comparison of algorithms to estimate chlorophyll-a in

Comparison of algorithms to

estimate chlorophyll-a in

inland and coastal waters of

Brazil

Milton Kampel, Dr.

Head of Remote Sensing Division

National Institute for Space Research (INPE)

Brazil

1st CoastColour User Consultation Meeting, Frascati 2010

WHO WE ARE

Main civilian organization for space activities in Brazil

Staff of ~2500

49 years in 2010

INPE headquaters, 2004, São José dos Campos, SP

MISSION

Fostering science and technology in earth and space context and

be able to offer products and regular services in benefit of Brazil.

VISION

Become a National and international reference in both space and

earth environment fostering knowledge and attending and

anticipating demands on Brazilian society life quality progress.

Facilities

São Luís MA

Eusébio CE

Atibaia SP

CRAAM SP

Cuiabá MT

Santa Maria RS

Brasília DF

Alcântara MA

Sao Jose dos Campos SP

Headquarters

Cachoeira Paulista SP

CPTEC

Natal RN

Role of Remote Sensing

Long-term:

Time-series

Change detection

Identify anthropogenic contributions

Cal/Val models

Short-term:

High quality information

Support for operational services

oceanography

land cover/change

planning

desasters

GEOSS: Social benefits

http://earthobservations.org/

GEOSS: Social benefits

http://earthobservations.org/

ChloroGIN/Antares

Inland/Coastal RS WG

SAFARI

CEOS Virtual Constellations

Athmospheric Chemistry Air quality, CO2

Land Imaging (Brazil: CBERS, Amazonia-1)

Ocean Surface Topography Climatic variability

Precipitation (Brazil: GPM-BR)

Ocean Colour (Brazil: SABIA-MAR)

Ocean Surface Winds

Common requirements, independent satellites, comparable data (CEOS – Committee on Earth Observation Systems)

INPE´s ground systems

Satellite control center (São José)

Cuiaba image reception station

Oil spill from a ship

ASAR Campos Basin Brazil

Brazil Current

Oil seep

16

Study Area

Conceptual display of Water Masses-Currents with

mesoscale features for SW South Atlantic

AVHRR SST

Cape São

Tomé

V

ST CF

V – Vitória Eddy

ST – São Tomé Eddy

CF – Cabo Frio Eddy

18

FITOSAT I Cruise - Bathymetry & Stations

Kampel et al., 2009

São Tomé Cape

Brazil Current Frontal Eddy

19

In situ - Fluorometry

2 L surface water samples with Niskin bottles

18 stations, 21-25/NOV/2004

Turner Designs TD-700 (Parsons et al. 1984)

0,077 a 0,197 mg m-3

0

10

20

30

40

50

60

70

80

90

100

-1.19 -0.99 -0.78 -0.36 -0.15 0.06 Mais

Log [CSM insitu ]

Fre

qü

ên

cia

(%

)

CSM insitu = 0.121 (±0.039)

N = 18

20

In situ – Above-water Radiometry

29 stations

Hyperspectral Spectron SE590 radiometer

375-1075 nm 400-800 nm (5 nm)

Protocol suggested by Fougnie et al. (1999) with

polarizer

SeaWiFS and MODIS bands were simulated integrating

the radiometric data by the trapezoidal rule.

21

In situ - Above-water Radiometry

Remote sensiong reflectance

OC2v4

OC4v4

OC3M

)(

)()(

d

wRS

E

LR

cd fLE )()(

Eq.1

Eq.2

071,0)135,0879,0336,2319,0( 32

2220,10 SSS RRR

aC490

555102 log RR S

)532,1649,0930,1067,3366,0( 44

34

2440,10 SSSS RRRR

aC max104 log RR S

)403,1659,0457,1753,2283,0( 43

33

2330,10 MMMM RRRR

aC max103 log RR M

22

In situ - Above-water Radiometry

23

In situ - Above-water Radiometry

0

20

40

60

80

100

-1.19 -0.99 -0.78 -0.57 -0.36 -0.15 Mais

Log [Clorofila-a ]

Fre

qü

ên

cia

(%

)

OC4 RAD = 0,222 (±0,098)

0

20

40

60

80

100

-1.19 -0.99 -0.78 -0.57 -0.36 -0.15 Mais

Log [Clorofila-a ]

Fre

qü

ên

cia

(%

)

O3MRAD = 0,226 (±0,103)

0

20

40

60

80

100

-1.19 -0.99 -0.78 -0.57 -0.36 -0.15 Mais

Log [Clorofila-a ]

Fre

qü

ên

cia

(%

)

OC2 RAD = 0,293 (±0,116)

24

LIDAR

0

0,1

0,2

0,3

0,4

0,5

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Cl* Cloro Lab Ist 1-5-8-12-16 2-xoil-9-13-17 3-6-10-14 4-7-11-15

175,0)(*9,4 clLIDAR ICSM

CSMLIDAR = 0,124 (±0,04)

0

10

20

30

40

50

60

70

80

90

100

-1.19 -0.99 -0.78 -0.57 -0.36 -0.15 0.06 Mais

Log [CSM LIDAR ]

Fre

qü

ên

cia

(%

)

CSMLIDAR = 0.14 (±0.04)

N = 31

25

MODIS/Aqua

OC3M GSM01 Carder

St. Tome

Cape

St. Tome

Cape

St. Tome

Cape

0

20

40

60

80

100

-1.1

9-0

.99

-0.7

8-0

.57

-0.3

6-0

.15

0.06

Mais

Log [GSM01sat]

Freq

uenc

ia (%

)

GSM01sat = 0.11 (±0.02)

N = 21

0

20

40

60

80

100

-1.1

9-0

.99

-0.7

8-0

.57

-0.3

6-0

.15

0.06

Mais

Log [Cardersat]

Freq

uenc

ia (%

)

Cardersat = 0.28 (±0.06)

N = 21

0

20

40

60

80

100

-1.19 -0.99 -0.78 -0.57 -0.36 -0.15 0.06 Mais

Log [OC3Msat]

Fre

qu

enci

a (%

)

OC3Msat = 0.14 (±0.02)

N = 21

Comparisons

Algorithm/LIDAR rmse-L rmse

OC2v4RAD

1,36 0,40

OC4v4RAD

0,93 0,28

OC3MRAD

0,93 0,28

OC3MSAT

0,36 0,11

GSM01SAT

0,28 0,08

CarderSAT

1,14 0,34

CSMLIDAR

0,48 0,14

Comparisons

RADARSAT – Banda C

Res: 50 m ASTER b1 (0,5-0,6 µm)

Res: 15 m

CBERS-2 2r1g1b

b1 (0,63-0,69 µm)

b2 (0,77-0,89 µm)

Res: 260 m

Oil Seepage

MODIS – MERIS - AVHRR

MERIS x In situ Chl Comparisons

y = 0,0871x + 0,1012R² = 0,0745

y = 0,1751x + 0,063R² = 0,4659

0,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

0,40

0,45

0,50

0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 0,45 0,50

ch

l in

situ

MERIS chl

algal_1

algal_2

Linear (algal_1)

Linear (algal_2)

DEPROAS Project

DEPROAS Project

OC4 OC2

GSM NN

0

5

10

15

20

25

30

35

40

-1.6 -1.2 -0.8 -0.4 0 0.4 Mais

Log Cinsitu (mg.m-3

)

Fre

qü

ên

cia

re

lati

va

Kampel et al. 2009 - XIV SBSR - Natal

RESULTADOS – CSM Médias Mensais

Tempo (07/2002-09/2007)

jan fev mar abr mai jun jul ago set out nov dez

Clo

rofila

(m

g.m

-3)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

#1 SW

#1 MOD

#2 SW

#2 MOD

#3 SW

#3 MOD

#4 SW

#4 MOD

#5 SW

#5 MOD

#6 SW

#6 MOD

#7 SW

#7 MOD

#8 SW

#8 MOD

#1 – SW=0,99 MOD=0,79 mg.m-3

#5 – SW=0,34 MOD=0,30 mg.m-3

1 2

3 4

5 6

7 8

Kampel et al. 2009 - XIV SBSR - Natal

RESULTADOS

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

Clo

rofil

a (

mg

.m-3

)

0.0

0.5

1.0

1.5

2.0

#1 SW

#1 MOD

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

0.0

0.2

0.4

0.6

0.8

1.0

#2 SW

#2 MOD

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

Clo

rofil

a (

mg

.m-3

)

0.0

0.2

0.4

0.6

0.8

1.0

#3 SW

#3 MOD

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

0.0

0.2

0.4

0.6

0.8

1.0

#4 SW

#4 MOD

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

Clo

rofil

a (

mg

.m-3

)

0.0

0.2

0.4

0.6

0.8

1.0

#5 SW

#5 MOD

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

0.0

0.2

0.4

0.6

0.8

1.0

#6 SW

#6 MOD

Tempo (07/2002-09/2007)

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

0.0

0.2

0.4

0.6

0.8

1.0

#8 SW

#8 MOD

Tempo (07/2002-09/2007)

Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07

Clo

rofil

a (

mg

.m-3

)

0.0

0.2

0.4

0.6

0.8

1.0

#7 SW

#7 MOD

Caixas rmse-L RDP r2 Declividade Intersecção N

#1 0,41 -19,47 0.83 0.78 0.02 63

#2 0,24 -9,07 0.86 0.93 -0.01 63

#3 0,26 -12,93 0.83 0.84 0.00 63

#4 0,28 -10,13 0.89 0.96 -0.01 63

#5 0,27 -9,49 0.83 0.81 0.03 63

#6 0,27 -8,92 0.80 0.95 -0.01 63

#7 0,28 -13,68 0.85 0.99 -0.02 63

#8 0,24 -11,28 0.88 0.88 0.00 63

N=63

#2

#2

SeaWifs - MODIS

MODIS x in situ

Clorofila-a estimada por satélite (mg m-3

)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Clo

rofila

-a in

situ (

mg m

-3)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

CARDER

GSM01

OC3M

Algoritmo rmse rmse-L RDP r2 Declividade Intersecção N

OC3M 0,23 0,76 1,32 0,82 3,38 -0,71 31

CARDER 0,29 0,97 32,29 0,82 4,05 -0,82 30

GSM01 0,25 0,82 -22,19 0,81 1,62 -0,24 30

Chl = 0,14 e 2,55 mg.m-3

, avg=0,49 (± 0,46) mg.m-3

MERIS x In situ

y = -0,0466x + 0,5027R² = 0,0287

y = -0,1892x + 0,5359R² = 0,0083

0,0

1,0

2,0

3,0

4,0

5,0

6,0

0,0 1,0 2,0 3,0 4,0 5,0 6,0

ch

l in

situ

MERIS chl

algal_1

algal_2

Linear (algal_1)

Linear (algal_2)

38

ANTARES-Ubatuba TST

23o44’S

45o00’W

39

Above-water radiometry

0,000

0,001

0,002

0,003

0,004

0,005

0,006

0,007

0,008

0,009

0,010

400 450 500 550 600 650 700

Comprimento de onda (nm)

Rrs

(sr-1

)

Ub01

Ub02

Ub03

Ub04

Ub05

Ub07

Ub08

Ub09

Ub10

Ub11

Ub12

Ub13

Ub14

Ub01

Ub12

Ub02

40

Empirical algorithms comparison

0

5

10

15

20

25

30

35

40

45

-0,6 -0,4 -0,2 0 0,2 0,4 0,6

Log10 [Clorofila in situ] (mg m-3)

Fre

qü

ên

cia

(%

)

0.1

1.0

10.0

0.1 1.0 10.0

Clorofila-a in situ (mg m-3)

Clo

rofi

la-a

Alg

ori

tmo

OC

2 (

mg

m-3

)

0.1

1.0

10.0

0.1 1.0 10.0

Clorofila-a in situ (mg m-3)

Clo

rofi

la-a

Alg

ori

mo

OC

4 (

mg

m-3

)

0.1

1.0

10.0

0.1 1.0 10.0

Chlorophyll-a in situ (mg m-3)

Clo

rofi

la-a

Alg

ori

tmo

OC

3M

(m

g m

-3)

Satellite x in situ

11/11/05

29/0

9/0

5

16/1

0/0

5

6/1

2/0

511/0

8/0

5

13/0

7/0

5

08/0

6/0

5

25/0

5/0

5

06/0

5/0

5

29/0

3/0

5

17/0

2/0

5

01/02/05

11/1

2/0

4

0.00

0.50

1.00

1.50

2.00

2.50

4 13 1 1 15 28 8 23 31 7 12 18 1 6 15 27 3 8 13 26 4 15 30 4 11 22 21 5 14 3 23 7 14 22

Dias (Dez./2004 a Jan./2006)

Clo

rofi

la a

(m

gm

3)

Variability analisis

MODIS – MERIS – SeaWiFS - AVHRR

MERIS (ENVISAT) in Amazon River Mouth

Reflectance in the Amazon region

Rrs_agua

0.000

0.002

0.004

0.006

0.008

0.010

0.012400

420

440

460

480

500

520

540

560

580

600

620

640

660

680

700

rad_1rad_212_rad_a12_rad_b13_rad16_radrad_320_rad21_rad24_rad25_rad

Inland waters

Itumbiara reservoir

barragem

Sampling

Field campaigns

• May 2009

• September 2009

Hidrological cycle

Precipitation and air temperature

Fonte: INMET

Water Level.

Data and Methods

Dados in situ

Dados de Rrs

Dados limnológicos

POIs

Análise espacial - krigeagem

Diagrama triangular

Análise derivativa

K-médias

Análise de variância

Razão de bandas

Dados satélite

Dados satélite

K-médias MODIS

Processadores MERIS

Spectral unmixing

Dados de Rrs

Razão de bandas

Dados in situ PL2

MERIS processors

BOREAL

EUTROPHIC

C2R (Case 2 Regional)

Total C Total C

DOC DOC

DIC DIC

May Sep

mg/L

Turbidity

NTU

Temperature

ºC

May Sep

Detritos - maio

0.00

0.04

0.08

0.12

0.16

0.20

300 400 500 600 700

Comprimento de onda (nm)

ad (m

-1)

CDOM - maio

0

1

2

3

4

5

6

300 400 500 600 700

Comprimento de onda (nm)

aCD

OM (

m-1

)

Fitoplâncton - setembro

0

0.01

0.02

0.03

0.04

0.05

300 400 500 600 700

Comprimento de onda (nm)

aph (

m-1

)

Detritos - setembro

0.00

0.04

0.08

0.12

0.16

0.20

300 400 500 600 700

Comprimento de onda (nm)

ad (

m-1

)

CDOM - setembro

0

1

2

3

4

5

6

300 400 500 600 700

Comprimento de onda (nm)

aC

DO

M (

m-1

)

0,00

0,01

0,02

0,03

0,04

0,05

300 400 500 600 700

aph(m

-1)

Comprimento de onda (nm)

Fitoplâncton - maio

maio

setembro

maio

0.00

0.01

0.02

0.03

0.04

0.05

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

setembro

0.00

0.01

0.02

0.03

0.04

0.05

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

maio

P11

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITUC2R

BOREUL

P23

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITU

BOR

C2R

EUL

P12

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITU

C2R

BOR

EUL

setembro

P23

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITU

C2R

BOREAL

EUT

P5

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITU

C2R

BOREAL

EUT

P17

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

400 500 600 700 800

Comprimento de onda (nm)

Rrs (

sr-1

)

IN SITU

C2R

BOREAL

EUT

MERIS – 07/05/2009 MERIS – 12/09/2009

MODIS – 10/05/2009 MODIS – 10/09/2009

K-médias

bbp – 470 nm bbp – 470 nm

m-1

maio setembro

0

4

8

12

16

20

0 0.02 0.04 0.06 0.08 0.1 0.12

Bbp (470 nm)

Pro

fundid

ade (

m)

0

4

8

12

16

20

0 0.02 0.04 0.06 0.08 0.1 0.12

Bbp (470 nm)Pro

fundid

ade (

m)

Sediments

Chl Organic matter

Clear water

May

Spectral unmixing

Chl Organic matter

sediments Clear water

Sep

Open access data polices will enable the Global

Earth Observation System of Systems to succeed

Thank you

milton@dsr.inpe.br