One question about the results of HICO data using polymer

[southocean]

Hello!

I use polymer-v4.12 to run HICO data, and I've drawn the result as the relationship between each band and the Rrs.


I also used the l2gen tool in seadas to process the same HICO data and got the following results


So, I found that the trend line of Polymer results is obviously different from that of l2gen, which is normally impossible to make such a big difference.The Rrs is even negative in several bands.
I would like to ask whether it is the problem of the Polymer algorithm when it is applied to HICO.

Thank you in advance!

Southocean

[fsteinmetz]

Hello,

Thanks for your feedback.
A few comments about HICO processing with Polymer (which is considered experimental):

• About the evaluated physical quantity: if you are expecting remote sensing reflectances in sr-1, be aware that Polymer provides dimensionless water reflectance, which explains a factor of pi between both results you are comparing.
• Calibration coefficients applied by default in Polymer may not be optimal, in particular after 800 nm where it is clear that calibration is incorrect. This is why these bands are not used for atmospheric correction (as configured in params.py)
• There is no correction of absorption by O2 and H2O, which may explain incorrect results in the red/NIR bands. Again, the bands showing negative reflectances are not used for atmospheric correction.

So for those reasons, I would advise to ignore Polymer/HICO results after 650nm.