Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe...

27
Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1, , Gianluca Volpe 1 , Simone Colella 1,3 , Salvatore Marullo 2 , Maurizio Ribera D’Alcalà 3 , Vincenzo Vellucci 3 ENEA -CR Casaccia – Sezione Modellistica Oceanografica Stazione Zoolologica ‘A. Dohrn’ Laboratorio di Oceanografia Biologica 3 2 1

Transcript of Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe...

Page 1: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Development of ocean color algorithms in the Mediterranean Sea

Rosalia Santoleri1,, Gianluca Volpe 1, Simone Colella1,3,

Salvatore Marullo2, Maurizio Ribera D’Alcalà3, Vincenzo Vellucci3

ENEA -CR Casaccia – Sezione Modellistica Oceanografica

Stazione Zoolologica ‘A. Dohrn’ Laboratorio di Oceanografia Biologica 3

2

1

Page 2: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

ISAC Contribution to Ocean Color activity

Mediterranean high resolution surface chlorophyll mapping

• Use available bio-optical data sets to estimate the uncertainties of the existing ocean colour algorithms and to define an optimal chlorophyll algorithm for the Mediterranean Sea.

• Adapt the OC processing software to include selected regional algorithms and validate satellite chlorophyll estimates on the basis of in situ data.

• Evaluate the uncertainties of all current global satellite chlorophyll products available from public archive (e.g. DAAC) in the Mediterranean

• Reprocessing SeaWiFS dataset using the selected Mediterranean algorithm

• Prepare Mediterranean gridded data compatible with model requirements.

Page 3: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Why Mediterranean needs regional algorithm ?

Chlorophyll concentrations over the oligotrophic waters of the Mediterranean Sea are systematically overestimated when global algorithms (e.g. OC4v4) are used to convert blue-to-green reflectance ratios in to chlorophyll-a concentrations:

•Gitelson et al. (Journal of Marine System, 1996)

•D’Ortenzio et al. (SIMBIOS meeting January 2001)

•D’Ortenzio et al. (Remote sensing of the Environment, 2002)

•Bricaud et al . (Remote sensing of the Environment, 2002)

•Claustre et al. (Geoph. Res. Letters, 2002)

From these works it results that global algorithms cannot be applied to-court to the Mediterranean Sea but a specific cal/val activity is needed.

Page 4: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Bio-optical measurements: (143 chl/opt measurement points) to define the Mediterranean regional algorithm

• In water downwelling irradiance and upwelling radiance profiles using SATLANTIC SPMR• above water measurements using the SIMBAD and SIMBADA radiometer • In the bio-optical stations phytoplankton pigments distribution (HPLC and spectro-fluorometric

analysis) and ancillary biological data were also acquired following NASA protocols.

In situ chlorophyll-a data: (872 chl profiles) to validate SeaWiFS, Polder, MODIS, MERIS chlorophyll products and merged level 3 binned data produced by Mersea

Mediterranean Ocean Color CAL/VAL DATA SETS

10 Mediterranean cruises from 1998 up to now were organized by ISAC in the framework of Italian National Projects

Bio-optical stations

Page 5: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Cruise Period Zone# of

profiles

Chlorophyll Range# of

Satlantic# of

SIMBAD(A)

 

MIN MAX  

MATER 425/4/199815/5/1998

Sardinia ChannelSicily Channel

57 0.025 0.105 - -  

MATER 520/10/199827/10/1998

Sicily Channel 50 0.047 0.085 - -  

EMTEC 9920/4/19997/5/1999

Ionian Sea 125 0.039 0.137 20 -  

MATER 614/5/199930/5/1999

Sardinia ChannelSicily Channel

100 0.003 0.135 - -  

PROSOPE04/09/199914/10/1999

Western BasinIonian Sea

16 0.020 0.112 16 -  

SYMPLEX 9921/10/19996/11/1999

Sicily ChannelIonian Sea

212 0.039 0.176 12 -  

NORBAL 126/3/200019/4/2000

Gulf of Lions 81 0.113 2.289 - -  

NORBAL 25/12/2001

20/12/2001Gulf of Lions

Tyrrhenian Sea65 0.088 0.386 13 14  

NORBAL 46/3/2003

26/3/2003Gulf of Lions 115 0.428 7.061 16 28  

NORBAL 518/4/200325/4/2003

Gulf of Lions 40 0.605 2.096 4 7  

DINA*29/3/200128/8/2001

Gulf of Naples 11 0.079 0.455 11 -  

DYFAMED*5/2/1998

25/11/2002Liguro-Provencal 55 0.042 2.366 - -  

All cruises 1998-2003 Mediterranean 872 0.003 7.06192 49

141

Page 6: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Validation of Ocean Color Algorithms

)( 44

33

221010 RaRaRaRaaC

4)( 3

32

21010 cC RcRcRcc

1)/( 5554430dRRdC

OC4v4: R is log10 of either the 443/555 or the 490/555 or the 510/555 band reflectance ratios, depending on its value (the maximum is chosen)

D’Ortenzio et al. 2002 (DORMA): R is log10 of the 490/555 band reflectance ratios.

Bricaud et al. 2002: R is log10 of the 443/555 band reflectance ratios for Chl<0.2 (OC4v4 is used in the other cases)

REGIONAL

GLOBAL

Page 7: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 8: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

)( 44

33

221010 RbRbRbRbbC

MEDOC4 : R is log10 of either the 443/560 or the 490/555 or the 510/555 band reflectance ratios. The switch from one band ratio to another one is based on the chlorophyll concentration itself (the Maximum is Chosen)

NEW MEDITERRANEAN ALGORITHM

Page 9: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 10: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 11: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

SeaWiFS data validation

A Match-up dataset between concurrent SeaWiFS data and in situ measurements has been contructed

1. SeaWiFS L1A passes corresponding to in situ observation were selected

2. The regional algorithms (DORMA, Bricout et al, and the new MED OC4) were inserted in the SeaDAS Code

3. SeaWiFS L1A passes were processed using the different algorithms:OC4v4, DORMA, Bricout, MEDOC4

4. A Match-up dataset is contructed.

Page 12: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Map of the Satellite-in situ match-up stations

Page 13: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 14: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 15: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Merging OC data in the MED

• Define the error of global OC color products at regional scale and define a strategy to take into account the regional optical properties of the MED in the Merging procedure

• Define of the suitable intercalibration and merging tecnique for OC

• Define an optimal interpolation algorithm that takes in to account the different characteristics of ocean colour retrieval in case I/case II waters.

Page 16: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

10 March 2003

Page 17: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

25 March 2003

Page 18: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Primary production in the Mediterranean Sea from remote

sensing data: a model adaptation

Page 19: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Depth

(m

)

CHL (mg m-3)

Depth

(m

)

CHL (mg m-3)

Page 20: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

WEST MED: 172 gc m-2 y-1

EST MED: 123 gc m-2 y-1Bosc et al.,2003

Antoine et al.,1995WEST MED: 197 gc m-2 y-1

EST MED: 137 gc m-2 y-1

In situ C14 Method EST MED: 55-97 gc m-2 y-1

WEST MED: 78-150 gc m-2 y-1

Page 21: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Remote Sensed data of pigment

concentration

Trophic States e Morel and Berthon’s (1989) Correlations

Primary Production Maps

Page 22: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Morel and Berthon carried out relationships between Ctot vs Cpd and Ctot vs Csat. Cpd is defined as the mean pigment concentration in a layer named penetration depth Zpd=1/k where k is the attenuation coefficient for the downwelling PAR irradiance:

Zpd can be estimated from Ze on the base of the relation: Zpd=Ze/4.6 .Instead, Csat is defined as the weighted mean pigment concentration in

the layer Zpd:

Csat is very similar to the Optically Weighted Pigment (OWP) concentration considered as the best representation of what satellite sensors can measurement.

pd

Z

0pd Z

dz C(z)C

pd

pd

pd

Z

0

Z

0sat

dz exp(-2kz)

dz exp(-2kz) C(z)C

Page 23: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

In situ bio-optical data has been used to verify the relationship between the euphotic depth (Ze) and total pigment concetration insite this layer.Ze: depth where the PAR (Photosynthetic Active Radiation) become 1% of the surface value

Page 24: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.

Ctot=54.679 Cpd

0.6532 Ctot=48.13 Csat0.627≠

Ctot=40.6 Cpd0.46 Ctot=40.6 Csat

0.459=

Page 25: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 26: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.
Page 27: Development of ocean color algorithms in the Mediterranean Sea Rosalia Santoleri 1,, Gianluca Volpe 1, Simone Colella 1,3, Salvatore Marullo 2, Maurizio.