Bathymetry and water column correction
LANDSAT 8 OLIP 
at Mulroy Bay, Ireland
Image courtesy of the U.S. Geological Survey
 
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

Work done july 2018
GSD 15 m
 
 
 
 
go to April 10th 2018
ScreenCastify
use Google Chrome
to watch a 54 minutes FullHD record of my screen,
from Import of raw data           all the way to profile_black
complete with explanations and comments.

 


 
April 2nd 2018
PLEIADES
There is still a lot that needs to be understood/mastered about PLEIADES images.
I fear that  PLEIADES's operational agility is playing tricks that I do not master in 4SM.
Processing suitable Landsat 8 images  is one way you may want to explore for your project.
 COSTA TUSCANA 
We shall in the end succeed in our present undertaking
 CONSULTANT 
I am willing to pursue this line of action, as a consultant to Consorzio LAMMA,
 
  • Deep water radiances over this scene are confusing.
    • NIR and Red bands increase steadily from south to north, while Green and Blue bands are not affected.
  • Viewing angle is 10 degrees southward along track and 1 degree westward.
    • Path radiance would decrease northward in all bands (if at al,l as the angle is small)
  • Variations in the content in suspended mineral particles could explain that NIR and Red bands increase steadily from south to north, while Green and Blue bands are not affected.
    • this might be caused by discharge of suspended particles in the northern part of the image (Fiume Ceccina).

Deep water radiances
along Profile_purple,
from South to North
               <VIEWING_ANGLE_ACROSS_TRACK_unit="deg">-1.04
<VIEWING_ANGLE_ALONG_TRACK_unit="deg">-9.99
So: viewing 10 degrees southward
and 1 degree westward
Landsat 8 on July_29th 2017

Landsat 8
ctOLI_192030_20170729_TCC
Bad hydrological conditions
along the Tuscany coast
Landsat 8 on July_29th 2017
Landsat 8
ctOLI_192030_20170729_PAN
The hydrological conditions are
helpless on that day
The PLEIADES image was aquired
on August  3rd 2017
The PLEIADES image was aquired
on August  3rd 2017
Landsat 8 on August 5th 2017

Landsat 8
ctOLI_193030_20170805_TCC
Complex atmosphere  and hydrology.
The situation has improved along the coast of Tuscany
Landsat 8 on August 5th 2017

Landsat 8



ctOLI_193030_20170805_PAN
4SM modeling of a PAN-sharpened image
at 15 m GSD seems  promising

I have been struggling with "First things first".
In other words, with removing any "glint", and estimating
an operational set of spectral deep water radiance Lsw for the PLEIADES image.
This is not under proper control at this time.
Therefore, my results shall depend on bold assumptions in this regard.
Results in the 0-~10 m depth range should prove to be acceptable,
while things can get bad at depth.
I assumed that this sandy/silty highly dynamic environment
does not have much in the way of shallow vegetated substrates.

Final_Depth=Tide_height+Retrieved_Depth*coefZ






April 10th 2018
Progressively, I seem to piece together various bits of information. This takes time.
DEGLINTED-SMOOTHED

Full image,
water only, deglinted-smoothed
  • The display of smoothed BPL pixels is fairly linear, down to approx 20 m: water type OII of Jerlov
    • then, deeper than 20 m, waters would become clearer,
    • up to water type OIB
    • or is it caused by the excessive noise?
DEGLINTED-NOT SMOOTHED

Full image,
water only, deglinted-not_smoothed
  • I think the weird/strong noise observed earlier is the cause of such discrepancies.

Full image
The water gets progressively clearer,
from OII along the water line, to OIB at depth.

Location of Blue and Red ROIs

Blue ROI

Red ROI
The Red ROI
exhibits local hydrologic conditions
  • In this all-sand/gravel highly dynamic environment, and under the BPL assumption, the brightest shallow bottoms are not expected to be darker than anywhere else in that scene over the whole depth range.
  • Deeper than ~5 m, the ratio Kblue/Kgreen decreases markedly; this can be caused by an increase of dissolved organic matter which causes Kblue (and also Kgreen) to increase.
  • The presence of a sewer system in this surrounding would explain both
    • patches of lower deep water radiances, which are only visible in the Blue and Green bands,
    • the calibration diagram for the Red ROI,
    • and, subject to a long-shore southward current, the long stretch of yellow/brownish hues in bottom brightness image.
Patches of lower deep water radiance
They affect only the Blue/Green range:
Red and NIR bands are not affected.


 





Six beach profiles

I could keep fiddling like that for weeks.
More unanswered/hazardous querries/statements would come forth.
Whereas possibly, in the absence of prominent bottom substrate variations,
the main finding is there, in plain, in the beach profiles below.
I refer to Anfuso et al. 2011: Geomorphology 129 (2011) 204–214.
Six beach profiles along Profile Black
F is northern section
D is the longest section: 950 m
A is southern section, 600 m long
  • Subject to favorable sea-truth comparisons, this kind of information should prove suitable for beach profile long term monitoring.
  • Section A would span over 40 Landsat 8 pan-sharpened 15 m GSD pixels , possibly enough for this kind of beach profiling.

Retreived depth
 
ScreenCastify
use Google Chrome
to watch a 54 minutes FullHD record of my screen,
from Import of raw data           all the way to profile_black
complete with explanations and comments.

 
 



Créer un site
Créer un site