trws-Mask3-Calibrate2

What follows is a bit old
since the ScreenCastify tool is so/most practical.
Still worth reading though,
as you want to capture the spiritof 4SM's collection of tools.

Calibration under mBPL_3_to_20
for the whole lagoon

Now we want to cover the whole lagoon:
both mSE=3 and mSE=4

change for...............-Extract/v/mBPL3/mask_4
- this shall only select shallow pixels under mSE>=3 and mSE<=20
- in order to exclude shallow pixels under mSE<3
run the script
- see in X[1-2]vsX[3]
- that -LsM/180.7/169.2/156.3 now is much too low
change for ...............-extract/v/mBPL3/mask_4
change for.........-B/tclNe/Bmin0/LBref170_100/cLM1.15
run the script
- see that -LsM/203.4/192.1/178.5_cLM=1.150 would suit better
- find it on your screen
disable ...........@LsM/180.7/169.2/156.3_cLM=0.95
add.......................-LsM/203.4/192.1/178.5_cLM=1.15
don't forget to change back for cLM1.00...........-B/tclNe/Bmin0/LBref170_100/cLM1.00

What follows now provide more insight
into what we have just done
syntax_Extract

The first calibration was done using only shallow pixels under mask mSE=4.
- it applies to that part of the image which is coded under mSE=4 in channel_mSE_4:
- it is the Red area in the lagoon

Then we wanted to extend the calibration to all pixels inside the lagoon,
- that is areas coded at 3 or 4 (up to 20)
- it is the Red and Orange areas which cover the whole lagoon
For this , we changed for mBPL3 and ensured mask_3 is disabled :
- -Extract/v/mBPL3/mask_3

Now we are in for quite a surprise

We ran a new extraction of calibration data over the whole lagoon
- A new calibration file tarawa-subset_m0.cal was extracted
- it overwrites the previous file tarawa-subset_m0.cal
- now the BPL pixels account for all shallow pixels in the lagoon area,
  - whether under mSE=3 or mSE=4 all the way to mSE=mSOIL-1=20
- that is: pixels with mSE<3 are excluded.
The result is a new calibration diagram for the whole lagoon
- much Brighter: it is seen that some very shallow areas under mSE=3 exhibit fairly brighter BPL pixels in the 0-50 cm depth range.
- much Clearer: it is observed that these very bright BPL pixels appear to be located where some crystal-clear waters flow from the reef rim into the inner lagoon (hoas).
- deeper than ~50 cm, inner lagoon water optical properties seem to prevail
  - which we observed under mask 4

Very shallow and very bright BPL pixels are located at the outlet of the hoas which bring cristal-clear waters from the outer reef inside the lagoon	Yellow/Red tones denote BPL pixels for the pairs 13 and 23. Blue tone denotes BPL pixels for the pair 12
First calibration under mask 4 only	Second calibration under masks 3 and 4

Bulging BPL pixels
Sure you want to see where in the image those bulging BPL pixels are located

This is done by enabling .../MapBPL/... in the -Extract/... argument

NewLsM

It is quite obvious that very shallow BPL pixels (0-0.5 m)
exhibit a distinctly brighter bottom signature:
In order to avoid saturation upon modeling, we had to increase LsM.
We did this by way of CoefLM in -B... argument
NewLsM=La+(LsM-La)*CoefLM

CoefLM
Syntax B

For this reason, in order to avoid excessive saturation of the computed bottom reflectance

we used a CoefLM=1.15 in the -B... argument
-B/tclNe5.00/LBref170_100/Bmin0/cLM1.15
- this "slides" the LsM point up the Soil Line
  - (or down if CoefLM<1.00)
a new -LsM... argument is offered on your terminal

search for NewLsM in your terminal:
- we enabled -LsM/203.4/192.1/178.5_cLM=1.15
- then returned to -B/tclne_2.0/LBref170_100/1/cLM1.00
this new -LsM... argument now accounts for the very bright bottoms that just showed up at some very shallow locations under mSE=3

The calibration now accounts for this new maximum bottom refelctance

before increase of LsM

-LsM/179.6/173.6/174.2
-B/tclNe5.00/LBref170_100/Bmin0/cLM1.15

after increase of LsM

-LsM/202.3/197.3/199.1_CoefLM=1.15
-B/tclNe5.00/LBref170_100/Bmin0/cLM1.00

NewLsM and CoefLM

See that, if CoefLM is given a value >1 or <1,
- a NewLsM is computed and applied:
- it is obtained by sliding LsM up or down along the Soil Line
- its value is shown in the Postscript display,
If found satisfactory, it may be used/enabled
- to replace/override the existing -LsM... argument in the command line.

Why locally Brighter bottoms?

This very shallow extent of pure white bottoms is a very gently sloping lime mudflat
- calcium carbonate precipitation occurs: waters there can be milky at times
- Halimeda sands/muds are ubiquitous in this lagoon
- finer grained entails higher reflectance
- locally cleaner muds entails higher reflectance
- etc

Why locally clearer waters?

This is easier to explain:
- waters over the reef flat are cristal clear oligotrophic ocean waters
These cristal clear oceanic waters intrude into the lagoon
- through "hoas" across the reef rim:
- nice swimming in those streams!

CALIBRATION: CONCLUSIONS

It is really necessary that the practitioner undertakes to gain understanding of the optics of the shallow water scene.
A very nice calibration has been obtained,
- particularly thanks to the availability of a NIR band

Ready for final Modeling

Modeling stratified waters in 4SM
using -CP/... arguments

Very shallow
We want to fine-tune the optical model in the 0-0.50 m depth range

remove any @CP... arguments
run the script
on your terminal, look for the offer of calibration pixels at 0.25 and 0.50 m
add the following to your command Line_10, just below the -M... argument
- @CP/193.96/157.85/065.64_0.25m
- @CP/185.04/130.29/027.71_0.50m

0.25 m

enable................-CP/193.96/157.85/065.64_0.25m
run the script for 0.25 m
- see that........-CP/193.96/157.85/065.64_0.25m .....Green.....is too low
- see that........-CP/193.96/157.85/065.64_0.25m .....NIR........is too high
disable.........@CP/193.96/157.85/065.64_0.25m
add.................-CP/196.96/157.85/058.64_0.25m
run the script
- K1/K2=0.175: Jerlov water type becomes OI:
  - water is as clear as can possibly get!
- BPL fit is much improved in the 0-0.4 m depth range

0.50 m

enable....................-CP/185.04/130.29/027.71_0.50m
run the script for 0.50 m
- see that no change is needed
- deeper than ~1.0 m, BPL model is now parallel to BPL pixels
- K1/K2=0.2589: Jerlov water type OII+0.75 (i.e. almost OIII)

0.75 m

this is not needed anymore

Conclusion

Oceanic I: a very thin layer of extremely clear waters is present at very shallow depth (0-0.5 m)
- Oceanic I water type of Jerlov
- over the brightest bottoms in the area masked at 3 (orange in mask mSE),
- where the bottom substrates are ~15% brighter than anywhere else in the scene.
Oceanic III: waters inside the lagoon exhibit an OIII Jerlov water type

So, what's this fuss all about?

Not much indeed for this SPOT image
Worth trying anyway to show off how well we master the situation!
Might be useful with some complex cases:
- see stratified waters at Figure 2 at San Lorenzo Channel, Baja California

search your session screen display:

using a series of -CP... arguments, one can model "stratified" waters
here with two -CP... arguments, we have
- OI water type , from 0 to 0.30 m
- OIII water type , deeper than 0.5 m
this is presented here to show that 4SM has this capability
actually of not much interest here,
- as we are fiddling within the first 50 centimeters of the water column
- and this refinement seems apply only in the vicinity of a few hoa intlets

Straight BPL
no -CP/... argument

Curved BPL
two -CP/... arguments