400 and 681 nm peaks for OLCI L2 Rrs retrievals with l2gen

mdco1990 · by **mdco1990** » Fri Dec 14, 2018 7:49 am America/New_York

Hello,
I processed OLCI-A L1B data using l2gen under SeaDAS v7.5. I used the Multi-scattering with 2-band, RH-based model selection and iterative NIR correction option for aerosol mode.
I got high peaks at 400 and 681 nm for all spectra that I compared with in-situ Rrs spectra. Why such peaks at 400 and 681 nm ? Is it due to the fluorescence at 681 nm ?
Does l2gen include any florescence correction by default ?
Thanks.
Mohamed

OB.DAACx - SeanBailey · by **OB.DAACx - SeanBailey** » Mon Dec 17, 2018 2:58 pm America/New_York

Mohamed,
No, there is not fluorescence correction in l2gen.
Without seeing your data, I can't provide a comment about the 400nm peak, however the 681 is very likely due to an indexing bug in the "smile" correction code as implemented in the version of l2gen released in SeaDAS v7.5. The bug has been fixed and we anticipate an update soon.

Regards,
Sean

mdco1990 · by **mdco1990** » Mon Dec 17, 2018 5:30 pm America/New_York

Sean,
Here attached the spectra from l2gen.
The 400 nm peak is quite similar to 681 nm bug. What is 400 nm peak reason for all spectra ?
Best regards.
Mohamed

OB.DAACx - SeanBailey · by **OB.DAACx - SeanBailey** » Mon Dec 17, 2018 6:40 pm America/New_York

Mohamed,

Looks like just ASD spectra were included in the figure you attached - OLCI is not hyperspectral...

Sean

mdco1990 · by **mdco1990** » Tue Dec 18, 2018 5:39 am America/New_York

Sean,
Sorry about that. I attached the right figure.
Regards.
Mohamed

OB.DAACx - SeanBailey · by **OB.DAACx - SeanBailey** » Tue Dec 18, 2018 9:34 am America/New_York

Mohamed,

Yep, that 681 peak is the bug :grin:
There are a number of possible explanations for the 400nm peak, but the most likely is calibration. We have not implemented a vicarious calibration for OLCI yet.

Sean

andrew.meredith · by **andrew.meredith** » Thu Mar 28, 2019 3:16 pm America/New_York

I assume the offset i'm seeing in the attached scatter plot of two Rrs_681 bands created from an OLCI L1B file using SeaDAS 7.4 and SeaDAS v7.5 is related to this bug.

Was the fixed for this bug released in any of the recent SeaDAS v7.5 minor level releases? If not, is there any work around. I tried disabling the smile correction with "rad_opt=0" option but that had no impact of the output L2.

Thanks,
Andrew

OB.DAACx - SeanBailey · by **OB.DAACx - SeanBailey** » Fri Mar 29, 2019 8:31 am America/New_York

The implementation of the smile correction for OLCI went in with v7.5
The fix for the smile correction bug went in with v7.5.2

Sean

andrew.meredith · by **andrew.meredith** » Mon Apr 01, 2019 10:28 am America/New_York

Great. Thank you.

Andrew

andrew.meredith · by **andrew.meredith** » Tue Apr 09, 2019 9:39 am America/New_York

Upgrading to SeaDAS v7.5.3 resolved the issue with the Rrs_681, but i also noted differences in the values of blue bands within L2 files when processing with SeaDAS v7.4 and v7.5.3 (see attached). The release notes didn't mention changes to OLCI other than the smile correction, so I just wanted to confirm that the v7.5.3 values are correct.

Thanks,
Andrew