesn't affect runs using the vis/ir pathway as new daytime values overwrite the nighttime zero grids. But setting the grids back to None goes back to the previous net solar behavior.