BRDF correction S3 OLCI data

Post Reply
spardo
Posts: 2
Joined: Thu Jan 14, 2021 12:32 pm
company / institution: Plymouth Marine Laboratory
Location: Plymouth, UK

BRDF correction S3 OLCI data

Post by spardo »

Hi there,

I'm processing S3 OLCI data from level 1 to level 2 using POLYMER v.4.13. The idea is to compare the POLYMER outputs with the OLCI standard product which, from the user manual, are directional water-leaving reflectances - i.e. not corrected for BRDF effects.

From the literature I see that the POLYMER v.4.13 uses the Park and Ruddick (2005) water model by default, so I'm guessing the POLYMER outputs are normalized water-leaving reflectances i.e. corrected for BRDF effects. Is that right?

Thanks in advance!
User avatar
fsteinmetz
Site Admin
Posts: 306
Joined: Fri Sep 07, 2018 1:34 pm
company / institution: Hygeos
Location: Lille, France
Contact:

Re: BRDF correction S3 OLCI data

Post by fsteinmetz »

Hi Silvia,

Indeed, Polymer's output water reflectances are normalized by default to a zenith-zenith geometry using Park and Ruddick (2005) - the same model that is used for Polymer's atmospheric correction.
This normalization can be disactivated by passing a value to the argument normalize.
Please bear in mind that this normalization step is directional but also spectral : before this step, Polymer works at each detector wavelength.
From polymer/params.py:

Code: Select all

# water reflectance normalization
#   * no geometry nor wavelength normalization (0)
#   * apply normalization of the water reflectance at nadir-nadir (1)
#   * apply wavelength normalization for MERIS and OLCI (2)
#   * apply both (3)
self.normalize = 3
So, in order to compare with the standard product, you may want to disactivate the directional normalization but not the spectral normalization by overriding "normalize":

Code: Select all

run_atm_corr(..., normalize=2)
I hope this helps.
Cheers,
François
spardo
Posts: 2
Joined: Thu Jan 14, 2021 12:32 pm
company / institution: Plymouth Marine Laboratory
Location: Plymouth, UK

Re: BRDF correction S3 OLCI data

Post by spardo »

Hi François,

Many thanks for the information, that's very helpful. I had a quick scan looking for references to BRDF in the code, but I should have used 'normalization' instead! We might leave the POLYMER dataset as it is and correct the standard product, but it's good to have the two options.


Thanks again, take care,

Silvia
Post Reply