Download A METHOD USING CHEMICAL OXIDATION TO REMOVE LIGHT

J. Phycol. 35, 1090–1098 (1999)
A METHOD USING CHEMICAL OXIDATION TO REMOVE LIGHT ABSORPTION BY
PHYTOPLANKTON PIGMENTS1
Giovanni M. Ferrari 2 and Stelvio Tassan
Space Applications Institute, CEC, Joint Research Centre, Ispra Site, 21020 Ispra, Italy
overall efficiency of this new depigmentation method and to demonstrate its distinctive features relative to the currently used methanol extraction procedure.
Methanol and sodium hypochlorite eliminate pigment absorption by different actions. The methanol
treatment dissolves photosynthetic pigments, which
are extracted from the thylakoids of the phytoplankton cell, but it might also eliminate other methanolsoluble substances, such as the phospholipids. Sodium hypochlorite oxidizes the pigment molecules
to reduce their absorption of light to negligible levels. The selective action of this compound depends
on the fact that it reacts rapidly with the pigments,
which are easily oxidizible products, and much
more slowly with other organic matter (detritus).
Rinsing with water eliminates the excess NaClO,
whose absorption is trivial above 400 nm but increases steeply at shorter wavelengths.
The substantially different action of the two methods has practical consequences. Methanol extraction
is unlikely to produce light-absorbing compounds
but cannot be applied to particle suspensions or to
particles retained on filters that are damaged by the
solvent (such as the 0.22-mm cellulose Millipore
membranes). In addition, methanol fails to extract
phycobiliproteins, which are present in the ubiquitous cyanobacteria as well as other species, including
cryptophytes (Bricaud and Stramski 1990). Oxidation by NaClO is applicable to particles retained on
any type of filter, as well as in suspension, but the
induced chemical reactions might cause new lightabsorbing compounds to appear. Both treatments
can alter the light-scattering properties of the particles, although one might expect that solvent extraction of pigments from the cell can more generally modify particle size and thus light scattering.
No good alternative exists to NaClO for use with
suspensions and filters attacked by methanol. When
using glass-fiber filters, comparison of the methanol
and NaClO depigmentation procedures requires
consideration of the efficiency of pigment removal,
its dependence on phytoplankton type, the sensitivity to the treatment parameters (i.e. reagent amount
and reaction time), and the possible presence of artifacts (e.g. formation of new light-absorbing compounds).
The estimate of the pigment removal efficiency is
not straightforward. Ideally, pigment removal
should eliminate all pigment absorption without affecting absorption and scattering by other matter
The effectiveness of a new method for the removal of pigment absorption in phytoplankton cells
as required in optical measurements has been extensively tested and compared with the performance
of the currently used methanol extraction method.
Key index words: hypochlorite oxidation; optical
models; photosynthetic pigments; phytoplankton;
remote sensing
Optical models in support of water-quality surveys
carried out by remote sensors require a precise description of the individual contributions of the various substances contained in the upper water layers
(Prieur and Sathyendranath 1981, Sathyendranath
et al. 1989, Carder et al. 1991). This demands more
accurate measurements of optical properties such as
the light absorption coefficient of the particulate
matter suspended in water, which includes phytoplankton, and organic and inorganic detritus. These
measurements are performed currently on particles
retained in glass-fiber filters, using dual beam spectrophotometers that compare the light transmitted
through the particle-retaining filter and a fresh ‘‘reference’’ filter (Yentsch 1962, Mitchell and Kiefer
1988, Cleveland and Weidemann 1993, Arbones et
al. 1996). The raw experimental data are expressed
in terms of the dimensionless ‘‘absorbance,’’ A,
equal to the base-10 logarithm of the inverse transmittance. This procedure offers two basic advantages: 1) It allows for particle concentration so as to
meet the instrumental accuracy requirements regardless of the high dilution of the suspension occurring in situ, and 2) it provides information on
the cells absorption in vivo. Discrimination of the
absorption due to phytoplankton pigments versus
detritus is generally obtained by subtracting measurements performed before and after methanol extraction of pigments, according to the procedure
described in Kishino et al. (1985).
A more versatile method for removal of pigment
absorption has been established on the basis of the
oxidizing action of sodium hypochlorite (Tassan
and Ferrari 1995). Extensive use of this procedure
at the Marine Environment Unit, SAI, CEC Joint Research Centre, Ispra, Italy, permitted us to examine
thoroughly the action of NaClO. This paper presents a summary of the tests performed to assess the
1
2
Received 10 June 1998. Accepted 1 June 1999.
Author for reprint requests; e-mail [email protected].
1090
REMOVAL OF PIGMENT ABSORPTION
1091
FIG. 1. Absorbance spectrum of a suspension
of Phaeodactylum tricornutum treated with a sodium
hypochlorite solution. Solution amounts and reaction times as shown.
present in the sample. However, a precise criterion
to verify how closely this ideal result is approached
in the real case is not available because of the
complexity of natural situations. Frequently, an investigator checks that the 675-nm peak of chl a absorption has disappeared and decides that this is a
sufficient indication of pigment removal. Such a situation can be verified rather accurately because detritus absorption is low at that wavelength.
Unfortunately, the previously mentioned criterion
gives no information on the removal of other pigments. Thus one should add the condition that the
absorbance spectrum of the depigmented sample
does not exhibit any discernible residual pigment
absorption peak and increases smoothly toward the
lower limit of the wavelength range with the typical
concave slope of detritus absorption. Clearly, this
condition cannot be verified as accurately as the previous one. There is still another concern, as the pigment removal treatment can also remove the absorption of unpigmented organic matter present in
the sample. This occurrence is not detected by the
second condition, and a third check is required.
This check is to ascertain that the original sample
and the depigmented one have the same absorbance at the upper limit of the wavelength range
considered, usually 750 nm, where the pigment contribution is negligible (more precisely, that both absorbance values are contained within the experimental error band). When applied together, these
three criteria give a reasonable estimate of pigment
removal efficiency. This approach has been followed
throughout the present study.
MATERIALS AND METHODS
The use of sodium hypochlorite for pigment absorption removal was first proposed as part of a new method for determining
light absorption by aquatic particles retained on filters based on
the measurement of light transmitted and reflected by the particle sample (Tassan and Ferrari 1995). According to this method,
the correction for absorption amplification induced by multiple
scattering in the filter (the so-called b factor; see Butler 1962) is
measured by removing pigment absorption from both the particle
suspension and the particles on the filter (Tassan and Ferrari
1995:1362–3).
The original procedure consisted of adding a few drops of a
15% NaClO solution (1% active Cl) either to strips cut from particle-retaining 47-mm-diameter GF/F filters or to particle suspension samples contained in a 3-mL cuvette. An equal amount of
NaClO was added to the reference sample (fresh filter or cuvette
containing pure water) to balance the absorption of the NaClO
added to the sample. With the chosen quantity of NaClO, pigment oxidation was completed in a few minutes. Filter strips and
cuvettes were measured against the ports of an integrating sphere
attached to the double-beam spectrophotometer (Tassan and Ferrari 1995: fig. 1).This procedure was used originally for measurements carried out in coastal waters (Adriatic Sea and Baltic Sea)
characterized by rather large phytoplankton and detritus concentrations.
The entire procedure was revised, aiming for better control of
its effects on samples with low absorbance. The osmotic pressure
of the NaClO was closely matched to that of the sample. Thus,
for freshwater samples, one uses a solution of sodium hypochlorite in Milli-Q water. For marine samples, one uses a NaClO solution in Milli-Q water containing 60 g·L21 of Na2SO4 (the use of
seawater is not recommended because the excess NaCl slows
down the oxidation reaction). The concentration of NaClO has
been considerably reduced (i.e. from 1%–0.1% active Cl). The
corresponding increase of both the solution volume and the oxidation time used permits larger acceptable variations from the
optimal values of these parameters.
The previously used 47-mm-diameter GF/F filter has been replaced by a 25-mm-diameter filter (21-mm-diameter clearance
area) that is directly located against the integrating sphere port
(the larger filter had to be cut into strips to be accommodated
on the port). The oxidizing solution is added to the filter placed
on the filtration equipment, with sufficient volume to cover the
filter’s upper surface.
The oxidation products, NaCl, and the excess NaClO are soluble in water and are removed from the filter by gentle filtration
of 50 mL of water, either fresh or salt. Some rinsed filters have
been treated with strong acids to detect any traces of Cl in the
cell, and no residual Cl was found. The magnesium is leached
out from the tetrapyrrolic ring of chlorophyll as MgCl2, but the
reaction pH of 10.6 avoids formation of magnesium hydroxide,
which would affect the absorbance measurement. The NaClO is
no longer added to the fresh filter but only to the reference cell
for the particle-suspension measurement. These changes im-
1092
GIOVANNI M. FERRARI AND STELVIO TASSAN
FIG. 2. Effect of excess NaClO amount and
reaction time on the absorbance spectrum of
phytoplankton samples in suspension. (A) natural water sample collected in Lake Varese, Italy. (B) Laboratory culture of Microcystis sp. (C)
Laboratory culture of Pavlova lutheri. Solution
amounts and action times as shown.
REMOVAL OF PIGMENT ABSORPTION
proved control of the oxidation procedure and gave better reproducibility.
Measurements were made on laboratory-grown algal cultures
and on samples of natural waters. The selected results presented
in the paper concern the freshwater and marine species Phaeodactylum tricornutum Bohlin 1090-1a, Microcystis aeruginosa Kuetzing
B 46.80, Pavlova lutheri (Droop) Green B 926-1, Emiliania huxleyi
(Lohmann) Reinhardt B 33.90, Prorocentrum micans Ehrenberg B
41.80, and Synechococcus leopeliensis (Raciborski) Komarek B 14021 as well as water samples collected in Lakes Varese and Maggiore
(Italy) and in the Adriatic Sea. The selection was made to provide
a large variation range in parameters affecting the oxidation process. The method used for the determination of the particle light
absorption was that described in Tassan and Ferrari (1995). The
measurements were performed over wavelengths from 400 to 750
nm; the lower limit was occasionally set to 350 nm.
RESULTS AND DISCUSSION
Depigmentation of particle suspensions and filter-retained particles. The optimum amount of NaClO to
be added to the particle suspension was carefully
determined by absorbance measurements made after successive additions of small quantities of the oxidizing agent. An example of application of this procedure is presented in Figure 1 (Phaeodactylum tricornutum; the curves follow from top to bottom as
listed in the legend, and some intermediate steps
are not shown). Comparisons of the spectra according to the three criteria for pigment removal previously outlined show that 120 mL of the NaClO solution caused complete removal in 15 min; increasing the reaction time to 35 min had a negligible
effect.
The amount of NaClO required for complete removal of pigment absorption was related to the chlorophyll concentration of the sample. Combining the
results of tests carried out on natural phytoplankton
(Lake Varese, Italy), as well as on laboratory-grown
algal cultures (Synechococcus leopoliensis, Emiliania
huxleyi, Phaeodactylum tricornutum, Pavlova lutheri,
and Microcystis aeruginosa), yielded the value 0.1 6
0.05 mg Cl (mg chl)21, where chl is the total pigment
amount, as measured by high-performace liquid
chromatography (HPLC), and the quoted error is
the standard deviation of 11 measurements. The observed dispersion is probably due to the varying
qualitative/quantitative distribution of pigments in
the phytoplankton cell.
A series of tests was performed to estimate the
effect of NaClO amounts and reaction times largely
exceeding the appropriate values. For practical reasons, the tests were carried out on particle suspensions. For example, Figure 2A shows absorbance
spectra measured for a sample obtained by resuspending particles concentrated on a 0.22-mm Millipore membrane (water collected in Lake Varese, Italy). The sample was first treated for 11 min with 25
mL of the standard NaClO solution (treatment satisfying the criteria for pigment removal as defined
previously) and then for an additional 2 min after
increasing the amount of NaClO solution to 50 mL
and so on, as specified in the figure legend. Note
that in this case, because of the very large detrital
1093
absorption (i.e. 3.5 times the pigment absorption at
the 440-nm wavelength of the main chl a peak), estimation of the pigment absorption is very sensitive
to the detritus absorption uncertainty. Comparison
of the spectra shows that a threefold increase of the
NaClO concentration, coupled to a 34-min increase
of the reaction time caused only a modest reduction
of the oxidized sample absorption (i.e. a 4% decrease at 440 nm), yielding a 13% variation of the
pigment absorbance at the same wavelength.
A similar test (Fig. 2B) carried out on a sample
of a laboratory culture of Microcystis aeruginosa varied
the reaction time from 20 to 105 min. This caused
a 2.5% increase in the 440-nm peak of the pigment
absorbance, obtained by subtracting the measurements obtained after NaClO oxidation from those
obtained before oxidation. Analogous evidence was
provided by the test performed on a sample of Pavlova lutheri (Fig. 2C): Varying the reaction time from
15 to 100 min decreased A (440) of detritus by 4%
and resulted in a 0.8% increase of the pigment 440nm peak absorbance. When comparing measured
absorbance spectra, one must account for the experimental error. In the case of particle suspensions,
provided that the homogeneity of the suspension is
maintained by agitation before each measurement,
the absorbance error (as 1 s) is about 0.001. Several
variations in the spectra displayed in the previous
figures were of the same order of magnitude of the
experimental error.
The small slope of the detritus absorbance spectra
of Figure 2 in the low-wavelength zone excludes the
presence of significant artifacts that would have
caused an increase of the absorption in that zone.
Appreciable artifacts were never observed. A trial
carried out measuring with HPLC on some samples
treated with insufficient NaClO showed a small
amount of phaeophytin that had not been present
initially. This suggests that the pigment oxidation
follows an intermediate pathway of phaeophyitinization with the extraction of the Mg11 ions. It follows that the first stage of chlorophyll disappearance
is partly balanced by phaeophytin formation. Because phaeophytin absorbs in the 400- to 440-nm
region, its appearance might introduce an artifact
in the absorption spectrum that is, however, without
practical consequence if the correct or excess NaClO amount is used.
In contrast to methanol, NaClO can be used with
particles retained on any filter type because the oxidation reaction has no effect on the filter material.
In particular, it can be applied to 0.22-mm Millipore
cellulose membranes, permitting quantification of
the absorption of extremely small picoplankton cells
(Ferrari and Tassan 1996).
Because a large number of measurements on filter-retained particles are usually required, a rapidly
applicable criterion for choice of the NaClO
amount to be used with each sample must be available. The criterion adopted is based on the mea-
1094
GIOVANNI M. FERRARI AND STELVIO TASSAN
FIG. 3. Absorbance spectra of filter-retained phytoplankton
samples obtained by NaClO oxidation and methanol extraction.
Original, NaClO treated, and methanol treated samples are
identified by thick, medium, and thin lines, respectively; total,
detrital (i.e. depigmented sample), and pigment absorptions are
marked by T, D, and P labels, respectively. (A) Natural sample
from Lake Maggiore, Italy. (B) Natural sample from Adriatic
Sea. (C) Laboratory culture of Emiliania huxleyi. (D) Laboratory
culture of Scenedesmus obliquus.
REMOVAL OF PIGMENT ABSORPTION
1095
FIG. 4. Absorbance spectra of filter-retained phytoplankton samples, with methanol
extraction followed by NaClO oxidation. Original absorbance and the absorbance of methanol-treated and MeOH/NaClO-treated samples are identified by thick, thin, and medium
lines, respectively; total, detrital (i.e. depigmented sample), and pigment absorptions
are marked by T, D, and P labels, respectively.
(A) Phaeodactylum tricornutum. (B) Prorocentrum micans.
sured height of the 440-nm absorbance peak of the
sample. A series of tests yielded the following expression: mL of solution (0.1% active Cl) 5 3·103
A(440). This amount is in excess of the values determined by the tests made on particle suspensions,
so as to ensure complete depigmentation even with
the less reactive phytoplankton species. The suggested treatment time is in the 5- to 10-min range.
The oxidation of nonpigmented matter remains
trivial because such oxidation requires longer times.
As with the particle suspensions, a considerable increase of the NaClO reaction time over that recommended had only a minor effect. If depigmentation proves incomplete because of a low NaClO
amount or a short reaction time or both, the treatment can be repeated.
In contrast to methanol, NaClO can be used with
particles retained on any filter type because the oxidation reaction has no effect on the filter material.
In particular, it can be applied to 0.22-mm Millipore
cellulose membranes, permitting quantification of
the absorption of extremely small picoplankton cells
(Ferrari and Tassan 1996).
Comparison of NaClO and methanol treatments. A
large series of tests was carried out to compare pigment absorption spectra of filter-retained samples
obtained by methanol extraction and NaClO oxidation. Typically, pigment absorption yielded by the
two methods agreed within a few percent at the 440nm peak of chl a. Four peculiar situations are shown
in Figure 3, namely, water containing cyanobacteria
(Lake Maggiore, Italy), water with a large content
of detritus (Adriatic Sea), the coccolithophore Emiliania huxleyi, and the freshwater chlorophycean Scenedesmus obliquus. The detritus absorbance of the
Lake Maggiore sample measured after methanol extraction showed residual peaks in the 500- to 600nm region (Fig. 3A). The peaks were ascribed to
cyanobacterial pigments that are known to be removed poorly by methanol. This leads to an over-
1096
GIOVANNI M. FERRARI AND STELVIO TASSAN
FIG. 5. Absorbance spectra of a hydroalcoholic solution of chl a before and after addition of
sodium hypochlorite (medium and thin lines, respectively).
estimate of detritus absorption and an underestimate of pigment absorption (both significant in
magnitude) in the 500- to 600-nm wavelength interval. The 440- and 670-nm chl a absorbance peaks
obtained by the two depigmentation treatments are
equal within the experimental error limits. Despite
the large detritus absorption of the Adriatic seawater
sample (45% of the total absorption at 440 nm), the
pigment absorbance spectra obtained by NaClO oxidation and methanol extraction are quite close to
each other (Fig. 3B); pigment removal seemed to
be complete with both procedures (no evidence of
residual peaks). Typically, pigment absorption yielded by the two methods agreed within a few percent
at the 440-nm peak of chl a. For Emiliania huxleyi,
the pigment absorbance spectra obtained by the two
pigment-neutralization methods were equivalent
over the entire wavelength range (Fig. 3C). Figure
3D displays a situation in which methanol extraction
was not effective (Scenedesmus obliquus).
In the first three cases, NaClO oxidation yielded
detritus absorption spectra with slope equal or lower
than the slope of the spectra obtained by methanol
extraction. Again, this evidence indicated that appreciable artifacts induced by oxidation were not
present.
Artifacts. Extensive comparison with the results of
methanol extraction permitted to exclude the formation of significant artifacts in the absorbance
spectrum of filter-retained particle samples over the
400- to 750-nm wavelength interval that could be
linked to the NaClO treatment. Because the light
absorption by NaClO increases steeply below 400
nm, the use of sodium hypochlorite for absorption
measurements below this wavelength, even followed
by repeated rinsing, is a priori not as accurate as for
measurements in the 400- to 750-nm range. This is
not a limitation with regard to satellite sensing of
water bodies, as this does not require data below 400
nm. For example, the recent SeaWiFS scanner by
NASA (Hooker et al. 1992) has its lowest channel
center at 411 nm.
Nevertheless, the occurrence of NaClO induced artifacts down to 350 nm was explored, treating the
filter-retained particle sample first with methanol and
then with NaClO. Figure 4A displays the original absorbance spectrum of Phaeodactylum tricornutum, together with the detritus absorbance spectra obtained
by methanol extraction (20-min treatment, thin line)
and by adding another 10 min oxidation by NaClO
(thick line). Also shown are the corresponding pigment absorbance spectra (thin and thick line, respectively; P label). Despite the saturated treatment
time, methanol failed to remove completely the pigment absorbance at 400 to 450 nm. No evidence of
excess absorption induced by oxidation was seen in
the 350- to 400-nm interval. The pigment absorbance
spectra obtained by the two methods were practically
equal, except for the slight underestimate given by
methanol in the zone of incomplete pigment removal. A different result was obtained in the case of an
aged, likely contaminated culture of Prorocentrum micans (Fig. 4B). Methanol caused incomplete removal
of pigment (detritus absorbance with a broad residual peak of about 600 nm, yielding a significant underestimate of the pigment absorbance), but the subsequent NaClO treatment produced a clear artifact,
namely, an absorption increase below 400 nm. Because of the relatively low detritus absorption, this
increase had a modest effect on pigment absorption
down to 375 nm. Other tests gave varying results, with
artifacts mostly absent. The cause of these artifacts
was not determined. The NaClO oxidation of heterotrophic bacteria was found to produce a steep increase of absorption below 450 nm (Ferrari and Tassan 1998); thus, the presence of these bacteria on
the filter could yield an artifact similar to that shown
in Figure 4B. One can conclude that appreciable ar-
REMOVAL OF PIGMENT ABSORPTION
1097
FIG. 6. Particle size distribution of laboratory culture of Emiliania huxleyi before and after treatment with sodium hypochlorite (medium and thin lines, respectively).
tifacts caused by NaClO reaction on filter-retained
phytoplankton samples are absent from 400 to 750
nm but might occur below 400 nm for some types of
algae. On the other side, the tests showed that methanol extraction often failed to remove some pigments
completely.
The presence of artifacts in the NaClO-oxidized
phytoplankton suspension was more difficult to identify because of the absence of data from another technique (solvent extraction cannot be applied to suspensions). Nevertheless, any significant artifacts in
the low-wavelength zone should be evidenced by an
anomalous absorbance increase, which the tests never displayed. For example, Figure 5 shows the result
of a test performed adding NaClO to an aqueous
50% ethanol solution of chl a (Sigma Chemical Co.,
St. Louis, Missouri). The absorbance of the chlorophyll oxidation products was low (2.5% and 0.1% of
the pigment peak absorbance at 425 and 665 nm,
respectively), with a spectrum similar in shape to that
of the absorption of CDOM produced by phytoplankton decay. These soluble oxidation products were removed from the filter-retained sample treated with
NaClO by subsequent rinsing with water.
Effect on light scattering. Depigmentation, either by
methanol extraction or by NaClO oxidation, can
modify the scattering of light by phytoplankton if it
causes a significant change in the particle size distribution. At a microscopic level, oxidation does not
seem to modify the size and internal structure of the
cells. This observation was confirmed by Coulter
counter measurements carried out on particle suspension samples for several algal species. An example of size distribution (as number of particles per
constant log[radius] interval) measured before and
after NaClO oxidation is presented in Figure 6 (Emiliania huxleyi). The minimum diameter reduction
shown in the figure was computed by Mie theory to
have a negligible effect on light scattering by the
particles. A somewhat larger decrease in size (mean
diameter change from 7–6 mm) was measured for
Phaeodactylum tricornutum treated with excessive
NaClO and was possibly due to the increase of the
osmotic pressure on the cell wall, which is less rigid
than that of Emiliania huxleyi. Also, in this case the
computed variation of light scattering was very low.
The main results of this study can be summarized
as follows: 1) NaClO-induced oxidation of pigments
was sufficiently faster than that of other organic matter to effectively discriminate pigment absorption.
Reagent amount and reaction time in excess of the
optimal values were well tolerated. 2) NaClO treatment was equally effective with particles retained on
any type of filter and in aqueous suspension, whereas methanol is applicable only to filter-retained particles and damages some filter types. 3) In the 400to 750-nm range, the agreement between pigment
absorption spectra yielded by methanol extraction
and NaClO oxidation was generally good. Appreciable artifacts caused by NaClO were not observed. 4)
In the 350- to 400-nm interval, detritus absorption
of some phytoplankton samples following sodium
hypochlorite treatment was higher than that obtained by methanol extraction. This was likely due
to an artifact caused by the NaClO-induced oxidation. The presence of artifacts below 400 nm is not
a limitation with respect to most remote-sensing applications. 5) NaClO was effective across a wide
range of phytoplankton types, including water-soluble pigments such as those contained in cyanobacteria. On the contrary, significant residual pigment
absorption peaks were observed frequently after
methanol extraction, which was also ineffective for
water-soluble pigments and with some taxa. 6)
NaClO-induced oxidation was found to cause only a
minor reduction of the equivalent radius of phytoplankton cells, with minimal effect on light scattering by the particles.
1098
GIOVANNI M. FERRARI AND STELVIO TASSAN
We conclude that chemical oxidation using a sodium hypochlorite solution is a valid alternative to
the currently used methanol extraction procedure
for the removal of pigment absorption in phytoplankton, over which it shows some distinctive advantages.
Arbones, B., Figueiras, F. G. & Zapata, M. 1996. Determination
of phytoplankton absorption coefficient in natural seawater
samples: evidence of a unique equation to correct the pathlength amplification of glass-fiber filters. Mar. Ecol. Prog. Ser.
137:293–304.
Bricaud, A. & Stramski, D. 1990. Spectral absorption coefficients
of living phytoplankton and nonalgal biogenous matter: a
comparison between the Peru upwelling area and the Sargasso sea. Limnol. Oceanogr. 35:562–82.
Butler, W. L. 1962. Absorption of light by turbid materials. J. Optical Soc. Am. 52:292–9.
Carder, K. L., Hawes, S. K., Baker, K. A., Smith, R. C., Steward,
R. G. & Mitchell, B. G. 1991. Reflectance model for quantifying chlorophyll a in the presence of productivity degradation products. J. Geophysical Res. 96:20599–20611.
Cleveland, J. S. & Weidemann, A. D. 1993. Quantifying absorption
by aquatic particles: a multiple scattering correction for glassfiber filters. Limnol. Oceanogr. 38:1321–7.
Ferrari, G. M. & Tassan, S. 1996. Use of the 0.22 mm Millipore
membrane for light-transmission measurements of aquatic
particles. J. Plankton Res. 18:1261–7.
1998. A method for the experimental determination of
light absorption by aquatic heterotrophic bacteria. J. Plankton
Res. 20:757–66.
Hooker, S. B., Esaias, W. E., Feldman, G. C., Gregg, W. W. &
McClain, C. R. 1992. An overview of SeaWiFS and ocean color. In Hooker, S. B. & Firestone, E. R. [Eds.], Vol. 1 of
SeaWiFS Tech. Rep. Series, NASA Tech. Memo. 104566,
NASA, Greenbelt, Maryland, p. 24.
Kishino, M., Okami, N. & Ichimura, S. 1985. Estimation of the
spectral absorption coefficients of phytoplankton in the sea.
Bull. Mar. Sci. 37:620–33.
Mitchell, B. G. & Kiefer, D. A. 1988. Chlorophyll a specific absorption and fluorescence excitation spectra for light-limited
phytoplankton. Deep-Sea Res. 35:639–63.
Prieur, L. & Sathyendranath, S. 1981. An optical classification of
coastal and oceanic waters based on the specific spectral absorption of phytoplankton pigments, dissolved organic matter and other particulate materials. Limnol. Oceanogr. 26:671–
89.
Sathyendranath, S., Prieur, L. & Morel, A. 1989. A three-component model of ocean colour and its application to remote
sensing of phytoplankton pigments in coastal waters. Int. J.
Remote Sens. 10:1373–94.
Tassan, S. & Ferrari, G. M. 1995. An alternative approach to absorption measurements of aquatic particles retained on filters. Limnol. Oceanogr. 40:1358–68.
Yentsch, C. S. 1962. Measurement of visible light absorption by
particulate matter in the ocean. Limnol. Oceanogr. 7:207–17.