JRC MERSEA Ocean Colour Portal by liuhongmei

VIEWS: 7 PAGES: 18

									     European COastal-shelf sea
      OPerational Observing and
    forecasting system Integrated
                Project

WP3: “Better use of remote
sensing and in situ observing
systems for coastal/regional
seas:

Task 3.2: “Improved ocean
colour algorithms and products
for Case-II waters”

Participants:
                       Bay of
                       Biscay
CNR-ISAC
Ifremer                             Adriatic Sea
JRC-EC
   Validation of ocean products Ocean Colour
          products in the Bay of Biscay
                          ECOOP WP3.2.1
In its use of Ocean Colour products, Ifremer has particularly developed the
                service to users, as encouraged by MarCoast
                 (GMES Service Element funded by ESA)



Therefore :

-2 parameters are targeted for validation and assimilation in biological
model : Chlorophyll and mineral SPM (for deriving KPAR)

- 3 parameters : SST, Chlorophyll and turbidity, are proposed for the
operational monitoring required by the Water Framework Directive
Examples of products and the covered area




Mean Chlorophyll-a            Mean Mineral SPM

MODIS 2003-2007               MODIS 2003-2007

weeks 17&18                   weeks 17&18
Validation at coastal station involved in coastal monitoring networks




                                     *
              Some stations are shifted
              for the matchups


                                                *

                                      National in situ networks
                                 *
                                      REPHY: phytoplankton & hydrology / Ifremer
                                      SOMLIT: hydrology / CNRS-INSU
                       *
Validation of the chlorophyll concentration : The 15-day
climatologies SeaWiFS (1998-2004) + MODIS (2005-2006) :
              the SRN Boulogne transect




                     P3
                      P2
                      P1
                      (Coastal)
Bell curve types for the chlorophyll seasonal cycle




   Mean Cabourg                               Percentile 90% Cabourg
   (nutrients from the Seine river)




   Mean Ouest Loscolo                       Percentile 90% Ouest Loscolo
   (nutrients from the Loire river)



 The validation of the satellite chlorophyll is not limited to the mean
 of the distribution but also to the variance. P90 is the parameter of
 interest for the WFD (eutrophic risk)
Spring and autumn peaks for the chlorophyll seasonal cycle




       Mean Men er Roue            Percentile 90% Men er Roue




                                   Percentile 90% Men Du
        Mean Men Du
Same systematic validations for Turbidity
                (NTU)


    Cancale
   Here the shift of 3 pixels has a strong
   effect on satellite turbidity (lower)     Ouest Loscolo




            Men er Roue                      Boyard
                                          ECOOP – Adriatic Sea

 The Adriatic Sea includes diverse water types,
 eutrophic to oligotrophic,
 for which the OC products are still affected by significant
 uncertainties.

 It is also covered by a wealth of field observations and

 is thus an ideal test bed
 for advanced remote sensing methods.

  The first year focused on
  validation of OC products.




NB: Full reprocessing of the SeaWiFS European
and global archive completed (Nov. 2007)
http://oceancolour.jrc.ec.europa.eu
                                   A unique site and data set
                                   for validation of OC products
AAOT
                            Acqua Alta Oceanographic Tower

                           - aerosol optical thickness τa
                             (AERONET site; Jul. 1996 – present)
                                                            Mélin & Zibordi, GRL, 2005
                                                            Mélin et al., JGR, 2006
                                                            Mélin et al., RSE, 2007a
                                                            Clerici & Mélin, submit.
- normalized water-leaving radiances LWN:
  in-water optical profiles (regular campaigns since 1995)
  autonomous above-water radiometry (May 2002 – present)

                                                          Zibordi et al., IJRS, 2004,
                                                          IEEE 2004, 2006,
- concentrations of optically significant constituents:   GRL, 2006, EOS 2006
  Chla, TSM (regular campaigns since 1995)                Mélin & Zibordi, Appl. Opt., 2007




- inherent optical properties (IOPs):
                                                          Mélin et al., RSE, 2005, 2007b

  phytoplankton, CDOM, detritus absorption, particulate back-scattering
  (regular campaigns since 1995)
                       Validation of SeaWiFS Radiometric Products

SeaPRISM




             SeaWiFS




                                       SeaPRISM
Similar analyses have been conducted
for MODIS and MERIS.                              Zibordi et al., GRL, 2006, EOS 2006
                                                  Mélin & Zibordi, Appl. Opt., 2007
                                                  Mélin et al., submit.
          Validation of Inherent Optical Properties

                      Phytoplankton absorption aph(λ)
SeaWiFS




                    Absorption by CDOM and NPP adg(λ)
SeaWiFS




                    Particulate back-scattering bbp(λ)


                                 encouraging results

                                     Mélin et al., RSE, 2005, 2007b
          Validation of OC Chlorophyll data
OC data are produced in NRT by CNR-ISAC in the framework
of the Adricosm Project for environmental assesment and data
assimilation in models
Aims/activities:
1. Assessment of the SeaWiFS and
   MODIS OC chlorophyll products
   in the Adriatic Sea with particular
   attention to the coastal waters
   (case 2)
2. Test      different     bio-optical
   algorithms for Chla (global and
   Mediterranean) to select the
   most appropriate one.
3. Define a strategy to improve the
   Adriatic CNR_ISAC operational
   regional     products in coastal
   waters to be implemented for
   ECOOP TOP phase
                                         http://gos.ifa.rm.cnr.it/adricosm/index.html
 The Adriatic Ocean Color
   CAL/VAL DATA SETS
                               Ancona-Pescara
                                                              Po river discharge heavily
          Po river discharge
                                                              influences the Western Adriatic
                                                              Current turbid waters
                                                                           In situ data set:
                                                                       3 oceanographic cruises
                                                                       Repeated stations
                                                                        acquired by regional
                                                                        environmental agencies
                                                              .
                                                                       634 SeaWiFS case 2
                                                                        water matchups
                                                                       340 case 2 water Modis
                                                                        matchups
                                            Projects/cruise    Stations          Area            Period

Modis Aqua                                  ADRICOSM_ER           1245     Nord Adriatic Sea    2002-2006

                                                REQUISITE         3528      Nord and Middle     2004-2006
6th july 2004                                    ADR06            150                          January 2006
                                                DART06a           159       Middle Adriatic    March 2006
                                                DART06b            57       Middle Adriatic    August 2006

                                                 TOTAL            5139
         Validation of SeaWiFS Chlorophyll products
 OC4v4                 OC4v4 (MUMM)             JRC                   CARDER



           R2=0.52                R2=0.42                  R2=0.45              R2=0.49
          RMS=0.49               RMS=0.49                 RMS=0.43             RMS=0.44
          BIAS=0.32              BIAS=0.27                BIAS=0.09            BIAS=0.09




MEDOC4                MEDOC4 (MUMM)          JRC (MUMM)               CLARK




           R2=0.52                R2=0.46                  R2=0.41              R2=0.44
          RMS=0.52               RMS=0.49                 RMS=0.40             RMS=0.50
          BIAS=0.34              BIAS=0.30                BIAS=0.02            BIAS=0.30



1. a general overestimation of the satellite Chla in all algorithms also when the
   regional Adriatic algorithm (JRC) is used
2. No improvements with MUMM atmospheric correction
3. Better results using Carder Algorithm (low Bias; uniform distribution)
        Validation of MODIS Chlorophyll products
                              OC3 (MUMM)               CARDER
       OC3




                  R2=0.42                   R2=0.35               R2=0.84
                 RMS=0.44                  RMS=0.42              RMS=0.24
                 BIAS=0.12                 BIAS=0.10             BIAS=-0.15




      MEDOC3                 MEDOC3 (MUMM)              GSM01




                 R2=0.44                    R2=0.39               R2=0.59
                RMS=0.44                   RMS=0.42              RMS=0.31
                BIAS=0.17                  BIAS=0.13             BIAS=-0.08




1. Chla overestimation using standard (OC3) and Med (MedOC3) algorithms
2. No improvements with MUMM atmospheric correction
3. Better results using Carder Algorithm (highest R2) but small bias
                 Conclusions
• An specific case 2 chla algorithm is required for the
  Adriatic Sea
• In meantime the Carder’s algorithm should be introduced
  in the CNR-ISAC Adriatic operational processing chain
  to estimate the chla in Case 2 while MEDOC3 should be
  maintained for case 1
• We need to develop of a method that produces a single
  chlorophyll map of the Adriatic with a different
  chlorophyll algorithm for case 2 and case 1 waters
  without introducing artificial gradients.
     European COastal-shelf sea
      OPerational Observing and
    forecasting system Integrated
                Project

 Deliverables and First-Year Status

D3.2.1.1: Report on comparison between R/S and
          in-situ data (Adriatic Sea)

D3.2.1.2: Report on comparison between R/S and
          in-situ data (Bay of Biscay)

D3.2.1.3: Report on multi-sensor merging and
          dynamic bio-optical algorithm selection (Adriatic Sea)

D3.2.1.4: Report on the merging technique between OC R/S
          and in-situ data (Bay of Biscay)

								
To top