Hi Quinten,
Do you have any insight to the thought process you used in this particular case to transition from the 0.24 AOT estimate to the 0.12 AOT estimate and particularly why you knew to not go below the 0.12 value? I have a belief that the AOT estimate also impacts the offset parameter for our SDB products during the calibration step. Knowing when and why this AOT estimate changes aside from the obvious would be good to know.
Should I be experimenting with reducing the AOT estimates in targets that return a preponderance of negative reflectances when TOA reflectance is not lower than the computed Rayleigh? Or should more development occur first?
The reason I mentioned previously that negative reflectances were an issue was that we are using the 740 band as a surrogate for turbidity (additional scattering) in very shallow water. Additionally, the ~833 band is often used when computing NDWI-- a useful mask for some of our products. Furthermore, the 704 band can be used in masking aquatic vegetation (Sargassum) using NDVI.
-Bryan
NEGATIVE rhos_ (bit flag 3 value 8) for entire scene Sentinel-2 DSF
Re: NEGATIVE rhos_ (bit flag 3 value 8) for entire scene Sentinel-2 DSF
Hi Bryan
I just processed using a couple AOT values lower than the estimate using "fixed". I did 0.10, 0.12, 0.15 I believe, and some more negatives started to appear at 0.15 so I went with 0.12.
A difference of 0.01 or 0.02 in the AOT will only have a small impact on the final output. This is not a very rigorous process, and could of course not be applied on bulk processing. Perhaps we can also have a look at the provided auxiliary data AOT.
If you have additional scenes with non-zero AOT and retrieved negatives please provide their identifiers as well so I have a collection of scenes for testing. Note that I will be away on field work soon with poor internet connection. I may have time to look into this, but not the ability do download imagery.
For turbidity mapping you could set the NIR negatives to 0, but this is not an option for NIR-SWIR with the NDWI.
Quinten
I just processed using a couple AOT values lower than the estimate using "fixed". I did 0.10, 0.12, 0.15 I believe, and some more negatives started to appear at 0.15 so I went with 0.12.
A difference of 0.01 or 0.02 in the AOT will only have a small impact on the final output. This is not a very rigorous process, and could of course not be applied on bulk processing. Perhaps we can also have a look at the provided auxiliary data AOT.If you have additional scenes with non-zero AOT and retrieved negatives please provide their identifiers as well so I have a collection of scenes for testing. Note that I will be away on field work soon with poor internet connection. I may have time to look into this, but not the ability do download imagery.
For turbidity mapping you could set the NIR negatives to 0, but this is not an option for NIR-SWIR with the NDWI.
Quinten
Re: NEGATIVE rhos_ (bit flag 3 value 8) for entire scene Sentinel-2 DSF
Hi Quinten,
Thank you for the information.
I sincerely apologize for the late reply. I will attempt to get a list of scenes to you by the end of this week.
-Bryan
Thank you for the information.
I sincerely apologize for the late reply. I will attempt to get a list of scenes to you by the end of this week.
-Bryan