Clarification on CAD Score Screening and Use of Extinction Coefficient Uncertainty in CALIPSO V4.51 Aerosol Product

[aakashmodis]

Hi,
My name is Aakash, and I am currently working on aerosol–cloud interactions. I would like to raise a query regarding the CALIPSO Version 4.51 aerosol product (CAL_LID_L2_05kmAPro-Standard-V4-51).

1. CAD Values for aerosol extinction coefficient screening:
In the user guide of the above mentioned product it is documented that the sign of the CAD score indicates the feature type: positive values signify clouds, whereas negative values signify aerosols. The absolute value of the CAD score provides a confidence level for the classification and larger the magnitude of the CAD score, the higher our confidence that the classification is correct (https://www-calipso.larc.nasa.gov/resources/calipso_users_guide/data_desc/cal_lid_l2_profile_v4-51_desc.php). The threshold values are reported as confidence levels based on |CAD score|: high (≥70), medium (50–69), low (20–49), and no confidence (<20)(https://www-calipso.larc.nasa.gov/resources/calipso_users_guide/data_desc/cal_lid_l2_vfm_v4-51_desc.php).

However in one of the clarification reply in the EarthData Forum I have noticed that CALIPSO team suggested
“ #Filter CAD_Score to get only values greater than -20
data_cad[data_cad <= -20] = np.nan “
(python code provided by ASDC, ASDC - cheyenne.e.land » Fri Jan 12, 2024 9:42 am).

So what threshold should I consider for confidence?



2. Valid range of Extinction_Coefficient_Uncertainty_532:
The Extinction_Coefficient_Uncertainty_532 is defined as uncertainty in the particulate extinction coefficients for 532nm reported for each profile range bin in which the appropriate particulates were detected; these are absolute uncertainties, not relative, thus the units are identical to the units of the particulate extinction coefficients. But the valid range is 0-99.99, which seems unphysical in the context of the valid range of Extinction_Coefficient_532 (0.0 to 1.25).
I tried to derived relative uncertainty (rel_un = {Extinction_Coefficient_Uncertainty_532 / Extinction_Coefficient_532}*100) and screening data by applying rel_un < 20%, but the number of observations are reducing extremely (more than 95%). Can you clarify how one can use this product to minimize the uncertainty associated with Extinction_Coefficient_532.

[ASDC - cheyenne.e.land]

Hello,

1. CAD Values for aerosol extinction coefficient screening:
This user particularly just wanted assistance in plotting the extinction coefficient profiles vs altitude with a CAD score greater that -20, so we provided them code to filter out anything less than or equal to -20.

2. Valid range of Extinction_Coefficient_Uncertainty_532:
We have contacted the CALIPSO Science team and will get back to you as soon as we have an answer.

Thanks,
ASDC

[aakashmodis]

Thank you for the clarification.
Understood — So, there's no rigid threshold for the CAD score for scientific use, the higher the CAD magnitude, the better the confidence.

Appreciate the guidance!