For dates 20190518 and 20190607, when specifying an area of interest on scihub, I get 2 different L2A and 2 L1C products, which have some overlap and seem to be coming from the same granule. The product and granule ids are the same but with different time stamps eg:
S2B_MSIL1C_20190518T085609_N0207_R007_T35SMD_20190518T110246.SAFE
S2B_MSIL1C_20190518T085609_N0207_R007_T35SMD_20190518T113032.SAFE
L1C_T35SMD_A011473_20190518T090715
L1C_T35SMD_A011473_20190518T090936
In the overlapping area, where my region of interest is located, i get slightly different values for the same pixel between the 2 different products. Is there a reason this is happening? and should i consider one or the other as the correct product? the latest one perhaps?
Your AOI is in the overlap between two Datastrips within a Datatake. You can see this in your screenshot; 2 of Tiles have no imagery at the bottom, and the other two have no imagery at the top. Here’s an image showing the orbital track; you’ll see the area of overlap of the two Datastrips is right over your AOI (Lesvos).
As to ‘values’, it’s not clear which values are of concern to you. However, all values in Sentinel-2 products are within the range specified in the Mission Requirements Document (MRD). So either one is good.
You can find a lot of information about the status and health of the instrument in the relevant Data Quality Report in the Sentinel-2 User Guide Document Library.
I can understand the difference in your case.
I also noted this difference analyzing the overlap area between two granules, acquired at the same date. On Radiance, the difference is 0 (with a few salt and pepper noise sometimes). However, when analysing sen2core corrected images (Reflectance), these values can be more noticed as in the attached image.
The central area is the overlap between two neighbor granules, being white close to 0. If atmosphere correction is causing this, shouldn’t it be performed using Sentinel whole image instead of granules?
