issue: chl_hu: incompatible sensor wavelengths for this algorithm for l7etm

[zhigang]

Greetings,

I have been trying used the l7etm bundle to process L7ETM+ data in Ubuntu 18.04, however the program always throws an exception:

....
Aerosol selection bands 835 and 2205
Loading aerosol models from /home/zgcao/seadas-7.5.1/ocssw/share/l7empm/aerosol/aerosol_l7etmp
Number of Wavelengths                          6
Number of Solar Zenith Angles                  33
Number of View Zenith Angles                   35
Number of Relative Azimuth Angles              19
Number of Scattering Angles                    75
Number of Diffuse Transmittance Wavelengths    6
Number of Diffuse Transmittance Zenith Angles  33

Limiting aerosol models based on RH.

Using Gordon & Wang aerosol model selection
Using bands at 835.0 and 2205.0 nm for model selection
Extrapolating from 2205.0 nm

Loading water-vapor correction coefficients.

Applying ocean BRDF including:
    Reflection/refraction for upwelling radiance.
    Reflection/refraction for downwelling radiance.
    Morel f/Q
chl_hu: incompatible sensor wavelengths for this algorithm
chl_hu: -1 1 2
....


The command is: l2gen ifile=/home/zgcao/Data/L2/tmp/LE07_L1TP_119038_20181216_20190111_01_T1_MTL.txt ofile=test.nc l2prod='Rrs_vvv' aer_opt=-1

I know the chl_hu needs the band of 443 nm, but ETM+ doesn't have it. Also, I confuse why the chl_hu is used here because I only generated Rrs_vvv.
I look over the common parameters under the ~/ocssw/share/l7etm/, but cannot find the relevant variables. So, I don't know how to solve it.

BTW, the l5tm has same problem.

Best Regards,
Zhigang

[zhigang]

Dear all,

I tried used the atmocor=0 to disable the atmospheric correction to evoid the bands problem of chl_hu before, finnaly, the l2gen did work for the l5tm and l7etmp. [l2gen ifile=LT05_L1TP_026027_20111022_20160830_01_T1_MTL.txt ofile=test.nc atmocor=0 l2prod="rhos_vvv Rrs_vvv" north=47.5 south=46.5 west=-91.5 east=-90]

Here, the results of rhos seemed normal, but Rrs values were NaN, as atmocor=0 disables the aerosol correction even though it makes the rayleigh correction.

Regards,
Zhigang

[OB SeaDAS - dshea]

I get NaNs also.
Since landsat 5 (and 7) are not supported, I can only say that I might look into it. 

don

[zhigang]

Hi Don,

Exactly. after all, the landsat 5 and landsat 7 is not very suitable for the ocean color, and I just wanna test these two bundles.:grin:
Also, the values of rhos_478 and rhos_560 derived the landsat 7 were negative for my tested images (LE07_L1TP_119038_20181216_20190111_01_T1 and LE07_L1TP_119038_20181029_20181124_01_T1).

Zhigang

[rakeshs]

Hello Zhigang,
Chl_hu needs three bands, 443 nm, 555 nm and 670 nm. L7ETM+ don't have 443 nm. Therefore, it is incompatible with this sensor.

You can edit the get_default_chl() function in get_chl.c file present in l2gen to avoid this problem.

In get_default_chl() function add a new case like:
case L7ETM:
        return (chl_oc2(l2rec,Rrs));
        break;
This will make atmocor2 to use chl_oc2 as the default chlorophyll algorithm for this sensor and you will be able to run atmospheric correction on L7ETM+.

Cheers,
Rakesh

[zhigang]

Thanks, Rakesh. I will try it, and I do believe that it can avoid the Chl_hu problem.
But, there are also some problems for Rayleigh-Luts of L7ETM, which results in uncorrect Lr (for example, Lr>Lt).
Maybe, for this problem, we should expect the next release of SeaDAS.:grin:

Best Regards,
Zhigang