Chl-a results for OLI data

[africaflores]

l2gen Chl-a results for OLI data

I processed a Landsat 8, OLI  image during an algal bloom event in an inland water body and I'm getting no- Chl-a data in what seems to be the highest Chl-a concentrations (based on RGB image). There is Rrs data for those pixels. Do you know why this is happening?

Thank you in advance for your support.

Africa F.

[OB SeaDAS - knowles]

Hover over the pixels and view the flags in the Pixel Info view window (click the "F" inside that window and perhaps toggle off the rest of the views within that window so you can see all the flags.  You can also drop a pin on the pixel of interest, select it and view the flags)

(See attached image for example)

Danny

[OB SeaDAS - knowles]

The flags will help reveal why you are getting no data.

Danny

[africaflores]

Thank you!
For one image I got two reasons for this, Chlorophyll algorithm failure and Chlorophyll out-of-bounds (<0.01 or >100 mg/m3).
For another image where Chl-a was higher, I got more no data values, this time due to atmospheric correction failure, exactly at the higher Chl-a concentrations points. In this second scene, with significantly higher Chl-a values I got no data for almost all the water body. Is there a way to prevent this in the configuration for l2gen?

Your support is really appreciate it.

Africa F.

[OB SeaDAS - knowles]

There are thresholds and atmospheric correction options in l2gen which you can adjust which might possibly yield more retrievals.  This would go outside of our  default scientific quality but are available.

Danny

[africaflores]

Thank you, Danny.

I did modify the chlthreshold to 200 e.g, but still got the same results. Do you have any recommendations about which thresholds to modify?

Bes regards,

Africa F.

[africaflores]

Hi again,
Using Math Band and following Hu et al 2012 Chl_OC4 algorithm I generated a Chl-a product. I got extremely high (no real) values in the areas with more presence of Chl-a, per optical image. Another disagreement is that in the areas with lower Chl-a concentration both products do not match, or at least have more similar values. Disagreements are in the order of 10 in those low Chl-a areas. What could be the reason(s) ?

In addition, can we assume and confirm that since this algorithm hasn't been calibrated for such high Chl-a concentrations retrievals are not appropriate and actually fails when retrieving values there? Which is in accordance to literature that suggest to use algorithms that include red bands when Chl-a concentrations are above 100 mg/m3.

I would still like to get some guidance about which parameters to modify in l2gen to at least get the Atmospheric correction work in OLI bands when there are algal blooms, so we can use that information to generate a Chl-a product using other algorithm.

Your support is greatly appreciate it.

Best regards,
Africa F.

[OB.DAACx - SeanBailey]

Africa,

First, the chlthreshold parameter has nothing to do with the chlorophyll algorithm, in fact isn't even used by l2gen (it is used by other programs in the l2gen "suite"...it's complicated...)

Yes, when you can walk on the water, or it looks like pea soup, it is very likely that most of our chlorophyll products will produce erroneous results - or fail to produce a result at all.
Typically, the part of the atmospheric correction that "doesn't work" in extremely turbid/eutrophic waters is the aerosol retrieval.  There are several alternatives to the default processing
methods that can be tried (via the aer_opt parameter):

aer_opt (int) (default=99) = aerosol mode option
      -99: No aerosol subtraction
      >0: Multi-scattering with fixed model (provide model number, 1-N,
           relative to aermodels list)
        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
       -4: Multi-scattering with fixed model pair
           (requires aermodmin, aermodmax, aermodrat specification)
       -5: Multi-scattering with fixed model pair
           and iterative NIR correction
           (requires aermodmin, aermodmax, aermodrat specification)
       -6: Multi-scattering with fixed angstrom
           (requires aer_angstrom specification)
       -7: Multi-scattering with fixed angstrom
           and iterative NIR correction
           (requires aer_angstrom specification)
       -8: Multi-scattering with fixed aerosol optical thickness
           (requires taua specification)
       -9: Multi-scattering with 2-band model selection using Wang et al. 2009
           to switch between SWIR and NIR. (MODIS only, requires aer_swir_short,
           aer_swir_long, aer_wave_short, aer_wave_long)
      -10: Multi-scattering with MUMM correction
           and MUMM NIR calculation

Options -99, -6 and -8 are one's I would play with. 

Sean