slcoli-combinedz-2

go to Combined Depth vs Fabio's DTM
go to Combined Depth vs Fabio's Sonar seatruth

Enforcing Combined Depth

Now that I have a new DTM derived for the whole study area by combining retrieved depths from nine scenes,
- I can enforce ZDTM while inverting the simplified radiative transfer equation.
So much for the Soil Line assumption, it is not needed anymore at this stage.
This allows for more contrasted water column corrected bottom reflectance signatures:
- this is a bonus for bottom typing.
- RMSE on retrieved depths now are really representative of the 4SM method, as this new DTM is near-perfect.

CC BOA WCC normalized CoefNorm=5	SAM Spectral Angle Mapper	Retrieved depth vs Combined Depth	Depth Histograms Red is Combined Depth Grey is Retrieved Depth
October 19th 2013	October 19th 2013	October 19th 2013 tide correction -0.6 m RMSE=0.78 m	October 19th 2013 tide correction -0.6 m
November 4th 2013	November 4th 2013	November 4th 2013 tide correction -0.5 m	November 4th 2013 tide correction -0.5 m
January 07th 2014	January 07th 2014	January 07th 2014 tide correction -0.2 m	January 07th 2014 tide correction -0.2 m
February 08th 2014	February 08th 2014	February 08th 2014 tide correction: -0.5 m	February 08th 2014 tide correction: -0.5 m
October 22th 2014	October 22th 2014	October 22th 2014 tide correction: 0.3 m	October 22th 2014 tide correction: 0.3 m
January 29th 2016	January 29th 2016	January 29th 2016 tide correction: 0.3 m	January 29th 2016 tide correction: 0.3 m
March 1rst 2016	March 1rst 2016	March 1rst 2016 tide correction: -0.3 m	March 1rst 2016 tide correction: -0.3 m
October 11th 2016	October 11th 2016	October 11th 2016 tide correction: -0.20 m	October 11th 2016 tide correction: -0.20 m puzzling!
October 27th 2016	October 27th 2016	October 27th 2016 tide correction: 0.00 m	October 27th 2016 tide correction: 0.00 m

Work finished
Andy

Champagne!
Fabio

Uncertainties: here you can appreciate that each of the 9 instances of this time series carries its own lot of unwanted/weird variations
- which need to be sorted out to produce a DTM
- Profile_black: see that combining depths removes most of the noise at depth: Goooooooooooooooood!
Existing nautical-chart/DTM?
- OK, 4SM does not require access to an existing map or depth dataset
  - But, in view of such adverse variations, it certainly helps the practioner to keep his cool
  - And chances are that some info may be gathered from the NET anyway
- Or a blind test has to be dead-blind: then make sure your scene is of exceptional/improbable quality, and be prepared to face the uggly consequences of a bad choice!

Breaking new ground here!
- You are the one to make sense of it all
- Please do not hesitate to let me know your thoughts:
  - what is usefull and needs to be investigated further
  - what is futile and should go to the back burner
Bay of La Paz
- I am surprised how good the combined depth results appear to be
- although I had gathered that this was a lost case, in view of so adverse (anthropogenic?) turbidity in the bay