seadas processing modifies modis L1A input files, is this normal?
Posted: Wed Jul 30, 2025 4:27 am America/New_York
Hi,
For few NASA sensors, we are processing L1A data to L1C, using SEADAS tools modis_GEO, modis_L1B and l2gen. I have noticed, our processing tool including those 3 SEADAS tools fails if input modis L1A files do not have write permission. When I added it, processing succeeded BUT! ... input L1A got modified: after very quick and surface inspection with SNAP browser, I detected at least one change: two new global attributes were added: `identifier_product_doi` and `identifier_product_doi_authority`. In general, I think this is a bad idea - changing input product, especially with this strange addition. However, maybe this was my fault and I unwillingly turned on this modification feature?
Tested the same with VIIRS data - no changes in the input file after applying the same tools exactly in same order. Note, the first in the line was modis_GEO tool where this collision happened, though there still may have been permission check for the benefits of next seadas tools in line.
My questions, summarizing:
1) is this an intended feature or a bug?
2) is there a way to configure processing avoiding this modification of inputs?
Thank you.
Best regards,
Andrei
For few NASA sensors, we are processing L1A data to L1C, using SEADAS tools modis_GEO, modis_L1B and l2gen. I have noticed, our processing tool including those 3 SEADAS tools fails if input modis L1A files do not have write permission. When I added it, processing succeeded BUT! ... input L1A got modified: after very quick and surface inspection with SNAP browser, I detected at least one change: two new global attributes were added: `identifier_product_doi` and `identifier_product_doi_authority`. In general, I think this is a bad idea - changing input product, especially with this strange addition. However, maybe this was my fault and I unwillingly turned on this modification feature?
Tested the same with VIIRS data - no changes in the input file after applying the same tools exactly in same order. Note, the first in the line was modis_GEO tool where this collision happened, though there still may have been permission check for the benefits of next seadas tools in line.
My questions, summarizing:
1) is this an intended feature or a bug?
2) is there a way to configure processing avoiding this modification of inputs?
Thank you.
Best regards,
Andrei
, the files need to be writeable. It is only metadata that is modified, not the DNs, so mostly harmless. Different code is used for VIIRS - one that doesn't modify the source L1 file 