Inland water algorithms
Candidates and challenges for Diversity 2
Daniel Odermatt, Petra Philipson, Ana Ruescas, Jasmin Geissler,
Kerstin Stelzer, Carsten Brockmann
1
Context
Chapter 3: Atmospheric Correction
Chapt
Processing step
Methods/Algorithms/Tools
GlobAlbedo atmospheric correction
3.6.1 Atmospheric Correction over Land
SCAPE-M for land
ATCOR 2/3
Durchblick
MERIS lakes and C2R
CoastColour
3.6.2 Atmospheric Correction over Inland Water
SCAPE-M for inland waters
Normalizing at-sensor radiances
MERIS lakes and C2R
3.6.3 Adjacency Effect Correction (Water)
ICOL
SIMEC
2
Context
Preclassification
Chapt
Processing step
Methods/Algorithms/Tools
Separation by Ecoregion
4.1 Differentiation of water types
Optical Water Type classification using Fuzzy Logic
Applicability ranges of water constituents
IOP retrieval by spectral inversion
4.2.1
algorithms
FUB
MIP
Quality
HYDROPT
Linear matrix inversion approaches
4.2.2 Feature specific band arithmetic
4.2.4 Valid Pixel Retrieval
FLH / MCI
Band ratio algorithms
Flagging
Statistical filtering
Quantity
Chapter 4: Lakes Processing
C2R / CC
4.2.3 Lakes Water Temperature
ARC-Lake Data Set
4.3.1 Water Extent
SAR Water Body Data Set
4.3.2 Lakes Water Level
ESA river and lake dataset
3
Content
 Introduction
– Algorithm requirements
– State of the art
 Atmospheric correction
– Normalizing TOA radiances
– Aerosol retrieval over water: C2R
– Aerosol retrieval over land:
Scape-M
 Water constituent retrieval
– Band ratios
– C2R
 Conclusions
4
Introduction
 Atmospheric correction requirements
– Unsupervised, automatic processing
– Computationally feasible
– Validated use with MERIS images
 Water constituent retrieval requirements
– … as above
– Applicable to optically deep inland waters
– Chlorophyll-a and suspended matter products
5
Introduction
1
2
3
mg/m3
0-3
3-10
>10
g/m3
0-3
3-30
>30
m-1
0-0.8
0.8-2
>2
6
Normalizing TOA radiances
 Over clear waters, Latm exceeds
Lw by an order of magnitude

Needs correction
 Over very turbid waters, signal
strength increases strongly

Uncorrected analysis possible
 Recent studies achieved better
results with FLH, MCI on Lw
Schroeder (2005), Binding et al. (2010), Matthews et al. (2010)
7
Aerosol retrieval over water: C2R
 Explicit atmospheric correction: MERIS Lakes ATBD
– TOA is corrected for pressure and O3 variations with MERIS metadata
– TOSA consists of aerosols, cirrus clouds, surface roughness variations
– BOA RLw is calculated by a forward-NN
– 2 different models for atmosphere and water training dataset
 C2R background reflectance training range:
– 0.01-100 g/m3 TSM
– 0.01-43 mg/m3 CHL
– 0.003-9.2 m-1 aCDOM
 Extended CoastColour training range:
– 1000 g/m3 TSM
– 100 mg/m3 CHL
Doerffer & Schiller (2008)
8
Aerosol retrieval over water: C2R
Odermatt et al. (2010)
9
Aerosol retrieval over land: Scape-M
 Scape-M
–
Modtran compiled LUTs
–
DEM input for topographic corrections
–
Retrieving rural aerosols and water vapour over land
–
Using 5 vegetation-soil mixture pixels per 30x30 km
–
Atmospheric properties are interpolated over lakes
–
Max. 1600 km2 lake area and 20 km shore distance
Guanter et al. (2010)
10
Aerosol retrieval over land: Scape-M
Guanter et al. (2010)
11
Summary: Atmospheric correction
 Automatic and accurate correction required in most cases
 Aerosol retrieval over water
–
Provides RT-based, accurate estimates for low reflectances
–
Application limited by training ranges
 Aerosol retrieval over land
–
Valuable backup for certain niches, e.g. retrieval of secondary chl-a peak
–
Limited by atmospheric, geographic and limnic constraints
12
Band ratio algorithms
 Derived through empirical regression or bio-optical modeling
 Retrieve CHL, TSM, CDOM individually
 Primary CHL-absorption (OC) algorithms (400-550 nm) not applicable
 Secondary CHL-absorption feature shifting with concentration (681 nm)
13
Red-NIR band ratios for CHL
Gitelson et al., 2011: Azov Sea
14
TSM sensitive bands
a
b
c
d
e
f
13 g/m3
23 g/m3
62 g/m3
355 g/m3
651 g/m3
985 g/m3
Doxaran et al. (2002): Gironde estuary, Bordeaux
15
CDOM absorption properties
Maritime vs. Baltic
sea aerosols
 Ambiguity can occur with all other optical parameters
 Angstrom variations over continents
 Band ratios make use of 2-4 bands of the visible spectrum
Maritime vs.
inland aerosols
 Methodological convergence is not significant
Watanabe et al., 2011; Kusmierczyk-Michulec & Marks, 2000
16
Algorithm validation ranges review
To which optically complex waters do recent “Case 2” algorithms apply?
The literature review includes:
–
Matchup validation studies
–
Constituent retrieval from satellite imagery
–
Optically deep and complex waters
–
Explicit concentration ranges and R2
–
Published in ISI listed journals
–
Between Jan 2006 and May 2011
These criteria apply to a total of 52 papers.
Odermatt et al. (2012)
17
Algorithm validation ranges review
The literature review aims to:
–
Quantify the use of recent algorithms and sensors
–
Derive algorithm applicability ranges for coastal and inland waters
–
Clarify the ambiguous use of attributes like “turbid” and “clear”
Authors
Chapra & Dobson (1981)
Oligotrophic Mesotrophic
Eutrophic
Hypereutr.
0-2.9
2.9-5.6
>5.6
n.a.
Wetzel (1983)
0.3-4.5
3-11
3-78
n.a.
Bukata et al. (1995)
0.8-2.5
2.5-6
6-18
>18
Carlson & Simpson (1996)
0-2.6
2.6-20
20-56
>56
Nürnberg (1996)
0-3.5
3.5-9
9-25
>25
0-3
3-10
>10
?
This study
Odermatt et al. (2012)
18
CHL band ratios
5 SeaWiFS
2 MODIS
1 GLI
8 MERIS
2 MODIS
1 HICO
2 TM/ETM+
1 MERIS
Odermatt et al. (2012)
19
TSM band ratios
5 empirical
5 semi-analytical
Odermatt et al. (2012)
20
CDOM band ratios
Odermatt et al. (2012)
21
Spectral inversion algorithms
Authors
Algorithm
CHL [mg/m3]
TSM [g/m3]
CDOM [m-1]
max
min
max
min
max
min
Binding et al. (2011)
NN algal_2
70.5
1.9
19.6
0.8
7.1
0.5
Cui et al. (2010)
NN algal_2
16.1
0.7
67.8
1.5
2.0
0.7
Validation of C2R/algal_2/(FUB):
Minghelli-Roman
et al. (2011)
NN algal_2
9.0
0.0
-
-
-
-
NN C2R
70.5
1.9
19.6
0.8
7.1
0.5
Giardino
et al. for
(2010)
NN C2R
– Adequate
low to medium concentrations
74.5
11.7
-
-
4.0
1.3
Matthews
et al. (2010)
NN C2R
– Inadequate for high concentrations
Odermatt et al. (2010)
NN C2R
247.0
69.2
60.7
30.0
7.1
3.4
9.0
0.0
-
-
-
-
Schroeder et al. (2007)
NN FUB
12.6
0.1
14.3
2.7
2.0
0.8
Shuchman et al. (2006)
coupled NN
2.5
0.1
2.7
1.3
3.5
0.0
MIM
2.2
1.3
2.1
0.9
-
-
Odermatt
– Limited
et in
al.number
(2008) and independence MIP
4.0
0.6
-
-
-
-
Santini
et al.
(2010) to “domestic” use
– Often
restricted
2 step inv
5.0
1.8
13.0
3.0
0.8
0.1
Van der Woerd & Pasterkamp (2008)
Hydropt
20.0
0.0
30.0
0.0
1.6
0.0
Binding
et al. (2011)
– Numerous
and independent
Validation
of (2007)
other algorithms:
Giardino
et al.
Odermatt et al. (2012)
validated | falsified | threshold R2=0.6
22
concentration
level
CDOM
Variability range scheme
high
TSM
type
contravariance
Reading example:
D‘Sa et al. (2006)
retrieve low
with 510, 565 nm bands
at 0.3-13.0 mg/m3 CHL
and 0.5-5.5 g/m3 TSM
Odermatt et al. (2012)
medium
510, 565 nm
D‘Sa et al., 2006
low
23
Variability range scheme
red-NIR band ratios
for very turbid TSM
red-NIR band ratios
for eutrophic CHL
NN for intermediate
concentrations
band ratios
for CDOM
OC band ratios
for oligotrophic CHL
Representing coastal
waters of mostly covarying constituents
Odermatt et al. (2012)
24
Diversity recommendations
wc retrieval:
- FLH, MCI
- Gitelson 2/3-band
atm. correction:
- none
- SCAPE-M
wc retrieval:
- FUB
- blue-green bands
atm. correction:
- C2R (+ICOL!)
- FUB (+ICOL?)
wc retrieval &
atm. correction:
- C2R
- FUB
25
Conclusions
Conclusions from the validation review:
–
Algorithm validity ranges are defined at high confidence (52 papers)
–
MERIS neural networks are sufficiently and independently validated
–
MERIS’ 708 nm band provides unparalleled accuracy for eutrophic waters
Open issues for use of the findings in diversity 2:
–
How is the required preclassification applied?
•
Based on previous knowledge or on-the-flight?
•
Spatio-temporally static or dynamic? – based on previous knowledge or iterative processing?
•
Should algorithm blending be applied as suggested by Doerffer et al. (2012)?
26
Thank you
27
Aerosol retrieval over water
 Implicit atmospheric correction: FUB
– Coupled water-atmosphere RT model MOMO
– Simulation of 5 optical thicknesses of 8 aerosol types, 4 rel. humidities
– Inversion of TOA reflectance
where
RSTOA is TOA reflectance in 12 MERIS bands
x, y, z are transformed observation angles
θs is the illumination angle
P is surface pressure for Rayleigh correction
W is wind speed
T is transmissivity
And x is a neural network learning pattern corresponding to a set of concentrations
–
(Originally not meant for use for inland waters, e.g. altitude))
Schroeder et al. (2007), Schroeder (2005)
28
CoastColour – Rio de la Plata
L1b RGB
CoastColour AC
L1b band 13 (865nm)
L2 3rd reprocessing
Case2R
SeaDAS l2gen
IGARSS * Munich * 24.07.2012
29
CoastColour neural network



Inv NN trained1.04
with 6 Mio. Hydrolight simulations
Rw(10 bands)  IOPs
3
CHL
=
21*apig
for
0.01-100
mg/m
kd calculated from IOPs for all 10 bands
5 IOPs: a_pig, a_gelbstoff, b_ particle; NEW: a_detritus, b_white
3
 SIOP
TSM
(bp1+bp2)*1.73
byofadditional
0.01-1000
rather
g/mhypothetically
accurate IOPs
z90 is=variations
the
average
the for
3 lowest
kd than



Accounting for T=0-36°C, S=0-42 ppt
-1
Corresponds
Secchi
discmdepth
CDOM= ad+agroughly tofor
0.01-4
Random variation spacing for bulk a and b instead of individual IOPs
Doerffer et al. (2012)
30
FUB neural network
 Some conceptual differences
–
Uses MERIS level1 B TOA
bands 1-7, 9-10, 12-14
–
Static 3-component IOP model
–
Forward simulations based on
coupled atmosphere-ocean model
(MOMO)
–
NN for direct RSTOA  IOP inversion
–
NN for indirect RSTOA  RSBOA  IOP
performed similarly
Schroeder, 2005; Schroeder et al., 2007
31
FUB neural network
 Alternative
Bio-optical model
architechture
training ranges
3
– CHL:
Uses MERIS level1
0.05-50
B mg/m
TOA bands
1-7, 9-10, 12-14
– TSM:
0.05-50 g/m3
– CDOM:
0.005-1 m-1
 Partially covarying constituents assumed
Schroeder, 2005; Schroeder et al., 2007
32
Validation examples
 Greifensee 2011 EUT/C2R/FUB
 16 MERIS images within 69 days
 CHL profiles acquired automatically
 Stratified cyanobacteria blooms
Odermatt et al., 2012
33