CoDICE: L3b Hi pitch angle distributions: Bug with unit vector computation - Card 3151

[mstarkey2158]

Since the latest updates to the l2_hi-sectored-intensity (fixed dimension ordering of spin angle vs elevation angle), the l3b_hi-pitch-angle data mismatch. The mismatch is traced to a misalignment of elevation angle and spin angle when computing the normal look direction unit vectors. In the code, the elevation_angle (1D array) is broadcasted to a 2D array to be compatible with the 2D spin_angle array. The resulting elevation_angle array should be transposed so that the appropriate dimensionality agrees with the spin_angle array.

I have recreated this issue in my own testing code and successfully reproduced the current L3b hi pitch angles (20260402_v023).

This is a blocker to the release of the hi-pitch-angle-distributions product.

[pleasant-menlo]

We've updated the L3 pitch angle logic to use transposed spin angles so that the calculations match the output prior to the L2 changes. We've attached output using the latest code for your reference; does this look good?

imap_codice_l3b_hi-pitch-angle_20260402_new_code_new_data.zip

[mstarkey2158]

We've updated the L3 pitch angle logic to use transposed spin angles so that the calculations match the output prior to the L2 changes. We've attached output using the latest code for your reference; does this look good?

imap_codice_l3b_hi-pitch-angle_20260402_new_code_new_data.zip

Instead of transposing the spin_angle array, you should transpose the elevation_angle array once you broadcast the 1D elevation_angle to a 2D array. If you transpose the spin_angle array then the actual data itself doesn't match up in dimensionality.
Indeed, we still mismatch with the file posted here.

I can jump on a quick call if needed to get this sorted.