MERIS Optical Measurement Protocols.mermaid.acri.fr/dataproto/CO-SCI-ARG-TN-0008_MERIS...Doc. no: CO-SCI-ARG-TN-0008 Issue: 2.0 Revision: 1.0 Date: July 2011 Document Signatures Name

MERIS

Optical

Measurement

Protocols

Doc : CO-SCI-ARG-TN-008

Name : MERIS Optical Measurement

Protocols. Part A: Reflectance

Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : i

All rights reserved, ARGANS Ltd 2011

MERIS Optical Measurement Protocols.

Part A:

In-situ water reflectance measurements

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : ii


Doc. no: CO-SCI-ARG-TN-0008

Issue: 2.0

Revision: 1.0

Date: July 2011

Document Signatures

Name Function Company Signature Date

Editor Kathryn

Barker Project Manager ARGANS

July 2011

Verification Jean-Paul

Huot MVT Coordinator ESA

July 2011

Approval Philippe

Goryl

Contract Manager,

ESA ESA

Updates

Issue Date Description

1.0 June 2010 Issue 1 Available online on MERMAID website:

http://hermes.acri.fr/mermaid

July 2010 Minor updates: redistributed to QWG and MVT

2.0 July 2011 Updates relating to MERMAID evolution (links with ODESA and new

web interface) and new matchup sites.

This is a public document, available for download on the MERMAID website: http://hermes.acri.fr/mermaid/dataproto

http://hermes.acri.fr/mermaid/dataproto

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : iii


Acknowledgement

To ACRI-ST (C. Mazeran; C. Lerebourg) and ESA (J-P. Huot) under whose contracts the production and

maintenance of this document falls, and from whom substantial input has been received. ESA Contract

numbers: 21091/07/I-OL and 21652/08/I-OL respectively.

To all MERIS Validation Team members for their interest in MERMAID and their feedback on the

database and the Protocols document, and to the MERIS QWG who have contributed to the MERIS Third

Reprocessing and provided inputs to this document where appropriate.

Protocol contributors

NAME AFFILIATION

S. AHMED City College of New York, USA

D. ANTOINE LOV, France

V. BRANDO CSIRO, Australia

P-Y. DESCHAMPS LOA, France

R. DOERFFER HZG, Germany

B. GIBSON Coastal Studies Institute, LSU, USA

B. HOLBEN NASA GSFC

A. HOMMERSOM Water Insight, Netherlands.

J. ICELY University of Algarve

M. KAHRU University of California, USA

S. KRATZER University of Stockholm, Sweden

H. LOISEL; C. JAMET Universite du Littoral Cote d'Opale, France

D. MCKEE University of Strathclyde, UK

K. VOSS; M. ONDRUSEK NOAA

K. RUDDICK MUMM, Belgium

D. SIEGEL; S. MARITORENA University of California, Santa Barbara, USA

K. SORENSEN NIVA

J. WERDELL (on behalf of NOMAD contributors) NASA/GSFC

G. ZIBORDI JRC, Italy

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : iv


Table of Contents

1. Introduction .......................................................................................................................................... 1

1.1 Document purpose and scope ....................................................................................................... 1

1.2 Applicable documents ................................................................................................................... 2

1.3 MERIS L2 water products overview ............................................................................................ 3

1.4 General radiometry and water colour ............................................................................................ 5

1.5 Requirements and recommendations for the validation of MERIS w ......................................... 8

2. The „MERIS MAtchup In-situ Database‟ (MERMAID).................................................................... 13

2.1 Introduction ................................................................................................................................. 13

2.2 Atmospheric parameters in MERMAID ..................................................................................... 18

2.3 MERMAID uncertainties ............................................................................................................ 18

2.4 ρw spectral correction .................................................................................................................. 18

2.5 Measurement and Processing flags ............................................................................................. 20

2.6 Data Access and Policy ............................................................................................................... 22

3. MERMAID PROTOCOLS I: SeaPRISM (AERONET-Ocean Color) .............................................. 24

3.1 Introduction ................................................................................................................................. 24

3.2 The Aqua Alta Oceanographic Tower (AAOT). PI: Giuseppe Zibordi ...................................... 28

3.3 Abu Al-Bukhoosh. PI: Giuseppe Zibordi ................................................................................... 29

3.4 CERES Ocean Validation Experiment (COVE_SeaPRISM). PI: B. Holben ............................. 29

3.5 Gloria. PI: G. Zibordi .................................................................................................................. 30

3.6 Gustav-Dahlen Tower. PI: Giuseppe Zibordi ............................................................................. 30

3.7 Helsinki Lighthouse. PI: Giuseppe Zibordi ................................................................................ 30

3.8 Long Island Sound Coastal Observatory (LISCO). PI: S. Ahmed, A. Gilerson ......................... 31

3.9 Lucinda Jetty Coastal Observatory (LJCO). PI: V. Brando ........................................................ 32

3.10 Martha‟s Vineyard Coastal Observatory (MVCO). PI: H. Feng................................................. 33

3.11 Pålgrunden Lighthouse, Lake Vänern. PI: Susanne Kratzer ....................................................... 33

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : v


3.12 WAVE_CIS_Site_CSI_6. PI: B. Gibson, A. Weidermann ......................................................... 34

3.13 AERONET-OC Key References ................................................................................................. 34

4. MERMAID PROTOCOLS II: Portable Radiometers ........................................................................ 36

4.1 SIMBADA. PI: Pierre-Yves Deschamps. ................................................................................... 36

5. MERMAID PROTOCOLS III: TACCS ............................................................................................ 42

5.1 North West Baltic Sea. PI: Susanne Kratzer ............................................................................... 42

5.2 Sagres, Algarve. PI: John Icely ................................................................................................... 49

6. MERMAID PROTOCOLS IV: Fixed-depth Moorings ..................................................................... 53

6.1 Buoy for the acquisition of long-term optical time-series (BOUSSOLE). PI: David Antoine ... 53

6.2 Marine Optical BuoY (MOBY). PI: Kenneth Voss .................................................................... 58

7. MERMAID PROTOCOLS V: Profiling Instruments ........................................................................ 62

7.1 Bristol Channel and the Irish Sea. PI: David McKee .................................................................. 62

7.2 California Current. PI: M. Kahru ................................................................................................ 65

7.3 Plumes and Blooms. PI: D. Siegel .............................................................................................. 66

8. MERMAID PROTOCOLS VI: TriOS Ramses .................................................................................. 69

8.1 English Channel. PI: H. Loisel, C. Jamet .................................................................................... 69

8.2 FERRYBOX. PI: K. Sørensen .................................................................................................... 71

8.3 French Guiana. PI: H. Loisel; C. Jamet ...................................................................................... 73

8.4 Helgoland/Cuxhaven Transect. PI: R. Doerffer .......................................................................... 74

8.5 MUMMTriOS. PI: K. Ruddick ................................................................................................... 77

8.6 Wadden Sea. PI: A. Hommersom ............................................................................................... 80

9. MERMAID PROTOCOLS VII: Miscellaneous datasets ................................................................... 83

9.1 NASA bio-Optical Marine Algorithm Data set (NOMAD). PI: Jeremy Werdell ....................... 83

10. Appendix 1: Values of the air-sea interface transmittance as function of wind speed, ws, view

angle, ', and salinity (salt and freshwater). ................................................................................................ 89

11. Combined References ..................................................................................................................... 94

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : vi


List of Figures

Figure 2-1: MERMAID Website ................................................................................................................ 13

Figure 2-2: ODESA online processing. (Figure courtesy of ACRI-ST). .................................................... 14

Figure 2-3: Samples from the MERMAID template; formatted PI in-situ data. This is a sample from

AAOT. ........................................................................................................................................................ 14

Figure 2-4: The MERMAID data extraction webpage and extraction options, on the MERMAID website

available (password restricted).................................................................................................................... 22

Figure 2-5: Example extraction items: CSV file of extracted data, descriptive plots and RBG. ................ 23

Figure 3-1: AERONET-OC sites in MERMAID. ....................................................................................... 24

Figure 3-2: The AAOT structure and instrumentation; a) the main tower and operational levels (from

Hooker et al., 2005), b) the CIMEL CE-318 (SeaPRISM) instrument (from

http://aeronet.gsfc.nasa.gov/new_web/photo_db/Venise.html). ................................................................. 28

Figure 3-3: COVE SeaPRISM site, 25 km East of Virginia Beach, Virginia: a) site location; b)

Lighthouse platform; c) AERONET sunphotometer. ................................................................................. 29

Figure 3-4: Gloria Platform, Black Sea. ..................................................................................................... 30

Figure 3-5 a) Gustav-Dahlen Tower in the northern Baltic Proper. The inset is a picture of the SeaPRISM

autonomous radiometer installed on the tower; b) The Helsinki Lighthouse in the Gulf of Finland.......... 31

Figure 3-6: LISCO site, Long Island Sound. .............................................................................................. 31

Figure 3-7: LJCO site, Eastern Australia. Images from: http://imos.org.au/ljco.html ................................ 32

Figure 3-8: The tower at MVCO ................................................................................................................ 33

Figure 3-9: The Pålgrunden lighthouse SeaPRISM platform in Lake Vänern, Sweden. ............................ 33

Figure 3-10: WAVE_CIS_Site_CSI_6. a) Platform; b) Instrumentation ................................................... 34

Figure 4-1: Spectral channels of the SIMBADA instrument ...................................................................... 37

Figure 5-1: The TACCS instrumentation rig. a) the in-water instrumentation shown being deployed by PI

Susanne Kratzer, and b) the above-water Ed sensor. ................................................................................... 42

Figure 5-2: Himmerfjärden area, NW Baltic Sea. Note that stations B1 and H2 do not differ optically

from the open sea station (Kratzer et al., 2008). STP: sewage treatment plant at the head of

Himmerfjärden close to station H5. ............................................................................................................ 43

Figure 5-3: Map of the Portuguese coast with the area of study indicated as a black box. Satellite image of

southwest coast of Portugal with the location of sampling sites A, B, C. .................................................. 50

CO-SCI-ARG-TN-0008_MERIS_Optical_Measurement_Protocols_Issue2_July2011.doc#_Toc299716304


MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : vii


Figure 6-1: Artist‟s view of BOUSSOLE (from Antoine et al., 2008), showing the above- and in-water

radiometers, and buoy structure. ................................................................................................................. 54

Figure 6-2: The NASA MOBY instrument set-up at Lanai, Hawaii .......................................................... 59

Figure 7-1: Errors in first two points of level 3 depth-averaged Lu values. ................................................ 64

Figure 7-2: CalCOFI transect locations, California Coast. (from: http://www.calcofi.org) ................ 65

Figure 7-3: Santa Barbara Channel, California USA. Plumes and Blooms stations are marked with an „x‟.

From Kostadinov et al. (2006). ................................................................................................................... 67

Figure 8-1: Location of the different stations visited in the eastern English Channel and southern North

Sea in 2004. The investigated area is bordered by (I) the mouth of the Seine River in the south and (II) the

mouth of the Escault River in the North. .................................................................................................... 69

Figure 8-2: Location of the stations sampled on 7-11 July 2006 (from Loisel et al., 2009). ...................... 74

Figure 8-3: Route and stations of the MERIS validation campaign "c30" on 13. July 2006 ...................... 75

Figure 8-4: a) TRIOS-Spectrometer for measuring upward directed radiance from water and sky radiance,

and b) TRIOS Spectrometer for measuring downwelling irradiance .......................................................... 76

Figure 8-5 Frame with three TriOS-RAMSES hyperspectral radiometers as installed on the research

vessel Belgica (Ruddick, 2006). ................................................................................................................. 78

List of Tables

Table 1-1: Documents pertinent to the MERIS Optical Measurement Protocols. ........................................ 2

Table 1-2: MERIS TOA Spectral Bands (from the MERIS Product Handbook, [AD 4]). ........................... 4

Table 1-3: Air-sea interface terms and values used in the datasets provided to MERMAID (where

provided and where relevant). Empty cells mean the information is not available. ................................... 10

Table 2-1: Datasets in MERMAID and details of the associated PI. Acronyms are fully expanded in the

Abbreviations and Definitions list on Page xii ........................................................................................... 15

Table 2-2: In-situ bandsets in MERMAID ................................................................................................. 19

Table 2-3: MQC flag criteria definition. Flag position is counted from the first numeric character after the

leading „M‟. Unless otherwise specified: 0 = No / Not done, 1 = Yes / done/ provided, 2 = Unknown / not

available / not applicable (N/A). ................................................................................................................. 20


MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : viii


Table 2-4: PQC flag criteria definition. Flag position is counted from the first numeric character after the

leading „P‟. Unless otherwise specified: 0 = No / Not done, 1 = Yes / done/ provided, 2 = Unknown / not

available. ..................................................................................................................................................... 21

Table 5-1: General sensor specifications of the TACCS 09 ....................................................................... 44

Table 5-2: Mean slope factors to derive spectral Kd for all TACCS channels from Kd490 in the north-

western Baltic Sea during summer (Kratzer et al., 2008) derived from AC9 data that was measured during

field campaigns in June 2001, August 2002 and July 2008 (Kratzer and Tett, 2009, Kratzer and Vinterhav

2010). The data set was divided up into I) outer fjord & open sea stations (B1-BY31 & H2), and II) inner

fjord stations (H3-H4). ................................................................................................................................ 45

Table 5-3: Percent uncertainties for TACCS Lu (lamda,0+). 1Moore, et al. (2011),

2Type B uncertainty:

educated guess ............................................................................................................................................ 48

Table 6-1: Nominal wavelengths at which BOUSSOLE provides in-water radiometric data to MERMAID

.................................................................................................................................................................... 53

Table 8-1: Number and dates of transect campaigns .................................................................................. 77

Table 10-1: Air-sea interface transmittance (35 psu). ................................................................................. 89

Table 10-2: Air-sea interface transmittance (0 psu). ................................................................................... 91

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : ix


List of Symbols

Symbol Definition Dimension / units

Geometry (see fig. 2.1)

Wavelength nm

s Solar zenith angle (s = cos(s)) degrees

v, Satellite or view zenith angle (v = cos(v)) degrees

Refracted view zenith angle (‟ = sin-1(n.sin(v))) degrees

π-θ degrees

Relative azimuth angle between the sun-pixel and

pixel-sensor directions degrees

Radiometric quantities

L(,s,v,) Spectral radiance W m-2

sr-1

nm-1

Inherent Optical Properties (IOPs)

),( Volume scattering function (VSF) sr-1

)(~ Normalised volume scattering function sr

-1 m

-1

a() Total absorption coefficient for wavelength m-1

apig(442) Pigment absorption coefficient at 442 nm m-1

b() Total scattering coefficient for wavelength m-1

c() Attenuation coefficient for wavelength m-1

bb() Backscattering coefficient m-1

Apparent Optical Properties (AOPs) and derived quantities

w(,s,v,) Water reflectance dimensionless

wn() Fully normalised water reflectance (i.e. the reflectance

if there were no atmosphere, and for s = v = 0) dimensionless

Eu() Upwelling irradiance W m-2

nm-1

Ed() Downwelling irradiance, above the surface W m-2

nm-1

Es (λ) Total downwelling irradiance just above the sea surface, W m-2

nm-1

denoted also as Ed (λ, 0+).

Lw (λ) Water-leaving radiance sr-1

Lwn (λ) Fully normalised water-leaving reflectance sr-1

Lwn_f/Q Normalised Water Leaving Radiance - f/Q corrected sr-1

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : x


R(, 0-) Diffuse reflectance at null depth, or irradiance reflectance dimensionless

(Eu / Ed)

F0() Mean extraterrestrial spectral irradiance W m-2

nm-1

f Ratio of R(0-) to (bb/a); subscript 0 when s = 0 dimensionless

f‟ Ratio of R(0-) to (bb/(a + bb)); subscript 0 when s = 0 dimensionless

Q(,s,v,) Factor describing the bidirectionality character of sr-1

R(, 0-) Subscript 0 when s = v = 0; Q = Eu/Lu

Other atmosphere and aerosol properties

α Angström exponent (α < 0). dimensionless

ε Eccentricity of the Earth‟s elliptic orbit dimensionless

τa() Aerosol optical thickness dimensionless

τray() Rayleigh (or molecular) optical thickness dimensionless

O3

() Ozone optical thickness dimensionless

Tray (λ) Rayleigh transmittance dimensionless

Ta (λ) Aerosol transmittance dimensionless

TO3 (λ) Ozone transmittance dimensionless

Td (λ) Total downwelling transmittance (diffuse + direct) dimensionless

Tu (λ) Total upwelling transmittance (diffuse + direct) dimensionless

Ps Surface pressure hPa

uO3 Ozone concentration cm-atm

RH Relative humidity percent

),( sdT Downwelling total transmittance at sea surface level dimensionless

Air-water interface

)'( Geometrical factor, accounting for multiple reflections and dimensionless

refractions at the air-sea interface (Morel and Gentilli, 1996;

defined further in Section 1.4.5)

n refractive index of sea water dimensionless

f() Fresnel reflectance at the air-sea interface for the scattering angle dimensionless

mean reflection coefficient for the downwelling irradiance at the

sea surface dimensionless

r average reflection for upwelling irradiance at the air-water interface dimensionless

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : xi


Root-mean square of wave facet slopes dimensionless

Angle between the local normal and the normal to a wave facet

p probability density function of facet slopes for the illumination dimensionless

and viewing configurations (s, v, )

Miscellaneous

ws Wind-speed just above sea level m s-1

ln Natural (or Neperian) logarithm

log10 Decimal logarithm

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : xii


Abbreviations and Definitions

AAOT Aqua Alta Oceanographic Tower

AD Applicable Document

AERONET Aerosol Robotic Network

AOP Apparent Optical Property

AOT Aerosol Optical Thickness

ARGANS Applied Research in Geomatics, Atmosphere, Nature and Space

ATBD Algorithm Theoretical Baseline Document

BBOP Bermuda Bio-Optics Project

BOUSSOLE BOUée pour l'acquiSition d'une Série Optique à Long termE

(Buoy for the acquisition of long-term optical time series)

BPAC Bright Pixel Atmospheric Correction

CalCOFI California Cooperative Oceanic Fisheries Investigations

CDOM Coloured Dissolved Organic Matter

Chl Chlorophyll-a concentration mg m-3

CHORS Center for Hydro-Optics and Remote Sensing

COVE Clouds and the Earth's Radiant Energy System (CERES)

Ocean Validation Experiment

CTD Conductivity Temperature Depth

DPM Detailed Processing Model

EO Earth Observation

ESA European Space Agency

EOLI-SA Earth Observation Link - Stand Alone

GDT Gustav Dahlen Tower

GPS Global Positioning System

HLT Helsinki Lighthouse Tower

HPLC High Performance Liquid Chromatography

IOP Inherent Optical Property

LOA Laboratoire d'Optique Atmosphérique

LISCO Long Island Sound Coastal Observatory

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : xiii


LISE Laboratoire Interdisciplinaire des Sciences de l'Environnement

LJCO Lucinda Jetty Coastal Observatory

LOV Laboratoire Océanographique in Villefranche sur mer

LUT Look-Up Table

MEGS MERIS Experimental Ground Segment

MERIS Medium Resolution Imaging Spectrometer

MERMAID MERis MAtch-up In-situ Database

MOBY Marine Optical Buoy

MODIS Moderate Resolution Imaging Spectrometer

MOS Marine Optical System

MQC Measurement Quality Control

MUMM Management Unit of the North Sea Mathematical Models

MVCO Martha’s Vineyard Coastal Observatory

MVT MERIS Validation Team

NaN Not a Number

NASA National Aeronautics and Space Administration

N/A Not Applicable

NCEP National Centre for Environmental Prediction

NIR Near Infrared

NOMAD NASA bio-Optical Marine Algorithm Dataset

OC Ocean Color

ODESA Optical Data Processor of the European Space Agency

OBPG Ocean Biology Processing Group

PAR Photosynthetically Available Radiation

PI Principle Investigator

PQC Processing Quality Control

QWG Quality Working Group

RMD Reference Model Document

RR Reduced resolution

RTC Radiative Transfer Code

SeaBAM SeaWiFS Bio-optical Algorithm Mini-Workshop

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : August 2011

PAGE : xiv


SeaBASS SeaWiFS Bio-Optical Archive and Storage System

SeaWiFS Sea-viewing Wide Field-of-view Sensor

SPM Suspended Particulate Matter

SPMR SeaWiFS Profiling Multichannel Radiometer

STP Standard Temperature and Pressure (To=273.5 K; Po=1013.25 hPa)

TACCS Tethered Attenuation Coefficient Chain Sensor

TBD To Be Determined

TOA Top Of Atmosphere

TOMS Total Ocean Mapping Scanner

TSM Total Suspended Matter (g m-3

)

UK United Kingdom

UTC Coordinated Universal Time

VSF Volume Scattering Function

WiSPER Wire Stabilized Profiling Environmental Radiometer

YS Yellow Substance absorption coefficient (m-1

)

YSBPA Absorptions of dissolved and bleached particulate matter (m-1

)

Case 2(S) water: Case 2 water dominated by SPM (see ATBD: PO-TN-MEL-GS-0005)

Case 2(Y) water: Case 2 water dominated by yellow substances (see ATBD: PO-TN-MEL-GS-

0005)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 1


1. Introduction

Within the framework of the MERIS Data Quality Working Group (QWG) and the MERIS Validation

Team (MVT) fall the validation activities essential for product assessment and quality assurance, such as

analysis and assessment of MERIS atmospheric correction over waters and vicarious adjustment. An

integral requirement for such activities is a reliable source of in-situ radiometric data, inclusive of the

metadata and parameters required for the validation research and decision making.

The MERis MAtchup In-situ Database (MERMAID) was created to satisfy validation aims, making

available an easy-to-use centralised database of merged in-situ optical measurements with concurrent

MERIS acquisitions to Ocean Colour researchers involved in the MERIS mission.

The long-term objectives of this database are to:

Enable the assessment of the MERIS marine Level 2 products delivered by the ENVISAT ground

segment.

Support the monitoring of these MERIS products over the lifetime of the mission by providing a

complete temporal coverage of the mission.

Provide support to atmospheric correction research.

Support vicarious adjustment of the MERIS instrument.

Provide a centralised validation resource to the ESA Optical Data Processor, ODESA.

MERMAID has become the only repository for ESA MERIS matchup data, for both in-situ and MERIS

acquisitions, and ESA-funded researchers have the potential to contribute to the development of this

valuable resource, by contributing their data for matchup with MERIS imagery. Moreover, data is also

sought from external, non-ESA funded sources. Data is provided to the database through agreement with

the Principle Investigator (PI), who has pre-processed their in-situ data to a standard where it can be

matched with the sensor.

MERMAID is an ever-developing and evolving database, and users of the data and readers of this

document are invited to provide feedback, comments and suggestions to [email protected].

1.1 Document purpose and scope

The purpose of this document is to describe the protocols followed by each of the PIs whose data has

been matched with MERIS acquisitions and is included in MERMAID. Along with in-situ data, PIs

provide information on how the in-situ radiometric and bio-optical measurements were taken and

processed prior to contribution. Many PIs already provide this information in the peer-reviewed literature,

so this document is an assimilation and condensation of these protocols, designed to accompany the

database with a comprehensive description of the data held within it. The measurements in the database

are made by a variety of techniques and instruments; the Optical Measurement Protocols provide a

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 2


guideline for the measurement of in-situ water reflectances (ρw: defined in section 1.4.2) by the means

used to acquire the data in the database.

The document does not aim to specifically promote any given technique; rather it confirms the adherence

to accepted protocols. It is important that the investigator describes the methods used for validation

because deviations from standards can cause significant problems in the interpretation of differences

between the in-situ and MERIS data.

The MERIS Optical Measurement Protocols are organised into three parts: marine reflectances,

atmospheric measurements and inherent optical property (IOP) measurements. This document is Part A,

and is concerned with the matchup of MERIS marine reflectances with in-situ measurements. It

documents measurement procedures, data processing by the PI and normalisation for MERMAID, and an

overview of the uncertainties in the data. The document is organised into several parts:

1) Introduction: the requirements for MERIS validation

2) Description of MERMAID

3) The Protocols of the MERMAID datasets currently in MERMAID.

1.2 Applicable documents

Table 1-1: Documents pertinent to the MERIS Optical Measurement Protocols.

Code Document Title

[1] N/A MERIS ATBDs

http://envisat.esa.int/instruments/meris/atbd/

[2] N/A MERIS Data Products Overview.

http://envisat.esa.int/support-docs/productspecs/

[3] PO-TN-MEL-GS-0026 MERIS Reference Model Document (RMD): Third Reprocessing

[4] N/A MERIS Product Handbook

http://envisat.esa.int/pub/ESA_DOC/ENVISAT/MERIS/

[5] MERIS ATBD 2.6 Algorithm Identification: Case II.S Bright Pixel Atmospheric

Correction (G. Moore and S. Lavender)

[6] N/A Tilt correction for irradiance sensors (J.-P. Huot).

[7] D9-b-final1 How a prediction of the radiance of the sky dome helps to improve

the water leaving reflectance measurements? (R. Santer)

http://envisat.esa.int/instruments/meris/atbd/

http://envisat.esa.int/support-docs/productspecs/

http://envisat.esa.int/pub/ESA_DOC/ENVISAT/MERIS/

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 3


1.3 MERIS L2 water products overview

MERIS provides a number of water products at Level 2. This document is concerned with the water

reflectance, ρw (λ), and primarily that measured in-situ and subsequently matched with MERIS ρw (λ) for

MERMAID. Parts B and C will address the atmospheric products and in-water constituents such as

chlorophyll-a (Chl), yellow substances (YS), also known as coloured dissolved organic matter (CDOM),

and total suspended matter (TSM) concentration.

This is a brief overview of the products available; further details of the L2 processing may be found in the

MERIS Product Handbook (AD [4]), ATBDs (AD [1]) and Data Products Overview (AD [2]).

w : a dimensionless term, water reflectance, valid in all waters. Defined fully in section 1.4.2.

Aerosol products:

o Aerosol optical thickness at 865nm, a(865) for the whole atmosphere (boundary layer +

troposphere + stratosphere),

o the slope of the spectral dependence of the aerosol optical thickness between 779nm and

865nm, α(779, 865) for the whole atmosphere.

Chl1: the algal pigment index 1 (Morel and Antoine, 1999), expressed as a chlorophyll concentration

in mg.m-3

, given in Case 1 waters.

Chl2: the algal pigment index 2, expressed as a Chl concentration in mg.m-3

. Chl2 is related in the

neural network algorithm via a scaling equation to pigment absorption at 442nm, apig(442), given in

all waters. As applied in the MERIS product, and defined in the MERIS RMD (AD [3]), we have:

04.1)]442([0.21][ pigaChl (1)

TSM, total suspended matter concentration, expressed as concentration in g.m-3

, given in all waters.

TSM is related in the neural network algorithm via a scaling factor to a particle scattering at 442 nm,

bp(442), given in all waters. As applied in the MERIS product, and defined in the MERIS RMD (AD

[3]), we have:

)442(.73.1)( 3

pbmgTSM (2)

YSBPA: proxy for the sum of absorptions of dissolved and bleached particulate matter at 442.5nm in

m-1

. In the rest of this document YS will be reserved for yellow substance (also known as CDOM)

absorption, and BPA will be used for bleached particle absorption. YSBPA = YS+BPA.

PAR: Photosynthetically Available Radiation.

Case 2_S: a flag indicating the presence of TSM in significant concentration.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 4


Case 2_Anom: a flag indicating abnormally high scattering in Case 1 water.

Case 2_Y: a flag indicating YS loaded water. This flag is at the moment inactivated in the ground

segment processing pending validation.

Marine reflectances, the subject of the present document (Part A), are given at 13 spectral bands: bands 1

to 10, 12, 13, 14. Band 11 (761.875nm) and 15 (900nm) are strong absorption bands by O2 and water

vapour, H2O, respectively. All bands are defined in Table 1-2 below. The characterised mean wavelengths

(over the 5 MERIS cameras) are used in the product evaluation procedure.

1.3.1 MERIS Atmospheric correction

Information and detail on the MERIS atmospheric correction, atmospheric parameters, definitions and

coefficients can be found in AD [3], the MERIS Reference Model Document (RMD) for the Third

Reprocessing. The Bright Pixel Atmospheric Correction (BPAC) is now applied as standard and detailed

in both AD[3] and AD[5], as listed in Table 1-1.

Table 1-2: MERIS TOA Spectral Bands (from the MERIS Product Handbook, [AD 4]).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 5


1.4 General radiometry and water colour

1.4.1 Definition of case 1 and case 2 waters

Natural waters can be classified into water types, Case 1 or Case 2, according to their optical properties

which vary according to what is present in the water (Gordon and Morel, 1983; Morel and Prieur, 1977).

In Case 1 water it is recognised that the principle agent responsible for variations in the IOPs is

phytoplankton and their degradation products only, and that the global variations in IOPs in these waters

can be represented by average models where the chlorophyll concentration is used as the unique index of

these changes.

In Case 2 waters it is often likely that the optical properties of substances other than phytoplankton

dominate the bulk optical properties; those substances vary independently of phytoplankton, particularly

TSM which presents a backscattering, bb (λ), component, and Coloured Dissolved Organic Matter

(CDOM, or yellow substances), presenting issues for Chl retrieval algorithms. Case 2 water bodies

display spatial and temporal variations in their organic and inorganic composition, often on small

temporal and spatial scales. This patchiness may cause issues when it comes to matchup with MERIS RR

imagery.

1.4.2 Definition of the product ρw (λ)

For the following and ensuring radiometric descriptions, all symbols are defined for reference in the

Symbols Table (Page xii).

Surface water is assumed to be Lambertian and its reflectance, w, is defined as:

),(

),,,(.)(

ss

vssvww

E

L

(3)

where Lw (λ) is the upward water leaving radiance (which does not account for any specular reflected

direct sun or sky radiance), and Es (λ) is the total downwelling irradiance at sea level, comprising of both

diffuse and direct components.

1.4.3 Downwelling solar irradiance at ground level, Es (λ)

Es is measured directly in-situ, at the sea level, with instrumentation or it may be derived by:

ISsISsdos TdFE __

2 cos),()()( (4)

where:

F0 is the Thuillier et al. (2003) mean extraterrestrial solar irradiance.

d2 is the corrective factor of the extraterrestrial sun irradiance, accounting for the Sun-Earth

distance as a

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 6


function of the Julian day (J) (see the MERIS RMD).`

θs (solar zenith angle) is computed for the time of the in-situ measurement (i.e. θs_IS)

Td is the total atmospheric transmittance for the downwelling path only and includes the

contribution of gaseous absorption. Formulations for computing Td are described for MERIS in

the MERIS RMD.

Another approach taken by some PIs, is to extrapolate Ed (0-) across the surface to derive Ed (0+), an

alternative term for Es. If used, the details are provided in the relevant protocol later in this document.

1.4.4 ρwn_ISME: correction for in-situ solar irradiance in w

Because the solar illumination used in the computation of the in-situ reflectance can be different from that

of the MERIS processing, MERMAID contains a complementary in-situ water reflectance consistent with

the MERIS formulation of Es, as a MERIS and in-situ quality control indicator. In-situ ρw (λ) is

renormalized to Es with the MERIS definition, using the ratio between the irradiance measured at sea

level and the MERIS-like irradiance estimated at ground level (Equation 5). The latter, termed „ρwn_ISME

(λ)‟ once normalised as usual (where „IS‟ is for „in-situ‟ and „ME‟ is for „MERIS‟), is calculated exactly

using the solar irradiance at TOA (as described in the MERIS DPM) and the total downwelling

transmittance, Td (λ), from MERIS LUTs.

Additional columns exist in MERMAID for ρwn_ISME (λ) at the same 13 bands as ρw_IS (λ).

)(

)()()( __

MERIS

s

IS

sISwnISMEwn

E

E (5)

To enable this, PIs provide to MERMAID their in-situ Es (whether computed as in Equation 4 or

measured directly with instruments). For the sites in MERMAID using CIMEL instruments (namely

AERONET-OC), an extra stage is added to the pre-matchup processing: Es is computed as in Equation (4)

using Gordon and Wang (1994) approximations and values of τray (Hansen and Travis, 1974) from the

RMD (Table 6-1 in the RMD).

1.4.5 Normalisation of Case 1 ρw (λ)

In Equation (3), neither quantities (ρw or Lw) are normalised to zenith sun or to nadir viewing angle, which

is a requirement for consistency with MERIS. The normalisation procedure applied to in-situ radiometric

data in MERMAID is the same as that used in the present MERIS Processor, MEGS (v8.0). The

normalisation utilises a look-up table for f /Q, and follows Morel and Gentilli (1993), Morel et al. (1995)

and Morel and Gentilli (1996). In MERMAID, the procedure for normalisation is as follows:

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 7


),',,,,(

),',,,,(

),,(

),,(

),'()()(

0

0

0

sa

sa

a

a

s

wwn

ChlQ

Chlf

ChlQ

Chlf

w (6)

where:

n

)sin(arcsin'

where n is the refractive index of sea water (dimensionless),

)0,0'(0 sw is the „Gothic R‟ factor as described in Equation (7).

),,(

),,(

0

0

chlQ

chlf

a

a

is the f/Q factor for an illumination/viewing configuration at the zenith/nadir,

Chl is known from in-situ or computation from ρw (λMERIS) using the MERIS

algal_1 Chl algorithm,

τa is an assumed constant value in MERMAID

The “Gothic R”, , factor (Morel and Gentilli, 1996) is defined by:

2

))'(1(

)1(

)1()'(

nfr

f

(7)

where:

)'( f is the Fresnel reflectance at the air-sea interface for the scattering angle ' (dimensionless)

is the mean reflection coefficient for the downwelling irradiance at the sea surface (dimensionless)

r is the average reflection for upwelling irradiance at the water-air interface (dimensionless)

are available in the MERIS RMD (AD [3]).

f (,s), is a function relating the apparent optical properties (and specifically the irradiance

reflectance) to IOPs (Morel et al, 2002):

)(

)(

),(

),(),(

bsd

sus

b

a

E

Ef (8)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 8


where Eu (λ) is the upwelling irradiance and Ed (λ) the downwelling irradiance.

a (λ) is the absorption coefficient

bb (λ) is the backscattering coefficient

Q (,‟,s,), is defined as:

),',,(

),(),',,(

su

sus

L

EQ

(9)

1.4.6 AERONET-OC (Ocean Color sites) normalisation

AERONET-OC data are fully normalised according to Morel et al. (2002), and relies on Chl-a determined

from the application of regional bio-optical algorithms to the normalised radiances, Lwn (λ), as in Zibordi

et al. (2009a). The full normalisation of Lwn also accounts for (cos(θs) . d2 .Td; as in Equation 4), thus

allowing for use of F0 to make the conversion to ρwn.

The LUT used for the normalisation is that of Morel, and are available only for a single value of τa, 0.2.

Normalisation is applied assuming case 1 waters only.

The processing of AERONET-OC data is documented in Zibordi et al., (2009b, 2004) and is consistent

with that applied for MERMAID (described in section 1.4.5).

For MERMAID, AERONET-OC radiances are converted into normalised water reflectance, ρwn (λ) by:

)(

)(.)(

0

/_

F

L Qfwn

wn

(10)

1.4.7 Normalisation of Case 2 reflectances.

The theory and method for determining exact normalised water reflectance is limited to Case 1 which are

homogeneous and relatively clear water masses with Chl ≤ 3 mg m-3

, due to the use of an empirical

function for remotely-sensed Chl. The algorithm for normalisation of water reflectance in Case 2 waters is

still in development for MERMAID and a Case 1 normalisation is applied for all sites at present.

1.5 Requirements and recommendations for the validation of MERIS w

In the validation of MERIS w a number of in-situ radiometric measurements are required, ideally at

MERIS bands (Table 1-2).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 9


1. Unnormalised reflectance, ρw to allow for consistency in the normalisation procedure.

2. Downwelling irradiance above the sea surface Ed0+

(λ), i.e. Es (λ).

Ed(0-,λ) is not sufficient due to fluctuations associated with focusing and defocusing of sunlight by

surface waves (Zaneveld et al., 2001) making such measurements noisier than Es (λ) made above the

surface (even though Ed(0-,λ) can be affected by ship or buoy motion).

3. Vertical profile of downwelling irradiance Ed (z) at MERIS wavelengths with reference of Ed (0+) at

surface (W m-2

nm-1

) taken simultaneously.

4. Vertical profile of upwelling nadir radiance Lu (z) with reference of Ed (0+) at surface (W m

-2 nm

-1 sr

-

1) taken simultaneously.

5. Vertical profile of upwelling irradiance Eu (z) at MERIS wavelengths with reference of Ed(0+) at

surface (W m-2

nm-1

) taken simultaneously.

6. Measurements of Ed and Eu or Lu at a minimum of 2 fixed depths in the water column, close to the

surface, and with reference Es taken simultaneously.

According to Mueller et al., (2003c), at present the most reliable in-situ method of determining water-

leaving radiance Lw (λ) is to extrapolate an in-water profile measurement of Lu (z, λ) to the sea surface to

estimate Lu (0-,λ). Using the Fresnel laws, the water-leaving radiance can be computed as:

2

)],'(1[),',,0(),,(

nLL

f

uw

(11)

where ρf (θ,θ‟) is the Fresnel reflectance of the sea-air interface for the associated directions θ (incident)

and θ‟(refracted), and n is the refractive index of seawater, and can be approximated with a value of 1.34

(Austin, 1974). The Fresnel transmittance is represented by the term [1-ρf] on the right hand side of

Equation (11), and can be approximated by the value of 0.975.

MERMAID PIs have used a variety of approaches and constants for these terms, as summarised in Table

1-3. In Section 10 (Table 10-1 and Table 10-2) are provided exact values of transmittance at the air-sea

interface, which may be used by the MVT (and presently used by two PIs, J. Icely and S. Kratzer).

Transmittance through the air-sea interface is a function of wavelength, temperature and salinity and can

be tabulated as function of these 3 parameters. It is a recommendation that LUTs of exactly computed air-

sea transmittance are made available by the MERIS QWG to the MVT and that PIs use them according to

measurement conditions (temperature and salinity).

If the PI's decide to use an approximated value for Fresnel transmittance and the refractive index for

seawater then an adjustment may be incorporated in the MERMAID processing: dividing Lw (λ) from the

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 10


PI by the approximation 0.543 and multiplying by the tabulated air-sea transmittance to account for its

spectral variability and the dependence with temperature and salinity.

Es (λ) is then used to determine RRS by:

)(

)()(

s

wRS

E

LR (12)

The RRS is then related to ρw (λ) by:

RSw R )( (13)

Table 1-3: Air-sea interface terms and values used in the datasets provided to MERMAID (where provided

and where relevant). Empty cells mean the information is not available.

MERMAID

PI

MERMAID

DATASET

Transmittance at

the air-sea

interface

2

)1(

n

f

Fresnel

reflectance at

the air-sea

interface

f

Refractive

index of

seawater

n

Fresnel

transmittance

at the air-sea

interface

f1

D. Antoine BOUSSOLE 0.97

R. Doerffer Helgoland 0.54

A. Hommersom Wadden Sea

J. Icely Sagres

M. Kahru California

Current 0.54

S. Kratzer NW Baltic Sea Factors in Table

10-1

H. Loisel

Fr. Guyanan

and E. English

Channel

0.543 0.025 1.34

D. McKee Bristol channel

& Irish Sea 0.021 1.345

K. Voss MOBY 0.543 0.02 1.34

K. Ruddick MUMMTriOS

D. Siegel Plumes and

Blooms

K. Sorensen Ferrybox 0.02

J. Werdell NOMAD 1.34 0.975

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 11


1.5.1 Measurement procedures and corrections

Ship and instrument position relative to sun

It is critical when making optical field measurements from a ship that due care is taken to avoid shading

and reflection due to or at the ship‟s body. All measurements must be carried out apart from the ship body

to avoid these effects and Mueller et al. (2003c) detail several methods to achieve this. This is an essential

criterion because a correction is not possible. However, in turbid Case 2 water the effect of the

underwater body of the ship is less critical due to the longer optical paths, although care must still be

taken for consistency. Free-falling profilers provide the possibility to operate the instrument sufficiently

away from the ship. Another source of uncertainty that appears in turbid waters is that the ship often

totally disrupts the vertical and horizontal structures of the water mass, so that the representativeness of

measurements is uncertain. Using as small ship as possible is a solution here (as far as logistics permit).

Furthermore, due care should be taken to avoid self-shading by the instrument in use; Mueller et al.

(2003c) provide a correction protocol for this, based on Gordon and Ding (1992) and Zibordi and Ferrari

(1995).

Dark readings

The dark current of optical sensors is frequently temperature dependent. As a consequence, accurate

radiometric measurements require that careful attention be given to dark current variability. Mueller et al.,

(2003c) recommend that each optical measurement be accompanied by a measurement of the instrument

dark current.

Patchiness

Optical measurements in patchy waters are of little value for validation because of the difficulty of

defining matchups. But they retain their value for the establishment of models. In that case a number of

IOP samples should be taken at the same time as reflectance samples to establish a statistical relationship.

Patchy waters may be sampled, although protocols for the matchup procedure are yet to be defined. The

strengthened focus on improved validation in Case 2 waters means that currently all data is accepted from

a PI. However extreme care must be exercised because it is impossible to relate concentrations to optical

measurements if a sample has been taken some 10 metres apart from the water body observed by the

optical instrument or if optical measurement and sample are not taken at the same time.

1.5.2 Ideal criteria for validation

There are a number of ideal criteria that should be met when making in-situ optical measurements:

Water dynamics low, to get data points which are valid for the geographical position for a few hours;

Water depth >> optical depths at wavelength of maximum transparency;

Vertical homogenous water;

Horizontal homogenous water;

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 12


Sufficiently far from land to avoid inclusion of land surface in pixel and influence of land reflectance

on atmospheric path radiance (> 5km);

Sun zenith angle s at the time of measurement /satellite pass should be < 60°. Note that pixels with a

larger s are flagged in the MERIS L2 product;

For a direct comparison between reflectance measurements and data derived from MERIS only clear

sky conditions with preferably low aerosol concentrations are acceptable;

Radiometric measurements should be discarded for validation when contaminated by thin stratus,

cirrus clouds or fog.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 13


2. The ‘MERIS MAtchup In-situ Database’ (MERMAID)

2.1 Introduction

The MERMAID database (Figure 2-1) was initiated in

2007 at ACRI-ST with the collation of radiometric

data from established time-series measurement sites.

In 2008 it officially became „MERMAID‟ (Barker et

al., 2008) and efforts were increased by ARGANS Ltd

to acquire more in-situ datasets. ACRI-ST and

ARGANS work together to improve the database and

extraction website. ARGANS‟ current efforts are

aimed at gathering further radiometric measurements,

and also now atmospheric parameters (to be addressed

in a subsequent protocols document). ARGANS is the

initial point of contact for the PI, whereby an

agreement of use is negotiated and the data is provided

in a suitable format. ACRI-ST is responsible for

matching the data with MERIS products and making

the database available online via a web interface.

The MERMAID data format document

(http://hermes.acri.fr/mermaid/format/format.php)

outlines the parameters in MERMAID. MERMAID is

hosted at http://hermes.acri.fr/mermaid. All queries,

suggestions and feedback can be made to

[email protected], and new PIs should make contact

via this address.

2.1.1 Minimum data requirements: Optical data and auxiliary information

MERMAID is flexible in how PIs submit their data, however it requested that PIs provide as a minimum:

Water reflectances, ρw (visible and NIR, at MERIS bands if possible); either multi or hyperspectral.

Or, hyperspectral convolved to the last definition of the 15 MERIS spectral filters;

The associated water-leaving radiances, Lw (λ) and downwelling surface irradiance, Es (λ), or Ed (λ

,0+), from which ρw were computed (or relevant CIMEL or TriOS parameters);

Associated Chl-a measurements (if available), with a description of the method to derive it;

Sun zenith angles if available;

Associated meta data (latitude, long, date, time in UTC);

A written protocol to be included in the MERIS Optical Measurement Protocols document; it is a

requirement of potential usage in matchups that adherence to an accepted protocol is confirmed.

Figure 2-1: MERMAID Website

http://hermes.acri.fr/mermaid/format/format.php

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 14


Additional information ideally consists of the sun zenith angle measurement, θs, and, if available, a

measurement of Chl-a, which is a requirement for the ρw (λ) normalisation procedure. If the θs is not

provided, it is computed as a function of latitude and longitude (both in decimal degrees), and time (UTC

time of observation). If in-situ Chl-a is not provided, it is computed from in-situ ρw using the MERIS

chlorophyll algorithm.

2.1.2 MERMAID as a validation tool for ODESA: Additional parameter requests

ODESA, the Optical Data processor of ESA

(Figure 2-2, http://earth.eo.esa.int/odesa/),

provides users a complete level 2 processing

environment for MERIS, as well as for the

future ESA optical sensors on board Sentinel 3.

ODESA supplies the user community with the

MERIS Ground Segment development platform

MEGS®, including source code, embedded in

an efficient framework for testing and for

validation activities. Such facilities include

match-up processing & analysis using

MERMAID, and to this end MERMAID can

now accept other data than optical.

If available MERMAID now accepts

concentrations such as TSM and CDOM, and

primary inherent optical properties (IOPs) i.e.

total absorption, at (λ); backscattering bb (λ)

and the component IOPs (i.e. those contributing

to at (λ) and bb (λ)).

2.1.3 MERMAID in-situ data format

MERMAID presently does not strictly specify a format for data submission. On receipt, the data are

formatted to a format suitable for the web interface (a section of which is exemplified in Figure 2-3). The

in-situ template consists, as minimum, of the geographical and temporal information, θs, Chl-a (if

available), depth and ρw (λ). Traceability of the data is essential and the template retains the site and the PI

name. Any additional parameters submitted are included in columns after these mandatory fields.

Figure 2-3: Samples from the MERMAID template; formatted PI in-situ data. This is a sample from AAOT.

Figure 2-2: ODESA online processing. (Figure

courtesy of ACRI-ST).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 15


2.1.4 Current datasets in MERMAID

Table 2-1 summarises all the datasets currently in MERMAID and the associated contact details of the respective PIs.

Table 2-1: Datasets in MERMAID and details of the associated PI. Acronyms are fully expanded in the Abbreviations and Definitions list on Page xii

DATASET LOCATION PARAMETERS

IN MERMAID PI PI Affiliation PI email contact

AE

RO

NE

T-O

C

AAOT N. Adriatic Sea

45.31oN, 12.50

oE

Lwn (λ), a (λ) G. Zibordi Joint Research Centre, Ispra,

Italy (JRC)

guiseppe.zibordi

@jrc.it

Abu Al-

Bukhoosh

Arabian Gulf

25N 53E Lwn (λ), a (λ),

Chl

G. Zibordi Joint Research Centre, Ispra,

Italy (JRC)

guiseppe.zibordi

@jrc.it

COVE_

SeaPRISM

Cove, Virginia

36.9oN, 75.7

oE

Lwn (λ), a (λ),

Chl

B. Holben NASA GSFC [email protected].

nasa.gov

Gloria Black Sea

44.3oN, 29.2

oE

Lwn (λ), a (λ),

Chl


Italy (JRC)

guiseppe.zibordi

@jrc.it

Gustav-

Dahlen Tower

Baltic Sea

58N, 17E Lwn (λ), a (λ),

Chl


Italy (JRC)

guiseppe.zibordi

@jrc.it

LISCO Long Island Sound, USA

40.5oN, 73.2

oE

Lwn (λ), a (λ) ,

Chl

S. Ahmed/

A. Gilson

City College of New York ahmed / gilerson

@ccny.cuny.edu

LJCO Queensland, NE Australia

18.3oS, 146.2

oE

Lwn (λ), a (λ) ,

Chl

V. Brando CSIRO, Australia Vittorio.Brando

@csiro.au

Helsinki

Lighthouse

Baltic Sea

59N, 24E Lwn (λ), a (λ) ,

Chl


Italy (JRC)

guiseppe.zibordi

@jrc.it

Pålgrunden Lake Vänern, Sweden

58N, 13E Lwn (λ), a (λ) ,

Chl

S. Kratzer University of Stockholm,

Sweden

Susanne.kratzer

@ecology.su.se

MVCO Martha‟s Vineyard, USA.

41N, 70W Lwn (λ), a (λ) ,

Chl

D.

Vandemark

University of New Hampshire,

USA

doug.vandemark

@unh.edu

WAVE_CIS_

Site_CSI_6

Gulf of Mexico

28.8oN, 90.4

oE

Lwn (λ), a (λ) ,

Chl

B.Gibson

A.

Weidermann

Coastal Studies Inst. Louisiana,

USA & NRL, Mass. USA

[email protected] &

Alan.Weidemann

@nrlssc.navy.mil

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 16


DATASET LOCATION PARAMETERS

IN MERMAID PI PI Affiliation PI email contact

Algarve Sagres, 36/37N. 8E ρw (λ), Es (λ) J. Icely Sagremarisco Lda,

Portugal

alicely2

@gmail.com

BOUSSOLE W. Med. 43.367oN, 7.9

oE ρw (λ), Chl,

Es (λ)

D. Antoine Laboratoire d'Océanographie de

Villefranche, France (LOV).

Antoine

@obs-vlfr.fr

Bristol

Channel and

Irish Sea

Bristol Channel &

Irish Sea

Variable: 51/54N, -3/-4E

ρw (λ), Ed (λ) D. McKee University of Strathclyde, UK. david.mckee

@strath.ac.uk

California

Current

California Coast. Variable:

32-34oN, 120-121

oW

Lu (λ), Es (λ). M. Kahru University of California, San

Diego

[email protected]

English

Channel

English Channel ρw (λ) H. Loisel/

C. Jamet

laboratoire d'oceanologie et de

geosciences (LOG)

Cedric.Jamet

@univ-littoral.fr

French

Guiana

French Guiana ρw (λ) H. Loisel/

C. Jamet

laboratoire d'oceanologie et de

geosciences (LOG)

Cedric.Jamet

@univ-littoral.fr

MOBY Lanai, Hawaii

20.822oN, 157.187

oW

Lw (λ), Es (λ). K. Voss National Oceanic and

Atmospheric Administration,

USA (NOAA).

Kenneth Voss

@noaa.gov

MUMM_

TriOS

Variable ρw (λ), Ed (λ), Lse (λ),Lsk(λ), Lw

(λ), SPM, Chl-a.

K. Ruddick Management Unit of the North

Sea Mathematical Models,

Belgium. (MUMM)

K.Ruddick

@mumm.ac.be

NOMAD Variable Lw (λ), Es (λ). J. Werdell NASA Goddard Flight Centre,

USA.

Jeremy.werdell

@gsfc.nasa.gov

N.W. Baltic

Sea

NW Baltic

Variable: 58N, 17E ρw (λ), Es (λ) S. Kratzer University of Stockholm,

Sweden

Susanne.kratzer

@ecology.su.se

Plumes and

Blooms

D. Siegel

SIMBADA Variable ρw (λ) a (λ)

P-Y.

Deschamps

Laboratoire Optique

Atmospherique - Université de

Lille (LOA).

Deschamp

@univ-lille1.fr

Wadden Sea Wadden Sea

52-53N, 4-6W RRS (λ) A.

Hommersom

University of Amsterdam annelies.hommersom

@ivm.vu.nl

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 17


2.1.5 Matchup with L2 MERIS data

MERIS Level 0 data are received by ACRI-ST through DDS (Data Dissemination System) and locally

archived. From the geographic and temporal information, ACRI-ST processes the relevant L0 products

with the MERIS Ground Segment data processing prototype, (MEGS 8.0) up to L1. MEGS is developed

and maintained in ACRI-ST and is in line with the MERIS IPF (Instrument Processing Facility). A

custom processing of the L1 up to Level 2 is required as some of the MERMAID data are intermediary

processing products and are therefore not available in standard L2 products. The generation of these

intermediary products is handled through ODESA. From these MERIS data is extracted a range of

products coincident with the in-situ information.

Extraction is achieved on 5x5 reduced resolution (RR) pixels around the site corresponding to in-situ

acquisition. The default extraction criteria selected on MERMAID web page follows the procedure

described by Bailey and Werdell (2006) when applicable. This procedure concerns time elapse between in

situ and satellite measurement, flags selection and statistical screening:

Time elapse: Difference in time between MERIS and the in-situ measurement does not exceed 3

hours. In the web interface the user can specify their preferred time difference up to +/-12 hours. +/-

12 hours is the search limit for matching MERIS overpasses.

Flags: At least 50% of the pixels in the box (selected macropixel) are not flagged as land, cloud,

high/medium-glint, ice haze, PCD-1-13 (uncertain normalised surface reflectance) or PCD-19

(uncertain aerosol type/optical thickness). The latter two flags correspond to a failure in atmospheric

correction and thus depend highly on the algorithm itself. However, Bailey and Werdell (2006)

recommend their inclusion.

Statistical screening: For a given wavelength, the mean, and standard deviation of ρw ( w , σ) is

computed over non-flagged pixels. Based on these two parameters, two statistical tests are performed

on a band per band basis, allowing some reflectances to be selected and other not:

o The “filtered mean”: A value in a macropixel is rejected if *)( FCww .

FC the filtering coefficient is set to 1.5 but it can be specified.

o The “coefficient of variation” (CV criteria): the entire macropixel is rejected if

CVw . The CV coefficient is set to 0.15 but it can be specified band per

band.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 18


2.2 Atmospheric parameters in MERMAID

Included in the MERMAID database are a number of atmospheric parameters: τa(870), τa(665) and α. If

the PI has these to provide, τa(865) and τa(665) are included directly in MERMAID. Presently, just

AERONET-OC sites provide these parameters to MERMAID, and in the MERMAID processing, α is

computed from τa following the Gordon and Wang (1994) approximation:

865.)865()( aa (14)

In the case of AERONET-OC 870 nm is received, therefore it is used instead of 865 nm.

The bands used to compute α are either 870 nm and 665 nm, or 870 nm and 675 nm, depending what is

received from the AERONET-OC site. This is indicated in the PQC column of an extraction (Processing

Quality Control), in the 7th and 8

th bit of the flag string (see Section 2.5 and the relevant extraction file for

more information).

2.3 MERMAID uncertainties

The various measurement protocols attached to each of the datasets in MERMAID will introduce

uncertainties specific to the measurement system and which should be assessed in order to qualify the

measurements in question as being usable for validation purposes. PIs are therefore encouraged to provide

traceability of their measurements and credible uncertainty estimates.

2.4 ρw spectral correction

For the AERONET-OC sites Aqua Alta Oceanographic Tower (AAOT), Gustav-Dahlen Tower (GDT),

Helsinki Lighthouse Tower (HLT) Zibordi et al. (2009a) have developed a series of algorithms based on

local measurements of IOPs to band-shift correct radiance measurements at the AERONET-OC bands to

MERIS (or any other sensor) bands. G. Zibordi performs the correction for AAOT, but for the present

time the GDT and HLT data remain at AERONET-OC bands.

Table 2-2 summarises the bands at which in-situ data is received and included in MERMAID. For those

not at MERMAID bands, the current method for MERMAID matchup is to use the data at the bands

nearest to the MERMAID bands (but not > 5nm).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 19


Table 2-2: In-situ bandsets in MERMAID

DATASET CENTRE BANDS

MERMAID 412.5 442.5 490 510 560 620 665 681.25 708.75 753.75 778.75 865 885

AE

RO

NE

T-O

C

AAOT 413 443 490 - 560 -- 665 -- -- -- -- -- --

Abu-Al-Bukhoosh 413 440 500 -- 555 -- -- 675 -- -- -- 869 --

COVE_SeaPRISM 413 441 489 -- 551 -- 668 -- -- -- -- 869 --

Gloria 412 441 491 -- 555 -- -- 675 -- -- -- 870 --

Gustav-Dalen Tower 412 439 500 -- 554 -- -- 675 -- -- -- 870 --

Helsinki-Lighthouse 413 441 491 -- 555 -- 668 -- -- -- -- 870 --

LISCO 413 442 491 -- 551 -- 668 -- -- -- -- 870 --

LJCO 412 441 491 -- 551 -- 668 -- -- -- -- 870 --

MVCO 412 439 500 -- 555 -- -- 674 -- -- -- 870 --

Pålgrunden 412 440 490 -- 555 -- 668 -- -- -- -- 868 --

WAVE_CIS 411 442 491 -- 555 -- 668 -- -- -- -- 869 --

Algarve 412.69 442.56 489.88 509.81 559.59 619.60 664.57 680.82 708.32 753.37 778.41 -- --

BOUSSOLE 412 443 490 510 560 -- 665 683 -- -- -- -- --

Bristol Ch. & Irish Sea 412 443 489 510 554 665 -- 700 -- -- -- --

California Current 412 443 490 510 555 625 665 -- 710 -- -- -- --

English Channel 412 442 490 511 559 619 664 -- -- -- 775 -- --

French Guiana 412 442 490 510 560 620 664 -- -- -- 776 864 --

MOBY 412.5 442.5 490 510 560 620 665 681.25 708.75 753.75 778.75 865 885

MUMMTriOS 412.5 442.5 490 510 560 620 665 681.25 708.75 753.75 778.75 865 885

NOMAD 411 443 489 510 555/560 619 665 683 -- -- -- -- --

NW Baltic Sea 412 443 490 510 560 620 665 681 708 -- -- -- --

Plumes and Blooms 412 443 490 510 555 665 -- --

SIMBADA 410 443 490 510 560 620 670 -- -- 750 -- 870 --

Wadden Sea 412 443 490 510 560 620 665 681 708 -- -- -- --

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 20


2.5 Measurement and Processing flags

Flags are required to define the level of quality control (QC) applied to the in-situ data. Two QC indicator

columns are included in MERMAID, each made of a string of 18 characters pertaining to a “flag”, i.e. a

particular aspect of the measurement protocol processing procedure.

MQC (Measurement Quality Control): defines the quality control checks made by the PI, prior to

submission. It pertains to the provision, or not, of a clearly defined measurement and processing protocol

by the in-situ data provider. An example MQC string is: M120000001101111210; the numbers are the

string options in the position indicated in Table 2-3, in this example protocol is provided, indicated in

position 1, no correction for straylight is made (position 3), and so on.

Table 2-3: MQC flag criteria definition. Flag position is counted from the first numeric character after the

leading ‘M’. Unless otherwise specified: 0 = No / Not done, 1 = Yes / done/ provided, 2 = Unknown / not

available / not applicable (N/A).

Flag

ID

Flag

position

String

options

Conditions and criteria

MQ

C

1 0 1 Protocol provided by PI

2 0 1 2 Correction of self-shading (2 = N/A as self-shading avoided).

3 0 1 2 Correction for straylight (2 = N/A)

4 0 1 2 Made dark measurements (and used in processing)

5 0 1 2 Measured immersion coefficients (and used in processing)

6 0 1 Instrument calibration history provided

7 0 1 Data processed to MERIS band characterisation

8 0 1 Hyperspectral integration done

9 0 1 Error budget provision

10 0 1 2 (L1.5) In-situ data filtering (PI‟s QC checks)

11 0 1 In-situ ρw already normalised or f/Q and corrected

12 0 1 2 Tilt measurement made

13 0 1 2 Calibration of tilt sensor

14 0 1 2 3 Type of Es: Es or Ed(0+) (0 = N/A, 1 = Es measured in-situ, 2 = Ed(0+)

measured in-situ/derived in-situ, 3 = Es computed)

15 0 1 Es tilt corrected

16 0 1 2 Type of Lu: Lw or Lu(0-) (and extrapolated to Lw(0+). (0 = N/A, 1 = Lw, 2 =

Lu(0-) (and extrapolated to Lw(0+).

17 0 1 Lu tilt corrected

18 0 1 2 (L1.5) (AERONET-OC only) Data quality level: 0 = N/A, 1 = L1.5, 2 = 2.0 (see

AERONET website http://aeronet.gsfc.nasa.gov, for more details)

http://aeronet.gsfc.nasa.gov/

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 21


PQC (Processing Quality Control): defines the post-submission quality control performed on the in-situ

data by ARGANS, and includes information on the normalisation procedure. The flag provides

information on the MERMAID in-situ processing for both the optical data received (e.g. ρw) and the

atmospheric parameters (e.g. α) newly added to the database. An example PQC is: P1010100 i.e. the ρw

passed the QC test (position 1), and it got Case 1 normalisation, not Case 2 (positions 3 and 4).

Table 2-4: PQC flag criteria definition. Flag position is counted from the first numeric character after the

leading ‘P’. Unless otherwise specified: 0 = No / Not done, 1 = Yes / done/ provided, 2 = Unknown / not

available.

Flag

ID

Flag

position

String

Options Conditions and criteria

PQ

C

1 0 1 Passed in-situ ρw QC

2 0 1 2 Hyperspectral integration

3 0 1 Case 1 Normalisation by MERMAID *

4 0 1 Case 2 Normalisation

5 0 1 Band shifted correction [AERONET-OC data only; presently only AAOT]

6 0 1 Nearest neighbour (refer to MERIS Optical Protocols):

0 = data at bands greater than ±5nm from MERIS

1 = data at bands less than ±5nm from MERIS

**NOMAD only: Flag is 0 when data is at 560 nm and 1 when at 555 nm

7 0 1 AlphaNIR (1 & 2) derived from 870-675 nm

8 0 1 AlphaNIR (1 & 2) derived from 870-665 nm

*See MQC flag #11 to check if normalisation has already been performed by PI.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 22


2.6 Data Access and Policy

2.6.1 User login and Password

MERMAID is subject to a strict data access policy viewable at

http://hermes.acri.fr/mermaid/policy/policy.php.

The database is made available to the MERIS QWG, the MVT and the contributing PIs through an

access-restricted data extraction page, for which a unique password is provided. PIs are given access if

they have submitted in-situ data and matchups are confirmed. Restricted access such as this allows for

better security and for site-use monitoring.

The password and login details must not be passed on to others; the MERMAID team must be contacted

and the colleague in question will be considered but not guaranteed access.

We welcome use of MERMAID outside the scope of the MERIS maintenance and evolution project.

Interested users who are not part of the MQWG, MVT or are not PIs, can request access with a unique

password through a Service Level Agreement. Please email [email protected] to express interest and

provide a description of your project.

2.6.2 MERMAID web interface.

The web interface (Figure 2-4) is versatile,

allowing users to specify their own extraction

criteria. In addition to selecting sites and dates, the

user can, for instance, extract matchups for a 1,

3x3 or 5x5 pixel grid, and for a time difference of

up to 3 hours. Geometry options are available to

the user, as are flag selection and statistical

screening options, all allowing for adaptable and

flexible matchup selection.

Figure 2-4: The MERMAID data extraction

webpage and extraction options, on the MERMAID

website available (password restricted).

Processing version

Site selection and date range

Physical screening and flag acceptance options

Statistical screening

Correction on ρw for theoretical Es

http://hermes.acri.fr/mermaid/policy/policy.php

mailto:[email protected]

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 23


2.6.3 Data submission

PI‟s should write to [email protected] to enquire about data submission. ARGANS is the next point of

contact for the PI, who is requested to submit data in any format (e.g. ASCII, HDF), as long as it is

adequately labelled and is accompanied by a protocol describing measurement and processing methods.

2.6.4 MERMAID extraction package

Once selections have been made, MERMAID extractions are provided as „Filzip‟ files: zip files

containing the extracted data, statistics and uncertainties files, regression plots and histograms. An

example is shown in Figure 2-5.

Figure 2-5: Example extraction items: CSV file of extracted data, descriptive plots and RBG.

2.6.5 Acknowledgements and Proprietary Rights

The MERMAID data policy (http://hermes.acri.fr/mermaid/policy/policy.php) requires that when

MERMAID extractions are used in publications, the Principal Investigators of in situ data (PIs) should

always be contacted for approval, be offered co-authorship and acknowledged. The PIs and their contact

details are listed on the website and in Section 2.1.4.

ACRI-ST and ARGANS should always be acknowledged too, as quality control, post-processing, MERIS

processing, extractions, database system and web facility are proprietary and operated on behalf of ESA.

Appropriate acknowledgement suggestions are made on the data policy page.

Proc. Version, Site, PI

mailto:[email protected]

http://hermes.acri.fr/mermaid/policy/policy.php

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 24


MERMAID Optical Protocols

This section presents the in-situ measurement protocols accompanying datasets supplied to MERMAID,

according to the MERMAID data policy. Also included here are protocols received from PIs from whom

data is expected; updates to this document will occur as new data and protocols are received.

3. MERMAID PROTOCOLS I: SeaPRISM (AERONET-Ocean Color)

3.1 Introduction

The Aerosol Robotic Network (AERONET) has an ocean colour (OC) component (AERONET-OC),

making use of autonomous above water radiometers (SeaPRISM) fixed on platforms located in coastal

regions. The network standardises measurements performed at different sites, with the same

instrumentation and data processing.

Twelve sites exist, of which MERMAID presently has 11 (Figure 3-1; USC_SEAPRISM is not yet

included):

Figure 3-1: AERONET-OC sites in MERMAID.

3.1.1 Products in MERMAID

AERONET-OC provides processed, quality-controlled data accessible online with a specified data policy

(http://aeronet.gsfc.nasa.gov/new_web/ocean_color.html). The products in MERMAID are the normalised

water reflectance, ρwn (λ) as computed from AERONET-OC normalised water-leaving radiance, Lwn (λ),

aerosol optical thickness a (λ) and Chl.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 25


3.1.2 Standardised measurement and processing procedure

SeaPRISMs are calibrated just before and after each field deployment (generally lasting 6-12 months). On

average, changes of less than 1% have been observed in instrument sensitivity during successive

deployments in 2006 and 2007. Data transmission is made through the METEOSAT meteorological

satellite that ensures almost real-time data handling. Collected data are processed and quality- assured at

the Goddard Space Flight Centre (GSFC) of the National Aeronautics and Space Administration (NASA).

Each SeaPRISM measurement sequence is executed every 30 minutes within ±4 hours around the local

noon, and comprises:

1. A series of direct sun measurements E as a function of λ, solar zenith and azimuth angles

(respectively θ0 and 0) for the determination of the aerosol optical thickness τa (λ), an ancillary

quantity required for the normalisation of water-leaving radiance, Lwn (λ);

2. A sequential set of NT sea-radiance measurements for determining total radiance from the sea, Lt (λ, θ,

) and of Ni sky-radiance measurements for determining the downwelling radiance at ground level,

Lsky (λ, θ′′,), serially repeated for each λ (where θ′′ =π-θ and is the relative azimuth with respect to

the sun). Lt (λ, θ, ) and Lsky (λ, θ′′, ) values are determined at θ = 40° and = 90°.

If the sun is cloud covered then E (λ, θ0, 0) measurements cannot be performed and the whole acquisition

sequence is cancelled. The sky and sea measurements for determining Lsky (λ, θ′′, ) and Lt (λ, θ, ) are

performed with Ni = 3 and NT = 11.

The total radiance measured at the uncontaminated sea surface, Lt (λ), is a combination of Lw (λ) plus two

sources of reflected light or glint: the sky and the sun. Sun glint is avoided by pointing the instrument

away from the sun by at least 90° away from the solar plane but not into any perturbations associated with

the platform. Aggressive filtering can be used to remove any glint spikes caused by oblique wave facets

(Hooker et al., 2002; Zibordi et al., 2002). Then, the only quantity needed for retrieval of Lw (λ) from Lt

(λ) is an estimate of the contribution of the sky radiance, Lsky (λ).

Additional measurements are performed at 870 and 1020 nm for quality checks, turbid water flagging,

and for the application of alternative above-water methods (Zibordi et al., 2002).

Data Processing of Lwn (λ)

A set of criteria is defined for the processing of AERONET-OC measurements and determination of Lwn

(λ). Principally, processing is only applied to measurement sequences fulfilling the following criteria

(Zibordi et al., 2009b):

1) there is no missing value;

2) dark values are below a given threshold;

3) measurements are performed with 0 values included within site-dependent limits to minimise

superstructure perturbations in Lt (λ, θ, ,), i.e. tower shading;

4) aerosol optical thickness data have been determined;

5) the wind speed is lower than 15m s-1

.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 26


For each measurement sequence qualified for the data processing, Lsky (λ, θ′′, ) is determined by

averaging the Ni sky-radiance data. Lt (λ, θ, ) is determined from the average of a fixed percent of the NT

sea-radiance measurements exhibiting the lowest radiance levels. This approach has been suggested by

independent studies (Hooker et al., 2002; Zibordi et al., 2002) highlighting the need for an aggressive

filtering of above-water measurements to minimise the perturbing effects of surface roughness in Lt (λ, θ,

).

From Lt (λ, θ, ) and Lsky (λ, θ′′, ), the water-leaving radiance Lw (λ, θ, ), i.e. the radiance emerging from

the sea quantified just above the sea surface, is computed as:

) ,'',(),,,(),,(),,( 0 skysftw LwLL (15)

where: ρf (θ, θ0, , ws) is the sea surface reflectance as a function of the measurement geometry identified

by θ, θ0, , and of the sea state expressed through wind speed, ws. The value of ρf (θ, θ0, , ws) at a given θ

and , can be theoretically determined as a function of θ0 and ws (Mobley, 1994).

Normalisation is carried out as part of the AERONET-OC processing scheme which is the same that that

used for MERMAID (and as described in section 1.4.6). For AERONET-OC, normalisation is performed

using F0 and formulations (as in Zibordi et al., 2004) requiring a series of transmittance computations.

The quantity td (λ) used in the AERONET-OC normalisation procedure is computed utilizing τa (λ)

determined from SeaPRISM measurements (the exponent being -0.16 in this instance). The Rayleigh

optical thickness (τray) used for computation of AAOT Rayleigh Transmittances is derived from Bodhaine

et al. (1999), which differs from nominal MERIS usage of Hansen and Travis (1974).

As Es is not available from AERONET-OC website, for the MERMAID pre-matchup processing for

AERONET-OC, Es (needed for computation of ρw_ISME(λ); section 1.4.4) is computed using Gordon and

Wang (1994) approximations and Hansen and Travis (1974) τray (values found in the MERIS RMD).

3.1.3 Quality assurance

All AERONET products are classified at three different quality assurance (QA) levels. Data at Level 1.0

only include Lwn (λ) determined from complete measurement sequences satisfying the basic criteria

addressed in data pre-processing. Level 1.5 Lwn data are derived from Level 1.0 products. Zibordi et al.

(2009b) detail the differences between levels 1.0 and 1.5, but for MERMAID, L2 is used. Fully quality-

assured Level 2.0 data refer to Lwn (λ) determined from Level 1.5 products for which:

1) Level 2 aerosol optical thickness data exist;

2) pre- and post-deployment calibration coefficients for LT and Lsky

measurements were determined and

exhibit differences smaller than 5%;

3) the Lwn (λ) spectral shapes are shown to be consistent through tests based on statistical approaches;

4) the Lwn (λ) passing all former tests do not exhibit dubious values during a final spectrum-by-spectrum

screening performed by an experienced scientist.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 27


While most QA tests rely on the application of thresholds, the methodology applied for the assessment of

the spectral consistency makes use of statistical methods effective in detecting artefacts in the shape of

Lwn (λ) spectra (D'Alimonte and Zibordi, 2006). In particular, the applied scheme rejects spectra

exhibiting:

1) low statistical representativeness within the data set itself („self-consistency’ test);

2) anomalous features with respect to a reference set of quality-assured data („relative-consistency’ test).

3.1.4 Uncertainties

Lwn uncertainties (quadrature sum) for AAOT are currently stated by Zibordi et al. (2009b) as 5.1% (421

nm), 4.5% (443), 4.7% (488 nm), 4.7% (551 nm) and 7.8% (667 nm). Zibordi et al. (2009b) go into

further detail on how these values are derived; the quadrature sum percentages are the result of

uncertainties in different contributing terms such as: absolute calibration; sensor sensitivity change

between calibrations; correction applied for removing dependences to the viewing angle and anisotropy of

light field in seawater; determination of Td; determination of ρf due to wave effects and data filtering;

value of ws; environmental effects.

Normalisation of all AERONET-OC data (as described in section 1.4.6) is applied assuming case 1 waters

only, and this is taken into account when defining the uncertainty budget of the normalised data.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 28


3.2 The Aqua Alta Oceanographic Tower (AAOT). PI: Giuseppe Zibordi

3.2.1 Introduction

The Aqua Alta Oceanographic Tower (AAOT) is part of AERONET-OC, a framework supporting ocean

colour validation activities through standardised radiometric measurements in coastal water (Zibordi et

al., 2009b; Zibordi et al., 2002; Zibordi et al., 2006; Zibordi et al., 2004). The AAOT is located in the

Northern Adriatic Sea (45o 18.51 N, 12

o 30.30 E), east of Italy, 8 nautical miles off the coastline of

Venice, and of 17 m depth, and at a site characterised by both Case 1 and Case 2 water types, and of a

mostly continental (sometimes marine) marine aerosol type

3.2.2 Measuring system and configuration

The AAOT (Figure 3-2) consists of four levels. At the fourth level the above-water instrumentation is

positioned, and includes the SeaPRISM system which is an adapted CE-318 autonomous sun-photometer.

The accurate sun-tracking required for SeaPRISM measurements imposes that the deployment platform is

a grounded structure. This structure needs to be at a distance from the mainland suitable to assume that

the adjacency effects are negligible in satellite data (Zibordi et al., 2002).

Figure 3-2: The AAOT structure and instrumentation; a) the main tower and operational levels (from

Hooker et al., 2005), b) the CIMEL CE-318 (SeaPRISM) instrument (from

http://aeronet.gsfc.nasa.gov/new_web/photo_db/Venise.html).

b) a)

http://aeronet.gsfc.nasa.gov/new_web/photo_db/Venise.html

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 29


3.3 Abu Al-Bukhoosh. PI: Giuseppe Zibordi

Abu Al-Bukhoosh is located at 25.49 ºN, 53.15 ºE in the Arabian Gulf. The PI of this site is G. Zibordi.

3.4 CERES Ocean Validation Experiment (COVE_SeaPRISM). PI: B. Holben

The Clouds and the Earth's Radiant Energy System (CERES) experiment is one of the highest priority

scientific satellite instruments developed for NASA's Earth Observing System (EOS). The CERES Ocean

Validation Experiment (COVE) provides continuous world-class measurements at the Chesapeake

Lighthouse for validation of CERES and other satellite products. This instrument site is located on the

Chesapeake Lighthouse ocean platform, a Coast Guard platform located 25 km East of Virginia Beach,

Virginia (Figure 3-3).

COVE is located outside of the surf zone and far enough away from shore to make it an excellent

validation site for space-borne retrievals of cloud and aerosol microphysics. The platform itself is small

(25x25 meters) and aerosol climatology at this location indicates optical depths and Angstrom exponents

that are consistent with polluted urban aerosols. Not all airmasses at Chesapeake Lighthouse are polluted,

however, as Easterly winds from occasional synoptic systems and frequent sea breezes provide a marine

aerosol source.

Instrumentation is located at the housing level and on the roof of the accompanying lookout tower at 37

meters above the surface, and includes the AERONET sunphotometer, uplooking and downlooking

multifilter rotating shadowband radiometers (MFRSRs), as well as pressure, temperature, relative

humidity, global positioning system integrated precipitable water vapor (GPS-IPW), wave height and

period. Much of the current instrumentation has been providing continuous data since March, 2000. A

local area network (LAN) at the facility and a microwave link to shore provide robust digital

communications.

Figure 3-3: COVE SeaPRISM site, 25 km East of Virginia Beach, Virginia: a) site location; b) Lighthouse

platform; c) AERONET sunphotometer.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 30


3.5 Gloria. PI: G. Zibordi

The Gloria site (Figure 3-4) is an oil platform 12 nautical miles from the Romanian coast in the Black

Sea, south of the Danube plume and north of Constanta (44.56o N; 29.36

o E).

Figure 3-4: Gloria Platform, Black Sea.

3.6 Gustav-Dahlen Tower. PI: Giuseppe Zibordi

AERONET-OC station Gustav-Dahlen Tower (GDT) is located in the Northern Baltic Sea, 10 nautical

miles off the Swedish coast, and in an average depth of around 16 m (Figure 3-5 a). This location is at

58.59 ºN, 17.47 ºE. The GDT is owned and managed by the Swedish Maritime Administration and the

AERONET PI of this site is G. Zibordi. In addition to SeaPRISM measurements, marine observations are

carried out 25 nautical miles east of the tower site within the framework of the National Monitoring

Program managed by the Swedish Environmental Protection Agency, which is focused on human impact

on seas and coastal areas, supports monthly monitoring of various environmental parameters including

Chl (Zibordi et al., 2009a).

3.7 Helsinki Lighthouse. PI: Giuseppe Zibordi

The Helsinki Lighthouse (Figure 3-5 b) is owned and managed by the Finnish Maritime Administration,

and located at 59.949° N and 24.926° E, in the Gulf of Finland, and approximately 12 nautical miles

south east of the harbour of Helsinki in an average water depth around 13 m. The PI of this site is also G.

Zibordi. Bio-optical data are also collected close to the tower through autonomous systems operated on

ferries (see Zibordi et al., 2009a and references therein). The Chl-a data included for this site in

MERMAID are determined from regional algorithms; they do not come from field measurements.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 31


Figure 3-5 a) Gustav-Dahlen Tower in the northern Baltic Proper. The inset is a picture of the SeaPRISM

autonomous radiometer installed on the tower; b) The Helsinki Lighthouse in the Gulf of Finland.

3.8 Long Island Sound Coastal Observatory (LISCO). PI: S. Ahmed, A. Gilerson

The LISCO site (Figure 3-6) is situated in Western Long Island Sound, 2 miles offshore (40.95o N; 73.34

o

E).

Figure 3-6: LISCO site, Long Island Sound.

a) b)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 32


3.9 Lucinda Jetty Coastal Observatory (LJCO). PI: V. Brando

The Lucinda Jetty Coastal Observatory (LJCO, Figure 3-7) is located on the end of the 5.8 km long

Lucinda Jetty (18.52 S, 146.39 E) in the coastal waters of the Great Barrier Reef World Heritage Area

close to the Herbert River Estuary and the Hinchinbrook Channel. It is operated by CSIRO.

Two different data streams are acquired: above water measurements of the water radiance and in water

measurements of the optical properties. A CIMEL SeaPRISM and Satlantic HyperOCR are used for each

type of measurement, respectively. An in situ water optical package is also deployed to measure:

Conductivity, temperature, pressure, dissolved oxygen, chlorophyll fluorescence and turbidity

(WETLabs WQM);

Coloured dissolved organic matter fluorescence (WETLabs WETStar fluorometer);

Particulate and dissolved absorption and attenuation spectral coefficients (WETLabs ac-s);

Total backscattering coefficients (WETLabs BB9).

All instruments were commissioned on 28 October 2009.

The data acquisition is managed with a Data Acquisition and Power Conditioning System (DAPCS)

developed by WETLabs. The DAPCS is specifically designed to enable high speed/bandwidth

communications and power control to support data collection from a broad suite of environmental

sampling instruments.

Data is acquired in real time by the Linux Server and PC installed on site. The raw data-stream is

uploaded via broadband to CSIRO‟s data storage in Canberra where is pre-processed and QA/QC-ed.

Figure 3-7: LJCO site, Eastern Australia. Images from: http://imos.org.au/ljco.html

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 33


3.10 Martha’s Vineyard Coastal Observatory (MVCO). PI: H. Feng

Among other facilities, Martha‟s Vineyard Coastal Observatory (MVCO) includes a tower 3 km from

shore and in 15 m depth, with a mounted SeaPRISM contributing to the AERONET objectives for this

region. Janet Fredericks is the MVCO manager and Hui Feng the PI of the SeaPRISM operation.

Figure 3-8: The tower at MVCO

3.11 Pålgrunden Lighthouse, Lake Vänern. PI: Susanne Kratzer

The SeaPRISM instrument is placed on a lighthouse platform („Pålgrunden‟ Lighthouse) in Lake Vänern,

Sweden (Figure 3-9), 58.76 ºN, 13.15 ºE, about 2 miles north of the town of Granvik, Lake Vänern. The

PI of this site is Susanne Kratzer, from Stockholm University, Sweden, but the managing organisation is

the Swedish National Maritime Administration. Pålgrunden Lighthouse stands at 30 m in height. This

CIMEL was previously deployed at the Swedish Meteorological and Hydrological Institute (SMHI),

Norrköping (AERONET station number 194), and prior to deployment at Pålgrunden the filters were

changed to 412, 443, 488, 551, 667, 870, 1020 nm.

Figure 3-9: The Pålgrunden lighthouse SeaPRISM platform in Lake Vänern, Sweden.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 34


3.12 WAVE_CIS_Site_CSI_6. PI: B. Gibson, A. Weidermann

The WAVE_CIS site is located on the roof of ST52B Quarters Platform (Figure 3-10a). This is part of a

triple oil rig platform in which the other two structures make up the production platforms.

Figure 3-10: WAVE_CIS_Site_CSI_6. a) Platform; b) Instrumentation

3.13 AERONET-OC Key References

Bodhaine, B. A., Wood, N. B., Dutton, E. G. &Slusser, J. R. (1999). On Rayleigh Optical Depth

Calculations. Journal of Atmospheric and Oceanic Technology 16: 1854-1861.

D'Alimonte, D. & Zibordi, G. (2006). Statistical Assessment of Radiometric Measurements From

Autonomous Systems. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 44(3): 719-

728.

Hooker, S. B., Lazin, G., Zibordi, G. & McClean, S. (2002). An evaluation of above- and in-water

methods for determining water leaving radiances. Journal of Atmospheric and Oceanic Technology 19:

486-515.

Mobley, C. D. (1999). Estimation of the remote-sensing reflectance from above-surface measurements.

Applied Optics 38: 7442-7455.

Morel, A., Voss, K. J. & Gentilli, B. (1995). Bidirectional Reflectance of Oceanic Waters: A Comparison

of Modeled and Measured Upward Radiance Fields. Journal of Geophysical Research 100: 13143-13150.

Morel, A., Antoine, D. & Gentilli, B. (2002). Bidirectional reflectance of oceanic waters: accounting for

Raman emission and varying particle scattering phase function. Applied Optics 41(30): 6289-6306.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 35


Morel, A. & Gentilli, B. (1993). Diffuse reflectance of oceanic waters. 2. Bidirectional aspects. Applied

Optics 32: 6864-6872.

Morel, A. & Gentilli, B. (1996). Diffuse Reflectance of Oceanic Waters. 3. Implications of

Bidirectionality for the Remote-Sensing Problem. Applied Optics 35: 4850-4862.

Zibordi, G., Hooker, S. B., Berthon, J.-F. & D'Alimonte, D. (2002). Autonomous above water radiance

measurements from stable platforms. Journal of Atmospheric and Oceanic Technology 19: 808-819.

Zibordi, G., Mélin, F. & Berthon, J.-F. (2006). Comparison of SeaWiFS, MODIS and MERIS

Radiometric Products at a Coastal Site. . Geophysical Research Letters 33: L06617.

Zibordi, G., Holben, B., Slutsker, I., Giles, D., D'Alimonte, D., Mélin, F., Berthon, J.-F., Vandemark, D.,

Feng, H., Schuster, G., Fabbri, B. E., Kaitala, S. & Seppälä, J. (2009b). AERONET-OC: a network for

the validation of Ocean Color primary radiometric products. Journal of Oceanic and Atmospheric

Technology (Accepted): 57.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 36


4. MERMAID PROTOCOLS II: Portable Radiometers

4.1 SIMBADA. PI: Pierre-Yves Deschamps.

4.1.1 Introduction

This protocol comes from the User Manual on the SIMBADA website (http://www-loa.univ-

lille1.fr/recherche/ocean_color/src/); there is no publication available although those interested may refer

to Deschamps et al., (2004) for information about an earlier version of the dataset, SIMBAD, for which

the processing and instrument configuration is very similar. This protocol is concerned only with

SIMBADA. SIMBAD had only five spectral bands; this dataset has 11 spectral channels. There is a slight

overlap with NOMAD, due to an alternative processing scheme used by Robert Frouin for NASA of some

of the SIMBADA data for NOMAD. The quality of the processed data is presumed to be the same (Pers.

Comm: P-Y Deschamps, 13th November 2008). The duplicates discovered when formatting for

MERMAID were kept.


The SIMBADA instrument is an above-water radiometer designed and manufactured by the Laboratoire

d'Optique Atmosphérique (LOA) of the University of Lille, France. It measures both water-leaving

radiance and aerosol optical thickness in 11 spectral bands (each bandwidth of 10 nm), centred at 350,

380, 412, 443, 490, 510, 565, 620, 670, 750, and 870 nm, (see

Figure 4-1) by viewing the sun (sun-viewing mode) and the ocean surface (sea-viewing mode)

sequentially. The same optics, with a field-of-view of about 3°, the same interference filters, and the same

detectors are used in both ocean-viewing and sun-viewing mode. A different electronic gain, high and

low, is used for each mode, respectively. The optics are fitted with a vertical polariser, to reduce reflected

skylight when the instrument is operated in ocean-viewing mode. Pressure, temperature, and viewing

angles are also acquired automatically. Attached in the front of the instrument, a GPS antenna acquires

automatically the geographic location at the time of measurement and a display indicates various

information.

4.1.3 Measurement Protocol

The SIMBADA radiometer measures direct sunlight intensity by viewing the sun, and water-leaving

radiance by viewing the ocean surface at 45° from nadir and 135° from the sun's vertical plane. It is

powered by batteries, which allow about 8 hours of continuous use.

Measurements have to be made in clear sky conditions (<2/8 of clouds and not obscuring the sun disk),

outside the glitter region (relative angle between solar and viewing directions of 135°), and at a nadir

angle of about 45°. For those angles, reflected skylight is minimised as well as residual ocean polarisation

effects. The measurements can be made on a steaming ship; there is no need to stop the ship to make

measurements. To normalise water-leaving radiance, incident solar irradiance is not measured, but

computed using the aerosol optical thickness data. The operator can select, in addition to ocean-viewing

and sun-viewing modes, dark current and calibration modes. Each series of measurements lasts 10

seconds. Frequency of measurements is about 8 Hz.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 37


Figure 4-1: Spectral channels of the SIMBADA instrument

The experimental procedure is to make, consecutively, one dark measurement, three sun measurements,

three sea measurements, three sun measurements, and one dark measurement.

DARK mode: Cardboard is placed at the end of collimator and/or dark cloth, so that no light can

enter the instrument. The measurement lasts 10 seconds.

SUN mode: The instrument is aimed at the sun, and the measurement lasts about 10 seconds. A

beep indicates the end of the measurement. The sun's azimuthal angle is stored in memory.

SEA mode: Measurement is made at the side of the ship and aimed at the ocean, after having

positioned the instrument at 135° from the sun's vertical plane by using the relative azimuthal

angle from the sun's vertical plane displayed on the top middle and after having positioned the

instrument at nadir angle of 45°. The measurement lasts 10 seconds. To avoid viewing the ship

trail or foamy sea, it is better to scan continuously the sea between 30° and 60°. It is important the

tilt is not more than about 20°, so that the polariser remains in a suitable position.

The SIMBADA measurements should be made during daytime, when the sun disk is not obscured by

clouds, outside foam and whitecaps. Ideally, weather permitting, the measurements should be made 1) at

each station during daytime (if the ship stops offshore), and 2) while the ship is moving around local noon

(time of SeaWiFS overpass). The best ship location to make the measurements is the bow. Since

SIMBADA does not like sea water, "en route" measurements should be made only when there is no risk

of wetting the instrument.

The following meteorological data should be acquired concomitantly, whenever possible: date, time,

latitude, longitude., cloud cover and type, air temperature, dew point (or wet bulb) temperature, surface

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 38


pressure, visibility, wind speed, wind direction, whitecaps (none, low, moderate, or high), water

temperature, surface Chl, phaeophytin, etc. Some of these data may be available from the bridge log.

Radiometric Calibration

In the following equations the numerical count for Sun and sea measurements, CN, is considered to be

corrected for dark current. The dark-current count, therefore, is omitted for clarity. These two

measurements require different calibration methods, which are described below.

Sun-Viewing Mode

The instrument in sun-viewing mode is calibrated by use of the Bouguer–Langley method: measuring the

sun intensity through a stable atmosphere as a function of air mass and extrapolating the measurements to

zero air mass. After passing through the atmosphere, the Sun intensity (irradiance) in each spectral band

can be expressed as:

)]()(exp[)(),( 2

sos mdFI (16)

Where τ (λ) is the total optical thickness of the atmosphere (assumed to be constant during the

calibration), m is the air mass, and d is the corrective factor for the sun-Earth distance. The F0 (λ) values

are obtained from Thuillier et al. (2003), (formerly Neckel and Labs, 1984), m is computed as a function

of θs following Kasten and Young (1989), and d is computed according to Paltridge and Platt (1977).

Equation (16) is valid only in the absence of absorption by water vapour and minor gases.

Sea-Viewing Mode

The instrument in sea-viewing mode is calibrated by use of an integrating sphere, whose output spectral

radiance is calibrated with equipment and methods that are traceable to the National Institute of Standards

and Technology and that are further controlled in radiometric inter-comparison activities. The equivalent

radiance of the sphere in each band, L, is first computed as follows:

])(/[])()([)(

dRdRLL (17)

Where λ is wavelength, L is the radiance delivered by the sphere, R is the spectral response of the

SIMBAD instrument, and the integral is over the spectral range of each band. The calibration coefficients

(k) are computed from L and the digital counts (CN).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 39


Aerosol Optical Thickness and Angstrom Coefficient

Aerosol optical thickness from the measurements in sun-viewing mode is deduced from the total optical

thickness of the atmosphere τa (λ) (computed from the digital counts, m and θs). τa (λ), is then corrected

for molecular scattering and gaseous absorption, due mostly to ozone. This gives the aerosol optical

thickness in each band, τa (λ), as;

),(),()()( 330 uOOPraya (18)

Where τray (λ) and τO3 (λ) are the Rayleigh and ozone optical thickness in band λ, respectively. The τray

(λ) depends on surface air pressure, P0. It is computed by use of a depolarisation factor of 0.0279. The

ozone contribution is computed from the vertically integrated ozone amount, uO3, obtained from

climatology (Keating et al., 1989) or derived from Total Ozone Mapping Scanner (TOMS) observations.

The Angstrom coefficient, α, defined by the law τa (λ) ≈ λ-α

, is determined by regressing on a log–log

scale τa (λ) versus the equivalent wavelength of a band, λ, by using the instrument‟s five spectral bands.

The determination of α is more difficult at low-aerosol optical thickness simply because the uncertainty

on τa (λ) is rather constant in absolute value but becomes increasingly large in relative value as τa (λ)

decreases.

Total Atmospheric Transmittance (downwelling)

The total (i.e. direct plus diffuse) downwelling atmospheric transmittance, Td (λ), must be estimated to

normalise the water-leaving radiance measurements into reflectance, ρw (λ). It is expressed as the product

of the transmittance due to gaseous absorption, mostly of ozone, TO3 (λ), and the transmittance due to

molecular (Tray (λ)) and aerosol scattering (including aerosol absorption), Ta (λ).

Td (λ) is computed by use of a radiative transfer model based on the successive orders of scattering

method (Deuzé et al., 1989) with θs, τray, and τa as variable input (where „ray‟ denotes „Rayleigh‟). The

effect of aerosol type is small and can be neglected.

Marine Reflectance

Because the variable of interest is diffuse marine reflectance (Equation 20), the polarised ρw (λ) measured

by viewing the surface (sea-viewing mode), is obtained from the recorded numerical count (CN) by using

the following formula:

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 40


)cos()(

d)CN()k( =)(

0

2

w

sF

(19)

ρw should be measured in a specific viewing geometry, i.e., at a nadir angle of 45° and at an azimuth angle

of 135° with respect to the sun, in order to minimise skylight reflection effects. The inclinometer and

magnetometer angles are used to select the optimum viewing geometry. Measurements made at a nadir

angle outside the range 45° ± 5° and at a relative azimuth angle outside the range 135° ± 10° are not

processed. The reflectance ρw is then corrected for residual skylight and atmospheric transmittance to

yield the vertically polarised diffuse water reflectance, ρw‟:

)()/T( -)(=)'( doww (20)

Where ρo (λ) is the reflectance due to skylight reflection. This reflectance is computed accurately from the

τray (λ), τa (λ) and type (i.e. α), and the surface wind speed, according to Fougnie et al. (1999).

The radiometric measurements might be contaminated by whitecaps caused by wind action on the surface

or by foam and bubbles generated by the ship or by residual glitter. Sunlight scattered by clouds may also

be reflected by the surface in the instrument‟s field of view. Thresholds on ρw(865) are applied to

eliminate the most perturbed measurements, and then ρw at the other bands are iteratively corrected for,

making the assumption that extra reflectance due to whitecaps, clouds, etc., does not depend on

wavelength in the spectral range 443–870 nm. Thus only ρw‟ in spectral bands 1–4, i.e. the bands centered

at 443, 490, 560, and 670 nm, respectively, is obtained after correction. Equation (20) can be applied

effectively because the radiometric measurements are acquired simultaneously in the instrument‟s five

spectral bands (cloud effects strongly depend on surface-wave slope, and whitecaps may be changing

quickly with time). However, treating whitecaps as gray bodies even though they are not white spectrally

(Fougnie and Deschamps, 1997; Frouin et al., 1996; Nicolas et al., 2001) is sufficient because only the

less-perturbed measurements are selected. Over turbid coastal waters, the diffuse reflectance is not null at

870 nm; consequently, this correction is not valid for those waters.

Molecules and hydrosols polarise the light scattered by the water body. Because SIMBADA

measurements are made through a vertical polariser, polarisation effects must be corrected to yield the

total ρw.

Reprocessing to MERIS bands

The SIMBADA data were reprocessed in July 2009; they were spectrally convolved with the MERIS

band filters, and converted into reflectances using extraterrestrial solar irradiances from the Thuillier

(2003) database as recommended by the QWG (no longer the Neckel and Labs, 1984 values).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 41


4.1.4 Uncertainties

TBD

4.1.5 Key References

Deuzé, J.-L., Herman, M. & Santer, R. (1989). Fourier series expansion of the transfer equation in the

atmosphere–ocean system. Journal of Quantitative Spectroscopy and Radiative Transfer 41: 483-494.

Deschamps, P.-Y., Fougnie, B., Frouin, R., Lecomte, P. &Verwaerde (2004). SIMBAD: A Field

Radiometer for Satellite Ocean-Color Validation. Applied Optics 43(20): 4055-4069.

Fougnie, B. & Deschamps, P.-Y. (1997).Observation et mode´lisation de la signature spectrale de

l‟e´cume de mer. In Proceedings of the 7th International Colloquium on Physical Measurements and

Signatures in Remote Sensing, Vol. 1, 227-234 (Eds G. Guyot and T. Phulpin). Rotterdam.

Fougnie, B., Frouin, R., Lecomte, P. & Deschamps, P.-Y. (1999). Reduction of skylight reflection effects

in the above-water measurements of diffuse marine reflectance. Applied Optics 38: 3844-6856.

Frouin, R., Schwindling, M. & Deschamps, P.-Y. (1996). Spectral reflectance of sea foam in the visible

and near-infrared: insitu measurements and remote sensing implications. Journal of Geophysical

Research 101: 14361-14371.

Kasten, F. & Young, A. T. (1989). Revised optical air mass tables and approximation formula. Applied

Optics 28: 4735-4768.

Keating, G., Pitts, M. C. & Young, D.-F. (1989).Improved reference models for middle atmosphere ozone

_New CIRA_. In Middle Atmosphere Program Handbook for MAP, Vol. 31, 37-49 (Ed G. Keating).

Urbana, Illinois: Scientific Committee on Solar-Terrestrial Physics Secretariat, University. of Illinois.

Neckel, H. & Labs, D. (1984). The solar radiation between 3300 and 12500 Å. Solar Physics 90: 205–

258.

Nicolas, J.-M., Deschampes, P.-Y. & Frouin, R. (2001). Spectral reflectance of oceanic whitecaps in the

visible and near infrared: aircraft measurements over open ocean. Geophysical Research Letters 28:

4445–4448

Paltridge, G. W. & Platt, C. M. R. (1977).Radiative processes in meteorology and climatology. In

Development in Atmospheric Science New York: Eslevier.

Thuillier, G., Hersé, M., Labs, D., Foujols, T., Peetermans, W., D., G., Simon, P. C. & Mandel, H. (2003).

The solar spectral irradiance from 200 to 2400 nm as measured by the SOLSPEC spectrometre from the

ATLAS and EURECA missions. Solar Physics 214: 1-22.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 42


5. MERMAID PROTOCOLS III: TACCS

5.1 North West Baltic Sea. PI: Susanne Kratzer

5.1.1 Measuring systems

The Tethered Attenuation Coefficient Chain Sensor (TACCS 09, Satlantic Inc., Canada,

http://www.satlantic.com/) is a multi-channel radiometer including 7 channels (SeaWiFS + 620 nm) for

upwelling radiance (Lu: 412, 443, 490, 510, 555, 620 and 670 nm; seen in Figure 5-1a), 3 channels for

downwelling irradiance above the surface (443, 491, and 670 nm; seen in Figure 5-1b). The spectral

attenuation coefficient, Kd (490) can be estimated from a chain of four sensors for Ed (0+, λ) at 490 nm, Ed

(490). The Ed (490) sensors are fixed on a cable at 2, 4, 6 and 8 m depth (Kd (490) chain). The natural

logarithm of the measured Ed (0+, λ) is plotted against depth and the slope of the line taken as Kd (490).

All channels have a bandwidth of 10 nm. Note that when the TACCS was deployed in the water, as

delivered by the manufacture in 2000, the instrument was not sitting straight in the water. Extensive trials

had to be made with adding on additional diving weights close to the bottom of the instrument in order to

make it look straight up to the sky, and also in order to make it more stable and more resistant to wave

movements. 2008, the Lu channel at 555 nm was exchanged to 560 nm (the instrument was initially built

for sea-truthing of SeaWiFS). When the instrument was acquired in 2000 the 620 nm channel was chosen

to have one additional channel matching MERIS, and because this band may be influenced by

phycocyanin from cyanobacteria.

The TACCS was deployed in the Himmerfjärden, North West Baltic Sea during three campaigns during

August 2002, July 2008 and May 2010. The stations of deployment in 2002 are shown in

Figure 5-2, and the Associate Professor Susanne Kratzer from Stockholm University, Sweden, is the PI

(seen deploying the TACCS in Figure 5-1).

Figure 5-1: The TACCS instrumentation rig. a) the in-water

instrumentation shown being deployed by PI Susanne Kratzer,

and b) the above-water Ed sensor.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 43


Figure 5-2: Himmerfjärden area, NW Baltic Sea. Note that stations B1 and H2 do not differ optically from

the open sea station (Kratzer et al., 2008). STP: sewage treatment plant at the head of Himmerfjärden close to

station H5.

In order to derive reflectance from the TACCS, spectral Kd must be modelled. This we can do with AC9

data measured at the same time. The AC9 plus (WETLabs), is a state-of-the-art instrument to measure

Askö

STP

STP

H5

H4

H3

H2

B1

MERIS FR, 300 m

MERIS RR, 1.2 km

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 44


spectral attenuation and absorption in nine channels, from which spectral scattering can be derived. It is

also fitted with a CTD and an ECO VSF3, a volume scattering function metre, which is used to derive

backscatter. If measured AC9 data is not measured at the specific station the spectral slopes published in

Kratzer et al. (2008) can be used. Measurements were made with the AC9 in 2002, 2008 and 2010, but

for the present time, the AC9 data from 2008 and 2010 have not yet been processed. Table 5-2 shows the

spectral slopes for the north-western Baltic Sea (summer months 2001-2002), categorised into coastal and

open sea data, and used for the processing of the 2008 TACCS data included in MERMAID. Kratzer et

al. 2008 describe the division into coastal and open sea data. The data will be updated at a later time, as

the AC9 processed data becomes available.

Table 5-1: General sensor specifications of the TACCS 09

In-air In-water

Es

downwelling irradiance

sensor

Lu

upwelling radiance

sensor

Ed

downwelling irradiance

(K-chain) sensor

Sensor Model ED-50 OCR-100 ED-20

Spatial Characteristics

Field of view cosine response 10o (0.025 steradians) in

water

cosine response

Entrance aperture 4.78 mm diameter

Collector area 86.0 mm2

86 mm2

Detectors Custom 13 mm2 silicon photodiodes

Spectral Characteristics

Wavelength range 443-670 nm 400-700 nm

Number of channels 3 7 1

Spectral bandwidth 10 nm 10 nm 10 nm

Filter type Custom low fluorescence interference

Discrete

wavelengths(centers)

443, 490, 670 nm 412, 443, 490, 410, 560,

670, 620 nm

490 nm

Optical characteristics

Out of band rejection 10-4

10-4

10-4

Out of field rejection 5x10-4

Cosine response within 3% 0 -60o

within 10% 60-89o

within 3% 0 -60o

within 10% 60-89o

Temporal characteristics NA NA NA

System time constant 0.015 seconds 0.015 seconds 0.015 seconds

-3dB frequency 10 Hz 10 Hz 10 Hz

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 45


Table 5-2: Mean slope factors to derive spectral Kd for all TACCS channels from Kd490 in the north-western

Baltic Sea during summer (Kratzer et al., 2008) derived from AC9 data that was measured during field

campaigns in June 2001, August 2002 and July 2008 (Kratzer and Tett, 2009, Kratzer and Vinterhav 2010).

The data set was divided up into I) outer fjord & open sea stations (B1-BY31 & H2), and II) inner fjord

stations (H3-H4).

TACCS

Band

Slope factors,

outer stations,

B1-BY31 & H2

Slope factors,

inner fjord,

H3-H4

Wavelength (λ) Mean St. Dev. Mean St. Dev.

412 2.424 0.17 2.043 0.10

443 1.673 0.08 1.502 0.04

490 1.000 0.00 1.000 0.00

510 0.838 0.03 0.878 0.01

555 0.652 0.05 0.713 0.03

620 1.110 0.06 0.930 0.06

670 1.611 0.11 1.204 0.11

5.1.2 Measurement Protocols

System units and software

A power/telemetry cable runs from a deck unit, MDU-100, to the TACCS instrument. The deck unit is

connected to a 12V-battery and a computer; thus serves as a power source to the TACCS and as a RS-422

to RS-232 level converter for data transmission. The computer runs the acquisition software for raw data

logging and display, SatView version 2.9.2 from Satlantic, and complements the data acquisition with a

GPS with NMEA data stream (RS232) connected to the computer to fix the location and GPS-GMT time

of the cast. At each station, the instrument was set to sample for 2 minutes at a rate of 1 sample per

second, having first been allowed to float 10-20 m away from the boat in order to avoid shading, and

having allowed the instrument to adjust to the surrounding water temperature for at 10 min. The data was

converted from binary to calibrated engineering units using the Satlantic SatCon software. In order to

derive reflectance from the TACCS data, the spectral shape of the diffuse attenuation must be derived,

given that only a measure of Kd (490) derived from the Kd chain was available. This was done by using

AC9 data measured during 2001-2002 (Kratzer et al., 2008).

5.1.3 Data Processing

The AC9 data was corrected for salinity and temperature, and processed according to WETLabs method 2

(Wetlabs, 2009), which assumes that the scattering correction is a fixed proportion of the scattering

coefficient . Spectral scattering, b (λ), was derived as difference between spectral beam attenuation, c (λ)

and spectral absorption, a (λ), for all AC9 channels, i.e. 412 nm, 440 nm, 488 nm, 510 nm, 532 nm, 555

nm, 630 nm, 676 nm, and 715 nm. For deriving spectral Kd, a (λ) and b (λ) were first derived at TACCS

channels by linear interpolation between the AC9 channels (TACCS channel at 443 nm, 488 nm, 630 nm,

and 676 nm). The following algorithm from Kirk (1994) was then used to estimate spectral Kd:

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 46


5.0

201

21

0 ))()()()(()( baggaKd

(21)

Where the mean cosine of the refracted solar beam just below the surface µo= 0.86 is assumed; and the

constants g1= 0.425 and g2= 0.19.

The data set was divided into i) outer fjord & open sea stations (B1-BY31 & H2), and ii) inner fjord

stations (H3-H4). The reflectance, ρw, was estimated from the TACCS radiometer in the following way:

The upwelling radiance, Lu, just below the surfaces was estimated from Lu at 50 cm depth by propagating

each reading to the surface using the estimated Kd values for each radiance channel. The radiance above

the surface was derived from the Lu below the surface by multiplying the radiance values with the factors

provided in the NASA protocols (Mueller and Austin, 1995). Then, ρw was estimated according to

Equation (3).

5.1.4 Deployment from AAOT and processing

The following description of uncertainties is based on a field inter-comparison and validation of in-water

radiometer and sun photometers for MERIS validation in Portugal during February 2010 (Moore et al.,

2011) and on the MVT intercalibration during the Arc2010 in July 2010.

The TACCS was secured to the AAOT (see Section 3.2) via the power/telemetry cable, and to reduce the

strain on the electrical connection during sampling and recovering phase of the instrument (note there is a

deployment cable strengthening the power/telemetry cable). The instrument was taken 30 meters from the

AAOT location by a zodiac to avoid shading effects from the tower. The TACCS was deployed by first

lowering the k-chain carefully into the water, followed by the buoy. Careful measures were taken to avoid

splashing of water onto the Es sensor during instrument deployment. After the buoy is deployed, the

instrument is allowed to drift up to 40 m away from the AAOT and to acclimatize to the surrounding

water temperature for 10 minutes before the acquisition software is started.

At the beginning of a station, the station name and position were input into the station log file and the file

was named accordingly using the PC-time stamp as sequential file order. Handwritten field notes were

taken regarding the visible sea state and sky conditions and any visible behaviour of the instrument in the

water (buoyancy above/below normal, extreme tilt, etc). The SatView software was programmed to

continuously log all casts for each station at a rate of 1 sample per second during 3 minutes with one

minute pause between each log. Dark readings were taken on deck immediately after recovery, to

minimize temperature dependent effects onto the dark readings. The instrument was deployed and

operated by two people, and each deployment took approximately 15-20 minutes per station.

Data processing

The raw binary log files were converted into calibrated physical units, using the calibration file of the

ARC2010 inter-calibration experiment. The data conversion program used was SatCon v1.5 from

Satlantic. A fixed location of the AAOT (Latitude =45.3139o N, Longitude=12.5083

o E) was used as

location of the TACCS during all stations. It was assumed that within a radius of 40 meters the TACCS is

measuring the same water body as the WiSPER (Wire Stabilized Profiling Environmental Radiometer),

and that the water column around the AAOT is vertically homogeneous in order to derive the spectral

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 47


shape of the diffuse attenuation. A type B uncertainty of 1% is assumed due to the relative position of the

TACCS from the AAOT.

The mean time between the start and the end time of the cast were used as a reference time for inter-

comparison with the TACCS logs. The acquired radiometric quantities (in physical units) at time t, Lu (λ,

t), Ed (490, z, t) and Es (λ, t) were visually screened for stability over time series (spikes and records

indicating instrument error are removed). If extreme noise or a high count of errors were found the

TACCS log was discarded. This was in the case of WiSPER cast v950603, when the instrument had been

taken out of the water to compensate buoyancy/tilt effects by adding 1.5 kg of extra weight to the buoy.

The robust-mean estimate -H15 (AMC, 1989) was calculated both for light and dark measurements of the

acquired radiometric quantities and then dark correction was applied. The TACCS data processing has

been summarized by Moore et al. (2011). In order to obtain Kd (490) and Ed (490, 0-) a log linear

regression was applied to the four Ed (490,z). The Ed (490, 0+) is estimated by propagating Ed (490,0

-) up

to the surface according to:

eueddd factorEE _)0,490()0,490( (22)

Where: ), , state, sea( 490_ weued ffactor (23)

The predicted relationship between KLu (490,0-) and Kd (490,0

+) at the depth of the Lu sensor and the four

k-chain depths was modeled using Hydrolight, and the uncertainty was found to be in the order of 1%,

depending on water type. The Kd spectrum is calculated from the AC9 data, according to Kirk (1994) as

mentioned above (Equation 21). To extrapolate the subsurface Lu(λ) at 0.5m to Lu(λ,0-), first Kd (λ) is

normalized by Kd (490):

)(

)490(_

d

dnormd

K

KK (24)

According to Moore et al. (2011), it is assumed that to the first order the diffuse attenuation coefficient,

KLu (λ), is equal to Kd (λ), thus:

normdK

uu

e

mLL

_5.0

)5.0,()0,(

Lu (λ,0+) is estimated from a surface interface term without taking into account the wind speed that is

dependent on the refractive index of water (Moore et al., 2011). Self-shading corrections are applied to

the Lu (λ,0+) according to NASA protocols (Mueller, 2003a).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 48


The Es (λ) spectrum is calculated by deriving the spectral shape of an interpolated irradiance model, Emod

(λ), from a modified Gregg and Carder (1990) model, with an uncertainty of 1%. The Es (λ) is normalized

to the Es (490) to correct for tilt and roll:

)()490(

)0,490()( s

s

schains E

E

EE

Uncertainties assessment

The TACCS assessment of uncertainties is described in Moore et al. (2011). During the Arc2010 an

uncertainty of 10% is assumed from the fact that only one AC9 cast per station is used to derive the Kd

spectrum, thus it is assumed that the derived Kd spectrum is constant during all casts for each station.

Another source of uncertainty may be caused by using the AERONET Level 1.5. Real Time Cloud

Screened data, where the solar azimuth angle is used among other variables as reference for TACCS

calculations. It was found that the median time difference is about 03:39 minutes, the average time

difference is 05:40 minutes with standard deviation of 06:50 and a maximum time difference of 31:05

minutes. An additional uncertainty of 10% is assumed for using the AERONET data and a 2% uncertainty

due to the TACCS measurement off-set from the WiSPER time. Note that during standard TACCS

deployments a one to one relation with the AC9 cast is expected. When the AC9 data is not available,

published spectral slopes published can be used (Appendix 1). The Percentage of the difference between

the calculated Es (λ), Es (λ)chain and Es(mod) from Gregg and Carder (1990) are given in Table 5-3.

Table 5-3: Percent uncertainties for TACCS Lu (lamda,0+). 1Moore, et al. (2011),

2Type B uncertainty:

educated guess

Source Value measured (%)

K-chain_lamda

measured (%)

other_wavelengths

Source

K-chain, sensor position error (m) 0.01 0.004 0.005 1

K-chain, float offset (m) 0.01 0 0 1

K-chain, Ed Radiometric Error (%) 3 0.09 0.12 1

Lu absolute error (%) 3 1.5 1.5 1

Lu position error (m) 0.02 0.3 0.4 1

Es, relative spectral error (%) 2 0 0 1

Kd spectral estimate (%) from AC9 10 0 1.05 1

TACCS time offset from AERONET closest

time to WiSPER time (%)

10 1 1 2

TACCS time offset from WiSPER time (%) 2 0.5 0.5 2

TACCS GPS location (%) 1 0.5 0.5 2

Hydrolight bio-optical assumptions 2 2 1

OVERALL 5.89 7.08

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 49



AMC (1989). Robust Statistics - How Not to Reject Outliers, Part 1: Basic Concepts. Analyst 114: 1683-

1697.

Gregg, W. &Carder, K. L. (1990). A simple spectral solar irradiance model for cloudless maritime

atmospheres. Limnology and Oceanography 35: 1657-1675.

Kirk, J. T. O. (1994). Light and Photosynthesis in Aquatic Ecosystems (Second Edition). Cambridge

University Press.

Moore, G. F., Icely, J. I. &Kratzer, S. (2011).Field Intercomparison and Validation of In-water

Radiometer and Sun Photometers for MERIS Validation. In ESA Living Planet Symposium,

Special Publication SP-686. In press.

Mueller, J. L. &Austin, R. W. (1995).Ocean Optics Protocols for SeaWiFS Validation, Revision 1. In

NASA Tech. Memo., Vol. 25(Eds S. B. Hooker, E. R. Firestone and J. Acker). Greenbelt,

Maryland: NASA Goddard Space Flight Center.

Mueller, J. L. (2003a).Chapter 6: Shadow Corrections to In-Water Welled Radiance Measurements: A

Review. In Ocean Optics Protocols For Satellite Ocean Color Sensor Validation, Vol. Revision

5, 32 (Eds J. L. Mueller, G. Fargion and C. McClain). Greenbelt, Maryland: NASA GSFC.

Werdell, P. J., Bailey, S. W., Fargion, G., Pietras, C. M., Knobelspiesse, K., Feldman, G. C. &al., e.

(2003). Unique data repository facilitates ocean color satellite validation. EOS Transactions

84(3): 379.

Wetlabs (2009).ac Meter Protocol Document. http://www.wetlabs.com/products/pub/ac9/acproto.pdf.

5.2 Sagres, Algarve. PI: John Icely

5.2.1 Introduction

The Portuguese team on the ESA MVT have started validation measurements at a study site off Cape

Sagres on the south-west coast of Portugal (Figure 5-3). This region has distinct winter and summer

oceanic conditions that reflect the large scale wind patterns induced by Meridional displacements of the

Azores High (Relvas, 1999). Spectra were measured of water reflectance, w(), along a transect

perpendicular to the coast that includes both coastal and oceanic sites (A,B, C in Figure 5-3). These data

have been used by Cristina et al. (2009) to validate MERIS satellite products for both Case 1 and Case 2

water which can then be used to optimise current algorithms relating in-situ to remote sensing data. The

data have been provided to MERMAID for matchup to MERIS RR, according to MERMAID protocols.

The principle reference for these data is Cristina et al. (2009) from which this information is taken.

5.2.2 Measuring system and measurement Protocol

Sampling campaigns for the validation of MERIS data products occurred on 8 September; 4, 13 and 26

October and on the 8 and 17 November 2008 to coincide with the passage of the ENVISAT satellite every

three days over the Iberian Peninsula. The online EOLI-SA programme and the offline ESA catalogues

were used to predict the dates and passage times of ENVISAT, but field campaigns were generally

http://www.wetlabs.com/products/pub/ac9/acproto.pdf

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 50


restricted to days when meteorological sites predicted clear skies and relatively calm sea conditions.

Figure 5-3 shows the approximate locations off the coastline at Cape Sagres of Site A, B and C at 2, 10

and 18 km, respectively. A wide range of physical, biological and optical variables were sampled, but this

protocol is restricted to the reflectance measurements that were timed to coincide with the MERIS

overpass within +/-30 minutes at Station A and B and within +/- 1 hour at Station C.

The Satlantic hyperspectral radiometer was set to record for 2 minutes at a rate of approximately 1 sample

per second (depending on integration time), at a distance of 20 m away from the boat in order to avoid

irradiance interference from the shadow of the boat. The data collected from the radiometer were taken

from underwater measurements of spectral upwelling irradiance, Lu (), at 0.62 nm, and spectral

downwelling irradiance at 490 nm, Ed (490) with sensors that were fixed on a cable at 2, 4, 8 and 16 m

depth. Above the sea surface, downwelling incident irradiance, Es () was also measured. These three

parameters provided data from dark and light signals between wavelengths (λ) of 348.34 and 803.45 nm.

The radiometric measurements were consistent with the protocols for the validation of MERIS water

products (this document).

Figure 5-3: Map of the Portuguese coast with the area of study indicated as a black box. Satellite image of

southwest coast of Portugal with the location of sampling sites A, B, C.

There is concurrent logging of GPS data to provide an exact tile and estimate of the station position for

matchups. There is some variation in exact station position, since by the nature of the radiometer these are

Lagrangian observations.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 51



The data from each sensor was converted from binary to calibrated engineering units using Satlantic

SatCon software. The data from the sensors of Ed (490), Es () and Lu () were then corrected for the dark

signal; explicit dark readings were taken for the Ed (490) sensors and both Es (λ) and Lu (λ) sensors have

internal dark shutters. The two hyper-spectral sensors were co-aligned, by linear interpolation, to a 1nm

grid, since the spectrographs have slightly different spectral sampling points.

For the Ed (490) sensors the natural logarithm of the measured Ed (490) was regressed against depth and

the slope was used to determine attenuation coefficient, Kd at 490 nm, and Kd‟ according to (25) below;

the intercept is used to determine Es (490).

)490()490()490(' wdd KKK (25)

Where: Kd is the diffuse attenuation coefficient. Values of Kw are taken from Morel and Antoine (1994) in

the MERIS and in Table 4.5 of the RMD.

Kd’ at other wavelengths was calculated by converting the Kd (490) obtained from the Ed (490) depth

profile into the apparent chlorophyll. The chlorophyll value obtained was used to estimate Kd’ at other

wavelengths using the coefficients of Morel and Antoine (1994) and Equation (26):

e

d ChlChlK ][)(,' (26)

The calculation of Kd’ (λ) is simply an inversion of (25).

In order to determine the values of w (0+) there are a number of necessary corrections/transformations.

The Lu (), the upwelling irradiance, has first to be extrapolated up to the sea surface. The measurements

of Lu () are taken at an offset depth of 0.62 m to avoid interference from surface waves. These values are

extrapolated to just below the sea surface using Equation (27):

dK

uu

e

LL

62.0

,0 (27)

Where: Kd is the spectral diffuse attenuation coefficient obtained from (26).

A self-shading correction is applied to Lu (0-,λ), based on the model of Gordon and Ding (1992) and this

correction also follows the ocean optics for satellite protocols for satellite ocean colour sensor validation

(Mueller, 2003a). The correction requires the spectral absorption coefficient, a (λ) which was

approximated as Kd (λ), and the ratio of diffuse to direct irradiance. The direct and diffuse irradiance were

calculated following Gregg and Carder (1990), using the ozone concentration, water vapour concentration

and aerosol optical thickness from the MERIS matchup pixel. The total irradiance was used as a check on

the Es () derived from normalisation, whereby Es () are normalised to the Es (490) obtained by the from

the Kd analysis in order to correct for tilt and roll.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 52


Finally, self-shading corrected radiance, Lu„(0

-, λ), is interpolated up to the level of the surface sensor

using Equation (11). For Sagres processing, the term 2

)1(

n

fis approximately 0.54, but exact values for

a nadir viewing profiler in seawater are given in Table 10-1.

These measurements were used to estimate w, based on Equation (3). The radiometric data were

determined at the MERIS wavelengths using linear interpolation of the hyper-spectral data. This is

appropriate since the Satlantic bandwidth is approximately 10-12 nm. It should be noted that there is a

small error at the 681.25nm band, since the MERIS bandwidth is different here.

5.2.4 Uncertainties

TBD

5.2.5 Discussion and conclusion

The study by Cristina et al.(2009) shows that there some agreement between w the MERIS products and

the in-situ data sets which increases from the station A to the station C, probably due to the decrease of

the influence of the land adjacency effects on the satellite data. Scatter plots and their match-up statistics

show a large spread of data, especially in the blue band of the spectrum. Cristina et al. conclude that the

algorithms for relating these in-situ data with MERIS data could still be improved. However, however it

will be worth comparing the matchups derived from MEGS with the Cristina et al. matchups and

discussions had regarding differences and similarities and future validation campaigns.


Cristina, S., Goela, P., Icely, J. I., Newton, A. &Fragoso, B. (2009). Assessment of water-leaving

reflectance of the oceanic and coastal waters using MERIS satellite products off the southwest

coast of Portugal. Journal of Coastal Research Special Issue (56): 5.



Gordon, H. R. &Ding, K. (1992). Self-shading of in-water optical instruments. Limnology and

Oceanography 37(3): 491-500.

Relvas, P. (1999).The Physical Oceanography of the Cape São Vicente Upwelling Region Observed

From Sea, Land and Space. In School of Ocean Sciences. Menai Bridge: University of North

Wales, Bangor.

Morel, A. &Antoine, D. (1994). Heating Rate Within the Upper Ocean in Relation to its Bio-Optical

State. Journal of Physical Oceanography 24: 1652-1665.

Mueller, J. L. (2003a).Chapter 6: Shadow Corrections to In-Water Welled Radiance Measurements: A

Review. In Ocean Optics Protocols For Satellite Ocean Color Sensor Validation, Vol. Revision

5, 32 (Eds J. L. Mueller, G. Fargion and C. McClain). Greenbelt, Maryland: NASA GSFC.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 53


6. MERMAID PROTOCOLS IV: Fixed-depth Moorings

6.1 Buoy for the acquisition of long-term optical time-series (BOUSSOLE). PI: David

Antoine

6.1.1 Introduction

The BOUSSOLE Project is composed of three complementary elements: a) a monthly cruise program, b)

a permanent optics mooring, and c) a coastal Aerosol Robotic Network (AERONET-OC; Holben et al.,

1999) station , detailed in Antoine et al. (2006). The specific aim of the project is to provide a

comprehensive time series of near-surface (0–200 m) oceanic and atmospheric inherent optical properties

(IOPs) and apparent optical properties (AOPs).

The BOUSSOLE BUOY (Figure 6-1) is deployed in the Ligurian Sea, one of the sub-basins of the

Western Mediterranean Sea. Water depth varies between 2,350–2,500m in this area, and it is 2,440m at

the mooring point, which is located at 7o

54E, 43o

22N. The BOUSSOLE cruise and mooring programme

(Antoine et al., 2006; Antoine et al., 2007; Antoine et al., 2008) was specifically designed to provide a

time-series of optical properties in the Mediterranean Sea, in support of MERIS.

For MERMAID, D. Antoine provides Chl, ρw (λ) at the bands in Table 6-1 (therefore requiring only

normalisation), θs, and the relevant metadata.

Table 6-1: Nominal wavelengths at which BOUSSOLE provides in-water radiometric data to MERMAID

Centre-Bands

MERMAID 412 443 490 510 560 620 665 681 708

BOUSSOLE 412 442 490 510 560 -- 665 683 --

6.1.2 Measuring system and configuration, and measurement protocol

One-minute acquisition sequences are performed every 15 minutes, with all instruments working

simultaneously. The buoy radiometer suite is made of Satlantic 200-series radiometers measuring Ed (λ),

Eu (λ), and Lu (λ) (nadir) at two depths (4 and 9 m on horizontal arms) and at the following seven discrete

wavelengths: 412 (alternatively 555), 443, 490, 510, 560, 670 and 681 nm. A Satlantic Multichannel

Visible Detector System (MVDS) 200-series radiometer measures Es (λ) at 4.5m above the water surface

and at the same seven wavelengths. Other instrumentation provides parameters for the processing of

radiometry measurements; an Advanced Orientation Systems, Inc. (AOSI, New Jersey, USA) two-axis tilt

and compass sensor at 9m (EZ-Compass-dive), and a Sea-Bird Electronics (Bellevue, Washington) 37-SI

CTD measuring conductivity, temperature, and pressure at 9 m. Other instrumentation is described in

Antoine et al. (2006), and includes two WETLabs ECOFLNTU Chl fluorometers at 4 and 9 m.

Platforms developed for oceanographic purposes are susceptible to problems in deploying radiometers,

due to the perturbations induced by the instruments themselves and often more significantly by the

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 54


platform, onto which they are installed, i.e. shading. Stability is also a problem; necessary for

measurements of nadir radiance or planar irradiance. BOUSSOLE was designed specifically to overcome

such problems. The platform minimises shading effects by reducing the cross-sectional area of structural

components, and its wave-interaction characteristics ensure the stability of the instruments. The design

constraints for the new platform were:

1. Measure Eu, Ed, and Lu (nadir) at two depths, plus Es at the surface;

2. Minimise the shading of the instruments;

3. Maximise the stability of the instruments; and

4. Permit deployment at a site with a water depth of 2440 m, and swells up to 8m (small horizontal

currents).

Figure 6-1: Artist’s view of

BOUSSOLE (from Antoine et al.,

2008), showing the above- and in-

water radiometers, and buoy

structure.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 55


The basic design principle for the buoy is that of a reversed pendulum, with Archimedes thrust replacing

gravity. A large sphere (with a diameter of about 1.8 m) is stabilised at a depth out of the effect of most

surface waves, and connected at the end of a long cable anchored on the sea bottom. This sphere creates

the main buoyancy of the system. A rigid, tubular, structure is fixed above the sphere, which hosts the

instrumentation on horizontal arms (at 4 and 9m depths). With the BOUSSOLE design there is no large

body at the surface generating shade and the stability of the instruments is ensured even for quite large

swells.

6.1.3 Data Processing of Lwn (λ)

BOUSSOLE processing is detailed in Antoine et al. (2008) with the salient details provided here.

The initial step of the data processing is a data reduction that derives one representative value of Es (λ), Ed

(λ), Eu (λ) or Lu (λ) for each of the 1-minute acquisition periods; during which about 360 measurements

are taken (the acquisition frequency of the radiometers is 6 Hz). The procedure consists of taking the

median of the 360 measurements (details in Antoine et al., 2007), and allows getting rid of the

perturbations caused by the wind-roughened air-sea interface. Therefore, it provides a value that would

ostensibly be measured if the sea surface was flat. In addition, it is verified that the coefficient of variation

within the 360 Es (λ) measurements is below 5%, which ensures that the above-surface irradiance was

stable during the 1-min acquisition sequence.

From the two values of Lu (z, λ), the upwelling nadir radiance at null depth z = 0- (immediately below the

sea surface) is then obtained as (omitting the wavelength dependence for brevity):

),,()4()0( ChlzfnezLL s

K

uudLz

(28)

where z is the measurement depth (not exactly 4 m when the buoy is lowered or when swell goes through

the superstructure). This stage of processing is the largest source of uncertainty for the BOUSSOLE. KL is

the diffuse attenuation coefficient for the upwelling nadir radiance. The latter is computed from the

measurements of Lu collected at the two depths:

z

zLzLK uu

L

)))4((/))9((log(( (29)

Where Δz is exactly 5 m. The rationale for, and the implementation of, the function appearing in the right

hand side of Equation (29) are provided in Antoine et al. (2008).

The value of Lu (0-) is then corrected for instrument self-shading as per Gordon and Ding (1992). The

parameters entering into this correction are the instrument radius, which is 4.5 cm (common to all

Satlantic 200-series radiometers), the total absorption coefficient, which is computed following Morel and

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 56


Maritorena (2001) using the Chl concentration, and the ratio between the direct-sun and diffuse-sky

irradiances. This ratio is computed following Gregg and Carder (1990), using the atmospheric pressure

and relative humidity measured in the vicinity (2 nm) of BOUSSOLE by a meteorological buoy, the

ozone content provided by the US National Centre for Environmental Prediction (NCEP) SeaWiFS near

real-time ancillary data, and a horizontal visibility corresponding to a Shettle and Fenn (1979) maritime

aerosol with τa of 0.2 at 550 nm.

From the corrected value of Lu (0-), the water-leaving radiance at nadir, Lw, is obtained as in Equation

(11).

The remote sensing reflectance, RRS, is then obtained as in Equation (12), before which Es is corrected for

the buoy tilt. The correction is a function of the orientation of the two axes of the tilt measurement with

respect to the sun azimuth, and computes the ratio of the diffuse (unaffected by the tilt) to direct (affected

simply through the cosine of the sun zenith angle) light for clear-sky conditions (Gregg and Carder,

1990). RRS is further multiplied by π in order to get ρw, which is consistent with the definition of the

product delivered by the MERIS mission (AD [4]).

A diffuse attenuation coefficient for the downward irradiance in the upper layers is also computed as:

z

EzEK dd

d

))0((/))(log((

(30)

where: z is the deepest of the two depths (nominally 9 m), and Ed (0-) is simply Es reduced by transmission

across the air-water interface, i.e., Es * 0.97 (Austin, 1974).

The final processing step for the buoy data consists in either eliminating or correcting data corrupted by

bio-fouling. The growth of various types of marine organisms, such as algae and bacteria, is unavoidable

with moored instruments, although it is much less severe in the clear offshore waters at BOUSSOLE than

it can be, for instance, in turbid coastal environments. The cleaning of the instruments takes place every

two weeks.


BOUSSOLE data is processed especially for MERMAID.

6.1.5 Uncertainties

There are several sources of potential uncertainty, explained in detail in Antoine et al. (2006) and Antoine

et al. (2008). Uncertainties on all terms are considered as random, Gaussian distributed, independent one

of each other, so that the final error budget is computed as the square root of the sum of the squares of the

individual error terms. Antoine et al. (2006) reach an overall uncertainty of 5.41% for BOUSSOLE and

revise this to a quadratic error budget of 6% in Antoine et al. (2008), due to the various uncertainties due

to radiometric calibration of field radiometers (3%), calibration decay over time (2%), toward surface

extrapolation (3%), self-shading (3%), and bidirectional effects (2%).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 57



Antoine, D., Chami, M., Claustre, H., d'Ortenzio, F., Morel, A., Becu, G., Gentilli, B., Louis, F., Ras, J.,

Roussier, E., Scott, A., Tailliez, D., Hooker, S. B., Guevel, P., Deste, J.-F., Dempsey, C. & Adams, D.

(2006).BOUSSOLE: A Joint CNRS-INSU, ESA, CNES, and NASA Ocean Color Calibration and

Validation Activity. Greenbelt, MD.: NASA/GSFC.

Antoine, D., Guevel, P., Deste, J.-F., Becu, G., Louis, F., Scott, A. & Bardey, P. (2007). The

'BOUSSOLE' Buoy - A New Transparent-to-Swell Taut Mooring Dedicated to Marine Optics: Design,

Tests and Performance at Sea. Journal of Atmospheric and Oceanic Technology In Press.

Antoine, D., Ortenzio, F., Hooker, S. B., Bécu, G., Gentilli, B., Tailliez, D. & Scott, A. (2008).

Assessment of uncertainty in the ocean reflectance determined by three satellite ocean color sensors

(MERIS, SeaWiFS and MODIS-A) at an offshore site in the Mediterranean Sea (BOUSSOLE project).

Journal of Geophysical Research 113(C07013, doi:10.1029/2007JC004472): 22.

Austin, R. W. (1974).The remote sensing of spectral radiance from below the ocean surface. In Optical

Aspects of Oceanography, 317-344 (Eds N. G. Jerlov and E. Steemann-Nielsen). New York: Academic

Press, London.

Bailey, S. W. & Werdell, P. J. (2006). A multi-Sensor Approach for the On-Orbit Validation of Ocean

Color Satellite Data Products. Remote Sensing of the Environment 102: 12-23.

Gordon, H. R. & Ding, K. (1992). Self-shading of in-water optical instruments. Limnology and


Gregg, W. & Carder, K. L. (1990). A simple spectral solar irradiance model for cloudless maritime


Morel, A. & Gentilli, B. (1996). Diffuse Reflectance of Oceanic Waters. 3. Implications of

Bidirectionality for the Remote-Sensing Problem. Applied Optics 35: 4850-4862.

Morel, A. & Maritorena, S. (2001). Bio-optical properties of oceanic waters: A reappraisal. Journal of

Geophysical Research-Oceans 106(C4): 7163-7180.

Shettle, E. P. & Fenn, R. W. (1979).Models for the aerosols of the lower atmosphere and the effects of

humidity variations on their optical properties. In Environment Research Papers, Vol. 676, 31

Massachusetts: Air Force Geophysics Laboratory, Hanscom AFB.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 58


6.2 Marine Optical BuoY (MOBY). PI: Kenneth Voss

6.2.1 Introduction

The Marine Optical Buoy (MOBY, Figure 6-2) (Clark et al., 1997; Clark et al., 2003) is a fixed mooring

system providing continuous time series of normalised water-leaving radiances Lwn (λ) since 1996 for the

purpose of on-orbit calibrating ocean colour sensors. Although MOBY is primarily designed for the

NASA sensors SeaWiFS and MODIS, the hyperspectral data is also processed for MERIS bands.

MOBY is a moored tethered buoy, consisting of two complete systems, one of which is moored and

operational at any given time. The MOBY operations site is located in Honolulu, on the south shore of the

island of Oahu, at the University of Hawaii‟s Marine Centre. The MOBY site is located in 1200 m of

water approximately 18 km from the west coast of the Hawaiian Island of Lanai. The mountains on the

islands of Molokai, Lanai, and Maui provide a lee from the dominant trade winds, reducing the sea swell

and cloud cover at the site. Full details on the MOBY system can be found in Clark et al. (2003).

From the password-protected Gold directory providing the MOBY data, MERMAID receives: Lw (λ), and

Lwn (λ), and Es (λ). However, to ensure consistency with MERMAID, Lw (λ) and Es (λ) are used to

calculate ρw (λ) rather than directly using the Lwn (λ). Normalisation proceeds as described in section

1.4.5. The main reference for MOBY is the NASA Ocean Optics Protocols, Volume 3 and Volume 6

(Mueller et al., 2003); the key points pertinent to MERMAID are summarised here.


MOBY is a 12 m spar buoy (including the lower instrument bay) designed as an optical bench for

measurements of Ed (z, λ) and Lu (z, λ) at depths of 1 m, 5 m, 9 m, and 12 m (see Figure 6-2). The MOBY

spar is tethered to a second surface buoy, which is slack moored, i.e. isolated by subsurface floats, to an

anchor on the sea floor. Sensors for wind speed, wind direction, air temperature, relative humidity, and

barometric pressure are mounted on the main mooring buoy. The Marine Optical System (MOS), the

heart of MOBY, consists of two single-grating CCD spectrographs connected via an optical multiplexer

and fibre optic cables to the Ed (z, λ) and Lu (z, λ) optical heads mounted at the ends of the buoy‟s 3

standoff arms. To provide low-loss transmission at ultraviolet wavelengths, 1 mm diameter silica fibre-

optic cables are used to connect the optical heads to MOS. Lu (12, λ), at z = 12 m, is measured through a

window in the bottom of the MOS housing itself. A seventh fibre optic cable connects a surface irradiance

Es (λ) cosine collector, mounted 2.5 m above the surface float, to the spectrographs. Each pair of in-water

optical heads is mounted on a standoff arm, to minimise radiometric artefacts due to shadows and

reflections from the buoy, and to minimise self-shading the Lu radiometer housings are small in diameter

(7cm) (Clark et al., 2003; Gordon and Wang, 1994).

6.2.3 Measurement Protocol

Processing follows the description provided in Mueller et al. (2003b). A single MOBY observation

comprises a sequence of four to seven spectral radiance and irradiance measurement cycles for the optical

collectors located at the different depths on the spar. Datasets are acquired daily for the nominal satellite

equatorial crossing times for SeaWiFS and MODIS Aqua overpasses.

On MOBY, Lu (z, λ) is measured at 4 depths that are rigidly separated at fixed intervals on the buoy.

These depths are nominally z1 = 1 m, z2 = 5 m, z3 = 9 m, and z4= 2.5 m. The radiance measurement at 2.5

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 59


m is not currently used to determine water-leaving radiance, Lw (λ). Ed (z, λ) is measured only at nominal

depths z1, z2 and z3.The MOS measures radiation input from one Lu (z, λ), Ed (z, λ) or Es (λ) head at a

time. A typical sequence would be to measure Lu (λ, z) from a depth, preceded and followed by Es (λ)

surface reference spectra and associated dark spectra. Then this sequence is repeated at the 2nd and 3rd

depths to complete the profile for Lu (λ, z). Note that there are a total of 35 measurements for radiances at

the 3 depths, surface irradiance Es (λ) and sensor dark spectra. The 35 measurements are grouped into

overlapping subsets of 15 measurements, representing the cycle associated with upwelled radiance

measurements at each depth. This entire procedure requires between 30 min and 1 hr to complete.

Figure 6-2: The NASA MOBY instrument set-up at Lanai, Hawaii

6.2.4 Data processing of Lwn (λ) and quality assurance

MOBY is designed to take measurements at 0.6 nm intervals, to suit the NASA ocean colour sensors, and

weighted band-averaged values, of Lwn are computed. Details of temporal averaging (unit integration

time) and the spectral response functions are detailed in Clark et al. (2003), however the parameters

pertinent to MERMAID are described here.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 60


Determining Lw (λ)

MOBY makes measurements at three depths near the ocean surface. The shallowest measurements (at 1

m) are propagated upward to just below the ocean surface by calculating the upwelling spectral radiance

attenuation coefficient, KL (λ), using:

),()(

),(ln

1

2

1

12

2

1

zL

E

E

zL

zzK

u

t

s

t

s

uL (31)

where z1 and z2 are the two depths at which measurements are made (z1< z2). Es (t1) and Es (t2) are the

averages of the Es measurements before and after the in-water Lu measurements to remove the effects of

solar irradiance changes from those measurements. The convention for this calculation is that the depth at

the surface is zero and that the values of z increase with depth.

Then the radiance change between depth z1 and the surface is calculated using:

1)(

1 ),(),0(zK

uuLezLL

(32)

The depth zi is selected according to the following hierarchical rules:

1. If the data from the top arm are valid, then that depth is selected.

2. Else, the data from the middle arm, if valid, are selected.

3. Else, the data sequence is rejected entirely.

To determine Lw (λ) from Lu (λ), the measurement of upwelling radiance from a selected depth zi is

propagated to the surface using Equation (11). The upward transmittance through the interface, the

Fresnel transmittance, for nadir viewing radiance, is approximately constant with the value 0.543.

For MOBY measurements, the propagation of light to the surface is made from the topmost arm of the

buoy (z1=1 m). However, there are periods when the topmost arm is either broken or missing. For these

periods, the centre arm (z1=5 m) is used in its place.

Tilt effect considerations and Es (λ)

There are no pressure transducers in each sensor head; there is only one, recording with each scan in the

MOS system. Additionally, the fixed arm depths are used to extrapolate Lu (λ) to the surface, and

therefore tilt is not accounted for (Pers. Comm. Flora, 28th Jan 2010). Es (λ) from MOBY is from

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 61


measurements; Es (λ) is not computed via Gordon and Wang (1994) approximations (Pers. Comm. Flora,

28th Jan 2010).

6.2.5 Quality Control

MOBY has sensors for dark readings and depth variation, as well as those for wind velocity, surface

barometric pressure, air temperature, relative humidity, water temperature and conductivity and

chlorophyll-a fluorescence. A weakness of MOBY is the amount of time it takes to go through a complete

measurement cycle. Pressure is measured just prior to or during the Lu(λ) measurement; however, the

integration time for an Lu(λ) measurement is in the order of 30 seconds for the full spectral range, so the

depth at time of measurement is only an approximation (Pers. Comm. Franz, 13th Oct. 2009).

MOBY data are corrected for dark readings, as specified in the NASA protocols, and also for straylight.

6.2.6 Uncertainties

The current claim of uncertainty for MOBY is, for so-called 'good scans' and not including a shadowing

correction, approximately 5 % for MODIS channels 8 through 12, increasing to 12.5 % for channel 13

due to a large shadowing correction (Brown et al., 2007). If only data from good scans labelled good days

after quality control checks are applied and a shadowing correction is applied to the data set, the

uncertainty is expected to reduce to less than 3 % for MODIS channels 8 through 12, increasing slightly,

to 3.3 %, for channel 13. The shadow-uncorrected uncertainties should be used for the present.




Press, London.

Clark, D. K., Gordon, H. R., Voss, K. J., Ge, Y., Broenkow, W. & Trees, C. (1997). Validation of

Atmospheric Correction Over the Oceans. Journal of Geophysical Research 102D: 17209-17217.

Clark, D. K., Yarborough, M. A., Feinholz, M. E., Flora, S., Broenkow, W., Kim, Y. S., Johnson, B. C.,

Brown, S. W., Yuen, M. & Mueller, J. L. (2003).MOBY, A Radiometric Buoy for Performance

Monitoring and Vicarious Calibration of Satellite Ocean Colour Sensors: Measurements and Data

Analysis Protocols. In Ocean Optics Protocols for Satellite Ocean Colour Sensor Validation, NASA

Technical Memo. 2003-211621/Rev4, VolVI, 3-34 (Eds J. L. Muller, G. Fargion and C. McClain).

Greenbelt, MD.: NASA/GSFC.

Gordon, H. R. & Wang, M. A. (1994). Retrieval of water-leaving radiances and aerosol optical thickness

over the oceans with SeaWiFS: A preliminary algorithm. Applied Optics 33(3): 443-452.

Mueller, J. L., Fargion, G. & McClain, C. (2003b).Ocean optics protocols for satellite ocean color sensor

validation, Revision 4. Vol. I - VII, 141 Greenbelt, Maryland: NASA.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 62


7. MERMAID PROTOCOLS V: Profiling Instruments

7.1 Bristol Channel and the Irish Sea. PI: David McKee

7.1.1 Introduction

D. McKee has provided a dataset of reflectances from a study of IOPs and AOPs in the Bristol Channel

and the Irish Sea in UK and Irish waters. A Satlantic freefall equipped with seven wavelengths (412, 443,

491, 510, 554, 665, 700 nm) is used. The main references for these data are the NASA Ocean Color

Protocols (Mueller et al., 2003a) the ProSoft 7.7 User Manual (Satlantic, 2007), and McKee (Pers.

Comm., 2008).

Note: the dataset on the MERMAID website is affected by the ProSoft processing bug which has now

been fixed. Explanation of the bug follows in this protocol and the updated dataset will be available

online soon. An advisory note is displayed prior to extraction of this dataset.

7.1.2 Measuring system and measurement protocol

The Satlantic (Satlantic Inc., Nova Scotia, Canada) SeaWiFS Profiling Multichannel Radiometer (SPMR)

is a multispectral instrument designed specifically for the purposes of SeaWiFS validation. It generates

measurements of upwelling radiance, Lu, and downwelling irradiance, Ed, across the visible spectrum

(412, 443, 489, 510, 554, 665 and 700 nm), with associated bandwidths of 10 nm. Designed as a

freefalling profiler, the SPMR instrument decouples measurements from ship motion and minimises ship

shadowing; the instrument orientate in the vertical position with the instrument deployed far enough from

the research vessel to avoid ship shadow effects. On board inclinometers provide orientation information

allowing data to be removed that does not satisfy predetermined criteria. The original system was

supplied with a surface irradiance (Es) sensor which was mounted high on the superstructure to give high

frequency, concurrent surface irradiance data during profiles. Unfortunately this unit failed after the first

couple of cruises and could not be replaced. The SPMR was calibrated by Satlantic on two occasions; the

original calibration for Irish Sea cruises and a second calibration for later cruises in the Bristol Channel.

The SPMR was deployed well away from the side of the vessel to minimise ship-shadow effects,

generally 20m or more. The free-fall nature of the instrument means it can be used like an underwater kite

- paying out enough cable and letting it move up and down the water column a few times usually

establishes a reasonable distance between the ship and the sensor.

The θs was not provided but computed from the provided dates, times, latitudes and longitudes.

Radiometric measurements were made during normal working hours (09:00 to 18:00 local) and under a

variety of atmospheric conditions, often cloudy; clear skies are relatively rare in this region.

Also provided to MERMAID are measurements of Chl, measured using 90% acetone extraction and

spectrophotometric analysis. There are two sets of equations used: (a) trichromatic equations of Jeffrey

and Humphrey (1975) to give Chla (and b, c, carotenoids) - this is „A1‟, and a second method gives Chla

and phaeopigment (after acidification) - this is „A2‟. A1 is used for MERMAID.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 63


7.1.3 Satlantic SeaWiFS Profiling Multichannel Radiometer Reprocessing

Data provided to MERMAID consist of single casts for each station. Each cast was selected as the

optimum available for that station taking into account: stable surface irradiance (where this information

was available), minimal tilt angle, and quality of derived diffuse attenuation products (for extrapolation to

/ through sea surface).

SPMR data were processed using Prosoft v7.7.11.

The program converts raw signals detected by the instrument into higher level products such as water

leaving radiance and reflectance profiles. At the first levels (1a/1b/2), calibration and reference data are

applied to the collected signals. Any measured points greater than 10 times the cast standard deviation

were removed, as well as points where the tilt of the instrument was greater than a defined value (10

degrees). Processing to level 2s involves defining a pressure coordinate system. The Ed and Lu sensors,

located at the top and bottom of the instrument respectively, are separated on the SPMR by a distance of

1.25 m. The measurements per depth are corrected to a common depth defined by the position of the Ed

sensor. To progress to level 3, the data was sorted into depth-averaged bins of 0.5 m. Optical sensor data

(Lu etc.) were natural log transformed to “straighten” the data prior to averaging.

The final stage incorporates radiometric equations to calculate level 4 products. Linear polynomial fits

were applied to data using a 5 point window for depth profiles of lnLu and lnEd. The slope of best fit lines

through log-transformed Ed and Lu versus depth gives Kd and KLu as a function of depth. Values of Ed and

Lu immediately beneath the sea surface (Ed (0-,λ) and Lu (0

-,λ)) are obtained as the intercepts of the best fit

line passed through the 5 depth profile values nearest the surface. Accounting for air-sea interface effects,

Lu and Ed are extrapolated through the surface to give Lw(0+,λ) as in Equation (11), and Ed(0

+,λ) as in:

1

),0(,0 d

d

EE (33)

where α is the Fresnel reflection albedo for irradiance from sun and sky (0.043). The above-surface values

of radiance and irradiance then provide inputs for calculating apparent optical properties, for example

remote sensing reflectance, RRS, as in Equation (12).

During the course of data processing (ProSoft version 7.7.11), an error was uncovered with the procedure

responsible for extrapolating optical values to the surface. The first two entries of Lu profiles derived at

level 3 were marginally lower than those presented in the level 2s data. An example of this is shown in

Figure 7-1. The error was found to be consistent across all wavelengths and was only present for the Lu

sensor. The binning algorithm used by ProSoft for L3 Lu data was determined as the cause of the error,

whereby the program was unable to effectively deal with NaN values. The SPMR configuration places Lu

sensors at the bottom of the profiler, about 1.25m deeper than Ed sensors mounted at the top of the unit. Ed

data profiles therefore routinely start ~1.25m shallower than Lu profiles. These “blank” Lu data points

were subsequently being replaced by NaNs (not a number – used by MATLAB) which were inadequately

handled by the depth-average binning. A correction routine was developed to address this issue, which

ignored the first two data points when performing extrapolations on Lu. Removal of the two Lu bins

closest to the surface unfortunately extends the range for extrapolation, but we consider the assumption of

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 64


a uniform surface layer over these depths to be a reasonable approximation. This approach has the merit

of at least removing known problematic points.

Figure 7-1: Errors in first two points of level 3 depth-averaged Lu values.

Water leaving radiance, Lw (λ) was calculated using Equation (11), where a default value of 0.021 was

used for f and default value of 1.345 used for n. Ed (0+, λ) was used as a substitute for Es (λ) in

MERMAID to convert the Lw (λ) to ρw (λ).

7.1.4 Uncertainties

No uncertainties have been provided to date.


Jeffrey, S. W. &Humphrey, G. F. (1975). New spectrophotometric equations for determining chlorophylls

a, b, c1 and c2 in higher plants, algae and natural phytoplankton. Biochemie und Physiologie der Pflanzen

167: 191-194.

Mueller, J. L., Bidigare, R. R., Trees, C., Balch, W. M., Dore, J. & Drapeau, D. T. (2003a).Ocean Optics

Protocols for Satellite Ocean Colour Sensor Validation, Revision 5, Volume V: Biogeochemical and Bio-

Opitcal Measurements and Data Analysis Protocols., 36 Greenbelt, MD: NASA/GSFC.

Satlantic (2007).Prosoft 7.7 User Manual. Vol. Revision E.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 65


7.2 California Current. PI: M. Kahru

Vertical profiles of downwelling spectral irradiance and

upwelling radiance were measured with underwater

radiometers (Biospherical Instruments MER-2040 and MER-

2048) as part of the California Cooperative Oceanic Fisheries

Investigations (CalCOFI; Figure 7-3) bio-optical program

(Kahru and Mitchell, 1999; Mitchell and Kahru, 1998), and

following SeaWiFS bio-optical protocols (Mueller and Austin,

1995). Mitchell and Kahru (1998) provide detailed explanations

of the measurement procedures followed. Downwelling spectral

irradiance (Ed) and upwelling radiance (Lu) at the following

nominal wavelengths were measured by the MER-2040: 340,

380, 395, 412, 443, 455, 490, 510, 532, 555, 570, and 665 nm. A MER-2041 deck-mounted reference radiometer also measured

downwelling irradiance at the following nominal wavelengths:

340, 380, 395, 412, 443, 490, 510, 555, 570, 665, 780, and 875

nm, PAR. Instrument self-shading correction (Kahru and

Mitchell, 1998; Gordon and Ding, 1992) was routinely applied

and profile with the ship shadow and/or variable illumination

were eliminated. Nearshore stations with increased reflectance

due to suspended sediments were also excluded. The

advantage of this dataset compared to the heterogeneous

SeaBAM dataset is that it has been collected with the same

well calibrated instruments and processed using similar

procedures. Measurements of Chl-a were taken in the CalCOFI program using the fluorometric method

(Holm-Hansen et al., 1965; Venrick and Hayward, 1984) consistent calibration protocols.

The MER-2040/2041 system described in had detailed system characterization and radiometric

calibration performed by the manufacturer, Biospherical Instruments, Inc. (BSI), and the Center for

Hydro-Optics and Remote Sensing (CHORS) of San Diego State University according to procedures

specified by the SeaWiFS Protocols (Mueller and Austin, 1995). Further details are provided in Mitchell

and Kahru (1998).

The MER unit was generally deployed immediately before or immediately after the CalCOFI water bottle

cast to ensure minimal offset in time and space for the optics and the pigment data set. Immediately

following each cast, a dark scan of the MER radiometer was run by attaching opaque PVC caps on the

radiometer heads and recording the data for several minutes.

The CalCOFI bio-optical profiles were processed with a modified version of the Bermuda Bio-Optics

Project (BBOP) data-processing system (Siegel et al., 1995). Mitchell and Kahru (1998) also explain in

detail the BBOP processing and pre-processing employed on the CalCOFI dataset.

Remote sensing reflectance, RRS, was computed as in Equation (12), using air-sea interface transfer

coefficients 0.54 and 1.04 for, respectively, Lu and Ed (Austin, 1974).

Figure 7-2: CalCOFI transect

locations, California Coast. (from:

http://www.calcofi.org)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 66




Aspects of Oceanography, 317-344 (Eds N. G. Jerlov and E. Steemann-Nielsen). New York:

Academic Press, London.



Holm-Hansen, O., Lorenzen, C. J., Holmes, R. W. &Strickland, J. d. H. (1965). Fluorometric

Determination of Chlorophyll. J. Cons Perm Int Expl Mer 39: 3-15.

Kahru, M. &Mitchell, B. G. (1998). Spectral Reflectance and Absorption of a massive Red Tide off

Southern California. Journal of Geophysical Research 103(21): 21601-21610.

Kahru, M. &Mitchell, B. G. (1999). Empirical Chlorophyll Algorithm and Preliminary SeaWiFS

Validation for the California Current. International Journal of Remote Sensing 20: 3423-3429.

Mitchell, B. G. &Kahru, M. (1998).Algorithms for SeaWIFS Standard Products Developed with the

CALCOFI Bio-Optical Data Set. In CALCOFI Report, Vol. 39, 15pp.

Mueller, J. L. &Austin, R. W. (1995).Ocean Optics Protocols for SeaWiFS Validation, Revision 1. In



Siegel, D. A., O'Brian, M. C., Sorensen, J. C., Konnoff, D. A. &Fields, E. (1995).BBOP Data Processing

and Sampling Procedures., Vol. 19, 77 pp.

Venrick, E. L. &Hayward, T. L. (1984).Determining Chlorophyll on the 1984 CalCOFI surveys. . In

California Coorperative Oceanic Fisheries Investigations Reports., Vol. 25, 74-79.

7.3 Plumes and Blooms. PI: D. Siegel

The following protocols are taken from and expanded upon in more detail in Kostadinov et al. (2007),

Toole and Siegel (2001) and Toole et al. (2000). The Plumes and Blooms program follows the NASA

Ocean Optics Protocols (e.g. Mueller, 2003b).

The primary goal of the Plumes and Blooms ocean color observational program is to assess the spatial

and temporal structure of sediment plumes and phytoplankton blooms in the Santa Barbara Channel.

Twice monthly, seven station transect cruises across the Santa Barbara Channel from Goleta Point to

Carrington Point off Santa Rosa Island are conducted (Figure 7-3). This cross-channel transect permits

the sampling of a wide range of oceanic conditions with chlorophyll concentrations ranging from 0.05 to

7.0 mg m-3

and total suspended sediments ranging from 0.0 to 3.4 mg l-1

. At each station, the Plumes and

Blooms program routinely makes three independent estimates of remote sensing reflectance, RRS.

7.3.1 Reflectances

A Biospherical Instruments Profiling Reflectance Radiometer, PRR-600 (Toole et al., 2000) was used to

obtain profiles of upwelling radiance, Lu (λ), and downwelling irradiance, Ed (λ), at 412, 443, 490, 510,

555, and 656 nm. The free-falling PRR-600 was cast separately on a loose tether to minimize ship

shadowing. Data in the upper 12 m were used to extrapolate the Lu (λ), and Ed (λ) data to just below the

surface. RRS just below the surface was then computed as in Equation (12). Conversion to above water

was done with the air-sea interface terms in Table 1-3. Further details on data processing are given in

Toole et al. (2000).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 67


7.3.2 Additional parameters

In addition to the optical sampling, a complementary data set is collected at each station. This data set

includes conductivity-temperature and depth profiles; particulate absorption by the filter pad method;

coloured dissolved organic absorption spectra; and concentrations of inorganic nutrients, phytoplankton

pigments, total suspended materials. All samples are collected, stored, prepared and analysed using

existing techniques recommended by the U.S. JGOFS and SeaWiFS programs (see

www.icess.ucsb.edu/PnB/MethodsManual.html).

Figure 7-3: Santa Barbara Channel, California USA. Plumes and Blooms stations are marked with an ‘x’.

From Kostadinov et al. (2006).

Surface chlorophyll a concentrations were obtained by fluorometry from Niskin bottle samples following

the study by Strickland and Parsons (1972) and using a Turner Designs 10AU fluorometer. A Shimadzu

UV2401-PC (a Perkin-Elmer Lambda 2 before mid-2003) spectrophotometer was used to obtain the

spectra of the phytoplankton absorption coefficient aph (λ), the CDOM absorption coefficient ag (λ), and

the detrital absorption coefficient ad (λ) at each station from the surface bottle samples.

A HobiLabs Hydroscat-6 was used to obtain profiles of the backscattering coefficient bb (λ) at each

station for λ = 442, 470, 510, 589, 671, and 870 nm. Pure water calibrations (done at the factory and

UCSB semiannually) were applied. The HS-6 measures the total volume scattering function β at 140o. β is

then converted to the total backscattering coefficient using bb (λ) = 2πχp(β - βw) + bb w (λ) where bw and bb

w (λ) come from the study by Morel (1974) and χp = 1.18 (Boss and Pegau, 2001). The upper 15 m of data

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 68


from the downcasts were filtered and averaged to obtain a surface backscattering value. The σ (λ)

correction was then applied to correct for light attenuated in the measurement path of the instrument

(Maffione and Dana, 1997) using concurrent AC-9 surface data. Spectra which were not monotonically

decreasing were rejected as unreliable.

A WetLabs AC-9 absorption and attenuation meter (Moore et al., 1992) was used to obtain profiles of in-

situ absorption and beam attenuation coefficients at each station [a (λ) and c(λ) at 412, 440, 488, 510,

555, 630, 650, 676, and 715 nm].

7.3.3 Key references

Boss, E. & Pegau, W. S. (2001). The Relationship of Light Scattering at an Angle in the Backward

Direction to the Backscattering Coefficient. Applied Optics 40: 5503-5507.

Maffione, R. A. &Dana, D. R. (1997). Instruments and Methods for Measuring the Backward-scattering

Coefficient of Ocean Waters. Applied Optics 36(24): 6057-6067.

Moore, C., Zaneveld, J. R. V. &Kitchen, J. C. (1992).Preliminary Results from an In-situ Spectral

Absorption Meter. In SPIE Society of Optical Engineering, Vol. 1750, 330-337.

Morel, A. (1974).Optical Properties of Pure Seawater. In Optical Aspects of Oceanography., 1-24 (Eds N.

G. Jerlov and E. Steeman-Nielsen).

Mueller, J. L. (2003b).In-water Radiometric Profile Measurements and Data Analysis Protocols. Vol.

211621 Revision 4, Volume 2, 7-20: NASA.

Kostadinov, T. S., Siegel, D. A., Maritorena, S. A. &Guillocheau, N. (2007). Ocean Color Observations

and Modeling for an Optically Complex Site: Santa Barbara Channel, California, USA. Journal

of Geophysical Research 112(C07011): doi: 10.1029/2006JC003526.

Strickland, J. D. H. &Parsons, T. R. (1972). A Practical Handbook of the Sea Water Analysis. Fisheries

Research Board Canada Bulletin: 167-311.

Toole, D. &Siegel, D. A. (2001). Modes and Mechanisms of Ocean Color Variability in the Santa Barbara

Channel. Journal of Geophysical Research 106(C11): 26985-27000.

Toole, D., Siegel, D. A., Menzies, D., Neumann, M. J. &Smith, R. C. (2000). Remote-Sensing

Reflectance Determinations in the Coastal Ocean Environment: Impact of Instrumental

Characteristics and Environmental Variability. Applied Optics 39(3): 456-469.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 69


8. MERMAID PROTOCOLS VI: TriOS Ramses

8.1 English Channel. PI: H. Loisel, C. Jamet

Intensive field cruises were conducted from March to June 2004 to characterize the bio-optical and

radiometric variability of the eastern English Channel and southern North Sea. Lubac and Loisel (2007)

describe fully the campaign and results. The stations sampled during the cruise periods (15–18 March,

11–15 April, 10–15 May, 25–30 June) are displayed in Figure 8-1.

The investigated area is bordered by the mouth of the Seine River in the south, and by the mouth of the

Escault River in the north. This region is affected by typical coastal process such as strong tide ranges,

river inputs, resuspension due to the low bathymetry, and mixing of various water masses. The studied

area is also characterized by relatively intense spring blooms of phytoplankton species, such as the

prymnesiophyceae Phaeocystis globosa (Breton et al., 2006; Rousseau, 2000).

Figure 8-1: Location of the different stations visited in the eastern English Channel and southern North Sea

in 2004. The investigated area is bordered by (I) the mouth of the Seine River in the south and (II) the mouth

of the Escault River in the North.

Hyperspectral (every 3 nm) radiometric measurements were performed in the 350–750 nm spectral range

from three TriOS radiometers. The first (on the deck) and second (in water) radiometers were equipped

with an optical fibre and a cosine collector that were pointed upward to measure the above surface

downward irradiance, Es (λ), and the in-water irradiance profile, Ed (z,λ), respectively. The third

radiometer is equipped with a field of view of 7° in air, and is pointed downward to measure the nadir

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 70


upward radiance profile, Lu (z,λ). Wavelength dependent correction factors, also called immersion

factors, are applied to correct the reduction of the solid angle and the decrease of the light transmissivity

due to the water immersion of the sensor (Mueller and Austin, 2003). To limit impact of external factors,

93 stations were selected for their favourable environmental conditions (clear sky, low winds, and smooth

water surfaces). The sun zenith angle values range between 30° and 80° for these selected stations (with a

mean value of 51°, and a standard deviation of ±16°).

The remote sensing reflectance, RRS(λ), is calculated as in Equation (12). Lw(λ) was determined as in

Equation (11) with the air-sea interface terms given in Table 1-3.

To determine Lu(0−,λ), Lubac and Loisel (2007) used a near surface vertical profile of Lu (z,λ), which

had previously been corrected from the instrument self-shading effect according to the procedure

described in (Leathers and Downes, 2004). The attenuation coefficient for upward radiance, KLu (z0,λ),

is then computed as the local (around the depth z0=5 m) slope of the ln(Lu (z,λ)) self-shading corrected

profile, and is used to determine Lu (0−,λ) from the upward radiance measured in the upper layer (Hooker

and Morel, 2003; Mueller, 2003b; Smith and Baker, 1984):

)),(( 0),(),0(zzK

uuuLezLL

(34)

8.1.1 Error estimations

The error between self-shading corrected Lu (0-,λ) and self-shading uncorrected Lu (0−,λ) is less than 4%

in the visible. This weak impact is due to the relatively high sun zenith angles (50°) and to the small

instrument radius (3 cm). Relative error due to instrument self-shading is calculated as: (1−e) / e, where e

=exp(−κ·a(λ) · r) (Leathers and Downes, 2004), with r the instrument radius (r =0.03 m), κ function of

the in-water solar zenith angle, and a(λ) the sum of aw(λ), aCDOM (λ), and ap (λ), not measured but

estimated from the sum of ap (λ) and aNAP (λ) calculated from the models of Bricaud et al. (1995) and

Babin et al. (2003), respectively.


Babin, M., Stramski, D., Ferrari, G. M., Claustre, H., Bricaud, A., Obolensky, G. &al., e. (2003).

Variations in the Light Absorption Corefficients of Phytoplankton, Non Algal Particles, and

Dissolved Organic Matter in Coastal Waters Around Europe. Journal of Geophysical Research

108: 3211 (doi: 3210.1029/2001JC000882).

Breton, E., Rousseau, V. &Parent, J. Y. (2006). Hydroclimatic Modulation of Diatom/Phaeocystis blooms

in nutrient-enriched Belgian Coastal Waters (North Sea). Limnology and Oceanography 100:

13321-13332.

Bricaud, A., Babin, M., Morel, A. &Claustre, H. (1995). Variability in the Chlorophyll-Specific

Absorption-Coefficients of Natural Phytoplankton - Analysis and Parameterization. Journal of


MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 71


Lubac, B. &Loisel, H. (2007). Variability and classification of remote sensing reflectance spectra in the

eastern English Channel and southern North Sea. Remote Sensing of the Environment 110: 45-58.

Hooker, S. B. &Morel, A. (2003). Platform and Environmental Effects on Above-Water Determinations

of Water-Leaving Radiances. Journal of Atmospheric and Oceanic Technology 20: 187-205.

Mueller, J. L. (2003b).In-water Radiometric Profile Measurements and Data Analysis Protocols. Vol.


Mueller, J. L. &Austin, R. W. (2003).Characterisation of Oceanographic and Atmospheric Radiometers.

Volume II: Instrument Specifications, Characterisation and Calibration. In Ocean Optics

Protocols For Satellite Ocean Color Sensor Validation, Revision 4Greenbelt, MD: NASA/GSFC.

Smith, R. C. &Baker, K. S. (1984).The Analysis of Ocean Optical Data. In Ocean Optics VIISPIE, Vol.

478, 119-126 (Ed M. A. Blizard).

8.2 FERRYBOX. PI: K. Sørensen

8.2.1 Introduction

Ferrybox is the name of a coordinated project aimed at collecting in-situ marine scientific data for marine

observation and monitoring. Funded by the EU Science Framework 5 Ferrybox enabled the cooperation

of 11 organisations and established the coordinated use of commercial ferry ships for the collection of

scientific data. The 11 partners operated on 9 shipping routes around Europe, from the eastern

Mediterranean to the Baltic. The Ferrybox systems consist of a fully automated flow-through system with

sensors and automatic analysers for the measurement of physical, biological and chemical parameters,

which uses ferryboats and other ships of opportunity as the carrier system, and these route and vessels

have been exploited by the optics community in Norway (NIVA) to mount TriOS Ramses instruments for

optical measurements. Some of the advantages of using these Ferrybox carrier vessels to mount TriOS

systems include easy maintenance in the harbour (no additional ship time is needed) and that the

information from a transect is often better than from a single location. For more information on Ferrybox

go to http://www.ferrybox.com/index.html.en.

Presently the data from the Norwegian TriOS is undergoing analysis and processing, and has not yet been

submitted to MERMAID. The measurement protocol is available in advance, as described in Høkedal and

Sørensen (2007).

8.2.2 Measuring system and measurement Protocol

Mounted on ferries covering most of the Norwegian coast are sets of optical sensors. Each set consists of

two radiance sensors (one viewing the surface of the sea, one at the sky) and one irradiance sensor, in

addition to GPS-sensor controlling the system. The collected data are used in calculating radiance

reflectance which is also one of the MERIS-products. The systems only require maintenance on weekly or

longer intervals, and are hardly affected by deposits of salt or airborne particles.

The radiance and irradiance sensors are from the German company TriOS, mounted on two ships, „Color

Festival‟ in the Skagerrak and „Trollfjord‟ along the Norwegian from 60ºN to the Russian border. Each

system set consists of one irradiance sensor (for downward irradiance) and four radiance sensors, two for

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 72


upward and two for downward radiance. These 5 sensors are together with a GPS receiver connected via

an interface box to a computer which records the data.

For ocean colour data to be used it is an absolute demand that the footprint is not disturbed by the ship,

either by wakes and the shadow of the ship. Additionally the azimuth angle between the sun and the

sensor should ideally be 135º to minimise the sun glint. Given this, and together with the route of the

ferries the sensors are installed in the following way: On „Color Festival‟ the downward viewing sensors

are mounted slightly forward at the starboard and port side (which is optimal with a ship heading

northward during daytime). The route of „Trollfjord‟ is more complicated; however the sensors are

mounted at the port side of the ship, forward and aft viewing respectively.


The parameter derived from the TriOS system processing is ρw (λ) as defined by Equation (3).

Lw is calculated from measurements of nadir-radiance below the surface (which are extrapolated to the

surface, then the water-leaving part is determined from the refractive index of water). This gives the right

hand side of Equation (35) which is compared with the reflectance given by MERIS.

On a continuously moving platform ferry subsurface measurements for determining Lw are, for obvious

reasons, impossible. Instead, radiance is measured above surface with a downward viewing radiometer.

However, this measures not just Lw, but the sum of this and the downward direct and diffuse light

reflected at the surface, i.e. Lt = Lw + Lwrefl

. To estimate the latter addend of this sum the downward

diffuse contribution Lsky at the nadir angle (π- θ) is commonly measured, assuming this is the radiance

reflected into radiometer (or proportional to it). Thus, we get:

),,(),,( skyf

refl

w LL (35)

Where, for a flat surface the Fresnel reflection f takes a value around 2%. With a wind roughened surface

the value of 2.8% has been determined for an overcast sky.

However, these values of f are not plausible because the reflected light does not only consist of Lsky (π-

); due to the rough surface the entire upper hemisphere Lwrefl

is dependent on Lsky (Ω) (which depends on

the solar elevation, where Ω denotes the solid angle); additionally Lwrefl

also depends on viewing direction

and wind speed. Such functionality is far more complicated than the scope of this protocol, so instead a

simpler solution is to presume that Lsky (Ω) is better correlated with Es than Lsky (λ, π- , ). Therefore, we

write:

)(),,( sfe

refl

w EL (36)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 73


where fe is the equivalent Fresnel reflection, expressed as sr-1

. It is independent of wavelength, but will

vary amongst the measured spectra. To determine fe we presume that due to high absorption the water

reflection coefficient is assumed to be nul somewhere in the wavelength domain from 800 to 900 nm.

With the observed values of upward radiance and downward radiance it is then straight forward to find

the value of fe for each measurement.

On the ferry the radiometers are mounted in fixed angles relative to the ship. As a consequence of this θs

(between the sun and the radiometer) depends on the sun‟s position as well as the course of the ship. Here

we will not go into details but mention that the measurements are filtered for shadow of the ship at the

footprint of the radiance sensor, θs less than 90o, and finally solar elevation less than 30

o.

8.2.4 Uncertainties

If the ferry follows an optimal transect combined with a careful choice of observing angles (to avoid ship

shadow on the sensors footprint as well as sun-glint from the surface) it is shown that data can be

collected most of the day, including during satellite passage. Satellite coverage and cloud conditions limit

the number of possible match-ups, not the ship-time which usually is the limiting factor. Results show

typically 20% deviation between reflectance determined from MERIS and in-situ data.

8.2.5 Key Reference

Høkedal, J. & Sørensen, K. (2007).Validation of MERIS-reflectance from Ferries. In ENVISAT

Symposium. Montreux, Switzerland, 23-27 April 2007.

8.3 French Guiana. PI: H. Loisel; C. Jamet

The location of the sampled stations in French Guiana is indicated in Figure 8-2; the campaign and results

are fully described in Loisel et al. (2009). Remote sensing reflectance, RRS (), measurements were

performed at each station using two TriOS hyperspectral (every 3 nm) radiometers: one measuring the

downwelling irradiance on the deck, Es (0+,λ), and one measuring the in water upwelling radiance just

below the sea surface, Lu (0-, ). The immersion factors, as well as the impact of the self-shading are

accounted for as in Lubac and Loisel (2007). These measurements were performed from a very small flat

bottomed boat, far away from any perturbations of the main boat. RRS () is calculated as in Equation (12)

with the air-sea interface transmittance term value (to convert Lu to above water) given in Table 1-3.

The remote sensing reflectance spectra exhibited a great variability in shape and amplitude; however, all

these RRS () spectra fall into the five classes defined by Lubac and Loisel (2007).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 74


Figure 8-2: Location of the stations sampled on 7-11 July 2006 (from Loisel et al., 2009).


Loisel, H., Mériaux, X., Poteau, A., Artigas, L. F., Lubac, B., Gardel, A., Caillaud, J. &Lesourd, S.

(2009). Analyze of the inherent optical properties of French Guiana coastal waters for remote

sensing applications. Journal of Coastal Research ICS Proceedings.



8.4 Helgoland/Cuxhaven Transect. PI: R. Doerffer

8.4.1 Introduction

This protocol (Doerffer and Schönfeld, 2009) describes data from the ferry cruises between Cuxhaven

and Helgoland, which are carried out for the validation of MERIS data (Figure 8-3). These matchups are

not yet in the MERMAID database, but are the subject of investigation into matchups over transects.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 75


Figure 8-3: Route and stations of the MERIS validation campaign "c30" on 13. July 2006

The basic idea is to sample data along a transect with a strong gradient of the concentration of water

constituents, in particular of yellow substance and suspended matter. Such a gradient can then be

compared with the corresponding transect extracted from MERIS data. The main problem of the

validation of case 2 waters is the strong patchiness and dynamics of case 2 coastal waters due to tidal

currents. A direct comparison between a water sample and a pixel is not possible because of the different

spatial and temporal relationships between these two measurements. When using a transect one can

compare at least the shape as well as the concentration range along the transect, although details may also

differ. Also the comparison between path radiance and water leaving radiance indicate if the atmospheric

correction over turbid water was successful. During the summer month there is a daily ferry connection

between Cuxhaven at the mouth of the Elbe River and the island of Helgoland (Figure 8-3). Shortly

before arriving Helgoland the overpass of MERIS happens. A further advantage is that it can be decided

to carry out a validation cruise on the basis of the weather forecast on short notice.

8.4.2 Measuring system and measurement protocol

During the entire cruise we operate a TRIOS RAMSES radiometer system to determine the water

reflectance spectrum (Figure 8-4). The instruments are mounted at the bow of the ship to avoid the ship

induced foam and to minimize shading / reflections by the ship‟s hull. One radiometer points to the sea

surface under an angle of 45 degrees with an azimuth angle of about 130-140 degrees with respect to the

sun. The second radiometer point to the sky under the same angles. The third radiometer measures the

downwelling irradiance. The water reflectance ρw (λ) is then computed according to:

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 76


)(

),,(),,()(

s

skyft

wE

LL (37)

With Lt (λ) the total upward directed radiance from the water, Lsky (λ) the sky radiance, Es (λ) the

downwelling irradiance (above surface) and ρf the Fresnel reflection of water surface, computed according

to the Snell laws using a wavelength dependent on the refractive index for the mean salinity along the

transect.

Note: For our purpose we compute the directional water leaving radiance reflectance, RLw, without the PI

factor, then the unit is [sr-1], but in the delivered files the PI-factor is included for comparison with MERIS

data. Furthermore, water is sampled along each transect from the surface at about 6 stations, with the last

station (close to Helgoland) at the time of ENVISAT overpass. Water samples are processed at the end of

the cruise in the laboratory of the Biologische Anstalt on Helgoland.

Table 8-1 lists the transect dates for which data was provided to MERMAID, and the following variables

are determined:

1) phytoplankton pigment,

2) dry weight of total suspended matter,

3) dry weight of the inorganic and organic fraction of TSM,

4) absorption spectrum of the water after filtration with a pore size of 0.2 μm,

5) absorption spectrum of the filter pad before and after bleaching.

6) Furthermore, water temperature and salinity are recorded.

Figure 8-4: a) TRIOS-Spectrometer for measuring upward directed radiance from water and sky radiance,

and b) TRIOS Spectrometer for measuring downwelling irradiance

a) b)

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 77


Table 8-1: Number and dates of transect campaigns

Station Ship Date Station Ship Date

C01 Wappen Von Hamburg 29/07/2002 C18 Südfall 20/04/2005

C02 Wappen Von Hamburg 30/07/2002 C19 Wappen Von Hamburg 21/04/2005





C09 Wappen Von Hamburg 05/08/2003 C24 Funny girl 06/10/2005







C16 Ludwig Prandtl 03-06/08/2004 C31 Wappen Von Hamburg 25/07/2006


8.4.3 Uncertainties

TBD


Doerffer, R. & Schönfeld, W. (2009).Validation Transect between Cuxhaven and Helgoland: GKSS

Technical Note (18th February 2009). GKSS, Germany.

8.5 MUMMTriOS. PI: K. Ruddick

The following protocol is extracted from Ruddick (2006), wherein further details can be found.

Measurements were performed with three TriOS-RAMSES hyperspectral spectroradiometers, two

measuring radiance, L, and one measuring downwelling surface irradiance Es. The instruments were

mounted on a steel frame as shown in Figure 8-5. Zenith angles of the sea- and sky-viewing radiance

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 78


sensors were 40o. The frame was fixed to the prow of the ship, facing forward to minimize ship shadow

and reflection. The ship was maneuvered on station to point the radiance sensors at a relative azimuth

angle of 135o away from the sun. Lenses were checked and, if necessary, cleaned prior to each

measurement. Measurements were made for 10 min, taking a scan of the three instruments every 10 s.

The sensors measured over the wavelength range of 350–950 nm with a sampling interval of

approximately 3.3 nm and a spectral width of about 10 nm. Position was measured simultaneously by

global positioning system (GPS). Data were acquired with the TriOS GmbH MSDA software using the

file recorder function and radiometrically calibrated using nominal calibration constants. Calibrated data

for downwelling irradiance, Ed0+

(hereafter termed Es, by the definition provided in Section 1.4), sea

radiance, Lsea0+

, and sky radiance, Lsky0+

, were interpolated to 2.5 nm intervals. The sensors were

calibrated in a MERIS Validation Team laboratory every year, after which the definitive spectra were

obtained.

Water reflectance, ρw, is calculated from simultaneous above-water measurements of Es, total upwelling

radiance (i.e., from the water and from the air-sea interface) at a zenith angle of 40o, Lsea

0+; and sky

radiance, Lsky0+

, sea sky in the direction of the region of sky that reflects into the seaviewing sensor, by:

s

skyfsea

wE

LL

00 (38)

where ρf is the air-water interface reflection coefficient for radiance equal to the Fresnel reflection

coefficient in the case of a flat sea surface. This corresponds to „„Method 1‟‟ of the NASA protocols

(Mueller et al., 2000).

Figure 8-5 Frame with three TriOS-RAMSES hyperspectral radiometers as installed on the research vessel

Belgica (Ruddick, 2006).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 79


ρf is expected to vary strongly with wind speed for clear sky conditions because of reflection of brighter

parts of the sky in the case of higher waves (Mobley, 1999), but is approximately independent of wind

speed for cloudy conditions. This can be accounted for by switching between clear sky (Equation 39) and

cloudy sky (Equation 40) models for ρf, according to the ratio Lsky0+

/Es at 750 nm, which takes a value of

about 0.02 in the clear sky simulations of Mobley (1999) but can reach much higher values (e.g., of order

0.3 for fully overcast conditions):

05.0)750(

)750(000034.000039.00256.0

0

2

s

sky

fE

LforWW (39)

05.0)750(

)750(256.0

0

s

sky

fE

Lfor (40)

The sunny sky formula of Equation 39 is derived as a function of wind speed, W, in meters per second at

height 10 m from the model simulations of Mobley (1999) based on the wave slope statistics of Cox and

Munk (1954). Although Mobley (1999) reports a slight sun zenith angle dependency for ρf, Equation 39

fits all simulations for the range 30o ≤ θ0 ≤ 70

o to within 1% for W 5 m

-s and to within 3% for W 10 m

-s.

For the intermediate case of partially cloudy skies, whether obscuring the sun or patchy near the sky-

viewing direction, neither of these formulations is entirely appropriate. Although problematic for many

above-water reflectance measurements, this intermediate case is not relevant because such data are

removed from the analysis, as are the fully cloudy data where measurement uncertainties are more

significant.


Ruddick et al. (2006a) describe data processing procedures for MUMM optical measurements. The

measurement sequence of scanning every 10 s for 10 min produces a time series of 60 scans. ρw calculated

from these 60 scans will vary in time for a number of reasons such as wave and sunglint, bidirectionality,

water optical properties, tilt, floating material, clouds and skyglint, solar zenith angle. After inspection

and analysis of the time series recorded for stations in a wide variety of atmospheric and marine

conditions, the following approach for calculation of the ρw was adopted. Scans are flagged for rejection if

any of the following cases occurred:

Inclination from the vertical exceeded 5%;

Es, Lsea0+

, Lsky0+

at 550 nm differs by more than 25% from either neighbouring scan; or

Incomplete or discontinuous spectra (occasional instrument malfunction).

In practice, scan rejection is low or zero for calm sea and clear sky conditions, but increases rapidly with

wind speed and/or if cloud cover is scattered and may reach 80% or more in the worst conditions. Next,

the ρw measured from the first five scans passing these tests are mean-averaged to yield the ρw and its

standard deviation for each station.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 80


Only stations meeting the following quality requirements were used:

Clear, sunny skies as denoted by the relation Lsky0+

/Es (750) < 0.05,

Standard deviation of the five scans of reflectance was,10% of the mean average at 780 nm,

Wind speed, 10 m-s

8.5.2 Uncertainty Estimates for the NIR

Ruddick et al. (2006b) describe measurement uncertainty analyses and estimates. The uncertainty of

measurements of ρw(λ) and of the derived similarity spectrum )(780 wn arising from the method

described in the main text is considered to arise from three main sources (Zibordi et al., 2002; Zibordi et

al., 1999): Instrument calibration and performance, correction for air-sea interface reflection, and optical

changes of the water induced by the measurement platform.


Cox, C. &Munk, W. (1954). Measurements of the Roughness of the Sea Surface from Photographs of the

Sun's Glitter. Journal of the Optical Society of America 44: 834-850.

Mueller, J. L., Davies, C., Arnone, R., Frouin, R., Carder, K. L., Lee, J. P., Steward, R. G., Hooker, S. B.,

Mobley, C. D. &McLean, S. (2000).Above-water Radiance and Remote Sensing Reflectance

Measurements and Analysis Protocols. In Ocean Optics Protocols for Satellite Ocean Cor Sensor

Validation. Rev 2.(Eds G. Fargion and J. L. Mueller). NASA.



Ruddick, K. (2006). Seaborne measurements of near infrared water-leaving reflectance: The similarity

spectrum for turbid waters. Limnology and Oceanography 51(2): 1167-1179.

Ruddick, K., De Cauwer, V., Park, Y. &Moore, G. F. (2006a). Web Appendix 1. Data Processing: Scan

Selection and Averaging. Limnology and Oceanography 51(2): 1167-1179.

Ruddick, K., De Cauwer, V., Park, Y. &Moore, G. F. (2006b). Web Appendix 2. Measurement

Uncertainty Analysis. Limnology and Oceanography 51(2): 1167-1179.

Zibordi, G., Doyle, J. P. &Hooker, S. B. (1999). Offshore Tower Shading Effects on In-water Optical

Measurements. Journal of Atmospheric and Oceanic Technology 16: 1767-1779.

Zibordi, G., Hooker, S. B., Berthon, J.-F. &D'Alimonte, D. (2002). Autonomous above water radiance

measurements from stable platforms. Journal of Atmospheric and Oceanic Technology 19: 808-

819.

8.6 Wadden Sea. PI: A. Hommersom

8.6.1 Introduction

The Wadden Sea, a shallow coastal area bordering the North Sea, is optically a complex area due to its

shallowness, high turbidity and fast changes in concentrations of optically active substances. Hommersom

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 81


et al. (2009) describe the site and methodologies in detail, but here provide a summary of the

measurements contributed to MERMAID.

8.6.2 Measurement protocol and data processing

Reflectance

Reflectance, ρw (λ), spectra were calculated from measurements with three multi-spectral radiometric

sensors (TriOS). Two radiance sensors (RAMSES ARC; field of view of 7 o) were employed at the front

of the ship in azimuth ~135 º away from the sun. One sensor measured light that entered the water at an

angle of 41 º in zenith direction (sky radiance, Lsky). The other sensor measured at angle of 41º from nadir

direction, the light escaping the water (surface radiance, Lt). Foam, shadow and reflectance from the ship

were avoided; if necessary the 135 º azimuth angle was adjusted in an angle between > 90 ° and < 180 °.

The third sensor (RAMSES ACC, cosine) was employed on top of the ship and was used to measure

downwelling irradiance (Es). Measurements were carried out according to the Ocean Optics protocols

(Mueller et al., 2003a). The water-leaving radiance (Lw) was calculated as:

)()()( skyftw LLL (41)

Where: ρf is the surface reflectance, estimated with the information from Mobley (1999).

ρw (λ) was calculated according to Equation (3).

Chl

All samples were taken with a bucket. For Chl concentration measurements GF/F filters were used. After

filtration the filters were frozen at -20 °C and transferred to -80 °C in the lab within two weeks of taking

the first sample. Chl samples were analysed on HPLC, mainly according to the Ocean Optics protocol

(Mueller et al., 2003b), except for the solvent gradient program, which was modified to improve

separation. Peak areas were measured relative to the peak areas of a Chl standard in fresh water.

Concentrations of the standard were determined in acetone with a spectrophotometer. A correction is

applied for the amount of water that remains in a filter following Mueller et al. (2003b). In an experiment

the amount of water retained in a 47 mm GF/F filter was found to be 0.58 ml.

8.6.3 Uncertainties

Not yet available




MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 82


Muller, J. L., Morel, A., Frouin, R., Davies, C., Arnone, R. & Carder, K. L. (2003).Ocean Optics

Protocols for Satellite Ocean Color Sensor Validation, Revision 4, Volume III: Radiometric

Measurements and Data Analysis Protocols., 78 Greenbelt, MD: NASA/GSFC.

Tilstone, G. H., Moore, G. F., Sørensen, K., Doerffer, R., Røttgers, R., Ruddick, K., Pasterkamp, R. &

Jørgensen, P. V. (2002).REVAMP Regional Validation of MERIS Chlorophyll products in North Sea

coastal waters., 78 Plymouth: Plymouth Marine Laboratory.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 83


9. MERMAID PROTOCOLS VII: Miscellaneous datasets

9.1 NASA bio-Optical Marine Algorithm Data set (NOMAD). PI: Jeremy Werdell

9.1.1 Introduction

The NASA bio-Optical Marine Algorithm Data set (NOMAD) is a compilation of data provided to the

NASA SeaWiFS Bio-optical Archive and Storage System (SeaBASS), a repository for in-situ radiometric

and phytoplankton pigment data used by the NASA Ocean Biology Processing Group (OBPG). The data

is used in the satellite validation activities of the OBPG (Werdell and Bailey, 2002; Werdell et al., 2003),

and in particular for remote sensing studies. The NOMAD dataset is made available on the SeaBASS

website, and version 2.0 of the dataset is that included in the MERMAID database. Werdell and Bailey

(2005) describe in detail the NOMAD database, but here the information pertinent to MERMAID is

presented.

For NOMAD, the relevant parameters for MERMAID are Lw (λ) and Es (λ), from which ρw (λ) is

computed.

9.1.2 Measuring systems and protocols

Approximately 15,000 radiometric observations were acquired from SeaBASS for NOMAD. The

radiometric data sources include in-water profiling instruments and handheld or platform mounted above-

water instrumentation. Measurements from both multi- and hyperspectral resolution instruments are

included; see Werdell and Bailey (2005) and Werdell (2005) and references therein for details.

Consistency in the data was facilitated by a requirement to adhere to pre-specified in-situ data

requirements and sampling strategies (Mueller et al., 2003a), namely the NASA Ocean Optics Protocols,

or similarly appropriate methods documented by the provider.

The data received by the OPBG is fully processed to depth-registered, calibrated geophysical values by

the data contributor prior to inclusion in SeaBASS and, apart from outliers (queried of the contributor) or

uncertain methods, were considered accurate (Werdell and Bailey, 2005).

9.1.3 Data processing to Lwn (λ)

NOMAD consists of radiometric profiles limited to those with coincident observations of upwelling

radiance, Lu (λ, z) and downwelling irradiance, Ed (λ, z), and if available measurements of surface

irradiance, Es (λ,0+), usually collected near-simultaneously either on the deck of the vessel or nearby

buoy.

The OBPG use especially-designed software (Werdell and Bailey, 2002) to visualise and process the data.

Measurements without near-surface profiler data (less than 5m) or without significant overall stability in

the reference surface irradiance were excluded. All remaining measurements of radiance and irradiance

were corrected for variations in solar irradiance using a surface reference value if available and not

performed prior to data submission. Contaminated observations, for example those resulting from wave

focusing near the surface or high tilt, were removed from the profiles. Werdell and Bailey (2005) provide

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 84


no further detail regarding exclusion of tilt-affected data. Data collected under non-ideal, or cloudy, sky

conditions are not excluded from NOMAD.

Near-surface diffuse attenuation coefficients were calculated by the OBPG from the radiance and

irradiance profiles via a linear exponential fit to the corrected data. These coefficients were used to

propagate the radiances and irradiances to just beneath the surface, Lu(0-) and Ed(0

-), respectively.

Water leaving radiances were then determined using Equation (11), and using a value of 0.975 for the

Fresnel transmittance term and 1.34 for n.

Similarly, extrapolated surface irradiances were computed as

)0(1

dds EtE (42)

Where td is the downward Fresnel irradiance transmittance across the air-sea interface (0.96; Mueller et

al., 2003c).

Observations were considered questionable and discarded if extrapolated surface irradiances could not be

reconciled with reference surface irradiances (extrapolated Es). Intermediate processing details were

logged, including the extrapolation depth intervals, which were occasionally variable as a function of

wavelength, extrapolation statistics, cast direction, and processor-defined comments.

The remote sensing diffuse attenuation coefficient

Gordon and McCluney (1975) demonstrated that 90% of remotely sensed radiance originates in the upper

layer, defined by depth z90, corresponding to the first optical attenuation length as defined by Beer‟s Law.

Measurements of Ed (k, z) were smoothed using a weighted least-square polynomial fit. Using the

smoothed data and the previously calculated subsurface irradiance, values for z90(k) were identified as the

depth which satisfied the condition

1

90 )0,(),( ekEzkE dd (43)

Remote sensing diffuse attenuation coefficients, Krs(k), were calculated from the original irradiance

profiles by applying a linear exponential fit over the depth range from z = 0- to z90(k). Radiometric

profiles with retrieved Krs(k) values less than the value for pure water (Kw(490)=0.016 m-1; Mueller,

2000) were considered questionable and discarded. Otherwise, both Krs(k) and z90(k) were recorded.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 85


Above-water radiometry

Above-water measurements of surface and sky radiance under known geometric conditions may be used

to derive water-leaving radiances (Deschamps et al., 2004; Mobley, 1999). Contributors of above-water

radiometric observations performed this derivation prior to submission to SeaBASS, thus eliminating the

need for additional data preparation. For 12% of these field campaigns (relating to 39% of all above-water

observations), the data contributor provided remote sensing reflectances, RRS (λ), in lieu of Lw (λ). In these

cases, water-leaving radiances were estimated from remote sensing reflectances via:

)()()( srsw ERL (44)

If Es (λ) was not explicitly provided, they were derived using a clear sky model based on Frouin et al.

(1989); this is an operation requiring an assumption of ideal sky conditions at the time of data collection.

The uncertainty introduced by such an assumption will be minimal when developing algorithms using

radiance ratios, as the modeled Es (λ) are, in general, spectrally flat. That is, the ratio of any two discrete

modeled irradiance values is approximately unity, and, following, the errors associated with the

magnitude of the modeled irradiances are mathematically cancelled. Uncertainty will also be negligible

for satellite validation activities, where only the clearest days are considered (Werdell et al., 2003). While

uncertainties associated with above-water radiometry may be significant even under ideal conditions

(Hooker et al., 2002) data collected onboard underway research vessels are included in the dataset if

located in an otherwise under-sampled bio-regime (Werdell and Bailey, 2005).

To mimic the band passes inherent to most commercially available multispectral radiometers, data

collected with hyperspectral instruments were degraded to 11-nm averages centred on λc, as defined by:

n

)(X)(X

5

5i i

c

c

(45)

where: X is some radiometric quantity, such as RRS, and n is the number of wavelengths considered (i.e.

11).

Radiometric data reduction

All replicate radiometric measurements were individually viewed and reduced via analysis of coincident

Lw (λ), Es (λ), and RRS (λ), the combined evaluation of which provides simultaneous insight into

processing artefacts, changing sky conditions, and water-mass variability that results from erroneous

replicate identification. For example, for a given station with multiple measurements, comparable surface

irradiance spectra indicate stable sky conditions and similar RRS (λ) implies a consistent water mass.

Under ideal circumstances, when the statistical variance of all three products was low, the geometric

mean was calculated. If only the RRS (λ) were stable, we retained the single observation with the highest

Es (λ), an indicator of the clearest sky conditions. Stable Es (λ) and highly variable Lw (λ) suggested errors

in data processing or replicate evaluation (e.g. insufficiently small spatial thresholds for frontal regions).

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 86


Data with these symptoms were re-evaluated, and discarded upon unsuccessful reconciliation. For all of

the above, the average observation time, latitude, and longitude were recorded.

Exceptions

Naturally, the volume of data considered and the wide range of sources prompted several exceptions to be

made in how specific data were treated. SeaBASS data contributors occasionally provided Lw (λ) that

included a correction for instrument self-shading (Zibordi & Ferrari, 1995), either without providing the

radiance profiles, or with a documented perspective in favour of the correction for their particular field

campaign. NOMAD therefore excludes all radiometric data collected solely on tethered buoys (e.g., the

Satlantic, Inc. Tethered Spectral Radiometry Buoy) and moorings, as these data are predominantly scarce

in SeaBASS and, when available, rarely includes supporting radiometric information for use in the

extrapolation of Lu (k, z) to Lu (k, 0-).

Wavelength generalisation

Lw (λ), Es (λ) and Kd coefficients retain their native instrument-resolution in the RDBMS, for example,

Lw(411.8) is not rounded to Lw(412), yielding approximately 250 uniquely catalogued wavelengths. In

general, such exact radiometric precision is not required for algorithm development (O'Reilly et al., 2000;

O'Reilly et al., 1998) so to simplify the data for generalised and efficient use, wavelengths are rounded in

each data export process. A series of 21 nominal wavelengths are predefined after both reviewing the

spectral resolution of past, present, and future ocean colour satellites and considering the frequency of

occurrence of centre wavelengths in the merged data set. When exported from the database, radiometric

data are assigned the predefined wavelength, λpd, that satisfies the condition (λpd-2-nm)_kn_(λpd+2-nm),

where λn is the native instrument wavelength.


The volume of data considered and the range of sources prohibit comprehensive description of data

sampling in this article. However, Werdell and Bailey (2005) provide the following description. Data are

fully processed to depth-registered, calibrated geophysical values by the data contributor prior to

inclusion in SeaBASS, thus eliminating the need for any additional calibration or normalisation efforts.

Data contributors were queried when outliers and questionable measurements were identified, or when

data processing methods or instrument calibrations were uncertain. Otherwise, the data were considered

accurate as is, following acquisition from SeaBASS. To facilitate the post-processing evaluation of

uncertainties resulting from varying observation types and measurement resolutions (such as in-water

versus above-water radiometry, or analysis of water samples collected via profiling rosettes versus

shipboard sea chests), the OBPG established a series of binary flags to record collection and processing

details for each measurement. As such, a flag field accompanies every measurement in the final compiled

data set.

9.1.5 Uncertainties

Upwelling radiances were not corrected for instrument self-shading (Gordon and Ding, 1992; Zibordi and

Ferrari, 1995), as the required supporting data were often inadequate (such as the absorption coefficient of

the water mass and the ratio between diffuse and direct sun irradiance). The uncertainty introduced by

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 87


omitting this correction varies geographically and temporally and by instrument. Uncertainties are

generally between 5 and 9%.




Press, London.

Deschamps, P.-Y., Fougnie, B., Frouin, R., Lecomte, P. & Verwaerde (2004). SIMBAD: A Field

Radiometer for Satellite Ocean-Color Validation. Applied Optics 43(20): 4055-4069.

Frouin, R., Longner, D. W., Gautier, C., Baker, K. S. & Smith, R. C. (1989). A simple analytical formula

to compute clear sky total and photosynthetically available solar irradiance at the ocean surface. Journal

of Geophysical Research 94: 9731-9742.

Gordon, H. R. & Ding, K. (1992). Self-shading of in-water optical instruments. Limnology and


Gordon, H. R. & McCluney, W. R. (1975). Estimation of depth of sunlight penetration in sea for remote-

sensing. Applied Optics 14: 413-416.

Hooker, S. B., Lazin, G., Zibordi, G. & McClean, S. (2002). An evaluation of above- and in-water

methods for determining water leaving radiances. Journal of Atmospheric and Oceanic Technology 19:

486-515.



Mueller, J. L. (2000).SeaWiFS algorithm for the diffuse attenuation coefficient, K(490), using water-

leaving radiances at 490 and 555 nm In SeaWiFS postlaunch calibration and validation analyses: Part 3.

NASA Technical Memorandum, Vol. 11, 24-27 (Eds S. B. Hooker and E. R. Firestone). Greenbelt,

Maryland, USA.: NASA Goddard Space Flight Centre.

Mueller, J. L., Morel, A., Frouin, R., Davies, C., Arnone, R. & Carder, K. L. (2003c).Ocean Optics

Protocols for Satellite Ocean Color Sensor Validation, Revision 4, Volume III: Radiometric

Measurements and Data Analysis Protocols., 78 Greenbelt, MD: NASA/GSFC.

O'Reilley, J., Maritorena, S. A., O'Brien, M. C., Siegel, D. A., Toole, D. & Menzies, D. (2000).Ocean

Color Chlorophyll a algorithms for SeaWiFS, OC2 and OC4: Version 4. . In SeaWiFS Postlaunch

Calibration and Validation Analysis:, 9-23 (Eds S. B. Hooker and E. R. Firestone). Greenbelt, MD:

NASA.

O'Reilly, J. E., Maritorena, S., Mitchell, B. G., Siegel, D. A., Carder, K. L., Garver, S. A., Kahru, M. &

McClain, C. R. (1998). Ocean Color Algorithms for SeaWiFS. Journal of Geophysical Research 103:

24,937-924,953.

Werdell, P. J. &Bailey, S. W. (2002).The SeaWiFS Bio-optical Archive and Storage System (SeaBASS):

Current Architecture and Implementation. Greenbelt, Maryland.: NASA Goddard Space Flight Centre.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 88


Werdell, P. J. &Bailey, S. W. (2005). An Improved In situ Bio-Optical Data Set for Ocean Colour

Algorithm Development and Satellite Data Product Validation. Remote Sensing of Environment 98: 122-

140.


(2003). Unique data repository facilitates ocean color satellite validation. EOS Transactions 84(3): 379.

Zibordi, G. & Ferrari, G. M. (1995). Instrument self-shading in underwater optical measurements:

experimental data. Applied Optics 34: 2750-2754.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 89


10. Appendix 1: Values of the air-sea interface transmittance as function of

wind speed, ws, view angle, ', and salinity (salt and freshwater).

Table 10-1: Air-sea interface transmittance (35 psu).

Wind force 1 2 3 4

ms-1 0.25 1.00 2.75 5.00

'

Band 1

412.691

0 0.534647 0.537108 0.532993 0.533324

15 0.534633 0.537071 0.532907 0.533184

30 0.534622 0.537041 0.532841 0.533079

45 0.534611 0.537012 0.532777 0.53298

60 0.534608 0.537004 0.532761 0.532955

Band 2

442.559

0 0.534641 0.537093 0.53296 0.533274

15 0.534623 0.537045 0.532852 0.533098

30 0.53461 0.537009 0.532773 0.532974

45 0.534597 0.536975 0.532697 0.532857

60 0.534593 0.536965 0.532678 0.532827

Band 3

489.882

0 0.540606 0.542336 0.541859 0.53747

15 0.540583 0.542275 0.541721 0.537255

30 0.540567 0.542233 0.541627 0.53711

45 0.540551 0.542193 0.541537 0.536974

60 0.540547 0.542182 0.541515 0.536941

Band 4

509.819

0 0.540603 0.542328 0.54184 0.537441

15 0.540578 0.542262 0.541692 0.537211

30 0.540561 0.542218 0.541594 0.53706

45 0.540544 0.542175 0.5415 0.536918

60 0.54054 0.542164 0.541477 0.536884

Band 5

559.694

0 0.540602 0.542326 0.541834 0.537431

15 0.540575 0.542255 0.541677 0.537188

30 0.540557 0.542209 0.541575 0.537032

45 0.54054 0.542164 0.541478 0.536886

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 90


60 0.540536 0.542153 0.541454 0.53685

Band 6

619.601

0 0.544808 0.543715 0.542217 0.543843

15 0.544778 0.543641 0.542055 0.543587

30 0.54476 0.543596 0.541953 0.543426

45 0.544743 0.543551 0.541855 0.543276

60 0.544738 0.54354 0.541831 0.54324

Band 7

664.5731

0 0.544809 0.543716 0.542219 0.543846

15 0.544778 0.54364 0.542053 0.543584

30 0.54476 0.543594 0.54195 0.543422

45 0.544742 0.543549 0.541851 0.54327

60 0.544737 0.543537 0.541827 0.543234

Band 8

680.821

0 0.544809 0.543717 0.54222 0.543847

15 0.544778 0.543639 0.542052 0.543583

30 0.54476 0.543594 0.541949 0.543421

45 0.544742 0.543548 0.54185 0.543269

60 0.544737 0.543537 0.541826 0.543233

Band 9

708.329

0 0.549592 0.545889 0.546687 0.545824

15 0.549559 0.54581 0.546515 0.545561

30 0.54954 0.545764 0.546411 0.545399

45 0.549521 0.545718 0.54631 0.545248

60 0.549517 0.545706 0.546285 0.545211

Band 10

753.371

0 0.549592 0.545889 0.546686 0.545823

15 0.549559 0.54581 0.546515 0.54556

30 0.54954 0.545764 0.54641 0.545399

45 0.549521 0.545717 0.54631 0.545247

60 0.549517 0.545706 0.546285 0.545211

Band 12

778.4091

0 0.549592 0.545889 0.546686 0.545823

15 0.549559 0.54581 0.546515 0.54556

30 0.54954 0.545764 0.54641 0.545399

45 0.549521 0.545717 0.54631 0.545247

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 91


60 0.549517 0.545706 0.546285 0.545211

Band 13

864.876

0 0.549592 0.545889 0.546686 0.545823

15 0.549559 0.54581 0.546515 0.54556

30 0.54954 0.545764 0.54641 0.545399

45 0.549521 0.545717 0.54631 0.545247

60 0.549517 0.545706 0.546285 0.545211

Band 14

884.944

0 0.549592 0.545889 0.546686 0.545823

15 0.549559 0.54581 0.546515 0.54556

30 0.54954 0.545764 0.54641 0.545399

45 0.549521 0.545717 0.54631 0.545247

60 0.549517 0.545706 0.546285 0.545211

Table 10-2: Air-sea interface transmittance (0 psu).

Wind force 1 2 3 4

ms-1 0.25 1.00 2.75 5.00

'

Band 1

412.691

0 0.540214 0.5427 0.538542 0.538877

15 0.5402 0.542663 0.538455 0.538735

30 0.540188 0.542632 0.538389 0.538629

45 0.540177 0.542603 0.538325 0.53853

60 0.540174 0.542595 0.538308 0.538504

Band 2

442.559

0 0.540163 0.54264 0.538464 0.538781

15 0.540145 0.542591 0.538355 0.538604

30 0.540131 0.542555 0.538275 0.538478

45 0.540118 0.54252 0.538199 0.53836

60 0.540115 0.542511 0.538179 0.53833

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 92


Band 3

489.882

0 0.546129 0.547876 0.547394 0.542961

15 0.546105 0.547815 0.547255 0.542743

30 0.546089 0.547773 0.54716 0.542596

45 0.546073 0.547731 0.54707 0.54246

60 0.546069 0.547721 0.547047 0.542426

Band 4

509.819

0 0.546103 0.547845 0.547352 0.542909

15 0.546077 0.547779 0.547204 0.542677

30 0.546061 0.547735 0.547104 0.542524

45 0.546044 0.547692 0.54701 0.542381

60 0.54604 0.54768 0.546986 0.542346

Band 5

559.694

0 0.546054 0.547795 0.547298 0.54285

15 0.546026 0.547723 0.547139 0.542605

30 0.546008 0.547676 0.547036 0.542447

45 0.545991 0.547631 0.546938 0.5423

60 0.545986 0.54762 0.546914 0.542264

Band 6

619.601

0 0.550253 0.549149 0.547636 0.549279

15 0.550223 0.549074 0.547472 0.54902

30 0.550205 0.549028 0.547369 0.548857

45 0.550187 0.548983 0.54727 0.548705

60 0.550182 0.548972 0.547246 0.548669

Band 7

664.5731

0 0.550222 0.549119 0.547608 0.54925

15 0.550191 0.549042 0.547439 0.548986

30 0.550173 0.548996 0.547335 0.548822

45 0.550155 0.54895 0.547236 0.548669

60 0.55015 0.548939 0.547211 0.548632

Band 8

680.821

0 0.550212 0.54911 0.547598 0.549241

15 0.550181 0.549032 0.547429 0.548975

30 0.550163 0.548985 0.547325 0.548811

45 0.550145 0.548939 0.547224 0.548657

60 0.55014 0.548928 0.5472 0.548621

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 93


Band 9

708.329

0 0.555027 0.551287 0.552093 0.551221

15 0.554994 0.551207 0.55192 0.550956

30 0.554975 0.551161 0.551814 0.550792

45 0.554956 0.551114 0.551712 0.550639

60 0.554951 0.551102 0.551687 0.550603

Band 10

753.371

0 0.555003 0.551263 0.552068 0.551197

15 0.554969 0.551183 0.551895 0.550931

30 0.55495 0.551137 0.551789 0.550768

45 0.554931 0.55109 0.551688 0.550615

60 0.554926 0.551078 0.551663 0.550578

Band 12

778.4091

0 0.554991 0.551251 0.552056 0.551185

15 0.554957 0.551171 0.551883 0.550919

30 0.554938 0.551124 0.551777 0.550756

45 0.554919 0.551078 0.551676 0.550603

60 0.554914 0.551066 0.551651 0.550566

Band 13

864.876

0 0.554954 0.551214 0.552019 0.551148

15 0.55492 0.551134 0.551846 0.550882

30 0.554901 0.551088 0.55174 0.550719

45 0.554882 0.551041 0.551639 0.550566

60 0.554877 0.551029 0.551614 0.55053

Band 14

884.944

0 0.554946 0.551207 0.552012 0.55114

15 0.554913 0.551127 0.551839 0.550875

30 0.554894 0.55108 0.551733 0.550711

45 0.554874 0.551033 0.551631 0.550558

60 0.55487 0.551022 0.551606 0.550522

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 94


11. Combined References

AMC (1989). Robust Statistics - How not to reject outliers, Part 1: Basic concepts. Analyst 114: 1683-

1697.

Antoine, D., Chami, M., Claustre, H., d'Ortenzio, F., Morel, A., Becu, G., Gentilli, B., Louis, F., Ras, J.,

Roussier, E., Scott, A., Tailliez, D., Hooker, S. B., Guevel, P., Deste, J.-F., Dempsey, C.

&Adams, D. (2006).BOUSSOLE: A Joint CNRS-INSU, ESA, CNES, and NASA ocean color

calibration and validation activity. Greenbelt, MD.: NASA/GSFC.

Antoine, D., Guevel, P., Deste, J.-F., Becu, G., Louis, F., Scott, A. &Bardey, P. (2007). The

'BOUSSOLE' Buoy - A new transparent-to-swell taut mooring dedicated to marine optics:

Design, tests and performance at sea. Journal of Atmospheric and Oceanic Technology In Press.

Antoine, D., Ortenzio, F., Hooker, S. B., Bécu, G., Gentilli, B., Tailliez, D. &Scott, A. (2008).

Assessment of uncertainty in the ocean reflectance determined by three satellite ocean color

sensors (MERIS, SeaWiFS and MODIS-A) at an offshore site in the Mediterranean Sea

(BOUSSOLE project). Journal of Geophysical Research 113(C07013,

doi:10.1029/2007JC004472): 22.


Aspects of Oceanography, 317-344 (Eds N. G. Jerlov and E. Steemann-Nielsen). New York:

Academic Press, London.

Babin, M., Stramski, D., Ferrari, G. M., Claustre, H., Bricaud, A., Obolensky, G. &al., e. (2003).

Variations in the light absorption coefficients of phytoplankton, non algal particles, and dissolved

organic matter in coastal waters around europe. Journal of Geophysical Research 108: 3211 (doi:

3210.1029/2001JC000882).

Bailey, S. W. &Werdell, P. J. (2006). A multi-sensor approach for the on-orbit validation of ocean color

Satellite data products. Remote Sensing of the Environment 102: 12-23.

Barker, K., Mazeran, C., Lerebourg, C., Bouvet, M., Antoine, D., Ondrusek, M. E., Zibordi, G.

&Lavender, S. J. (2008).MERMAID: The MERis MAtchup In-situ Database. In 2nd MERIS

(A)ATSR Users WorkshopFrascati, Italy. September 2008.: ESA.

Bodhaine, B. A., Wood, N. B., Dutton, E. G. &Slusser, J. R. (1999). On Rayleigh optical depth

calculations. Journal of Atmospheric and Oceanic Technology 16: 1854-1861.

Boss, E. &Pegau, W. S. (2001). The relationship of light scattering at an angle in the backward direction

to the backscattering coefficient. Applied Optics 40: 5503-5507.

Breton, E., Rousseau, V. &Parent, J. Y. (2006). Hydroclimatic modulation of diatom/phaeocystis blooms

in nutrient-enriched Belgian coastal waters (North Sea). Limnology and Oceanography 100:

13321-13332.

Bricaud, A., Babin, M., Morel, A. &Claustre, H. (1995). Variability in the chlorophyll-specific

absorption-coefficients of natural phytoplankton - analysis and parameterization. Journal of


Brown, S. W., Flora, S., Feinholz, M., Yarborough, M. A., Houlihan, T., Peters, D., Kim, Y. S., Mueller,

J. L., Johnson, B. C. &Clark, D. K. (2007).The Marine Optical BuoY (MOBY) radiometric

calibration and uncertainty budget for ocean color satellite sensor vicarious calibration. In SPIE

Europe Remote SensingFlorence, Italy: SPIE Europe.

Clark, D. K., Gordon, H. R., Voss, K. J., Ge, Y., Broenkow, W. &Trees, C. (1997). Validation of

atmospheric correction over the oceans. Journal of Geophysical Research 102D: 17209-17217.

Clark, D. K., Yarborough, M. A., Feinholz, M. E., Flora, S., Broenkow, W., Kim, Y. S., Johnson, B. C.,

Brown, S. W., Yuen, M. &Mueller, J. L. (2003).MOBY, A radiometric buoy for performance

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 95


monitoring and vicarious calibration of satellite ocean colour sensors: measurements and data

analysis protocols. In Ocean Optics Protocols for Satellite Ocean Colour Sensor Validation,

NASA Technical Memo. 2003-211621/Rev4, VolVI, 3-34 (Eds J. L. Muller, G. Fargion and C.

McClain). Greenbelt, MD.: NASA/GSFC.

Cox, C. &Munk, W. (1954). Measurements of the roughness of the sea surface from photographs of the

sun's glitter. Journal of the Optical Society of America 44: 834-850.

Cristina, S., Goela, P., Icely, J. I., Newton, A. &Fragoso, B. (2009). Assessment of water-leaving

reflectance of the oceanic and coastal waters using MERIS satellite products off the southwest

coast of Portugal. Journal of Coastal Research Special Issue (56): 5.

D'Alimonte, D. &Zibordi, G. (2006). Statistical assessment of radiometric measurements from

autonomous systems. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING

44(3): 719-728.

Deschamps, P.-Y., Fougnie, B., Frouin, R., Lecomte, P. &Verwaerde (2004). SIMBAD: A field

radiometer for satellite ocean-color validation. Applied Optics 43(20): 4055-4069.

Deuzé, J.-L., Herman, M. &Santer, R. (1989). Fourier series expansion of the transfer equation in the

atmosphere–ocean system. Journal of Quantitative Spectroscopy and Radiative Transfer 41:

483-494.

Doerffer, R. &Schönfeld, W. (2009).Validation transect between Cuxhaven and Helgoland: GKSS

Technical Note (18th February 2009). GKSS, Germany.

Fougnie, B. &Deschamps, P.-Y. (1997).Observation et mode´lisation de la signature spectrale de

l‟e´cume de mer. In Proceedings of the 7th International Colloquium on Physical Measurements

and Signatures in Remote Sensing, Vol. 1, 227-234 (Eds G. Guyot and T. Phulpin). Rotterdam.

Fougnie, B., Frouin, R., Lecomte, P. &Deschamps, P.-Y. (1999). Reduction of skylight reflection effects

in the above-water measurements of diffuse marine reflectance. Applied Optics 38: 3844-6856.

Frouin, R., Longner, D. W., Gautier, C., Baker, K. S. &Smith, R. C. (1989). A simple analytical formula

to compute clear sky total and photosynthetically available solar irradiance at the ocean surface.

Journal of Geophysical Research 94: 9731-9742.

Frouin, R., Schwindling, M. &Deschamps, P.-Y. (1996). Spectral reflectance of sea foam in the visible

and near-infrared: insitu measurements and remote sensing implications. Journal of Geophysical

Research 101: 14361-14371.



Gordon, H. R. &McCluney, W. R. (1975). Estimation of depth of sunlight penetration in sea for remote-

sensing. Applied Optics 14: 413-416.

Gordon, H. R. &Morel, A. (1983). Remote assessment of ocean color for interpretation of satellite visible

imagery. A review.: Springer-Verlag, New York.

Gordon, H. R. &Wang, M. A. (1994). Retrieval of water-leaving radiances and aerosol optical thickness

over the oceans with SeaWiFS: A preliminary algorithm. Applied Optics 33(3): 443-452.



Hansen, J. E. &Travis, L. D. (1974). Light scattering in planetary atmospheres. Space Science Review 16:

527-610.

Høkedal, J. &Sørensen, K. (2007).Validation of MERIS-reflectance from ferries. In ENVISAT

SymposiumMontreux, Switzerland.

Holben, B., Eck, T. F., Slutsker, I., Tanré, D., Buis, J. P., Setzer, A., Vermonte, E., Reagan, J. A.,

Kauffman, Y. J., Nakajima, T., Lavenu, F., Janowiak, I. &Smirnov, A. (1999). AERONET - A

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 96


federated Instrument network and data archive for aerosol characterization. Remote Sensing of the

Environment 66: 1-16.

Holm-Hansen, O., Lorenzen, C. J., Holmes, R. W. &Strickland, J. d. H. (1965). Fluorometric

determination of chlorophyll. J. Cons Perm Int Expl Mer 39: 3-15.

Hommersom, A., Peters, S., Wernand, M. &de Boer, J. (2009). Spatial and temporal variability in bio-

optical properties of the Wadden Sea. Estuarine, Coastal and Shelf Science 83: 360-370.

Hooker, S. B., Lazin, G., Zibordi, G. &McClean, S. (2002). An evaluation of above- and in-water

methods for determining water leaving radiances. Journal of Atmospheric and Oceanic

Technology 19: 486-515.

Hooker, S. B. &Morel, A. (2003). Platform and environmental effects on above-water determinations of

water-leaving radiances. Journal of Atmospheric and Oceanic Technology 20: 187-205.

Jeffrey, S. W. &Humphrey, G. F. (1975). New spectrophotometric equations for determining chlorophylls

a, b, c1 and c2 in higher plants, algae and natural phytoplankton. Biochemie und Physiologie der

Pflanzen 167: 191-194.

Kahru, M. &Mitchell, B. G. (1998). Spectral reflectance and absorption of a massive red tide off Southern

California. Journal of Geophysical Research 103(21): 21601-21610.

Kahru, M. &Mitchell, B. G. (1999). Empirical chlorophyll algorithm and preliminary SeaWiFS validation

for the California Current. International Journal of Remote Sensing 20: 3423-3429.

Kasten, F. &Young, A. T. (1989). Revised optical air mass tables and approximation formula. Applied

Optics 28: 4735-4768.

Keating, G., Pitts, M. C. &Young, D.-F. (1989).Improved reference models for middle atmosphere ozone

_New CIRA_. In Middle Atmosphere Program Handbook for MAP, Vol. 31, 37-49 (Ed G.

Keating). Urbana, Illinois: Scientific Committee on Solar-Terrestrial Physics Secretariat,

University. of Illinois.

Kirk, J. T. O. (1994). Light and photosynthesis in aquatic ecosystems (Second Edition). Cambridge

University Press.

Kostadinov, T. S., Siegel, D. A., Maritorena, S. A. &Guillocheau, N. (2007). Ocean color observations

and modeling for an optically complex site: Santa Barbara Channel, California, USA. Journal of

Geophysical Research 112(C07011): doi: 10.1029/2006JC003526.

Kratzer, S., Brockmann, C. &Moore, G. F. (2008). Using MERIS full resolution data (300 m spatial

resolution) to monitor coastal waters– a case study from Himmerfjärden, a fjord-like bay in the

north-western Baltic Sea. Remote Sensing of the Environment 112(5): 2284-2300.

Leathers, R. A. &Downes, T. V. (2004). Self-shading correction for oceanographic upwelling

radiometers. Optics Express 12: 4709-4718.

Loisel, H., Mériaux, X., Poteau, A., Artigas, L. F., Lubac, B., Gardel, A., Caillaud, J. &Lesourd, S.

(2009). Analyze of the inherent optical properties of French Guiana coastal waters for remote

sensing applications. Journal of Coastal Research ICS Proceedings.



Maffione, R. A. &Dana, D. R. (1997). Instruments and methods for measuring the backward-scattering

coefficient of ocean waters. Applied Optics 36(24): 6057-6067.

McKee, D. (2008).Dataset [Bristol Channel and Irish Sea]. (Ed K. Barker).

Mitchell, B. G. &Kahru, M. (1998).Algorithms for SeaWIFS standard products developed with the

CALCOFI bio-optical data set. In CALCOFI Report, Vol. 39, 15pp.

Mobley, C. D. (1994). Light and water. Radiative transfer in natural waters.: Academic Press Inc.



MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 97


Moore, C., Zaneveld, J. R. V. &Kitchen, J. C. (1992).Preliminary results from an in-situ spectral

absorption meter. In SPIE Society of Optical Engineering, Vol. 1750, 330-337.

Moore, G. F., Icely, J. I. &Kratzer, S. (2011).Field intercomparison and validation of in-water radiometer

and sun photometers for MERIS validation. In ESA Living Planet Symposium, Special

Publication SP-686. In press.

Morel, A. (1974).Optical properties of pure seawater. In Optical Aspects of Oceanography., 1-24 (Eds N.

G. Jerlov and E. Steeman-Nielsen).

Morel, A. &Antoine, D. (1994). Heating rate within the upper ocean in relation to its bio-optical state.

Journal of Physical Oceanography 24: 1652-1665.

Morel, A. &Antoine, D. (1999).Pigment index retrieval in Case1 waters. In ENVISAT-MERIS Algorithm

Theoretical Basis Document 2.9, 25: European Space Agency.

Morel, A., Antoine, D. &Gentilli, B. (2002). Bidirectional reflectance of oceanic waters: accounting for

Raman emission and varying particle scattering phase function. Applied Optics 41(30): 6289-

6306.

Morel, A. &Gentilli, B. (1993). Diffuse reflectance of oceanic waters. 2. Bidirectional aspects. Applied

Optics 32: 6864-6872.

Morel, A. &Gentilli, B. (1996). Diffuse reflectance of oceanic waters. 3. Implications of bidirectionality

for the remote-sensing problem. Applied Optics 35: 4850-4862.

Morel, A. &Maritorena, S. (2001). Bio-optical properties of oceanic waters: A reappraisal. Journal of


Morel, A. &Prieur, L. (1977). Analysis of variations in ocean color. Limnology and Oceanography 22:

709-722.

Morel, A., Voss, K. J. &Gentilli, B. (1995). Bidirectional reflectance of oceanic waters: A comparison of

modeled and measured upward radiance fields. Journal of Geophysical Research 100: 13143-

13150.

Mueller, J. L. (2000).SeaWiFS algorithm for the diffuse attenuation coefficient, K(490), using water-

leaving radiances at 490 and 555 nm In SeaWiFS postlaunch calibration and validation analyses:

Part 3. NASA Technical Memorandum, Vol. 11, 24-27 (Eds S. B. Hooker and E. R. Firestone).

Greenbelt, Maryland, USA.: NASA Goddard Space Flight Center.

Mueller, J. L. (2003a).Chapter 6: Shadow corrections to in-water welled radiance measurements: A

review. In Ocean Optics Protocols For Satellite Ocean Color Sensor Validation, Vol. Revision 5,

32 (Eds J. L. Mueller, G. Fargion and C. McClain). Greenbelt, Maryland: NASA GSFC.

Mueller, J. L. (2003b).In-water radiometric profile measurements and data analysis protocols. Vol.


Mueller, J. L. &Austin, R. W. (1995).Ocean optics protocols for SeaWiFS validation, Revision 1. In



Mueller, J. L. &Austin, R. W. (2003).Characterisation of oceanographic and atmospheric radiometers.

Volume II: Instrument specifications, characterisation and calibration. In Ocean Optics Protocols

For Satellite Ocean Color Sensor Validation, Revision 4Greenbelt, MD: NASA/GSFC.

Mueller, J. L., Bidigare, R. R., Trees, C., Balch, W. M., Dore, J. &Drapeau, D. T. (2003a).Ocean optics

protocols for satellite ocean colour sensor validation, Revision 5, Volume V: Biogeochemical and

bio-opitcal measurements and data analysis protocols., 36 Greenbelt, MD: NASA/GSFC.

Mueller, J. L., Davies, C., Arnone, R., Frouin, R., Carder, K. L., Lee, J. P., Steward, R. G., Hooker, S. B.,

Mobley, C. D. &McLean, S. (2000).Above-water radiance and remote sensing reflectance

measurements and analysis protocols. In Ocean Optics Protocols for Satellite Ocean Cor Sensor

Validation. Rev 2.(Eds G. Fargion and J. L. Mueller). NASA.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 98


Mueller, J. L., Fargion, G. &McClain, C. (2003b).Ocean optics protocols for satellite ocean color sensor

validation, Revision 4. Vol. I - VII, 141 Greenbelt, Maryland: NASA.

Mueller, J. L., Morel, A., Frouin, R., Davies, C., Arnone, R. &Carder, K. L. (2003c).Ocean optics

protocols for satellite ocean color sensor validation, Revision 4, Volume III: Radiometric

measurements and data analysis protocols., 78pp. Greenbelt, MD: NASA/GSFC.

Neckel, H. &Labs, D. (1984). The solar radiation between 3300 and 12500 Å. Solar Physics 90: 205–258.

Nicolas, J.-M., Deschampes, P.-Y. &Frouin, R. (2001). Spectral reflectance of oceanic whitecaps in the

visible and near infrared: aircraft measurements over open ocean. Geophysical Research Letters

28: 4445–4448

O'Reilly, J., Maritorena, S. A., O'Brien, M. C., Siegel, D. A., Toole, D. &Menzies, D. (2000).Ocean color

chlorophyll a algorithms for SeaWiFS, OC2 and OC4: Version 4. . In SeaWiFS Postlaunch

Calibration and Validation Analysis:, 9-23 (Eds S. B. Hooker and E. R. Firestone). Greenbelt,

MD: NASA.

O'Reilly, J. E., Maritorena, S., Mitchell, B. G., Siegel, D. A., Carder, K. L., Garver, S. A., Kahru, M.

&McClain, C. R. (1998). Ocean color algorithms for SeaWiFS. Journal of Geophysical Research

103: 24,937-924,953.

Paltridge, G. W. &Platt, C. M. R. (1977).Radiative processes in meteorology and climatology. In

Development in Atmospheric Science New York: Eslevier.

Relvas, P. (1999).The physical oceanography of the Cape São Vicente upwelling region observed from

sea, land and space. In School of Ocean Sciences. Menai Bridge: University of North Wales,

Bangor.

Rousseau, V. (2000).Dynamics of phaeocystis and diatom blooms in the eutrophicated coastal waters of

the Southern Bight of the North Sea., Vol. PhDBruxelles: Université Libre de Bruxelles.

Ruddick, K. (2006). Seaborne measurements of near infrared water-leaving reflectance: The similarity

spectrum for turbid waters. Limnology and Oceanography 51(2): 1167-1179.

Ruddick, K., De Cauwer, V., Park, Y. &Moore, G. F. (2006a). Web Appendix 1. Data processing: Scan

selection and averaging. Limnology and Oceanography 51(2): 1167-1179.

Ruddick, K., De Cauwer, V., Park, Y. &Moore, G. F. (2006b). Web Appendix 2. Measurement

uncertainty analysis. Limnology and Oceanography 51(2): 1167-1179.

Satlantic (2007).Prosoft 7.7 User Manual. Vol. Revision E.

Shettle, E. P. &Fenn, R. W. (1979).Models for the aerosols of the lower atmosphere and the effects of

humidity variations on their optical properties. In Environment Research Papers, Vol. 676, 31

Massachusetts: Air Force Geophysics Laboratory, Hanscom AFB.

Siegel, D. A., O'Brian, M. C., Sorensen, J. C., Konnoff, D. A. &Fields, E. (1995).BBOP data processing

and sampling procedures., Vol. 19, 77 pp.

Smith, R. C. &Baker, K. S. (1984).The analysis of ocean optical data. In Ocean Optics VIISPIE, Vol.

478, 119-126 (Ed M. A. Blizard).

Strickland, J. D. H. &Parsons, T. R. (1972). A practical handbook of the sea water analysis. Fisheries

Research Board Canada Bulletin: 167-311.

Thuillier, G., Hersé, M., Labs, D., Foujols, T., Peetermans, W., D., G., Simon, P. C. &Mandel, H. (2003).

The solar spectral irradiance from 200 to 2400 nm as measured by the SOLSPEC spectrometer

from the ATLAS and EURECA missions. Solar Physics 214: 1-22.

Toole, D. &Siegel, D. A. (2001). Modes and mechanisms of ocean color variability in the Santa Barbara

Channel. Journal of Geophysical Research 106(C11): 26985-27000.

Toole, D., Siegel, D. A., Menzies, D., Neumann, M. J. &Smith, R. C. (2000). Remote-sensing reflectance

determinations in the coastal ocean environment: Impact of instrumental characteristics and

environmental variability. Applied Optics 39(3): 456-469.

MERIS

Optical

Measurement

Protocols




Issue : 2.0 Rev.: 1.0

Date : July 2011

PAGE : 99


Venrick, E. L. &Hayward, T. L. (1984).Determining chlorophyll on the 1984 CalCOFI surveys. In

California Coorperative Oceanic Fisheries Investigations Reports, Vol. 25, 74-79.

Werdell, P. J. (2005).An evaluation of Inherent Optical Property data for inclusion in the NASA bio-

Optical Marine Algorithm Data set. 6pp.: NASA Ocean Biology Processing Group.

Werdell, P. J. &Bailey, S. W. (2002).The SeaWiFS Bio-optical Archive and Storage System (SeaBASS):

Current architecture and implementation. Greenbelt, Maryland.: NASA Goddard Space Flight

Center.

Werdell, P. J. &Bailey, S. W. (2005). An improved in situ bio-optical data set for ocean colour algorithm

development and satellite data product validation. Remote Sensing of Environment 98: 122-140.


(2003). Unique data repository facilitates ocean color satellite validation. EOS Transactions

84(3): 379.

Wetlabs (2009).ac Meter Protocol Document. http://www.wetlabs.com/products/pub/ac9/acproto.pdf.

Zaneveld, J. R. V., Boss, E. &Barnard, A. (2001). Influence of surface waves on measured and modeled

irradiance profiles. Applied Optics 40: 1442-1449.

Zibordi, G., Berthon, J.-F., Mélin, F., D'Alimonte, D. &Kaitala, S. (2009a). Validation of satellite ocean

color primary products at optically complex coastal sites: Northern Adriatic Sea, Northern Baltic

Proper and Gulf of Finland. Remote Sensing of the Environment doi:10.1016/j.rse.2009.07.013:

18.

Zibordi, G., Doyle, J. P. &Hooker, S. B. (1999). Offshore tower shading effects on in-water optical

measurements. Journal of Atmospheric and Oceanic Technology 16: 1767-1779.

Zibordi, G. &Ferrari, G. M. (1995). Instrument self-shading in underwater optical measurements:

experimental data. Applied Optics 34: 2750-2754.

Zibordi, G., Holben, B., Slutsker, I., Giles, D., D'Alimonte, D., Mélin, F., Berthon, J.-F., Vandemark, D.,

Feng, H., Schuster, G., Fabbri, B. E., Kaitala, S. &Seppälä, J. (2009b). AERONET-OC: a

network for the validation of Ocean Color primary radiometric products. Journal of Oceanic and

Atmospheric Technology (Accepted): 57.

Zibordi, G., Hooker, S. B., Berthon, J.-F. &D'Alimonte, D. (2002). Autonomous above water radiance

measurements from stable platforms. Journal of Atmospheric and Oceanic Technology 19: 808-

819.

Zibordi, G., Mélin, F. &Berthon, J.-F. (2006). Comparison of SeaWiFS, MODIS and MERIS radiometric

products at a coastal site. . Geophysical Research Letters 33: L06617.

Zibordi, G., Mélin, F., Hooker, S. B., D'Alimonte, D. &Holben, B. (2004). An autonomous above-water

system for the validation of ocean color radiance data. IEEE TRANSACTIONS ON

GEOSCIENCE AND REMOTE SENSING 42(2): 401-415.

http://www.wetlabs.com/products/pub/ac9/acproto.pdf