C2RCC, C2X and C2X-COMPLEX chl-a uncertainties suspicious output

Optical Toolbox
1

Hi everyone, dear Ana @abruescas

My team and I run the C2RCC variants aiming for the chl-a variable, and we have noticed a weird outcome in terms of the chl-a uncertainties. In particular, regarding the C2X-Complex we notice that the pixel values between the chl-a (conc_chl) and the respective chl-a uncertainty (unc_chl) are almost identical. This is not the case for the C2RCC and C2X variants.

We have checked on the github repo and confirmed (based on the comment) that the chl-a uncertainty units are in mg/m^3 and not a percentage or something else.

Does anybody have any insight on this? Is this a bug or something that we may miss?

Below are the examples of the pixel values (up: conc_chl, down: unc_chl):

C2RCC

image

C2X

image

C2X-Complex

image
image152Ɨ106 4.02 KB

Thank you,

Kostas

2

Dear Kostas,

thank you for reporting tis issue. I have been in contact with the science-team and their conclusion is that for the present code, the answer is:

chl ~= unc_chl because apig ~= unc_apig.

However, there is presumably a bug, which does not scale uncertainty from apig to chl according to the law of propagation of uncertainty.

This will be further investigated.

With best regards,

Tom

1 Like
3

Hi Tom,

Iā€™m really grateful for you answer. It really helps a lot with our decisions.

I will be following the post for any further news.

Kostas