• A to B in open ocean waters: the Blue bands increase gently: why is that?
• B is located over the sharp Looe Key reef
• B to C accross the clear waters of the lagoon
  • the NIR bands show adjacency effect building up
  • the Red band shows just one peak over LooeKey reef and then faint bottom detection in the deeer lagoon
  • the Green and Blue bands show strong bottom detection over the reef and in the deeer lagoon
  • Deglinting removes the adjacency effect
• C is located at the outer edge of the turbid water plume
• C to D the profile enters the turbid waters plume and proceeds northwards through several keys and channels
  • note a sudden increase of radiances in NIR to Green bands
  • note that the blue[483] band drops to its deep water radiance value, while the blue[443] drops well below its deep water radiance value, while the green band remains high
  • the turbid plume waters originate from inside this maze of keys
• D to F the profile now proceeds eastwards then northwards across the bay
  • first over bright bottoms
  • then in a dark-bottomed area where radiances in both blue bands exhibit "negative bottom contrast"
• F to G out again in the open over bright Florida bay
• G to H the profile now crosses shallow waters flowing from the Everglades marshes
  • bottoms are bright at the center of this estuary
  • but much darker at both sides
• H reaches the Everglades marshes

• K2/K3~=0.39: this is "blue waters" over the outer reef
• very low K values in the blue range are indicative of very clear "blue water": Sargasso Sea is not far away:
  • Water type is OIB: this is known from the Hyperion image above
• WL[green] at 573.2 nm
  • in order to obtain Water type is OIB,
  • we have to set WL[green] at 573.24 nm
  • instead of 565 nm at mid-waveband
• WL[green] not at mid-waveband in Landsat and ALI images:
  • the operational wavelength for the green waveband must be offset at longer wavelength than mid-waveband
  • this is quite an achievement, as it confirms what was observed with most TM, ETM and ALI images
  • possibly just because of the shape of the response curve

• K2/K3~=0.85: this is coastal waters over the Florida bay, under strong influence of waters flowing from the Everglades
• Water type: OIII (this is the same as Coastal 1)
  • WL[green] at 573 nm : OIII+0.08
  • WL[green] at 565 nm : OIII+0.00 at mid-waveband
  • shifting WL[green] is not required for coastal waters, because K[green] is very close to K[blue]
  • it does not hurt though

• Smoothing applied/required to Hyperion data,
  • because of adverse SN ratio
• No smoothing applied/required to ALI data,
  • thanks to excellent SN ratio
  • but ALI has adverse imbalance among its four stripes: this is quite bad for 4SM shallow water modeling

• Densham's profile is 650 m long
  • runs from ~3 m at Looe Key reef
  • to ~14 m along the outer reef slope
  • shows LIDAR measured depths
• That's approximately
  • from waypoint B at ~2 m
  • to waypoint C at ~15 m
  • of the profile on the left
• Note that the "shoulder" at ~10 m is present

This brings a nice support
to our 4SM calibration
using Jerlov's data

• We had to offset the ALI image by 90 m in X and 960 m in Y in order co co-register the ALI image with the HYPERION image
• The calibration of the ALI image uses only the the footprint of the HYPERION image
• The bottom typing training sites used for the HYPERION image were also used unchanged for the ALI image