ACOLITE for MERIS imagery

[matheushtavares]

Hello!

I'm developing a model to estimate water quality parameters for a turbid coastal lake using MERIS, and starting with the atmospheric correction, I coundn't find many comparative studies of AC processors for MERIS for inland waters, but I know that it is complex due to the lack of SWIR/thermal bands and NIR bands that were added to Sentinel-3. I've seen on the manual that ACOLITE can process MERIS imagery, so I was curious to see if there has been any validation for ACOLITE/MERIS since this implementation?

Thanks,
Matheus Tavares

[quinten]

Hi Matheus

I added the MERIS capabilities at the end of 2021/beginning of 2022, and I have not done any validation for the MERIS outputs myself. I expect them to be reasonably similar to the ones achieved for OLCI (https://doi.org/10.1016/j.rse.2021.112284), as the DSF seems rather robust to the available band set. Perhaps there are ACOLITE users that have looked at the absolute performance of the MERIS processing by ACOLITE.

Within the team we have some MERIS matchups in the Belgian coastal zone from 2002-2012, so perhaps it is worth to do an analysis already using these data. Neither these matchups nor the OLCI paper linked to above would really provide a comprehensive validation for inland or extremely turbid waters.

Which lake are you working on, and do you have any in situ measurements, radiometry or other?

Quinten

[matheushtavares]

hi Quinten

thank you for your quick and comprehensive response! One question I have is how the lacking of the 1020 nm band could impact the retrieval of consistent water-leaving radiance, as apparently it was the most important in correcting for the atmospheric effects on S3.
I'm working on lake Mangueira, it is an oligotrophic lake in Brazil with a high load of suspended matter, making it turbid in spite of its low nutrient/chla content. Unfortunately we only have radiometric measurements starting in 2021, but there is a large dataset of limnological measurements (chla, Secchi disc depth, temperature, TSS, etc) starting in 2000. At first, I'm testing the consistency of the water-leaving radiances provided by different AC processors. I already tested the Bright Pixel Atmospheric Correction (which was very inconsistent) and SeaDAS, which produced a relatively high amount of negative pixels and was also unsatisfactory. A second validation would be by deriving the water quality parameters with models and assessing their consistency with our data.
I've read a few papers that derived water quality parameters from the Rayleigh-corrected MERIS water radiance, so I thought no AC processor worked well in inland waters for MERIS. In the future I'll test ACOLITE and see how it goes. Any suggestions regarding the ACOLITE processing options?

Thanks a lot!
Matheus

[quinten]

Hi Matheus

I think you may get reasonable result using ACOLITE for your lake. I would recommend using a fixed aerosol optical depth and crop the scene to a region of interest around your lake, for example:

Code: Select all

limit=-33.875,-53.5748,-32.5213,-52.0148
dsf_aot_estimate=fixed

This seems to provide good reflectance data:

matheus_meris.jpg (361.13 KiB) Viewed 62877 times

Turbid water chlorophyll retrievel and turbidity also seem reasonable, but perhaps you have better algorithms than the default ones in ACOLITE:

EN1_MERIS_2011_01_14_12_57_16_L2W_chl_re_gons.png (556.96 KiB) Viewed 62877 times

EN1_MERIS_2011_01_14_12_57_16_L2W_TUR_Nechad2009_665.png (413.37 KiB) Viewed 62877 times

Quinten

[matheushtavares]

Awesome, thank you!!