C2RCC processor problem

Optical Toolbox
#1

Hi there,

I was wondering could anyone help me with that I used S2-L1C resampled data try to get the chl-a result. However after C2RCC the result is wrong as some of the area of my lake are missing (like gone…only show partial of the lake on the screen). How can i fix it???

Also, If i want to compare the in-suit data and the chl-a result which get from express (21.0*(iop_apig)^1.04 is that enough to find the relationship between observed data and modelling results?? Do i need to do other calculation （band math）in snap after C2RCC??

Thank you very much
Sincerely

#2

If part of your data is “gone”, probably it is because the data was not of good quality, and some flags have been raised and applied (clouds, sun glint…).
For comparison of chl predicted vs. in situ you can use a simple linear regression.

1 Like
#3

Hi abruescas,

Thank you for your reply.
However, the raw image (I mean the RGB image) shoed no clouds covered on top, but the C2RCC result still has some part is missing. Can I fixed it through some ways maybe？

Thank you so much.

#4

Please check the provided flags. This might tell you why there is not data.

image
image1360×606 54.3 KB

Maybe the valid-pixel expression does not work for your scene, and you need to adapt the expression.
The expression is meant to exclude land pixels from the processing.

image
image673×528 15 KB

#6

Hi Marpet,

I see what you mean now, some of the value show in the flag are “invalid pos”, and it might need to fixed before run the C2RCC.

BTW, when I compare the OLCI predicted CHL data with the in-situ data, can i used the conc_chl? or it is better to use the data after compute from the express (21.0*(iop_apig)^1.04?

Thank you again.

#7

conc_chl is the result of the expression.
But you might need to tweak the expression with different factors adapted to your region of interest.
See here:
C2RCC Conversion Coefficient for Chl-a and TSM - snap - STEP Forum (esa.int)

1 Like
#8

I see. So I can adapted the chl by change the factor according to my region of interest.
If there is no much change, then I can just use the conc_chl to compare with my observed data to find out the relationship between them.
Am I right and am I understood you clearly?
Thank you

#9

Yes. correct

#10

Hi Marpet,
sorry for bothering your again.
I am trying to fix the S2-L1C project missing area after C2RCC. As you told me that might need to change the valid-pixel expression, however, could you help me to understand what is B8>0 AND B8<0.1 stand for? is that within 0.1-0 range is exclude the land pixels?
thanks.

#11

Yes, only if pixel of Band B8 are within the range of 0 to 0.1 the pixels are considered valid and are processed. This shall exclude all the land pixels.

#12

So I do not need to change the expression, but need to check with the pixels value of Band B8 to double check if it is within 0 to 0.1?
If not, then I need to change the expression ?
Am I right?

#13

Yes.

For testing purpose, you can also set the expression to ‘true’. Then the whole scene will be processed

#14

I see.I will try it now. Thanks.

By the way, for the Ozone, the C2RCC unit is DU, but statistic is KG/M-2, how can I convert it?
For example, 0.006kg·m-2 equal to 600DU?

#15

No, this conversion is not correct. But I’m also no expert for this.
I think this conversation can help you:
How to convert Ozone units from µg/m3 to Dobson Unit? (researchgate.net)

#16

Hi Marpet,

I was wondering expect C2RCC, are there any other options for us to derive the concentration of CHL-a or algal density for S3 products?
Thanks.

#17

There are some other ways.

And more, I guess.