Landsat 8 from La Parguera, Puerto Rico
1757*1219, 30 m pixel size, UTM zone 19, october 30th 2015
scene LC80050482013361LGN00, December 27th 2013. Image courtesy of the U.S. Geological Survey

Work done may 2016
home

image october 30th 2015
Co-registration
of 2*2 sum-aggregated PAN band       and of LIDAR DTM
is now more securely implemented in 4SM

home
 
 



under mask_4_ROI
Depth Retrievals down to ~13 m
PAN solution is a confirmed substitute to the GREEN solution
under mask 4

TCC raw under mask 4
mask_4 is shown
profile_white is shown

black ROI is shown
under mask 4
  • mask 4 is too vast
  • for calibration and profiling,
    • we shall concentrate on the black rectangular ROI
 
this is shelf waters off Mayaguez
BLUE vs GREEN
under mask_4

BLUE vs GREEN
K
BLUE/KGREEN=0.82
This is OII+0.7 water type
Maximum depth of bottom detection is ~24 m
  • LsM is for coral sand
  • The brightest pixels are seen to be ~3 times darker than coral sands
COASTAL vs GREEN
under mask_4

COASTAL vs GREEN
Modeling
PAN solution

along profile WHITE

Modeling by the "PAN solution"
The RED profile shows LIDAR depth
 
  •   this plot also shows water column corrected signatures for Coastal, Blue and Green bands
  • turbid shelf waters off Mayaguez cause depth under-estimation
    • like on section A at 20 km
    • like on section C between 16 km and 26 km
PAN and GREEN solutions yield ......
Modeling
GREEN solution

along profile WHITE

Modeling by the "GREEN solution"
The RED profile shows LIDAR depth
  • this plot also shows water column corrected signatures for Coastal, Blue and Green bands





....approximately the same depth
GREEN solution  vs PAN solution
along profile YELLOW

PAN and GREEN solutions yield
approximately the same depth retrieval
  • Minute differences allow for spectral features to appear that the GREEN solution cannot resolve
  • The Landsat 8 GREEN solution mapth depth down to 24 m, while the Landsat 8 PAN solution maps depth down to 18 m in these waters
GREEN solution vs PAN solution
along profile YELLOW

This confirms that the PAN solution
is trustworthy to considerable advantage
under mask_4

ZLIDAR-ZPAN
over the whole scene
  • Calibration only used pixels under mask_4 and was applied to the whole scene
    • BLUE: depth over-estimation
    • RED: depth under-estimation
  • Using the PAN solution, coastal shelf waters  allow for modeling down to 10-13 m in this environment
    • where waters are more turbid than assumed, depth retrievals are badly under-estimated: deep platform waters are more turbid
    • where waters are distinctly clearer than assumed, depth retrievals are badly over-estimated: open ocean waters are clearer: this needs a special study under mask_3
  • Only the shallow shelf off Mayaguez and isolated pinacles exhibit acceptable depth retrievals down to 10-13 m
    • deeper than that: there is probably no bottom detection altogether, apart from the outer oceanic areas
 

ZLIDAR-ZPAN in centimeters
  • light tones for moderate error
  • darker tones for more severe error

GREY used the PAN solution
GREEN used the GREEN solution
 

Z4SM vs ZLIDAR
under mask_4_ROI

Z4SM vs ZLIDAR
under mask_4_ROI
smoothing by a 5 pixels kernel
  • most pixels in the 0-13 m depth range
    • display along the diagonal:
    • NO NEED FOR FIELD DATA for optical calibration
  • although seatruth suggests minor adjustments
    • slight increase of Lm for the PAN band
    • then estimate a tide height for correction

ZLIDAR-ZPAN
under mask_4_ROI

ZLIDAR-ZPAN
under mask_4_ROI
  smoothing by a 5 pixels kernel
 
  • depth underestimation caused by coastal shelf waters shows around 13-25 m
  • there is nothing we can do, unless maybe refer this case to a "semi-analytic" approach?


Z4SM vs ZLIDAR

under mask_5_ROI


Z4SM vs ZLIDAR
under mask_5_ROI

 
 

TCC water column corrected


Command line for the Mask_4 area
This has been a great work.
  • LIDAR: thanks to the LIDAR DTM, I finaly was able to see through this very nice real-life study case.
  • Tools: and it was time for fine-tuning a number of 4SM tools against an excellent seatruth dataset.
    • using these tools at will inside 4SM provides the practioner the greatest opportunity for try and error.
  • Co-registration: quite importantly, achieving adequate co-registration while importing the 2*2 sum-aggregated Panchromatic band, and also the 7.5*7.5 sum-aggregated LIDAR dataset, has been quite a challenge: this is now hard coded into 4SM.
Creating the database, importing the data,
and running the AutoCalibration process
nice -20  ./4SM.6.00 -Process    -Origin/USGS                                   
-DB/laparguera_20151030/40_3S_7S_0s/7_44/1726_1186/665.7900_2011.74000/0_0                   
-Mis/PuertoRico/LaParguera/Landsat/OLIP/TIF/bOA/UTM_+19_008/0.030_0.030/30_OCT_2015               
@MakePIX/Import/Scale/autoCAL                    @-Prepare/Scale                               
-import/v/dTM/data*LC80050482015303LGN00/dbnc_0_0s_1S_0s/R7561_L7731/Origin_599.700_2034.300_0_0/chIn1,7/chOut44,50/import        
@Import/v/dTM/data*LC80050482015303LGN00/dbnc_0_0s_1S_0S/R13169_L9135/Origin_599.700_2034.300_0.030_0/chIn1,1/chOut47,47/import    
@Import/v/DTM/data*PR_BATHY_4M_Mean_NNresample/dbnc_0_0s_0S_1S/R13169_L9135/Origin_665.327_2012.212_0.030_0/chIn1,1/chOut42,42/import    
-SCL/01083/01168/01383/01423/01557/02015/00896                                   
-LS30000/30000/30000/30000/30000/30000/30000                                   
-deglint/vRbaD/q/F1/L6/GlintM25                                           
-AutoCAL/Land_-5_1.3_6                                           

 
Running the project
nice -20  ./4SM.6.00   -Process  -Origin/USGS                       
-DB/laparguera_20151030/40_3S_7S_0s/7_44/1726_1186/665.790_2011.740/0_0       
-Mis/PuertoRico/LaParguera/Landsat/OLIP/TIF/bOA/UTM_+19_008/0.0300_0.0300/30_OCT_2015   
-import/v/dTM/data*LC80050482015303LGN00                           
-LS/30000/30000/30000/30000/30000/30000/30000           
-import/v/dTM/data*LC80050482015303LGN00           
-CP/127.7/095.8/045.9/000.0/015.5/004.8/000.5_15.00m       
-SCL/01083/01168/01383/01423/01557/02015/00896_SCL           
-LsM/160.7/148.3/115.6/117.2/107.0/089.4/199.4_cLM=0.250   
-LsMCS/379.4/396.7/358.6/385.5/381.7/343.4/796.0_CoralSand           
-Lsw/115.4/080.2/032.5/026.5/015.4/004.8/000.5_Lsw_AutoCal       
-dLsw/0-7.1/0-2.4/003.2/001.9/0-0.0/0-0.0/00.0_dLsw       
-La/0090.4/064.6/034.1/025.2/015.3/004.8/000.5           
-M/@000001/@0002/00003/00004/@0005/00006/00007-M       
-Lm/0001.0/001.0/000.3/000.1/000.3/255.0/255.0_Lm       
-Veg/103.5/075.3/050.1/041.9/027.9/126.9/074.4_Veg_AutoCal       
-KK2_3_0.64/Knir4.500/CoefK_1.0/Bm440_BG500_GR600_RN700/mask_3       
-KK2_3_0.82/Knir4.500/CoefK_1.0/Bm440_BG500_GR600_RN700/mask_3       
-Z/MSL0.00d/cSL1.50/N_2/nDR_43_1000_zDTM/mask_3                   
-B/TclNe/cNorm5.00/Bmin0/cLM 0.32 0.45 0.75 1.00                   
-deglint/vRbaD/q/F1/L6/FN/mDEGLINT1/GlintM255                   
-extract/v/rawBDH/FullBDH/NIRband6/NIRmax255/mSOIL21/MBPL4 /Mask_4           
-calibrate/vAI/BdSMpzg/plot_BPL/BC_6_5_3_2/BDh_12_23_24           
-Model/v/runWV2/DTM/makeClouds /Mask_4    -LL10           
-Smooth/1/D/Csmooth3/smart+           
-Tpsv/2/4/Text/Bmin150/RLD4/MSE_2/PanCode   
@-E1/1726/1/1185  @-E1/1726/1/1185               
@-DN300/100/3/comment @-DDshp @-DNshp @ -Eshp @-DD355/178           
@MASK 4=======MASK 4=======MASK 4=======MASK 4=======MASK 4=
@Classify/typerp_1.05/Classifier_40_2_scef_LastChType15/mode_5_16_25       
@ProfileAB/LzrVWb/profile_green/DTM/chAB_1_2_3_4_5_6                
@ProfileAB/lZrvwB/profile_white/DTM/chAB_1_2_3 _16_20 22_26            
@ProfileAB/lZrvwB/profile_white/DTM/3_4/0_201/0_50_5/*               
@ProfileAB/lzRvwC/profile_white/DTM/3_4/0_201/0_50_50/*               
-RegressZZ/v/p/DTM_2/CXY_1.0_0_0/CZR_1.000_0.000/ZR/WZ_3_4/B_0_201/ZC_0_30_50/ALL   
@Recode/GT28_InChannel_32_into_234_OutChannel_8/shallow               
@Recode/234_InChannel_29_into_248_OutChannel_29/248               
@Recode/248_InChannel_8_into_21_OutChannel_8/mSOIL /mask_5           
@Recode/234_InChannel_8_into_240_OutChannel_8/deep               
@Recode/2_InChannel_8_into_4_OutChannel_8/mask_4                
@Recode/1_InChannel_8_into_3_OutChannel_8/mask_3                
@Recode/LT1_InChannel_41_into_0_OutChannel_41                
@Recode/2_InChannel_8_into_3_OutChannel_8/mask_3 /land           
@Recode/2_InChannel_8_into_20_OutChannel_8/dLsw mask_3 /land           
@Recode/234_InChannel_8_into_1_OutChannel_8 /dLsw /land               
@Recode/LT28_InChannel_29_into_29_OutChannel_29/dLsw               
@Recode/GT6_InChannel_32_into_234_OutChannel_8/shallow               
@Recode/5_InChannel_8_into_20_OutChannel_8/dLsw                
@Recode/LT20_InChannel_29_into_20_OutChannel_29/dLsw                
@Recode/LT21_InChannel_8_into_248_OutChannel_8/land               
@Recode/2_InChannel_8_into_1_OutChannel_8/mask_1                

 




 



Créer un site
Créer un site