Wrong/No Cloudcoverpercentage and "Damaged" Band 10

Hi all,

I only use Sentinel 2-Products without clouds. Therefore I use the scihub to detect only products without clouds. But now I got a product with clouds (S2A_MSIL1C_20170316T102141_N0204_R036_T25CDT_20170316T102141.SAFE). Is there any place where I can report this kind of pictures thus there will be add to an other class or something like this?

Furthermore @julius and I recognized that in products with no clouds band 10 looks really noisy with kind of periodic streaky structure. I will add an picture.

1 Like

Hi @MaraM
What you’re seeing is an intra-detector noise pattern in the Across Track (AC) direction. This is covered under section 6.3 of the Data Qualtiy Report that can be found in the S2 User Guide Document Library. Although I note that the site is currently down for maintainance, so I’ll put the text here:

This feature can be observed in very dark images (typically on B10 or B12 over the sea). It is characterized by across-track lines covering a whole detector. The typical range of this noise pattern is a few digital counts, and therefore within the requirements of the mission.
This phenomenon is induced by the compression noise on “blind” pixels used for dark signal correction. A solution to filter out this noise has been identified and its operational implementation is currently under study.



S2 MPC/CC Operations Manager

1 Like

Only for clarifying a little more about the band 10: it is a band with very strong absorption of water vapour. The water vapour is concentrated in the lower layers of the atmosphere, so if the radiance is not reflected before it arrives to these low layers (by mountains or high altitude cirrus) it is going to be completelly absorbed and we will see only noise. It is a very useful band for the cirrus detection, but if there are no cirrus, it is not very useful. For this reason it is removed from the level 2 products.


Yeah, we use it for cloud-detection. But all in all thank you for clarification. :slight_smile: