Incorrect Axis Dimensions Set for 3D Sinking Flux Diagnostic Variables

[yichengt900]

Currently, the 3D sinking diagnostic variables aren't working because the axis dimensions were not set correctly (thanks to @feida6996).

Incorrect Axis Dimensions Set for 3D Sinking Flux Diagnostic Variables

Since we already have an integrated one (i.e., fcadet_arag_100), I'm not sure if we still want to keep those diagnostic variables. Any suggestions are welcome.

[charliestock]

Hello @yichengt900, @feida6996, I'm sorry I missed this when it was raised. We definitely do want to maintain 3D sinking flux diagnostics. I wasn't aware that we were having a problem with this. There is a somewhat cryptic note above this that seems to provide an explanation, but I'm not sure that I follow it. Do you have any output that illustrates the problem?

[feida6996]

Hi @charliestock and @yichengt900, the issue occurs only when writing out the 3D diagnostics. If the axis dimensions aren't fixed, the model won't start, so there's no outputs to compare. Once I fixed the axis dimensions, the model runs fine and outputs are generated. Correcting the axis dimension in cobalt_reg_diag.F90 should resolve this without affecting the model results.

[charliestock]

Thanks @feida6996. I will be diving into the diagnostics later this week and will address this.

[yichengt900]

As @feida6996 mentioned (and Niki's notes above this code section), the register_diag_field wrapper currently only requires the size of the axes, not the actual values. So, if you're trying to output a 3D diagnostic, the size of the axes input needs to be 3 instead of 1. This should be an easy fix—either by changing axesTi(1:1) to axesTi(1:3), or by making it consistent with other 3D diagnostics and using axes(1:3).