Skip to main content

Hi Users


I am getting low NDVI values from Jan 2022.


I am using Sentinel 2 L2A products to calculate the NDVI. Python and restario to read the band values and calculate NDVI from RED and NIR bands.


Could you please help us to resolve this issue.


Thanks

Hi,

Can you provide some more details on how you are calculating NDVI? e.g. your time period, area of interest

Are you doing this using Sentinel Hub Process API or another method?

Many Thanks


Hi

Please find the below information



  • I am using SentinelAPI.query from sentinelsat to query the product and SentinelAPI.download to download the Sentinel L2A data. (The url is https://scihub.copernicus.eu/dhus)




  • I am analyzing data from 2019-2023 mostly for Brazil and USA




  • I am calculating NDVI manually from RED and NIR band from the downloaded images. (NDVI= (NIR-RED)/(NIR+RED))




  • What I have observed is NDVI range for 2022 is less around 0.1-0.2 compare to the previous years (2019-2021)




  • Current resolution is 10m.



Please let me know if you need any other information.


Hi,

This is probably due to ESA updating the Sentinel-2 processing baseline to version 04.00 in January, 2022, which introduced breaking changes to the interpretation of digital numbers (DN).


Hi

Thank you for the information.

Is there any way to get similar information for the images before baseline 4 changes.

I mean can we get these negative radiometry, bias offset for images earlier than Jan 2022?


No, this is not possible.

That said, the ESA is reprocessing the data and by end of the year there should be complete archive available with the same baseline, i.e. data as you see it now.


Okay, thank you for confirming that.


Hi, I have noticed the same issue. Quite incredible when one thinks how critical up to date integral consistent time series for NDVI are, also that Landsat 8 is not working either. Am absolutely amazed.


Did you find any ‘solution’ ?


Reply