Water Column, Bottom

Type, Bottom Depth

Wesley J. Moses &wesley.moses@nrl.navy.mil

Naval Research Laboratory

Wesley J. Moses &Steven G. Ackleson

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Relevant NRL Research Emphases

Algorithm development to retrieve– Water column properties

– Bottom characteristics

Sensor design analysis– Spatial resolution

– Spectral resolution

– Signal-to-Noise Ratio (SNR)

3

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Algorithm Development

Types of algorithms:

– Spectral band ratios

– Look-Up Table (LUT)-based

approach

– Optimal Estimation

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

4

5

Water

0

0.8

1.6

2.4

3.2

400 500 600 700

Wavelength (nm)

Ab

s. C

oef

fici

ent.

CDOM

NAP

Chl-a

Total

Absorption Spectra

0

0.002

0.004

0.006

0.008

400 500 600 700 800

Wavelength (nm)

Rrs

(S

r-1)

Reflectance Spectra708

665

-aRR chl 21

1

(Gitelson 1992)

Algorithms – Band Ratios

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

chl-specific

features

Successful Performance of NIR-red Model

Water Body Chl-a Chlmeas vs. Chlest RMSEmg m-3 mg m-3

NE Lakes 2008 2.1 – 69.2 0.9315x + 1.7169 3.9NE Lakes 2009 4 – 53 0.9204x – 3.2713 5.7Chesapeake Bay 2006 6.2 – 35 1.0002x + 3.0447 3.4Kinneret 2009 4.6 – 21 0.9618x + 0.8356 1.46Azov Sea 2008 – 2010 1.1 – 66.5 1.075x - 2.1475 5.6

0

20

40

60

80

0 20 40 60 80

In Situ Measured Chl-a (mg m-3

)E

stim

ated

Ch

l-a

(m

g m

-3)

Lake Kinneret

Chesapeake Bay

Azov Sea

2008 NE Lakes

2009 NE Lakes

1:1 Line

6

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Band Ratios – still relevant in HSI era7

0

0.002

0.004

0.006

0.008

400 500 600 700 800

Wavelength (nm)

Rrs

(S

r-1)

Reflectance Spectra

low chl

high chl

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

0.005

0.01

0.015

0.02

400 500 600 700

Rrs

(sr-1

)Wavelength (nm)

Phycoerythrin Phycocyanin

HICO image of Sea of

Galilee; 11 Mar 2013

LUT-Based Approach

Coastal Waters Spectral Toolkit (CWST):

Extract subset of parametersexpected to be found in area

[index, parameters, RRS spectra]

PhytoplanktonSediment

CDOM, DepthBottom Type

Input into Radiative

Transfer Model

Parameters:Phytoplankton

SedimentCDOMDepth

Bottom Rrs

Database

3 Component Radiative Transfer

Model (EcoLight)Remote Sensing

Reflectance Spectrum (RRS)

Bottom Reflectance

Sand .1 .2 .3 .4

Brown Mud .2 .2 .2 .2

Yellow Clay .3 .2 .1 .5

Seagrass .1 .1 .8 .2

Pigment Absorption

Chlorophyll .1 .2 .3 .4

Diatom .2 .2 .2 .2

Dinoflagellat .3 .2 .1 .5

Cyanobacteri .1 .1 .8 .2

Metric is Euclidean Distance

Compare measured spectrum to selected spectra to find best

match - takes time

Calibrated At Sensor Radiance

Lee Stocking Island, The Bahamas

Atmospheric Correction

Index Depth Bottom Pigment

111 0.0 14 1

112 0.5 3 1

113 1.0 14 4

114 2.0 14 2

Depth (m)0.0 - 0.5

1.0 - 1.5

2.0 - 2.5

30 - 3.5

4.0 - 4.5

5.0 - 6.0

7.0 - 8.0

> 20.0

8

Courtesy: Jeffrey Bowles, NRL

Optimal Estimation9

Measured Reflectance, Rrs = F(X)

Bottom characteristics (depth, bottom type)

Water-column properties (chl-a, CDOM, SPM concentrations

Atmospheric characteristics (aerosol type, optical thickness, etc.)

X

F Radiative transfer model (e.g., Hydrolight-Ecolight)

Goal: Find the set of xi that correspond to Rrs

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Measured Data

(Rrs)

RT Model

(initial guess for

parameters, Xi)

Simulated Data

(F(X))

Levenberg-

Marquardt

Minimization

Squared

Difference

≤ Threshold?

Yes

Final

Estimated

Parameters

NoAdjust Parameters

Optimal Estimation

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

1012 Apr 2000 AVIRIS;

Kaneohe Bay, HI

Bottom Type Map

Bottom Depth from a Landsat-8

ImageBottom Depth from Lidar Data

Sensor Design Analysis

– Spatial resolution

– Spectral resolution

– Signal-to-Noise Ratio (SNR)

11

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

12

Spatial Resolution

11th Aug

2015

Landsat-8Baltic Sea

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

13

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

14

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

15

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

16

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

17

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

18

30 m

Spatial Resolution for Coastal Remote Sensing19

• Beyond GSDt, decreasing the GSD results in only marginal

gain in spatial information

• Below GSDt, there is significant gain in spatial information

• Ideally, spatial resolution should be much lower than GSDt.

Region of moderate change in

𝐶𝑉𝑎

Region of steep increase in 𝐶𝑉𝑎

Transition

Region, GSDt

~ 200 m

Ground Sampling Distance, GSD (m)

Signal-to-Noise Ratio20

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Trade-Off

Impact of Signal-to-Noise Ratio

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

21

SNR ~ 150

SNR ~ 500

Bottom Depth Retrieval Uncertainties Imposed by SNR22

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Water: Cchl = 0.1 mg m-3; ag,450 = 0.2 m-1; Ccal = 0.3 g m-3 Bottom: 100% CoralReference:

MaxMin

SNR = 100 MaxMin

SNR = 500 MaxMin

SNR = 1000

Unconstrained Water Constituents Constrained Water Constituents

Optical Depth

Shallow Water Autonomously Navigating Surveyor

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

23

Autonomous Kayak and Sensors

Satlantic HyperPro

(349 – 804 nm; D = 3.34 nm)

Articulating

Electric Motor

Navigation & Sensor

Control Electronics

Deck Aft-Looking

GoPro Camera

Down-Looking

GoPro Camera

Depth Sounder

Lowrance Multibeam

Side-Scan Sonar

(455 & 800 KHz)

25

Contact:

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard

Spectral Resolution for Bottom Detection at SNR = 40026

75% Coral Cover0.5 m Depth

Noise Equivalent

1.5 m Depth

1 m Depth

Noise Equivalent

2 m Depth

Noise Equivalent

58 nm

42 nm

28 nm13 nm

07 Aug 2018; Chesapeake Bay Workshop; NASA Goddard