Bathymetry and water column correction
on a series of 9 Landsat 8 OLI
images
 of Abu Dhabi, UAE
images courtesy of USGS

This is the Marawaah island study case
see also Shuwayat island 


work done in August 2013, then october 2013


 
1 - NO NEED for field data, nor for atmospheric correction
2 - this is demonstrated in this website, using a variety of hyper/multi spectral data
 
Requirements are
1 - homogeneous water body and atmosphere
2 - some coverage of optically deep water
3 - some coverage of dry land
 
Problems are
1 - the precision on estimated depth is found wanting, because the noise-equivalent change in radiance  of accessible data is too high for shallow water column correction work 
2 - radiance data should be preprocessed by the provider at level 1 in order to improve S/N ratio
3 - exponential decay: the deeper/darker the bottom, the poorer the performances
 
So
I keep digging
until suitable data
become available
 

 

The data:
Nine Landsat 8 OLI images
Complex and variable atmosphere
  • clouds, adjacency effect, path radiance increasing from spring to summer, streaks of industrial smoke, and sea-surface glitter all need to be thoroughly removed through the "deglinting" process, based on the observed minimum radiance in the SWIR1 band over optically deep waters
  • it is obvious here that sampling the image for glint regressions requires the whole image
 
SWIR1 band at 1610 nm
  • has extremely low water penetration
  • is very well correlated with other bands
  • therefore is an excellent base for deglinting

April 20th TCC
 

April 20th SWIR1
glint regression       TCC deglinted

May 22nd TCC

May 22nd SWIR1
May 22nd SWIR2
glint regression looks excellent
but deglinting is awfully bad

June 7th TCC
glint regression is excellent,
even the PAN vs SWIR1,
which usually tend to be fuzzy
see SWIR1 : this is a very homogeneous atmosphere==>it deglints very well
Calibration diagram for the PAN band: X2 vs X4

June 7th TCC deglinted
Deglinting is very good, apart from the electronic noise.
The problem at the far east can be corrected for.
Profile White is shown.

 
see deglinting of the PAN band along Profile White
  • Can you apply a formal atmospheric correction to this scene: I'd be interested to compare with my simple "dark pixel" solution.
  • and this potentially shall allow me to validate my estimation of the all-important water volume  reflectance: Lw=Lsw-La

June 23rd TCC
see the SWIR1

June 23rd deglinted
glint regression is excellent

July 9th TCC
see SWIR1
glint regression is excellent
Calibration diagram for PAN band_4
see that RM, Rw and Ra are displayed

July 9th TCC deglinted looks good
BUT WAIT: force the enhancement
 Deglinting can be very frustrating!
 Results can only be questionable

 

July 25th TCC
Stuffy atmosphere in july, no cloud at all
the glint regressions are good

July 25th TCC deglinted looks good
BUT WAIT: force the enhancement
 Deglinting can be very frustrating!
 Results can only be questionable

 

August 10th TCC
can't get suitable glint regressions
no results

There are several atmosphere types here,
each with its own disting set of glint regressions:
I give up.

August 26th TCC
glint regressions were quite difficuly to obtain

August 26th TCC deglinted
Deglintinng is found wanting!!!!

September 11th TCC
glint regressions are splendid
but quite difficult to obtain
atmosphere is very heterogeneous

September 11th TCC deglinted
"so to speak"!!!