Due to the timeline mentioned I believe this has something to do with this change:

Hi Users,
As some of you might be aware, ESA will be updating the Sentinel-2 processing baseline in January 2022. One breaking change is the definition of digital numbers (DN) in the source files. Negative reflectance values will be possible and DNs before and after the change are shifted and will no longer be comparable.
To keep this transition as smooth as possible, we are introducing the harmonizeValues parameter. This optionally goes is in the input.processing section of your request.
The…

Hi!

I use the available Sentinel-2A images made available within AWS, without needing to change any index values. Can you tell me if the images available after 01.25.2022 are with harmonizeValues ​​= false, because I’m currently verifying a difference in the indexes, with lower values ​​than the old ones. I don’t know where I need to introduce harmonizeValues ​​to correct the reflectance.

I wanted to find out where I have to change my search to have the indexes as before without reflectance or what I need to change in my database and what factor to put to revert the indexes I use to get closer to the old indexes.

Hi @marcusprado! I’m experiencing the exact same issue.

The related change description provides the formula (Copernicus Sentinel-2: Major Products Upgrade Upcoming - Copernicus Sentinel-2: Major Products Upgrade Upcoming - Sentinel Online):

L2A_BOAi = (L2A_DNi + BOA_ADD_OFFSETi) / QUANTIFICATION_VALUEi

The values for BOA_ADD_OFFSET and QUANTIFICATION_VALUE are provided in the metadata.xml file under s3://sentinel-s2-l2a/products/2022/3/1/S2B_MSIL2A_20220301T134209_N0400_R124_T21HWD_20220301T171440/datastrip/0/ (for example). But after applying this formula, the resulting NDVI values seem to be slightly higher than expected (by up to 0.1).

Is this the correct approach @gmilcinski?

The data stored on AWS are not modified in any way, so you need to perform harmonization yourself.

Thanks! Any suggestions/resources I could use to get the correct results?

I would recommend to use Sentinel Hub APIs, obviously, but I am quite a bit biased on this.

he data stored on AWS are not modified in any way, so you need to perform harmonization yourself.

gmilcinski, I still haven’t been able to access the indices as it was before February 2022.

Has there been any progress on the matter?

I really see that the NDVI values ​​are still lower than expected against the images before the update.

I am not sure what progress you are referring to here. In case you are using the data directly on AWS, you will have to implement a process to compensate for the change of the data structure. We cannot help you on this one.

If, however, you are using Sentinel Hub APIs, I would kindly ask you to open a new thread and describe your issue in more details.

Dear @gmilcinski,

according to your response on the data stored in AWS, I would apreciatte if you could explain me why it is not modified by AWS. Would you know if AWS could apply this correction in the future?

Nowadays, I am working with S2-L2A stored in AWS and I am applying the corresponding correction. My question is due to I would need to know if I am going to apply the correction for a specific range of time or until a new deployment of Sentinel-2 processing baseline.

I was looking for information on Sentinel-2 - Registry of Open Data on AWS and the contact address me to Sentinel@AWS - Forum@Sentinel Hub

Thanks in advance

AWS’s Open Data policy is to keep the data as close to the original as possible - ideally exactly the same. This way the users can trust the data is the same as if coming from the official source.

Therefore, there are no modifications done. And note that these modifications might suit one set of users, but not to the others.

There are no corrections/modifications planned to be done on original data.

Thank for your reply.