euroargodev / clean_ctd-rdb_for2021v01release Goto Github PK

Spotted by @Anto79124 @ OGS
To do:

• add default values to the functions in case that data.qclevel and data.source does not exist (Antonella's code snippet)
• add an option to box_copy.m to assign the same source value to all profiles if the variable does not exist.

Problem: The content duplicate check code has a step where the profiles that are suspicious of being content duplicates are found by comparing the deep portion of the profile (deeper than 800 db). Afterwards, the entire profiles are compared.
Since profiles are shallower in the MedSea, all profiles are check against all profiles and then the duplicate check takes too long.
Therefore the profile portion to be checked to find those suspicious profiles should be defined by the user.

Problem found by: @Anto79124

To do: Identify where this needs to be changed and add an argument to the deep content comparison function

We talked sometime in the past (Birgit, Christine and probably Cécile) that we would like to standardize the maximum vertical resolution of the profiles. A single profile with too many samples (resolution better than 1db) result in ctd_xxxx.mat matrices that are way to large (with most of the values equal to NaN) which slows down the OWC software run.