A time series of Landsat 8 OLIP images
at Nosy Be, Madagascar

work done in may 2020 at 15 m GSD
Using the Panchromatic band for water column correction
to derive water depth and spectral bottom signature:

Landsat 8 OLIP bandset used for this work

Purple=1Blue=2Green=3PAN=4Red=5NIR=6 and SWIR1=7

   peer-reviewed and published in 2017    

image aquired October 11th 2019
The data

notice the presence of and scattered clouds
and a thick layer of haze

some white caps are observed over open waters

TCC raw

Deglinting is excellent, but noisy
Glint is strong
 actually, it is caused by haze 
glint regressions are very good

Optical calibration

Calibration Blue, Green, Red and NIR
no stratification is observed

Calibration Blue, Green, PAN, and NIR
Hardly any stratification is observed
0-29 m Oceanic II+0.3 water type of Jerlov

X2,3 vs X5
Notice that a very good fit 
of the blue/red and green/red 
pairs of bands
distinctly requires 
that Kred be decreased
X2,3 vs X5
  • This is obtained here by setting CoefK=1.0 for the Red band
  • This might be related to the very diffuse downwelling light flux caused by the haze and scattered clouds


Computed depth in centimeters


image WZ
  • The PAN solution only reaches down to max 11 m
  • The Green solution reaches down to ~20 m 

image B: bottom brightness
This actually shows bottom brightness in the Blue band 

image WCC
UltraBlue, Green, and Blue (BGR)

Color composite of water column corrected reflectances

Radiances along Profile Blue
  • Radiances tend to increase eastward, as haze becomes thicker (or is it real glint?), and waters become choppy
  • Deglinting is quite efficient, although quite noisy
    •  smoothing shall be necessary


Location of profiles
Lsw radiance is sampled from the small blue rectangle

Modeling along Profile_red
Please note the prominent variations of the spectral bottom signature for the UltraBlue, Blue and Green bands

Green solution versus PAN solution
still needs minor adjustment

Créer un site
Créer un site