fix bug in spitter call (MAIN)

[har917]

CABLE

Description

Following analysis in CABLE4 and discussion in #354 we should use total SW down in the call to spitter and apply the resultant Fbeam equally to the visible and near-infrared bands. The original intent (cable1.4) of spitter is to evaluate the fraction of SW down that is diffuse in nature - via the beam fraction Fbeam. This was evaluated once based on the total shortwave down. From CABLE2 onwards spitter was called twice - once for the visible and once for the near-infrared components. However the function itself was not adjusted - the result is that beam fraction is universally smaller than the original intent with unknown impacts on the model trajectory.

This change set reverts the evaluation of Fbeam to operate on the total SW down and then applies Fbeam equally to the two radiation bands.

This has been implemented as a bug fix, not on a switch. This change may require recalibration of the CABLE model (notably stomatal conductance parameters, leaf nitrogen exponential and any other parameters that impact GPP).

Fixes #355

Type of change

• Bug fix
• New or updated documentation

Checklist

• The new content is accessible and located in the appropriate section.
• I have checked that links are valid and point to the intended content.
• I have checked my code/text and corrected any misspellings

---

📚 Documentation preview 📚: https://cable--357.org.readthedocs.build/en/357/

[har917]

As expected benchcab comparing MAIN and this branch detected differences in many (but interestingly not all) cases - this is expected. Closer inspection of these files indicates that this occurs in all output variables, in all output timestamps (except for some sites where run starts overnight) indicating exact replication up to the point of first non-zero SW_down which then triggers a change. Detailed output indicates that variability in output over time/correlation between outputs remains qualitatively similar as expected given a bug fix.

benchcab log at
benchmark_cable_qsub.sh.o122014778.txt

Proceeding to me.org for scientific assessment.

[har917]

modelevaluation.org results at

General impression is that not much has changed - there is

• a general shift from latent to sensible heat
• slight overall improvement on LE
• slight degradation on H [configuration dependent])
• a general decrease in magnitude of NEE and degradation overall
• no particular difference in impact by PFT type

[har917]

@ccarouge This looks safe enough to apply as is - however should this go in on a switch?

[ccarouge]

This is good to go.

We will have to do a bugfix release for this, it will come in a few weeks I suspect once we agree on what information we should include in.