SeaDAS atmospheric correction algorithms

[madhusmita]

In L2GEN , 'Processing Options' tab there is 'aero_opt' option (aer_opt (int) (default=99) = aerosol mode option).
There is various atmospheric correction to choose form in this option.
like,
        0: White aerosol extrapolation.
       -1: Multi-scattering with 2-band model selection
       -2: Multi-scattering with 2-band, RH-based model selection and
           iterative NIR correction
       -3: Multi-scattering with 2-band model selection
           and iterative NIR correction

  etc etc.....
Can any body please tell the functional form of all the correction algorithms :sad::sad:

Thanks in Advance.

[OB.DAACx - SeanBailey]

See:

https://oceancolor.gsfc.nasa.gov/forum/oceancolor/topic_show.pl?tid=7026

[disconomer]

Hi Sean,
Regarding your recommendation regarding aer_opt selection in the historical thread:
"Options with positive values are the extreme case for these, in that the user literally selects the aerosol model to use (by its number - not a trivial task to define now that there are 80 of them - in RH order ...a trick I use is to limit the aermodels list to the model I want and set aer_opt=1 that way I don't have to figure out what number in the list is the model I want :wink:) "

I thought it would be relatively straightforward to have a list of aerosol models (i.e. like what is returned when you run l2gen with no arguements, i.e. aermodels (string) (default=[r30f95v01,r30f80v01,r30f50v01 etc... ), and also what is echoed upon a proper run of l2gen with arguments where it lists them in the same order from "model 0" up to "model 79", with corresponding alpha values etc.

Can you please advise whether this list is in order of aer_opt i.e. does aer_opt=1 choose r30f95v01 (model 0) and does aer_opt=2 choose r30f80v01 (model 1) and so-on? I know that during processing it certainly does echo back according to these names in this order, however i'm questioning whether the actual model being used corresponds to this?

The reason why I ask is I have just done a bunch of processing in a FOR loop where i've incremented aer_opt from 20 to 59 and have been performing pixel matchups. We see huge differences between a given f95 and an f80 result, where the f95s are always pretty close to the truth, then f80 is way far away, then the results get progressively closer as we go towards f00 for a given RH.
This seems a bit counter-intuitive to me. Are the alphas listed in l2gen angstrom coefficients? Can you please point me to a reference for the models?

Cheers
Matt

[jared]

Hi Matt,

You are right, aer_opt=1 chooses r30f95v01, aer_opt=2 chooses r30f80v01......

alpha listed in the products tab means polarization rotation angle. angstrom in the tab means angstrom exponent.
I don't know if there is a good reference for those aerosols models.

Good luck,
Minwei

[OB.DAACx - SeanBailey]

In the l2gen STDOUT output, the "alpha" *is* the angstrom exponent for the model....the product 'alpha' is the polarization angle as Minwei described.

The reference for the aerosol models would be:

Ahmad, Z., Franz, B. A., McClain,C. R., Kwiatkowska, E. J., Werdell, P. J., Shettle, E.P., & Holben, B.N. (2010). New aerosol models for the retrieval of aerosol optical thickness and normalized water-leaving radiances from the SeaWiFS and MODIS sensors over coastal regions and Open Oceans. Applied Optics, 49(29). doi:10.1364/ao.49.005545

Sean

[disconomer]

Thanks for this rapid information Sean and Minwei,

So Sean - what is the "sd=" then referring to? How does this relate to the terminology in the Ahmad et. al. paper?
So it turns out that when analysing the matchups I had the good old "0'th vs 1st" index confused, so after the bugfix our matchup results are intuitive. We have excellent agreement for the lowest alphas. I should have caught that bug before I got on the keyboard, or waited till the next day!

Thanks again, and best wishes!
Cheers
Matt

[OB.DAACx - SeanBailey]

Matt,
The "sd" refers to the "size distribution number". The smaller the number, greater the fine mode aerosol contribution.

Sean