Hi,

I encounter problems with some MODIS L2_LAC files, for which I see suspects stripes inside.
I found this because the reprojection software I develop reproduce these stripes on the output reprojected images.
I believed this come from a bug in my reprojection software in the handle of the 10-lines scans overlap, but now after investigations I think the stripes are in the L2_LAC.
I know what the scan overlap effect can do when displaying L2_LAC, but I believe these stripes cannot be created by the scan overlap.

On some L2_LAC the stripes are on chlor_a, nLWxxx and other parameters, but on others there are stripes only on nLW_412.
I've 2 examples (I will post them in 2 other posts), one where the stripes (on chlor_a and nLW_412) seem aligned on the 10-lines scan limits, and another where they are only on nLW_412 and only every 20 lines...
I will attach for my 2 examples the JPGs of the interpolated LON/LAT which proove there is not a big scan overlap effect.

What do you think about this?

Thanks in advance,

Julien

First example:

L2 file: A2002178124000.L2_LAC
ROI center (in the swath): (row=485, col=295)
CHLORA min/max: 0.1226 => 0.2448
NLW412 min/max: 680 => 1044 (without scale/offset applied)
LON min/max: 7.7326 => 8.7556
LAT min/max: 43.1076 => 43.623

Warning, it's oriented in the swath, south-up and east-left...

chlor_a

nLW_412

longitude

latitude

Example 2:

L2 file: A2006073130000.L2_LAC
ROI center: (row=675, col=725)
CHLORA min/max: 0.2697 => 0.4463
NLW412 min/max: 522 => 653 (without scale/offset applied)
LON min/max: 0.06205 => 2.0457
LAT min/max: 37.3264 => 38.0399

chlor_a

nLW_412

longitude

latitude

Julien,

You are about 8 years behind on this one.  It is an artifact of the instrument design.  Any  MODIS image will show some level of detector striping and mirror-side banding.  The relative calibration is imperfect, and the relative response of each detector and each each mirror-side is dependent to some degree on the source strength, source spectral distribution, and source polarization.  Significant effort has been made to reduce these relative differences in detctor to detector and mirror-side to mirror-side response, but it will never be perfect.  It is a fact of life for instruments of this design. SeaWiFS does not have this problem because it has only one effective detector per band, and it always looks at the mirror through the same angle-of-incidence. OCTS has bigger problems. 

-- bryan