Measurements of light transmission in deep sea with the

Nuclear Instruments and Methods in Physics Research A 487 (2002) 423–434

Measurements of light transmission in deep sea with the AC9trasmissometer

A. Caponea,b, T. Digaetanoc, A. Grimaldid, R. Habele, D. Lo Prestid,E. Mignecof,g, R. Masulloa, F. Moroc, M. Petruccettib, C. Pettag, P. Piattellif,

N. Randazzod, G. Riccobenef,*, E. Salustib, P. Sapienzaf, M. Seditaf, L. Trasattih,L. Ursellac

aDipartimento di Fisica Universit !a La Sapienza, P.le A. Moro 2, 00185 Roma, Italyb INFN Sezione Roma-1, P.le A. Moro 2, 00185 Roma, Italy

cDipartimento di Oceanografia Fisica, Osservatorio Geofisico Sperimentale, Borgo Grotta Gigante 42C, 34016 Sgonico (TS), Italyd INFN Sezione Catania, Corso Italia 57, 95129 Catania, Italy

eDipartimento di Fisica Universit !a di Cagliari and Seione INFN Cagliari 09042, Monserrato (CA), ItalyfLaboratori Nazionali del Sud INFN, Via S.Sofia 44, 95123 Catania, Italy

gDipartimento di Fisica e Astronomia Universit !a di Catania, ItalyhLaboratori Nazionali di Frascati INFN, Via Enrico Fermi 40, 00044 Frascati (RM), Italy

Received 18 September 2001; received in revised form 14 November 2001; accepted 15 November 2001

Abstract

The NEMO Collaboration aims to construct an underwater Cherenkov detector in the Mediterranean Sea, able to

act as a neutrino telescope. One of the main tasks of this project, which implies difficult technological challenges, is the

selection of an adequate marine site. In this framework the knowledge of light transmission properties in deep seawater

is extremely important. The collaboration has measured optical properties in several marine sites near the Italian coasts,

at depths > 3000 m; using a setup based on a AC9, a commercial trasmissometer, manufactured by WETLabs. The

results obtained for the two sites reported in this paper (Alicudi and Ustica), show that deep seawater optical properties

are comparable to those of the clearest waters. r 2002 Elsevier Science B.V. All rights reserved.

PACS: 95.55.Vj; 29.40.Ka; 92.10.Pt; 07.88.+y

Keywords: Neutrino telescope; NEMO; Attenuation; Absorption; Deep sea

1. Overview

The observation of Ultra High Energy CosmicRays (UHECR) with energy higher than 1020 eV

has attracted the attention of the astrophysics andparticle-physics community on the most energeticphenomena taking place in the Universe. It issupposed that such energetic particles are acceler-ated in extra-galactic sources.

Gamma ray sources with energy up to tens ofTeV have also been observed. If high energyphotons are generated through the production and

*Corresponding author. Tel.: +39-095-542271; fax: +39-

095-714815.

E-mail address: [email protected] (G. Riccobene).

0168-9002/02/$ - see front matter r 2002 Elsevier Science B.V. All rights reserved.

PII: S 0 1 6 8 - 9 0 0 2 ( 0 1 ) 0 2 1 9 4 - 5

decay of neutral pions, it is reasonable to expect,from the same sources, an associated flux ofhigh energy neutrinos, generated through theproduction and decay of charged pions. Alongtheir journey in the universe, most part ofthe electromagnetic and hadronic emission isdeflected or absorbed by the electromagneticbackground and by the intergalactic and inter-stellar matter. Neutrinos, on the contrary, are notsignificantly absorbed by the intergalactic mediumand are not deflected by the intergalactic magneticfields. Already in 1960 Markov [1,2] proposed touse seawater as a huge target to detect UHEneutrinos, looking at their charged current weakinteractions. The outgoing lepton generates, alongits path in seawater, Cherenkov light that can bedetected by a lattice of optical sensors. Thereconstruction of the muon track, and thus ofthe neutrino direction, offers the possibility toidentify the neutrino sources opening the newexciting field of neutrino astronomy. The observa-tion of high energy neutrino fluxes expected fromastrophysical sources requires a detector with aneffective area close to 106 m2 instrumented along adistance comparable to the range in water (B km)of the high energy muons (B1000 TeV). Theidentified neutrino sources identified could becatalogued in the sky map and eventually com-pared with the known gamma sources. Theconstruction of a detector of such dimensions,usually called a km3 Neutrino Telescope, is one ofthe main challenges of astroparticle physics today.The Mediterranean Sea offers optimal conditions,on a worldwide scale, to locate an underwaterneutrino telescope. The choice of the km3 scaleneutrino telescope location is such an importanttask that careful studies of candidate sites must becarried out in order to identify the most suitableone. Along the Italian coasts several sites exist, atdepth 3300–3500 m; that are potentially interestingto host an undersea neutrino telescope. In thesesites we have studied deep seawater opticalproperties (absorption and attenuation) and en-vironmental properties: water temperature andsalinity, biological activity, water currents, sedi-mentation. In this paper we report light transmis-sion measurements carried out in two sites namedUstica (during November 1999) and Alicudi (on

December 1999), in the Southern Tyrrhenian Sea,located at:

* 391050 N 131200 E; North-East of Ustica is-land;

* 391050 N 141200 E; North of Alicudi island.

2. Optical properties of deep sea

Water transparency to electromagnetic radia-tion can be characterized by means of quantitativeparameters: the absorption length La and thescattering length Lb: Each length represents thepath after which a photon beam of intensity I0 atwavelength l; travelling along the emission direc-tion, is reduced to 1=e by absorption or diffusionphenomena. These quantities can be directlyderived by the simple relation:

Iðx; lÞ ¼ I0ðlÞe�x=LðlÞ; ð1Þ

where x is the optical path traversed by the beamand I0 the source intensity. In literature absorption(a ¼ 1=La) and scattering (b ¼ 1=Lb) coefficientsare extensively used to characterize the lighttransmission in matter as well as the attenuationcoefficient (c) defined as:

cðlÞ ¼ aðlÞ þ bðlÞ: ð2Þ

The main cause of light absorption in water isexcitation of vibrational states of the watermolecule by photons [3–5]: due to such processthe photon energy is entirely deposited in thetraversed medium. Scattering refers to processes inwhich the direction of the photon is changedwithout any other alteration. Scattering phenom-ena in which the photon wavelength changes (e.g.Raman effect) happen less frequently. Scatteringcan take place either on molecules (Rayleighscattering) or on dissolved particulate (Mie scat-tering).

In pure water, light absorption and scatteringare strongly wavelength dependent. In particularlight transmission in pure water is extremelyfavored in the range 350–550 nm; overlappingthe region in which PMTs usually reach thehighest quantum efficiency. In the visible regionof the electromagnetic spectrum light absorption

A. Capone et al. / Nuclear Instruments and Methods in Physics Research A 487 (2002) 423–434424

steeply decreases as a function of wavelength andreaches its minimum at about 420 nm (see Fig. 1).This is the reason why seawater has a blue-greencolor.

The optical properties of natural seawaterare functions of water salinity, water temperatureand of the concentration of dissolved organicand inorganic matter. Light absorption anddiffusion in water as a function of salinityand temperature have been extensively studied[7]. It has been noticed that, for lX400 nm;the dependence of scattering coefficient on tem-perature and salinity is negligible while thevariation of the absorption coefficient is signifi-cant, in particular at l > 710 nm (for details seeSection 7). The seawater diffusion and absorptioncoefficients can be parameterized as the sum of aterm due to optically pure water (i.e. water withoutdissolved particulate) at defined conditions oftemperature and salinity (aT;SW ; bT;SW ), and a termthat accounts for interaction of light with particu-late (ap; bp):

aSWðlÞ ¼ aT;SW ðlÞ þ apðlÞ ð3Þ

bSWðlÞ ¼ bT;SW ðlÞ þ bpðlÞ: ð4Þ

Optical measurements of deep seawater haveshown that the presence of particulate has anegligible effect on light absorption but it enlargesthe light diffusion coefficient. Since water tem-

perature and salinity and particulate concentrationmay vary significantly in different marine sites it isextremely important to measure optical para-meters in situ.

3. The AC9 trasmissometer

We performed attenuation and absorptionmeasurements of light in deep seawater by meansof a setup based on a trasmissometer: the AC9manufactured by WETLabs [8]. The device com-pactness (68 cm height �10:2 cm diameter) and itspressure resistance (it can operate down to 6000 mdepth) are excellent for our purposes. The AC9performs attenuation and absorption measure-ments independently using two different lightpaths and spanning the light spectrum over ninedifferent wavelengths (412, 440, 488, 510, 532, 555,650, 676, 715 nm). In our measurements we obtainan accuracy in aðlÞ and cðlÞ of about 1:5�10�3 m�1:

The AC9 attenuation and absorption measure-ment technique is based on the Lambert’s law ofcollimated beams (see Eq. (1)) where x is the beampath-length, I0ðlÞ is the intensity of the collimatedprimary beam, at a given wavelength l; andIðx; lÞ ¼ Ia;cðx; lÞ is the beam intensity measuredat distance x; as a result of absorption orattenuation effect, respectively.

10-2

10-1

1

10

10 2

10 3

10 4

10 5

10 6

10 7

10-13

10-12

10-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

1 10 102

103

sea water

pure water

wavelength (m)

a (1

/m)

Fig. 1. Absorption coefficient of electromagnetic waves for pure and seawater as function of wavelength. Data taken from Mobley [6].

A. Capone et al. / Nuclear Instruments and Methods in Physics Research A 487 (2002) 423–434 425

In order to produce collimated monochromaticlight beams, the instrument is equipped with anincandescence lamp and a set of collimators andnine monochromatic (DlB10 nm) filters. Twodifferent beams are available at the same time forindependent measurements of attenuation andabsorption. Each beam is split in two parts by amirror: the reflected one reaches a reference siliconphoton detector. The refracted one crosses aquartz window and enters inside a 25 cm longpipe. During deep sea measurements seawater fillsthe pipes (flow tubes). The flow tube used forattenuation measurements has a black inner sur-face in order to absorb all photons scattered byseawater. A collimated silicon photon detector(angular acceptance B0:71) is placed at the end ofthe path, along the source axis. Thanks to thisstrongly collimated layout the end-path detectorreceives only photons which have not interacted(neither absorbed nor scattered). The referencedetector measures the source intensity I0ðlÞ; theend-path detector measures the attenuated beamintensity Icðx; lÞ; x is the known beam path inwater (0:25 m). The attenuation coefficient istherefore calculated as

cðlÞ ¼1

xln

I0ðlÞIcðx; lÞ

: ð5Þ

In the absorption channel, the inner surface ofthe absorption flow tube behaves like a cylindricalmirror. The light scattered by seawater is reflectedand redirected towards a wide angular acceptancesilicon photon detector. In first approximation allscattered photons are detected and the ratiobetween the intensities I0ðlÞ and Iaðx; lÞ is only afunction of the seawater absorption coefficientaðlÞ:

Using AC9 data, the scattering coefficient canbe calculated by subtracting the absorption valuefrom the attenuation value at each given wave-length (see Eq. (2)).

4. AC9 measurements principles

In the interval of wavelength interesting for aCherenkov neutrino telescope, l ¼ 350–550 nm;the expected values of absorption and attenuation

coefficients in deep seawater are B10�2 m�1;very close to the pure water ones. This impliesthat the instrument should have sensitivityand accuracy of the order of 1–2� 10�3 m�1:The calibration of the instrument plays themost important role in determining theaccuracy in measurements. In the above definedwavelength range, pure water optical propertieshave been extensively measured, therefore purewater can be assumed as a reference medium[10,11].

Instrumental effects—such as the status ofoptical windows, of the electronics, etc.—can alsobe studied filling the flow tubes with a mediumwith a negligible light absorption and attenuation(e.g. dry air or N2).

The instrument calibration can be, thus, per-formed and tested any time filling the flow tubeswith a medium with known optical properties:either pure water (pure water calibration) or N2 (aircalibration). Filling the flow tubes with pure waterwe measure, for example in the absorptionchannel, the values:

aðlÞ ¼ aIðlÞ þ aPWðlÞ ð6Þ

and, in the case of N2;

aðlÞ ¼ aIðlÞ þ aN2ðlÞ: ð7Þ

The extra-term aIðlÞ takes into account the lightabsorption in the instrument optics (that isfunction of the status of quartz windows andmirror surfaces) and all other instrumental effects(see Section 7). This means that this term can varywith time and can be a function of the internalelectronics temperature. The same argument isvalid for the attenuation channel.

With pure water inside the flow tubes themeasurement of AC9 can be set equal to theknown values of aPWðlÞ and cPWðlÞ: The result ofthe water calibration procedure is a set of 18calibration constants (for the nine absorption andattenuation channels) that represent the workingstatus of AC9. The AC9 internal software sub-tracts these coefficients to the actual reading of theinstrument such that each AC9 output value, whenreference water fills the flow tubes, should be equalto zero.


Filling the flow-tubes with sea water wemeasure:

aðlÞ ¼ aIðlÞ þ aSWðlÞ ð8Þ

cðlÞ ¼ cIðlÞ þ cSWðlÞ: ð9Þ

Then the values of AC9 output corresponding tothe case of deep sea water filling the flow tubes are:

DaSWðlÞ ¼ aSWðlÞ � aPWðlÞ ð10Þ

DcSWðlÞ ¼ cSWðlÞ � cPWðlÞ: ð11Þ

5. AC9 calibration procedure

The AC9 manufacturer (WETLabs) providesthe instrument calibration performed with de-ionized and de-gassed pure water, at giventemperature (251C), as referenced in Refs.[10,11]. The optical properties of this medium, atthe nine wavelengths relevant for the AC9, arelisted in Table 1. Wetlabs provides also the resultsof the instrument calibration performed with dryair. The set of constants that relate the water

calibration values to the air calibration ones areprovided by WETLabs.

In principle, in order to test, from time to time,the validity of the used set of water calibration

constants, the user should check the AC9 responseafter filling the flow tubes with pure water inreference conditions. However, since pure water isnot easily available during cruises, we check thecalibration of the AC9 in dry air testing thevalidity of air calibration constants after havingfilled the flow tubes with high purity grade N2: Weextensively perform these operations during navalcampaigns before every deployment.

We have accurately studied the dependence ofair calibration constants as a function of AC9internal temperature. We noticed that during

measurement in the Mediterranean Sea, where atdepth > 1500 m the water temperature is B13–141C; the AC9 internal temperature stabilizes atTAC9B22:51C: To reduce the systematic error inthe knowledge of the instrument calibrationconstants, during checks of air calibration we keepthe internal AC9 temperature at B22:51C bymeans of a refrigerator. Fig. 2 shows some AC9measured raw values DaN2

ðlÞ;DcN2ðlÞ (analogous

to DaSWðlÞ and DcSWðlÞ as defined in Eqs. (10) and(11)), as a function of TAC9; obtained checking theair calibration just before the first deployment inAlicudi site. In the range 221CoTAC9o231C theDaN2

ðlÞ; DcN2ðlÞ average values are close to zero

and RMS are of the order of 1:5� 10�3 m�1:These average values (that we call aoffN2

ðlÞ andcoffN2

ðlÞÞ are used as offsets and subtracted duringoff-line data analysis, as described in the followingsections.

6. Deep sea setup

During deep sea measurements the AC9 isconnected to a standard oceanographic CTD(Conductivity Temperature Depth) probe, theOcean MK-317 manufactured by IDRONAUT. Apump is used to ensure re-circulation of seawaterinside the AC9 pipes. The AC9 and the pump arepowered by a 14 V battery pack. In Fig. 3 we showthe whole setup mounted on an AISI-316 stainless-steel cage before a deployment.

When the system is in operation, the RS-232

stream of the AC9 data is converted into FSK

stream by a modem card placed inside the CTD.Data are sent to sea surface through an electro-mechanical cable, that is also used to transmitpower to CTD (B1 A at 30 VDC). The dataacquisition system permits both the AC9 datatelemetry and the data storage on a PC onboard

Table 1

Absorption and attenuation coefficients (in units of 10�3) of pure water at T ¼ 251 referenced in the AC9 manual [8] and [9]

l (nm) 412 440 488 510 532 555 650 676 715

a ð10�3 m�1) 5.4 8.3 17.7 38.2 51.6 69.0 359.4 441.6 1049.2

c ð10�3 m�1) 9.7 11.9 20.0 40.2 53.3 70.4 360.1 442.2 1049.7


the ship. The CTD data (depth, water salinity andtemperature) are recorded on a local memory.Both instruments also record an absolute timeinformation, which is used to couple the AC9 andthe CTD data during off-line analysis. Thisprocedure allows to relate water optical propertiesto depth, water salinity and temperature. TheCTD-AC9 acquisition program gives about sixmeasurements per second, usually we deploy thesetup at speed of 1 m=s: In Fig. 4 we show, asfunction of depth, water salinity and temperaturetogether with the final absolute values of absorp-tion and attenuation coefficient for l ¼ 440 nmmeasured during the first (black dots) and thesecond (red dots) deployments in Alicudi site. As itappears in the figure, the layer composition ofTyrrhenian Sea, well studied by oceanologists interms of salinity and temperature, is also indicatedby the measurements of water optical properties:the AC9 sensitivity permits to distinguish layers of

water where absorption and attenuation coeffi-cients vary for B1� 10�3 m�1:

7. Data analysis

In order to obtain the values of deep seawaterabsorption and attenuation coefficients from themeasured raw values we need to apply fewcorrections. The first correction consists in remov-ing the set of calibration constants aoffN2

ðlÞ; coffN2ðlÞ

described in Section 5

Da0ðlÞ ¼ DarawðlÞ � aoffN2ðlÞ ð12Þ

Dc0ðlÞ ¼ DcrawðlÞ � coffN2ðlÞ: ð13Þ

Further corrections for the attenuation channelcould be needed to take into account that thesilicon photon detector in the attenuation channelhas a finite angular acceptance (0:71) and that the

-0.01

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0.008

0.01

20 20.5 21 21.5 22 22.5 23 23.5 24

-0.01

-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0.008

0.01

20 20.5 21 21.5 22 22.5 23 23.5 24

Fig. 2. Absorption DaN2ðlÞ and attenuation DcN2

ðlÞ raw values for l ¼ 412; 676 nm; measured, as function of the AC9 internal

temperature, during a dry air calibration before the first measurement in Alicudi site.


inner surface of the attenuation flow tube does notbehave as a perfect absorber. These two correc-tions have been evaluated to be much lower than1:5� 10�3 m�1 that we quote as systematic errorassociated to the result. Therefore DccorrðlÞ ¼Dc0ðlÞ:

The absolute values of the light attenuationcoefficients in seawater (as a function of depth) canbe finally obtained inverting Eq. (11).

Concerning the absorption channel, up to now,we considered that the inner surface of the flow

tube is perfectly reflecting. This assumption is validonly in first approximation and a proper correc-tion has to be applied to the measured raw values.If the inner mirror is not perfectly reflecting, inpresence of light scattering, a fraction of thediffused photons do not reach the end-pathdetector. We now illustrate how we have evaluatedthe amount of this effect using AC9 data collectedfor l > 650 nm:

Photon diffusion in the absorption channel isalso present when the tube is filled with pure water.

This implies that, with the described calibrationprocedure, part of the effect is already accountedfor at the calibration time, i.e. the effect of photonsdiffused at large angle by Rayleigh scattering onmolecules.

The presence of particulate in deep seawaterresults as an additional cause of absorptionand diffusion processes; but, at red and infraredwavelengths, the absorption due to the particulatepresent in deep seawater is negligible [12].It follows that the values Da0ðlÞ at l ¼ 676and 715 nm; measured by the AC9 in deepseawater, allow us to evaluate the effect of thenot perfect reflectivity of the absorption flow tube

mirror.Actually a value of Da0ðlÞa0 for seawater has to

be expected because of the presence of salts andbecause deep seawater temperature (B13–141C inthe deep Mediterranean Sea) is not equal to thecalibration water temperature (251C). The depen-dence of light absorption and diffusion in water asa function of salinity and temperature has beenextensively studied [7]. It has been noticed that, forl > 400 nm; the dependence of bT;SW ðlÞ on tempera-ture and salinity is negligible; on the contrary thevariation of the absorption coefficient can beexpressed by the equation:

DaT;SðlÞ ¼ ½CT � ðT � Tref Þ þCS � ðS � Sref Þ� ð14Þ

where Tref ¼ 251C; Sref ¼ 0 practical salinity units(p.s.u.), T and S are the actual values of seawater.The constants CT and CS are known as a functionof the wavelength [7,12]: for l ¼ 676 nm the valuesare CTð676Þ ¼ 1� 10�4 m�1

1C�1 and CSð676Þ ¼8� 10�5 m�1 p:s:u:�1: The slope of the tempera-ture corrections for l ¼ 715 nm is much larger:CTð715Þ ¼ 2:9� 10�3 m�1

1C�1 (while CSð715Þ ¼�8� 10�5 m�1 p:s:u:�1).

We evaluate the correction due to the internalmirror of the absorption flow tube only at l ¼676 nm since the uncertainty on the temperaturecorrection at this wavelength is smaller.

The evaluation of the contribution to the valueof Da0ð676Þ due to the photons diffused byparticulate and not reflected toward the end-pathdetector is determined by means of the equation:

Damirrorð676Þ ¼ Da0ð676Þ � DaT;Sð676Þ: ð15Þ

Fig. 3. Deployment of AC9 deep-sea setup.


The measurement of SSW and TSW; made by theCTD (see Section 6), allows to evaluate thecorrection Damirrorð676Þ as a function of depth.

It has been suggested [13] that the shape andmagnitude of the Mie volume scattering function,in first approximation, can be considered almostindependent on wavelength for the interval of l inwhich the AC9 operates: the correction due to themirror effect is, therefore, independent on wave-length. It turns out that Damirrorð676Þ can be usedto correct the measured values of absorptioncoefficients for all wavelengths. Applying thiscorrection we obtain:

DacorrðlÞ ¼ Da0ðlÞ � Damirrorð676Þ: ð16Þ

Finally, adding the pure water absorption (aPW)and attenuation (cPW) coefficients to the obtainedDacorr and DccorrðlÞ; we evaluate (as a function of

depth) the seawater inherent optical properties:

aSWðlÞ ¼ DacorrðlÞ þ aPWðlÞ ð17Þ

cSWðlÞ ¼ DccorrðlÞ þ cPWðlÞ: ð18Þ

Fig. 5, that refers to the second measurement inAlicudi site, illustrates the analysis procedure forl ¼ 412 nm: As function of depth, we show theraw measured values of the absorption andattenuation coefficients (black dots), the valuesobtained applying the offset correction (red dots)and the mirror correction (blue dots) for aSWð412Þand cSWð412Þ:

The same analysis has been applied to the valuesmeasured at l ¼ 440; 488, 510, 532, 650, 676 and715 nm: The values of aSWðlÞ and cSWðlÞ measuredin the four deployments carried out in Alicudi andUstica show very good agreement.

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

13.5 14T (C)

dept

h (m

)

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

38.5 38.75

S (psu)

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

0.02 0.04 0.06

c 440nm (1/m)

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

0.02 0.04

a 440nm (1/m)

Fig. 4. Temperature, salinity, attenuation and absorption coefficients (at l ¼ 440 nm) as a function of depth, measured in the first

(black) and in the second (red) deployment in Alicudi site. The values measured in the two deployments are nearly superimposed in the

figure.


We do not show results for l ¼ 555 nm due to atemporary hardware problem happened to theinterferometric filter during the naval campaign.

In the following we will quantify the opticalproperties of deep seawater averaging the absorp-tion and attenuation coefficients in the range ofdepth interesting for a km3 neutrino detector: a400 m wide interval, with its base B150 m abovethe seabed.

In Fig. 6 we show, as an example, the distribu-tion of aSWð412Þ and cSWð412Þ values, averaged foreach meter, in the interval of depth 2850–3250 m;related to the deployments in Ustica and Alicudi

(seabed depth B3400 m for both sites).In Table 2 we list the average values of the

distributions of DacorrðlÞ and DccorrðlÞ in the sameinterval of depth.

The statistical errors associated to these valuesare evaluated from the RMS of the distributions(see, for example, Fig. 6). The systematic errors,mainly due to the accuracy of the calibration check

procedure, have been evaluated to be equal to1:5� 10�3: The absolute (aSWðlÞ; cSWðlÞ) valuescan be obtained adding the values of attenuationand absorption of the reference water (seeEqs. (17) and (18) and Table 1).

Finally, we present in Fig. 7 the values ofabsorption and attenuation lengths, as a functionof wavelength, for the same interval of depth.

8. Discussion

The importance of measuring seawater opticalproperties in situ has been discussed by severalauthors. The setup we used permits to evaluateseawater light absorption and attenuation coeffi-cient as a function of depth and wavelength (in therange 412–715 nm). Accurate calibration checksallow us to obtain an accuracy in the evaluation ofaðlÞ and cðlÞ of the order of 1:5� 10�3 m�1:

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

0.01 0.02 0.03 0.04

raw valueoffset correctionmirror correction

a412 (1/m)

dept

h (m)

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

0.02 0.03 0.04 0.05 0.06 0.07

raw valueoffset correction

c412 (1/m)

Fig. 5. Raw values (black dots) of the absorption and attenuation coefficients at 412 nm: The figure also shows the values obtained

after offset correction (red dots) and, only in the case of absorption, the values after mirror correction (blue dots).


Table 2

Values of Dacorr and Dccorr measured in the interval of depth 2850–3250 m in Alicudi and Ustica during November–December 1999.

Negative values, when statistically significant, are due to the dependence of water optical properties on salinity and temperature

Coefficient Alicudi-1 Alicudi-2 Ustica-1 Ustica-2

�10�3 ðm�1Þ �10�3 ðm�1Þ �10�3 ðm�1Þ �10�3 ðm�1Þ

a412 12:370:6 12:470:5 17:870:8 14:270:5c412 23:671:8 23:371:7 28:571:7 22:872:0a440 10:170:5 9:970:5 13:270:6 10:970:5c440 18:171:8 17:871:7 19:871:8 17:771:9a488 4:470:4 4:370:4 5:570:5 4:570:5c488 13:871:8 13:771:7 14:171:6 14:271:6a510 �0:770:4 �0:670:3 0:470:5 �0:270:5c510 0:671:9 0:671:7 0:471:6 1:771:6a532 0:870:4 1:270:3 1:370:5 1:270:4c532 �3:171:9 �4:071:7 �3:771:6 �1:871:8a650 �3:170:4 �2:570:5 �2:670:4 �1:870:4c650 15:771:9 18:471:9 16:271:5 17:271:7a676 070:5 070:5 070:5 070:6c676 3:771:8 2:971:6 5:671:1 �0:971:6a715 �33:71:0 �32:71:0 �33:71:0 �33:71:0c715 �33:72:0 �33:71:6 �33:71:2 �33:71:5

0

5

10

15

20

25

30

0.015 0.02 0.025 0.030

5

10

15

20

25

30

0.015 0.02 0.025 0.03

0

2

4

6

8

10

12

14

16

18

20

0.025 0.03 0.035 0.04 0.0450

2

4

6

8

10

12

14

16

18

20

0.025 0.03 0.035 0.04 0.045

Fig. 6. Distributions of the absorption and attenuation coefficients, at l ¼ 412 nm; measured in Alicudi and Ustica in the depth

interval 2850–3250 m: The results for a412 and c412 for the two measurements in Alicudi are in excellent agreement and appear nearly

superimposed in the figure.


The values of aSWðlÞ measured in the depthinterval of interest in Alicudi and Ustica are veryclose to the ones reported by Smith and Baker aspure seawater absorption [14]. The discrepancy isless than B5� 10�3 m�1 at all wavelengths,except at 715 nm where the temperature effect isrelevant. Average absorption length for blue light(l ¼ 412 and 440 nm) is B50–55 m; the averageattenuation length is B30 m: These values areextremely good when compared to publishedseawater attenuation values obtained in conditionsof collimated beam and detector geometry [3]. Themeasured blue light attenuation length value isvery close to the ones measured by Khanaev andKuleshov [15] in the NESTOR site [16].

On the contrary, our results cannot be com-pared to the ones published by Bradner et al. [17]for the DUMAND project, Anassontzis et al. [18]for NESTOR and the ones measured by the

ANTARES collaboration [19]. These measure-ments were, indeed, carried out in conditions ofnot collimated geometry and the measured value isa quantity usually called effective attenuation

coefficient ceff ðlÞ: This quantity (an apparent

optical property) is defined as the sum of absorp-tion and only a fraction of the scattering coeffi-cient:

ceff ðlÞ ¼ aðlÞ þ ð1�/cosðWÞSÞ � bðlÞ ð19Þ

where /cos ðWÞS is the average cosine of thevolume scattering function distribution. Thisquantity strongly depends on the amount anddimension of the dissolved particulate. Measure-ments carried in ocean [6] gives/cosðWÞSB0:95: ceff ðlÞ for NEMO explored siteswill be evaluated only after an in situ precisemeasurement of the volume scattering functionscheduled for year 2001.

0.91

2

3

4

5

6789

10

20

30

40

50

6070

400 500 600 700

Alicudi 1Alicudi 2Ustica 1Ustica 2

wavelength (nm)

Abs

orpt

ion

leng

th (

m)

0.91

2

3

4

5

6

789

10

20

30

40

400 500 600 700

Alicudi 1Alicudi 2Ustica 1Ustica 2

wavelength (nm)

Att

enua

tion

leng

th (

m)

Fig. 7. Absorption and attenuation lengths as function of wavelength measured in Alicudi and Ustica; in the depth interval 2850–

3250 m: The error bars include the systematic error.


Acknowledgements

The NEMO collaboration wants to thank M.Astraldi, G.P. Gasparini (Istituto di OceanografiaFisica—CNR, La Spezia) and E. Accerboni, G.Gelsi, B. Manca, R. Mosetti (Istituto Nazionale diOceanografia e Geofisica Sperimentale, Trieste),M. Leonardi (Istituto Sperimentale Talassografi-co—CNR, Messina) C. Viezzoli (SOPROMAR)for the fruitful collaboration. We also want tothank Captains V. Lubrano and M. Gentile,officers and crew of the Urania OceanographicResearch Vessel, for their outstanding experienceand professionalism shown during the navalcampaign.

References

[1] M.A. Markov, Proceedings of the 10th International

Conference on High Energy Physics, Rochester, 1960.

[2] M.A. Markov, I.M. Zheleznykh, Nucl. Phys. 27 (1961)

385.

[3] S.Q. Duntley, J. Opt. Soc. Am. 53 (1963) 214.

[4] S.G. Warren, Appl. Opt. 23 (1984) 1206.

[5] C.L. Braun, S.N. Smirnov, J. Chem. Edu. 70 (1993) 612.

[6] C.D. Mobley, Light and Water, Academic Press, San

Diego, 1994.

[7] W.S. Pegau, D. Gray, J.R.V. Zaneveld, Appl. Opt. 36

(1997) 6035.

[8] WETLabs, AC9 manual in www.wetlabs.com.

[9] R.M. Pope, E.S. Fry, Appl. Opt. 36 (1997) 33.

[10] L. Kou, D. Labrie, P. Chylek, Appl. Opt. 32 (1993) 3531.

[11] R.M. Pope, Optical absorption of pure water and sea

water using the integrating cavity absorption meter, Ph.D.

Thesis, Texas A& M, College Station, 1993.

[12] M.S. Twardowski, et al., J. Atm. Ocean. Tech. 16 (1999)

691.

[13] J.R.V. Zaneveld, J.C. Kitchen, SPIE Vol. 2258, Ocean

Optics XII, Vol. 49, 1994.

[14] R.C. Smith, K.S. Backer, Appl. Opt. 20 (1981) 177.

[15] S.A. Kanaev, A.P. Kuleshov, Proceedings of the Third

NESTOR Workshop, Pylos, 1983.

[16] L.K. Resvanis, Proceedings of the Third NESTOR

Workshop, Pylos, 1983.

[17] H. Bradner, G. Blackinton, Appl. Opt. 23 (1984) 1009.

[18] E.G. Anassontzis, et al., Nucl. Instr. and Meth. A 349

(1994) 242.

[19] N. Palanque-Delabrouille, Proceedings of the XXVI

International Cosmic Ray Conference, HE 6.3.20, Salt

Lake City, 1999.


http://www.wetlabs.com

Documents

Measurements of light transmission in deep sea with the