Deep-Sea Research Part II xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Deep-Sea Research Part II

journal homepage: www.elsevier.com/locate/dsr2

Bio-optical characterization of the northern Antarctic Peninsula waters:

Absorption budget and insights on particulate backscattering
⁎
Amabile Ferreiraa, , Áurea M. Ciottia, Carlos A.E. Garciab,c
a
Centro de Biologia Marinha, Universidade de São Paulo, Rod. Manuel Hypólito do Rego,Rego, Km 131.5, Praia do Cabelo Gordo, São Sebastião, SP 11612-109, Brazil
b
Instituto de Oceanografia, Universidade Federal do Rio Grande, Rio Grande, RS 96201-900, Brazil
c
Programa de Pós-Graduação em Oceanografia, Universidade Federal de Santa Catarina, SC 88900-900, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: We utilize a comprehensive set of bio-optical properties collected in the northern Antarctic Peninsula to con-
Phytoplankton tribute to the poor documentation of in situ bio-optical measurements in the Southern Ocean. The relative
cdom contributions of phytoplankton, colored dissolved organic matter (cdom) and detritus to light absorption high-
Detritus light the importance of phytoplankton, but cdom contribution is often important in the blue range of the spectra
Absorption
despite its roughly constant values at 443 nm along the three orders of magnitude of total chlorophyll-a con-
Scattering
centration, [TChl a], (0.019–2.91 mg m−3) observed in the upper 100 m. The particulate backscattering coef-
Backscattering
ficient, bbp(λ), was remarkably low if compared to other oceanic waters, but agrees with previous studies in the
Southern Ocean. Even with very low absorption, detritus was the component better correlated with particulate
backscattering in the Antarctic Peninsula, while phytoplankton cells (dominant in the particles pool) mostly
covaried with particulate scattering. Particulate backscattering ratios (bbp(λ) divided by the particulate scat-
tering coefficient, bp(λ)) were also below values observed in other oceanic waters. The spectral diffuse at-
tenuation coefficient, Kd(λ), was highly correlated to [TChl a] (R2=0.90 at 443 nm) and showed no dependence
on bbp(λ). Indeed, Kd(443) and non-water absorption coefficients at 443 nm were related by a 1 to 1 dependence.
The shape of the spectral remote sensing reflectance varied little responding mainly to variability in [TChl a],
while [TChl a] vs. maximum band ratios dependence deviated from global trends in a very similar fashion as in
other studies of the Southern Ocean, likely due to very low bbp(λ).

1. Introduction
The optical properties of seawater can be reconstructed from the
additive contributions of water molecules, phytoplankton, nonalgal
particles (mainly detritus) and colored dissolved organic matter (cdom),
called bio-optical components. The magnitude and spectral shape of
light absorption and scattering, inherent properties (Preisendorfer,
1961) are altered by changes in concentrations and compositions of
these components (e.g., IOCCG, 2006). The characterizations of optical
properties in different portions of the ocean are necessary steps for
using optical measurements, performed by in situ and remote platforms
as tools for biogeochemical studies (e.g., Johnson et al., 2009).
In case 1 waters, where phytoplankton and its co-varying particu-
late and dissolved materials govern the optical signal (Gordon and
Morel, 1983; Morel and Prieur, 1977; Prieur and Sathyendranath,
1981), parameterizations of bio-optical properties as functions of
chlorophyll-a concentration have been useful for biogeochemical
modeling (e.g., Morel, 1988). Chlorophyll-a concentrations are used to
predict optical coefficients such as the particulate (e.g. Bricaud et al.,
1998) and phytoplankton (e.g., Bricaud et al., 2004; Bricaud et al.,
1995) absorption, the particulate beam attenuation (e.g., Behrenfeld
and Boss, 2006; Loisel and Morel, 1998), the particulate scattering (e.g.
Gordon and Morel, 1983; Huot et al., 2008) and the particulate back-
scattering (Antoine et al., 2011; Dall’Olmo et al., 2009; Huot et al.,
2008).
While variations in the phytoplankton absorption coefficient are
well documented, spectral absorption of cdom (e.g., Babin et al., 2003;
Bricaud et al., 1981) and detritus (e.g., Babin et al., 2003; Bricaud et al.,
1998; Bukata et al., 1995) have been less systematically investigated,
and because of that, frequently disregarded or simple assumed to
covary with chlorophyll-a concentration (Carder et al., 1991; but see
Ciotti et al., 1999). In situ measurements of the beam attenuation
coefficients have long been carried out routinely as an index for par-
ticles (see Bartz et al., 1978), in contrast to the backscattering coeffi-
cient that only recently became systematically measured (e.g., Antoine
et al., 2011; Slade and Boss, 2015; Huot et al., 2008; Reynolds et al.,

⁎
Corresponding author.
E-mail address: amabilefr@gmail.com (A. Ferreira).

http://dx.doi.org/10.1016/j.dsr2.2017.09.007

0967-0645/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Ferreira, A., Deep-Sea Research Part II (2017), http://dx.doi.org/10.1016/j.dsr2.2017.09.007
A. Ferreira et al.
Fig. 1. Geographic locations of the sampling stations in the Gerlache Strait (GS, triangles), Bransfield Strait (BS, stars), northwestern Weddell Sea (NWW, circles). NWWHT denotes “high
transparency” for two stations sampled in 2014 (asterisks). In GS, BS and NWW, most of the stations of 2013 were sampled in 2014 and are indicated with the edge markers in black.
2016; Twardowski et al., 2007). Backscattering measurements may
provide information on particulate organic matter (Stramski et al.,
2008), phytoplankton carbon concentration (Graff et al., 2015, 2016),
chlorophyll-a to autotrophic carbon ratios (Cetinić et al., 2015) and
particle size distributions (Kostadinov et al., 2009; Loisel et al., 2006;
Slade and Boss, 2015).
A large dispersion is generally observed in global empirical re-
lationships developed for bio-optical properties or parameter as func-
tions of chlorophyll-a concentration (e.g. Bricaud et al., 1998; O'Reilly
et al., 1998) that could be associated with methods, instruments, as well
as actual regional differences (Morel and Maritorena, 2001). Indeed,
the empirical relationships between reflectance ratios and chlorophyll-a
concentration, tools for estimating phytoplankton biomass from remote
sensing, are significantly different in samples observed in Antarctica
compared to the other oceanic regions (e.g., Dierssen and Smith, 2000;
Reynolds et al., 2001), also manifested as errors in the chlorophyll-a
retrievals when global algorithms of ocean color are applied (e.g.,
Garcia et al., 2005; Szeto et al., 2011). These differences have been
attributed to unique “packaged” phytoplankton absorption character-
istics (Dierssen and Smith, 2000; Mitchell and Holm-Hansen, 1991), but
the contribution of distinct particulate backscattering coefficients
cannot be overruled (Brown et al., 2008). In fact, in situ particulate
backscattering measurements in Antarctica have shown low magni-
tudes (Dierssen and Smith, 2000; Reynolds et al., 2001), but they re-
main too scarce to provide conclusive results (see Dierssen, 2010).
Retrievals of chlorophyll-a concentration in the Antarctic Peninsula by
global ocean color algorithms generally over and underestimate in situ
low and high chlorophyll-a, respectively, and this aspect is generally
related to a lower water leaving reflectance and a narrower variation in
the blue-green band ratio (Zeng et al., 2016) as well as low back-
scattering coefficients in the green part of the spectrum at high chlor-
ophyll-a (Dierssen and Smith, 2000). However, the mechanisms which
control the ocean color variability in the Southern Ocean need to be
further understood.
This work describes variations in the bio-optical properties in the
northern Antarctic Peninsula observed during late austral summers of
2013 and 2014. To the best of our knowledge, our work is the first to
describe concurrent measurements of absorption, scattering, back-
scattering, diffuse attenuation coefficients and remote sensing re-
flectance in the Antarctic Peninsula. The main goal of this study is to
report the relative contributions of phytoplankton, cdom and detritus to
light absorption and to access potential relationships between scat-
tering and backscattering coefficients and chlorophyll-a concentration,
2
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
comparing to previous relationships found elsewhere. A complementary
purpose of this study is to explore the prediction of chlorophyll-a
concentration and absorption coefficient from the diffuse attenuation
coefficient for downward irradiance, Kd(λ). This is of special interest
because Kd(λ) is a relatively simple measurement to obtain in the field
and provides absorption (e.g., Loisel and Stramski, 2000), justified by
the dominance of absorption in the diffuse attenuation process (Morel
et al., 2007), particularly in waters where backscattering is low. Finally,
we characterize the variability in the hyperspectral remote sensing re-
flectance, Rrs(λ), and evaluate the relationships between and Rrs(λ)
ratios and chlorophyll-a concentration.
2. Material and methods
2.1. Sampling
Data were collected during two oceanographic cruises in the
northern tip of the Antarctic Peninsula in the austral summers of 2013
(14 February–3 March) and 2014 (8–24 February). The study area
covered the Gerlache Strait (GS), which separates the Palmer
Archipelago from the Peninsula, the Bransfield Strait (BS), between the
southern Shetland Islands and the Peninsula, and the northwestern
Weddell Sea (NWW). A total of 89 and 70 stations were sampled in
2013 and 2014, respectively (Fig. 1). A detailed description of sampling
strategy can be found in Ferreira et al. (in press).
Discrete samples for determination of chlorophyll-a concentration
and spectral absorption coefficients were collected at 5, 15, 25, 50, 75
and 100 m at night, and only at 5 m during the day. Profiles of inherent
optical properties were acquired down to 100 m in all day and night
stations. Above water hyperspectral reflectance's were recorded during
the day. Radiometric profiles (see 2.5) to obtain the spectral Kd(λ) were
performed less often, because of the unfavorable sea conditions com-
monly found in Antarctic waters. Therefore, the number of data for
each dataset and analysis is variable and informed when pertinent. The
data of discrete samples from 5 to 100 m were grouped in layers ac-
cording to vertical variations of light. The first optical depth has im-
plications for remote sensing application, while the others keep con-
sistency with our previous work (Ferreira et al., in press). Briefly,
L > 37% corresponds to the layer between surface and the first optical
depth, at which the downward irradiance 37% of its value just below
the surface. The data within L > 37% contributes mostly to the ocean
color signal that corresponds only to the first portion of the photic zone
returning a weighted average value over this layer (Sathyendranath and
A. Ferreira et al.
Platt, 1989). L < 10% refers to the layer below the depth where the
downward irradiance falls to 10% of its value just below the surface.
The layer between the basis of L > 37% and top of L < 10% is referred to as
L37–10%. The layers L > 37%, L37–10% and L < 10% comprised mostly
samples at 5 m, 15 and 25 m, and 50–100 m, respectively. This in-
dicates that the light distribution along the vertical varied little among
stations. This separation in optical layers indicated bio-optical char-
acteristics very similar in L > 37% and L37–10%, while in L < 10% results
tended to depart from the main trends (see Ferreira et al., in press).
2.2. Chlorophyll-a concentration
Total chlorophyll-a concentration [TChl a], representing the sum of
divinyl chlorophyll-a, chlorophyll-a epimers and allomers, chlor-
ophyllide-a was measured by high-performance liquid chromatography
(HPLC), using the procedure described by Mendes et al., (2012, 2007).
Volumes of 1–2 L of seawater were filtered onto 25-mm glass fiber fil-
ters (Whatman GF/F) for post-cruise analysis.
2.3. Absorption coefficients
Volumes of 1–2 L were concentrated on 25 mm GF/F filters and
immediately preserved in liquid nitrogen. Absorption of particulate
material was determined following the TR method of Tassan and Ferrari
(1995) using a spectrophotometer (Lambda 35, Perkin Elmer) with an
integrating sphere. Optical density (OD, no unit) of samples and blank
filters were first measured against air from 390 to 750 nm at the en-
trance of the sphere and then placed directly against the exit of the
sphere, backed with a light trap, for a second scan. Blanks and sample
filters were treated with a few drops of 0.5% NaClO for 10–15 min and
then carefully washed with 0.2-μm filtered seawater to obtain the det-
ritus absorption (Tassan and Ferrari, 1995). The OD of extracted filters
were measured as described above. Using the correction of pathlength
amplification described in Tassan et al. (2000), the OD values were
converted into absorption coefficients after subtraction of values aver-
aged between 745 and 750 nm (null correction) and correction for the
filtered volume and the clearance area of the filter. The spectral ab-
sorption coefficients of phytoplankton, aph(λ) in m−1, were computed
as the difference between absorption coefficients before (total particu-
late) and after (detritus, adet(λ)) the NaClO extraction. Detritus ab-
sorption associated to OD lower than 0.0005 (instrumental noise
of ± 0.0005) in the blue wavelengths was removed from the dataset (n
= 89). In these cases, adet(λ) is considered as zero (and so the adet(λ)
contribution to non-water absorption), so apart(λ) equals to aph(λ). An
exponential function was fitted by nonlinear regression to each adet(λ)
spectrum between 350–600 nm, following the equation adet(λ) =
adet(λr) exp(-Sdet(λ – λr)), where λr (nm) is a reference wavelength,
443 nm, adet(λr) (m−1) is the absorption estimate at λr and Sdet(nm−1)
is the spectral slope of the adet(λ) spectrum (Babin et al., 2003).
The filtrates obtained from the filtration described above were re-
filtered through a Polycap AS capsule filter (0.2 µm) and collected in
amber Qorpak bottles previously cleaned with acid and sterilized. Light
absorption of cdom was measured throughout the UV and visible
spectral domains using a WPI UltraPath liquid waveguide spectro-
photometer along a 50 cm pathlength. The spectral absorbance of the
filtered water was measured between 320 and 750 nm immediately or
within 3 h after sampling. Freshly produced, ultraviolet (UV) irradiated
Milli-Q water was used as a reference. The spectral OD was converted
into absorption coefficients, acdom(λ) in m−1. The slope for each
acdom(λ) spectrum, Scdom in nm−1, was obtained as for adet(λ), following
the equation acdom(λ) = acdom(λr) exp(-Scdom(λ – λr)) (Bricaud et al.,
1981).
The relative contribution of each optical component to spectral
absorption was determined summing up the spectral absorption coef-
ficients aph(λ), adet(λ) and acdom(λ) and dividing each component by
that sum, at the same wavelength. Therefore, the relative absorption
3
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
contributions regard the spectral non-water absorption, anw(λ).
2.4. Particulate scattering and backscattering coefficient
The inherent optical properties, IOPs, were measured from just
below the water surface down to 100 m of depth by deploying sensors
in a frame that included: (i) one conductivity, temperature and pressure
sensor (SBE 37 SI, Sea-Bird Electronics Inc.); (ii) one multi-spectral
photometer (AC-9, WET Labs Inc.) equipped with 25 cm pathlength
tubes measuring non-water absorption and attenuation, at the following
wavelengths: 412, 440, 488, 510, 532, 555, 650, 676 and 715 nm and
(iii) one backscattering sensor (HS-6, HOBI Labs Inc.) measuring in the
wavelengths of 420, 442, 470, 510, 590 and 700 nm. Spikes of the
optical profiles were firstly removed considering values that exceeded
50 times (HS-6) and 30 times (AC-9) +1 the mean difference computed
between adjacent data along the vertical (function diff.m in MATLAB),
and then applying a third-order one-dimensional median filter (func-
tion medfilt.m, with the endpoint filtering ‘truncate’ in MATLAB). The
temperature, salinity and optical profiles were smoothed along depth
with a data window of 7 and finally binned to 1 m intervals.
Absorption and attenuation measurements recorded by the AC-9
(previously calibrated in the laboratory using pure water) were first
corrected for temperature and salinity effects following the AC-9 m
Protocol Document, WET Labs Inc (http://wetlabs.com/ac-s). The re-
sulting absorption values were corrected for residual scattering effects
using the proportional method of Zaneveld et al. (1994) and 715 nm as
the reference wavelength. The particulate scattering coefficients, bp(λ)
in m−1, were obtained by subtracting the absorption coefficients from
the attenuation coefficients.
The volume scattering function, β(λ), of water was computed using
the concomitant temperature and salinity measurements following
Zhang et al. (2009). Measured β(λ) values subtracted the β(λ) of water
were firstly interpolated linearly to match the AC-9 wavelengths and
extrapolated to particulate backscattering, bbp(λ), following Maffione
and Dana (1997) using the factor 1.13 for the angle of measurement of
140° of the HB-6. The initial bbp(λ) values were corrected for absorption
and scattering effects using the concurrent AC-9 absorption coefficients
following the iterative method of Doxaran et al. (2016).
2.5. Diffuse Attenuation Coefficient
Underwater radiometric measurements were obtained at stations
during favorable sea conditions (n=17) with a hyperspectral profiling
radiometer (HyperOCR, Satlantic, Inc.), which measures the upwelling
radiance, Lu(λ,z), and the downward irradiance, Ed(λ,z), over the
350–800 nm spectral range with a resolution of 3.3 nm (137 spectral
bands). The initial data processing was performed using the ProSoft ver.
7.7.19 software. Dark offsets and the manufacturer's radiometric cali-
brations were applied to the raw data, and the spectral radiometric data
were binned every 0.1 m. Data were further analyzed in MATLAB. The
diffuse attenuation coefficient of downwelling irradiance, Kd(λ), at 5 m
was calculated as the slope of a least-squares linear regression of the
log-transformed Ed(λ,z) within the depth interval of 2–7 m.
2.6. Remote Sensing Reflectance
At 51 stations, hyperspectral radiometry was acquired above sea-
water using a HyperSAS (Satlantic, Inc.) that measures the downward
irradiance, Es(λ), the water leaving radiance, Lt(λ), and the sky ra-
diance, Li(λ), over the 350–800 nm spectral range with a resolution of
3.3 nm (137 spectral bands). The radiance sensors viewed the water
and sky at 40° of zenith and 135° azimuth angles. The initial data
processing was performed using the ProSoft ver. 7.7.19 software to
obtain L3 level data, that contained averaging of Es(λ), Lt(λ) and Li(λ)
of the L2 level data (ProSoft manual). Data were further analyzed in
MATLAB. Radiometric measurements were converted to above-water
A. Ferreira et al. Deep-Sea Research Part II xxx (xxxx) xxx–xxx

remote-sensing reflectance, Rrs(λ), according to Rrs(λ) = (Lt(λ) - ρsky x
Li(λ)) / Es(λ), where ρsky represents the proportion of sky radiance that
is reflected off the surface of the water and is dependent upon wind
speed, solar zenith, sensor viewing angle, wavelength and cloud cover
in the sky radiance measurements. A look-up-table of ρsky was created
using Hydrolight.
The Rrs(λ) spectra were quality-controlled as following: removal of
data outside 10 degrees from the compass mean as well as spectra with
negative values; removal of spectra within 50% of the highest observed
Rrs(780) to reduce glitter effects on the air–water interface. The final
Rrs(λ) for each station represents the average of the Rrs(λ) spectra
within 1 standard deviation. All Rrs(λ) spectra were interpolated line-
arly to 1 nm intervals.
3. Results and discussion
3.1. Light absorption of phytoplankton, cdom and detritus
Phytoplankton frequently dominated non-water absorption in the
blue part of the spectrum in L > 37%, with contributions from 40% to
80% at 443 nm (mean of 65%), where the three components absorb
importantly and is near the primary peak of chlorophyll-a absorption
(Fig. 2a). It is noteworthy the appearance of absorption peaks of ac-
cessory pigments, centered at 460 and 490 nm which are bands corre-
sponded to the absorption by photoprotective pigments and fucox-
anthin, respectively (see Ferreira et al., in press). The presence of
fucoxanthin is also revealed at around 590 and 640 nm. The maximum
absorption contribution at 676 nm exhibits the secondary peak of
chlorophyll-a absorption in combination to a very low influence of non-
algal material in the red spectral portion.
The fraction of absorption due to cdom can be equal to that of
phytoplankton for blue wavelengths, reaching contributions up to 50%
below 443 nm in L > 37% (mean of 32%, Fig. 2b). Similar cdom relative
contributions are observed in the green part of the spectrum (minimum
phytoplankton absorption), from 10% to 70% at around 560 nm. The
cdom contribution progressively increases from 443 nm towards lower
wavelengths, making up 30–80% of non-water absorption near surface
and about 80% of absorption at the blue and green spectral ranges in
deeper waters. Because cdom absorption was relatively constant in all
depths (see later) and phytoplankton absorption varied over about
three orders of magnitude (0.006–0.135 m−1 at 440 nm; Ferreira et al.,
in press), cdom contributions surpass that of phytoplankton in relatively
lower [TChl a] mainly in L < 10%, as phytoplankton biomass decreases
in depth.
Detritus makes a small impact to the non-water spectral absorption
in all layers, with contributions below 20% throughout most of the
spectrum and values nearly at the detection limit in many samples
(mean of 4% in L > 37%, Fig. 2c). Values of adet(443) were very low
compared to observations in coastal waters (Babin et al., 2003), re-
maining relatively constant and below 0.004 m−1, excepted for few
cases exceeding this value (Fig. 3a). Two observed adet(443) values
above 0.014 m−1 were associated with low cdom absorption but when
[TChl a] were above 2 mg m−3, thus, are presumably related to the
increase in the phytoplankton biomass. Values of adet(443) showed no
trend with [TChl a] (Fig. 3a). Note that, however, values of adet(443)
increase with the increase of [TChl a] but no clear trend can be es-
tablished between these two variables even where [TChl
a] > 1 mg m−3. Furthermore, the very low adet(λ) and adet(λ)/anw(λ)

1 1
a b

0.8 0.8
( )
( )

nw

0.6 0.6
nw

( )/a
( )/a

0.4 0.4
cdom
ph
a

0.2 0.2

0 0
400 450 500 550 600 650 700 400 450 500 550 600 650 700
Wavelength (nm) Wavelength (nm)

1
c L>37%
L37-10%
0.8 L
<10%
adet ( ) / anw ( )

0.6

0.4

0.2

0
400 450 500 550 600 650 700
Wavelength (nm)
Fig. 2. Relative contributions of spectral absorption coefficients of a) phytoplankton, aph(λ) b) cdom, acdom(λ), and c) detritus, adet(λ), to non-water absorption, anw(λ), which refers to
aph(λ) + acdom(λ) + adet(λ), in L > 37%, L37–10% and L < 10%.

4
A. Ferreira et al. Deep-Sea Research Part II xxx (xxxx) xxx–xxx

0.02
L a b
>37%
L
0.04
37-10%
0.015 L <10% 0.035
(443) (m )
-1

Sdet (nm-1)
0.03
0.01 0.025
0.02
det
a

0.005 0.015
0.01

0 0.005
0.05 0.1 1 4 0.001 0.01 0.03
-3
[TChl a] ( mg m ) a (443) (m -1)
det
Fig. 3. (a) Detritus absorption coefficient at 443 nm, adet(443), as a function of total chlorophyll-a concentration, [TChl a], in L > 37%, L37–10% and L < 10%. (b) Slope of the spectral
absorption coefficient of detritus, Sdet, as a function of adet(443).

values observed here suggests minor terrigenous influence in the
northern Antarctic Peninsula.
Detritus corresponds to all non-algal particles, which may include
zooplankton, bacteria, and actual detrital particles and phytoplankton
detritus (e.g., Iturriaga and Siegel, 1989; Kishino et al., 1985; Morel and
Ahn, 1991; Stramski and Kiefer, 1998), and this pool may be also linked
to zooplankton grazing on phytoplankton in the Antarctic Peninsula
(Mendes et al., 2012).
Values of Sdet varied largely (0.0071–0.0484 nm−1) but most fall in
a range of 0.01–0.02 nm−1 (Fig. 3b), as reported in the literature (Babin
et al., 2003; Bowers et al., 1996; Ferreira et al., 2014), decreasing
continuously with adet(443), but this inverse trend is not normally ob-
served (Babin et al., 2003; Matsuoka et al., 2011) and its cause should
be investigated. The comprehension of causes in the variability of Sdet is
limited, but it probably reveals the nature of particles (Babin et al.,
2003; Babin and Stramski, 2004; Estapa et al., 2012), with a trend of
higher values for mineral particles (e.g., Bukata et al., 1995).
The cdom absorption varied little in our observations, and
acdom(443) values were typically around 0.011 and 0.070 m−1 (Fig. 4a)
in all sampled depths (excepted for three much higher values above
0.1 m−1). Therefore, there is no relationship between acdom(443) and
[TChl a] in < 100 m, contrasting with some observations for near-
coastal environments (e.g., Babin et al., 2003; Meler et al., 2016) and in
the Ross sea and Antarctic Polar Front Zone (Reynolds et al., 2001), but
agreeing with what is mostly reported for oceanic waters (Bricaud et al.,
1981; Nelson and Siegel, 2013; Nelson et al., 1998; and references
therein; but see e.g., Bricaud et al., 2010). This uncoupling reflects
different timescales associated with the creation and destruction of
cdom and the dynamics of phytoplankton blooms. Also, no dependence
was found between acdom(443) and salinity (not shown), so glacial
meltwater in summer (Dierssen et al., 2002) is not a notable source for
cdom in the Antarctic Peninsula. cdom is produced primarily as a by-
product of microbial metabolism and phytoplankton degradation
(Nelson et al., 2004; Yentsch and Reichert, 1962). One possible scenario
for nearly invariable cdom absorption is that the observed cdom consists
of a labile fraction, which is rapidly degraded by microbes and any
input that exceed the constant pool of cdom is not necessarily detectable
in the water column (Nelson et al., 2004). One must bear in mind,
however, that cdom represents only portion of dissolved organic
matter (Nelson and Siegel, 2002).
The spectral shape of cdom absorption hints changes in its compo-
sition. Scdom varied one order of magnitude in our dataset, from 0.001 to
0.022 but stayed often within 0.011 and 0.015 nm−1 (Fig. 4b), with a
general inverse trend with acdom(443) as observed elsewhere (e.g.,
Matsuoka et al., 2011; Vodacek et al., 1997), mainly in L > 37%. The
Scdom values observed here are generally on the lower end of values of
published data, that are 0.015–0.030 nm−1 for open ocean spectra
(e.g., Nelson and Siegel, 2013; Twardowski et al., 2004) and
0.014–0.020 nm−1 for coastal waters (Babin et al., 2003). Scdom is
generally higher in the open ocean because of modifications in com-
position caused by photo-bleaching (Del Vecchio and Blough, 2004;
Goldstone et al., 2004; Swan et al., 2012). One possibility for relatively

0.25 0.025
a b

0.2 0.02
acdom (443) (m )
-1

(nm )
-1

0.15 0.015
cdom

0.1 0.01
S

0.05 0.005

0
0 0.01 0.03 0.1 0.2
0.05 0.1 1 4
-1
[TChl a] ( mg m-3 ) a (443) (m )
cdom

Fig. 4. (a) The cdom absorption coefficient at 443 nm, acdom(443), as a function of total chlorophyll-a concentration, [TChl a], in L > 37%, L37–10% and L < 10%. (b) Slope of the spectral
absorption coefficient of cdom, Scdom, as a function of acdom(443).

5
A. Ferreira et al.
0.01
This study
MM01
R01-Ross Sea
R01-APFZ
b (442) (m -1)
b (442) (m -1)
B05
H08
bp
0.001
a 0.0003
0.0003
0.05 0.1 1 4
-3
[TChl a] ( mg m )
Fig. 5. (a) The particulate backscattering coefficient at 442 nm, bbp(442), as a function of total chlorophyll-a concentration, [TChl a], in L > 37%, L37–10% and L < 10%. (b) bbp(442) as a
function of the absorption coefficient by detritus at 443 nm, adet(443). The solid lines represent the best fits of the linear regression obtained for L > 37% and L37–10%. In (a), the power laws
proposed by Morel and Maritorena (2001), MM01, and Huot et al. (2008), H08, and an invariant bbp(λ) mean value (near 0.0012 m−1) for chlorophyll concentration below 0.14 mg m−3
observed by Behrenfeld et al. (2005), B05, using satellite data are shown for comparison.
lower Scdom reported here is the combination of the location close to the
coast (Fig. 1) and low levels of light in Antarctica relatively to other
areas in the globe (Bricaud et al., 2012), yet a general inverse trend of
Scdom against acdom(443) may be indicative of photo-bleaching near the
surface. Any thoughtful comprehension on the cdom sources and sinks
in Antarctica should be further addressed using chemical and experi-
mental analyses.
We concluded that phytoplankton dominate light absorption near
surface in the Antarctic Peninsula waters. In L > 37%, 80% of our sam-
ples (60 of 73) showed phytoplankton contribution of above 60% at
443 nm (Fig. 2). If we define case 1 waters as those where aph(443)
represent more than 60% of the total non-water absorption (IOCCG,
2000), the northern Antarctic Peninsula will be classified as case 1
waters. As expected, however, the relative contributions of absorption
are highly dependent on the considered wavelength as, for instance,
cdom absorption at the green spectral region can be important. The
partitioning of non-water absorption coefficients observed here is in
agreement to the proportion shown in IOCCG (2015) for the Southern
Ocean (their Fig. 4.8). Such absorption budgets and magnitudes of
absorption reported here could support the validation of current ap-
proaches that provide satellite estimates of non-algal material and
phytoplankton absorption for the considered area. This is noteworthy
since very few data of absorption coefficients are available in literature
for the Southern Ocean (Szeto et al., 2011).
The large and small importance of cdom and detritus, respectively,
to spectral absorption observed here is in agreement with observations
mostly for oceanic waters (Nelson and Siegel, 2013). One may consider
a background of cdom absorption in the northern Antarctic Peninsula
for the purpose of regional bio-optical modeling, with a constant value
of acdom(443) of about 0.030 m−1 and a general spectral slope, Scdom, of
0.0136 nm−1 (the averages observed here). Despite the very low con-
tribution of detritus absorption, one can consider an average of
0.0025 m−1 for adet(443) with a mean Sdet of 0.0151 nm−1. The
variability in the spectral absorption coefficients of phytoplankton is
described elsewhere, including their possible impacts on bio-optical
modeling (Ferreira et al., in press). The partition of the particulate
absorption into phytoplankton and detritus is useful for the inter-
pretation of backscattering sources (see later), since the particulate
backscattering cannot be partitioned into the contribution of both
components.
3.2. Particulate backscattering coefficient
The bbp(λ) values were low, ranging from 0.0004 to 0.0067 at
6
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
0.01
L >37% b
L 37-10%
L
<10%
bp
0.001
0.001 0.01 0.04
-1
adet(443) (m )
442 nm, bbp(442), against the values of 0.008 m−1 of pure sea water.
These magnitudes are within the values found by Dierssen and Smith
(2000) in the western Antarctic Peninsula and in the clearest waters
sampled in Arctic waters, bbp(λ) were below 0.005 m−1 in the visible
spectral region (Reynolds et al., 2016). Yet, the bbp(λ) values observed
here are about the same order of magnitude of very low bbp(λ) mea-
surements made by Boss et al. (2007) in Crater Lake and by Stramska
and Stramski (2005) in the Greenland Sea, and only higher (about one
order of magnitude) than the lowest particulate bbp(λ) (on the order of
10−5 m−1) reported in the literature for the central South Pacific gyre
(Twardowski et al., 2007).
Different optical configurations of the instruments, calibration
procedures and the processing of data can impact on final reported
magnitudes of bbp(λ) among studies. For instance, correction of the
measured bbp(λ) signal for absorption and scattering losses was applied
to our data following the method developed by Doxaran et al. (2016)
that accounts losses due to absorption and scattering. The ability to
resolve total backscattering structure in extremely low backscattering
waters with commercially available instruments has been demon-
strated, but these scattering levels push the limits of uncertainties in
pure seawater, so improvements on this matter remain highly necessary
(Twardowski et al., 2007). The effect of other components such as
bubbles can account for bbp(λ) (Stramski et al., 2004; Zhang et al.,
2002), which was predicted to contribute as much as 25% of the total
particulate backscattering in the Southern Ocean (Dierssen et al., 2017)
and deservers further attention.
When bbp(442) is presented as a function of [TChl a], a relationship
statistically significant is found when considering L > 37% and L37–10%
(Fig. 5a), but there is a large amount of scatter around the fit that
covered a restricted [TChl a] range (0.35–3.03 mg m−3). A power law
fit provided the dependence bbp(442) = 0.0013 [TChl a]0.864 (n = 109,
R2 = 0.32 for the log-transformed data; p < 0.00001), but a linear
model returned a similar degree of explanation (R2 = 0.30). In L < 10%,
bbp(442) ranged mostly from 0.0004 to 0.0012 m−1, but with a large
dispersion against [TChl a], but still with a significant dependence:
bbp(442) = 0.0013 [TChl a]0.169 (n = 104, R2 = 0.12 for the log-
transformed data; p = 0.0003). Our bbp(442) values in the low [TChl a]
range are below the prediction of an invariant bbp(λ) mean value (near
0.0012 m−1) for chlorophyll concentrations below 0.14 mg m−3
(Behrenfeld et al., 2005, their Fig. 5a). In accordance to the low
backscattering levels observed here, our bbp(442) values are con-
siderably below the relationship developed by Morel and Maritorena
(2001) for case 1 waters around the globe and in the eastern South
Pacific Gyre (Huot et al., 2008), but agrees with Reynolds et al. (2001)
A. Ferreira et al. Deep-Sea Research Part II xxx (xxxx) xxx–xxx

for the Ross Sea. This shows that particulate assemblages in the Ross
Sea and in our study region tend to backscatter less light than in the
Antarctic Polar Front Zone (Reynolds et al., 2001). The scatter around
bbp(λ) vs. [TChl a] observed in our study is somewhat larger if we
consider other statistical dependences of backscattering and chlor-
ophyll-a concentration found elsewhere in the global ocean (e.g.,
Antoine et al., 2011; Huot et al., 2008; Reynolds et al., 2016, 2001; but
see Loisel et al., 2001). In general, phytoplankton communities during
the surveys, estimated by pigment composition, were dominated by
diatoms and cryptophytes in the pico- and nanophytoplankton size
fraction (Ferreira et al., in press). This implies that phytoplankton of
small sizes were not major contributors to bbp(λ) considering this da-
taset. One hypothesis for low bbp(λ) and poor covariation with [TChl a]
would be that photoadapted phytoplankton cells are poor backscatters,
as phytoplankton adapted to low light are expected to have low carbon
to chlorophyll ratios (Antoine et al., 2011).
The covariation of bbp(442) with adet(443) (Fig. 5b) is considerably
stronger than that observed with [TChl a]. A linear model better de-
scribed the dependence with the following equation (bbp(442) = 0.320
x adet(443) + 0.0007, n = 104, R2 = 0.78; p < 0.00001; L > 37% and
L37–10%), compared to a power law (R2 = 0.46 for log-transformed
data). Lower adet(443) (by half) occurred in L < 10%, and this layer was
not considered in the statistical adjustment for consistency with the fit
for bbp(442) vs. [TChl a]. About 78% of the variability in bbp(442) is
explained by adet(443), against 32% by [TChl a] (Fig. 5a) and 45% for
apart(443) (adet(443) + aph(443), not shown). This is somewhat
intriguing if we consider that detritus is too low in Antarctica and not
covaried with [TChl a] (Fig. 3a), which is the main variable considered
for parametrizations of bbp(λ) for bio-optical modeling. Modeling re-
sults suggest that particles < 0.1 μm are responsible for the majority of
backscattering the oceans (see revision of Stramski et al., 2004). In
particular, these particles include small-sized non-living particles with
perhaps sizable contributions of heterotrophic microbes (Morel and
Ahn, 1991; Stramski and Kiefer, 1991). The bacteria activity in Ant-
arctica is generally high (e.g., Delille, 2004; Thomson et al., 2010) and
uncoupled from phytoplankton populations (Karl et al., 1996).
All considerations above and the reasonable agreement of adet(443)
and bbp(λ) (Fig. 5b) draw the conclusion that detritus is the main
component for bbp(λ) in the Antarctic Peninsula, despite of the low
adet(443) compared to phytoplankton absorption (Fig. 2) and [TChl a]
(Fig. 3a). This is an interesting result that contributes to the under-
standing of backscattering sources not only in Antarctica but also in the
global ocean. While detritus is negligible for non-water absorption
(Fig. 2c), it is the main contributor to bbp(λ) in Antarctic waters. Even
for intermediate to high [TChl a] found here (0.019–2.91 mg m−3), the
very low levels of detritus surpass the effect of phytoplankton biomass
in bbp(λ). The fact that bbp(λ) in not well correlated to [TChl a] partially
contrasts to numerical simulations (e.g., Bernard et al., 2009), labora-
tory measurements (e.g., Vaillancourt et al., 2004; Whitmire et al.,
2010) and field observational studies (e.g., Dall’Olmo et al., 2009,
Antoine et al., 2011) that have suggested phytoplankton may be a
significant contributor to bbp(λ). General better agreements have been

Fig. 6. (a) The particulate scattering coefficient at 650 nm, bp(650), as a function of total chlorophyll-a concentration, [TChl a], in L > 37%, L37–10% and L < 10%. The best fit obtained for all
layers in the form of a power law is also shown, along with the power laws provided by Loisel and Morel (1998), LM98, and Huot et al. (2008), H08, and the linear relationships derived
by Figueroa (2002), F02GR and F02Br. (b) bp(650) as a function of detritus absorption coefficient at 443 nm, adet(443). (c) The particulate backscattering ratio at 650 nm, bbp(650)/
bp(650), as a function of [TChl a], along with the best power law fit considering all the layers and the empirical relationships derived by Morel and Maritorena (2001), MM01,
Twardowski et al. (2001), T01, Whitmire et al. (2007), W07, and Huot et al. (2008), H08. (d) bbp(650)/bp(650) as a function of adet(443) and the best linear fit for L37% and L37–10% all
together.

7
A. Ferreira et al.
Fig. 7. (a) The diffuse attenuation coefficient at 443 nm, Kd(443), as a function of total chlorophyll-a concentration, [TChl a], in L > 37%. The best linear fit and the power law developed
by Morel and Maritorena (2001), MM01, are also shown. (b) The non-water diffuse attenuation coefficient, Kdnw(443) (Kd(443) subtracted the water value of 0.00885 m−1) as a function
of the non-water absorption, anw(443), which refers to aph(443) + acdom(443) + adet(443), following a 1:1 line (dashed line). The inset graph shows the Kdnw(λ) spectra in the
400–560 nm range.
found for bbp(λ) against particulate organic carbon, POC, than for
chlorophyll-a (e.g., Stramski et al., 2008; Reynolds et al., 2016). It
seems, therefore, that reasonable parametrizations for bio-optical
modeling of bbp(λ) can be conducted using detritus rather than [TChl a]
in the northern Antarctic Peninsula waters. This aspect needs to be
investigated in other areas of the Southern Ocean and global ocean.
3.3. Particulate scattering coefficient
Fig. 6a shows bp(λ) as a function of [TChl a] at 650 nm. This wa-
velength is analyzed in most of studies that rely on beam attenuation
measurements, cp(λ), to estimate particulate scattering, because the
non-water absorption is considered negligible at this spectral portion. A
good agreement between bp(650) and [TChl a] is observed and a power
law provided the best statistical fit in the form of 0.0313 x [TChl
a]0.7136 (R2=0.79 for the log-transformed data; n=455, p < 0.0001).
Differently to that for bbp(λ), the L < 10% data followed the same trend
as in L > 37% and L37–10% and were included to the fit (see Loisel and
Morel, 1998). Our data is above the relationship provided by Loisel and
Morel (1998), but are rather well represented by the equation proposed
by Huot et al. (2008). Intriguingly, the relationships found by Figueroa
(2002) in areas of the northern Antarctic Peninsula (Gerlache - F02GR -
and Bransfield - F02Br – Straits) are much above our data and less
dependent on chlorophyll-a variations.
Interestingly, no pattern is observed between bp(650) and adet(443)
(Fig. 6b). This is consistent with our results that detritus is the main
responsible for bbp(λ), while mostly of bp(λ) can be attributed to algal
cells. Theoretical simulations for homogeneous spherical particles (i.e.,
Mie theory) predict that bp(λ), cp(λ) and bbp(λ) are influenced by the
size of the particles (Morel and Ahn, 1991; Stramski and Kiefer, 1991).
As a specific example, Mie theory predicts that 50% of bp(λ) is due to
particles < 3.2 μm, while 50% of bbp(λ) is from particles < 0.2 μm, but
these proportions depend on the refraction index of particles (Stramski
and Kiefer, 1991). Thus, the relationship between bbp(λ) and [TChl a]
may not be as strong as that for bp(λ) and [TChl a] and would depend
on the conservative nature of the particle size distribution (PSD) (i.e.,
that small particles covary with phytoplankton-sized particles). Any
covariation of bbp(λ) and chlorophyll-a concentration as generally ob-
served (e.g., Antoine et al., 2011; Dall’Olmo et al., 2009; Huot et al.,
2008) would be due to covariation between chlorophyll-a and particles
that contribute directly to bbp(λ). The uncoupling between adet(443)
and [TChl a] (Fig. 3a), the poor and good agreements between bbp(λ)
and [TChl a] (Fig. 5a) and between bbp(λ) and adet(443) (Fig. 5b), re-
spectively, and the good relationship between bp(λ) and [TChl a]
(Fig. 6a) reinforces the conclusion that detritus is the main portion of
8
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
the particles pool contributing to bbp(λ) in the Antarctic Peninsula,
while phytoplankton cells (dominant in the pool of particles) mostly
contribute to bp(λ). Our results contrasts to the good agreement be-
tween bp(λ) and bbp(λ) observed in diverse open ocean environments
reported by Westberry et al. (2010).
3.4. Particulate backscattering ratio
The lack of a clear pattern between the backscattering ratio at
650 nm, bbp(650)/bp(650), and [TChl a] is not unexpected (Fig. 6c), as
phytoplankton is not the major component of the bbp(λ) but contribute
efficiently to bp(λ). A power law was fitted considering the data from all
three layers combined and provided: bbp(650)/bp(650) = 0.0037 x
[TChl a]−0.405 (n = 219, R2 = 0.48, p < 0.0001). No clear pattern is
observed between bbp(650)/bp(650) and adet(443) considering the three
layers (Fig. 6d), but a significant linear dependence was observed for
L37% and L37–10%: bbp(650)/bp(650) = 0.423 x adet(443) + 0.002 (n =
103, R2 = 0.56, p < 0.0001). Intuitively, if the bbp(λ) values observed
here are lower than normally in the other oceanic waters, which is not
the case for bp(λ) (Fig. 6a), than bbp(650)/bp(650) values will stay
below the observations found elsewhere. It is noteworthy, however,
that bbp(650)/bp(650) in L < 10% approximates to values found in other
studies (Fig. 6c). One explanation may be that the detritus in depth is
mostly composed of sunken mineral material and/or very small parti-
cles, resulting thus in higher backscattering ratios (e.g. Boss et al.,
2004).
The bbp(λ)/bp(λ) values observed in our data fall within the range
indicative of the “soft,” water-filled organic particles one would expect
in the open ocean (Twardowski et al., 2001). In fact, Twardowski et al.
(2006) concluded that the bbp(λ)/bp(λ) in particle assemblages domi-
nated by phytoplankton converge on a value around 0.5%, which clo-
sely agrees with the measurements here. Indeed, our data is consistent
to observations of bbp(λ)/bp(λ) to be spectrally flat (not shown) in as-
sociation to “soft” particles (e.g., Morel and Maritorena, 2001;
Whitmire et al., 2007).
3.5. Diffuse attenuation coefficient
Values of Kd(λ) at 443 nm, Kd(443) were linearly correlated with
[TChl a] (Kd(443) = 0.0449 x [TChl a] + 0.0384; n = 17, R2=0.90;
p < 0.00001, Fig. 7a). Interestingly, the offset obtained from our fit
(0.0384 m−1) minus the absorption coefficient of pure water described
in Pope and Fry (1997) of 0.00885 m−1 is 0.030 m−1, identical to the
mean acdom(443) found in this study. This implies that for extreme low
[TChl a], Kd(λ) is driven by light absorption of water and cdom (i.e.,
A. Ferreira et al.
detritus contribution to absorption is too small). The best fit obtained
with linear dependence, rather than a power law as that suggested by
Morel and Maritorena (2001), is probably due the small number of
Kd(λ) measurements (n = 17), unable to detect the nonlinearity of
Kd(λ) vs. [TChl a].
The Kd(λ) values for a given [TChl a] are appreciably below those
derived by the relationship of Morel and Maritorena (2001). In addi-
tion, no dependence of Kd(λ) on bbp(λ) was observed (not shown) and
Kd(443) minus the assumed diffuse attenuation coefficient due to water,
Kdbio(443), and non-water absorption coefficient, anw(443), are highly
correlated almost at a 1:1 correspondence (n = 15, Fig. 7b). This is
intuitive considering the theoretical relationship Kd(λ) = [a(λ) +
bb(λ)] / μd(λ) (where a(λ), bb(λ) and μd(λ) corresponds to total ab-
sorption, scattering and the average cosine for downward irradiance,
respectively, Kirk, 1994) and the very low magnitude of bbp(λ) relative
to that for anw(λ) observed here. Similarly, Organelli et al. (2017) found
Kd(λ) values generally lower for a given chlorophyll-a concentration in
the Southern Ocean (not including the Antarctic Peninsula). The au-
thors attributed these observations to the presence of a higher pigment
packaging and low specific absorption coefficients of phytoplankton
communities in the Southern Ocean than observed in other areas, in
association to large phytoplankton cells. On contrary, the phyto-
plankton communities during our surveys were mainly composed by
small cells and specific absorption coefficients of phytoplankton similar
to values found worldwide (Ferreira et al., in press), supporting low
bbp(λ) as the main justification of low Kd(λ) values in our dataset.
Two important points can be made from our results: 1) it corrobo-
rates the low levels of bbp(λ) observed in this study with independent
measurements, and 2) it reveals that measurements of Kd(λ) in the
sampled waters correspond essentially to total absorption. In phyto-
plankton-dominated waters, it essentially reveals the phytoplankton
absorption spectrum (inset in Fig. 7b). This is interesting because Kd(λ)
is obtained promptly in the field, compared to the very laborious
measurements of absorption at laboratory, often requiring different
instruments for particles and cdom. It also advantages absorption
measurements obtained with sensors such as AC-9 and AC-s, as Kd(λ)
uncertainties are less sensitive to calibration issues.
3.6. Remote sensing reflectance
Radiometric measurements in each station were mostly performed
under complete or partially overcast skies, typical conditions in
Antarctica. Note that the consequent spectral variations are minimized
by normalization of water leaving radiance by above-water irradiance
to obtain Rrs(λ). A small part of the variability seen in Rrs(λ) spectral
data can, therefore, be explained by the various illumination conditions
encountered. Overall, the Rrs(λ) data set is characterized by important
changes in magnitude (Fig. 8), for surface [TChl a] varying from 0.12 to
3.03 mg m−3. Two Rrs(λ) spectra considerably depart from the main
shape with the highest values in the blue, in association to the lowest
[TChl a] (0.120 and 0.137 mg m−3). A common spectral feature is
observed: Rrs(λ)a primary peak near 400 nm and secondary peak
around 500 nm, consistent with the maximum phytoplankton absorp-
tion in the blue range. Indeed, the spectra with the lowest magnitudes
(higher [TChl a]) exhibit pronounced shoal at around 440 nm, mir-
roring the phytoplankton absorption spectrum. Although noisy, the red
end of the visible spectra reveals the chlorophyll-a fluorescence, cen-
tered at about 683 nm in most spectra. Not surprisingly, no features as
maximums near 570 nm are observed, which would suggest an im-
portant role of bbp(λ) (e.g. Lubac and Loisel, 2007). The effect of cdom
absorption is not evidenced in the blue. As expected from the lack of
relationships between bbp(λ) and Kd(λ), no dependence was found be-
tween bbp(λ) and Rrs(λ) in any wavelength. This contrasts to the well
covariation of both variables observed by Stramski et al. (1999) for the
Southern Ocean. The authors stated that the covariation of bbp(λ) and
Rrs(λ) implies that backscattering and absorption coefficients are well
9
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
0.01
0.008
Rrs (sr )
0.006
-1
0.004
0.002
0
350 400 450 500 550 600 650 700
Wavelength (nm)
Fig. 8. Spectra of the remote sensing reflectance, Rrs(λ), within the 350 and 700 nm
range. The spectra associated to the lowest [TChl a] (0.120 and 0.137 mg m−3) are shown
as dashed lines.
correlated. In fact, anw(λ) and bbp(λ) in individual spectral bands were
not well correlated in our study (not shown), consistent with our ob-
servation that acdom(λ) and adet(λ) do not covary with [TChl a] as
bbp(λ) does (although not in a well-defined function).
Despite the rigorous processing of the data and the resulting con-
sistent shapes of Rrs(λ), some small features are observed, mainly below
400 and above 600 nm. These features were probably a result of loca-
lized noisy data combined to the smoothing procedures used in each
spectrum, and will not be discussed in here. The applied corrections for
radiometric data acquired above water (Section 2.6) may also promote
changes in magnitudes. While algorithms relying on spectral ratios of
reflectance or water-leaving radiance are probably less sensitive to
these errors, closure exercises may be challenging in those cases, but
may be essential to verify the consistency of measured IOPs with those
estimated by the inversion of the measured light field properties.
For consistency with previous studies, the best agreement between
[Chl a] (instead of [TChl a], and Rrs(λ) ratios was observed for
Rrs(490)/Rrs(555) (Fig. 9a), as similarly observed in previous studies
aimed to develop regional bio-optical models for the Southern Ocean
(Dierssen and Smith, 2000; Garcia et al., 2005). Our data fall very close
to those predicted by algorithms proposed for the Southern Ocean,
considering both the Rrs(490)/Rrs(555) (Fig. 9a) and Rrs(bluemax)/
Rrs(green) (Fig. 9b) ratios (see Fig. 9's caption). Therefore, [Chl a] va-
lues for a given Rrs(λ) ratio are above the global algorithms. This de-
viation, commonly observed for the Southern Ocean, was presumed to
be due to higher chlorophyll content per cell of phytoplankton in re-
sponse to low light (Dierssen and Smith, 2000; Mitchell and Holm-
Hansen, 1991; Szeto et al., 2011) and also due to the low backscattering
(Dierssen and Smith, 2000). Our previous work, however, showed that
in this dataset the absorption coefficients of phytoplankton and con-
centration of accessory pigments do not differ from the global trends
(Ferreira et al., in press). Our results of bbp(λ) distinctly low and the
deviations of reflectance ratios and [Chl a] from global trends (Fig. 9)
may partially agree with the sensitivity analysis that investigated the
impact of variations in the inherent optical properties on the maximum
band ratio of blue-to-green reflectance for the Southern Ocean (IOCCG,
2015). Their results showed that cdom absorption and bbp(λ) could
largely impact (> 70%) changes in the reflectance ratios, while phy-
toplankton has a significant but mostly constant influence (50%). Since
the phytoplankton absorption for our samples did not differ from global
trends (Ferreira et al., in press) and cdom absorption was relatively
much less variable (Fig. 4), low bbp(λ) is likely the primary cause ex-
plaining lower reflectance ratios for a given [Chl a] in our dataset
compared to the global algorithms (Fig. 9).
Another issue that deserves attention is the in situ [Chl a] data. Some
A. Ferreira et al. Deep-Sea Research Part II xxx (xxxx) xxx–xxx

100 100
a b

10 10
[Chl a] ( mg m )

[Chl a] ( mg m )
-3

-3
1 1

0.1 0.1
This study Szeto2011
FURG OC2L SPGANT-GLIv3
DS 2000 Johson2013
0.01 0.01
0.1 1 10 0.3 1 10
R (490)/R (555) R (blue )/R (green)
rs rs rs max rs

Fig. 9. (a) The chlorophyll-a concentration, [Chl a] as a function of the Rrs(490)/Rrs(555) ratio, and the best linear fit in the form (log10([Chl a]) = −2.121 x log10(Rrs(490)/Rrs(555)) +
0.5709, R2=0.62, n=51, p < 0.00001). The grey points in (a) and (b) represent the NASA's Bio-optical Marine Algorithm Data Set (NOMAD; Werdell and Bailey, 2005). The linear fits of
Garcia et al. (2005) (FURG OC2L) and Dierssen and Smith (2000) (DS 2000) are also shown. (b) [Chl a] as a function of the ratio between the maximum Rrs(λ) among the blue
wavelengths, Rrs(bluemax), and the green wavelength, Rrs(green). The fourth-polynomials of Szeto et al. (2011) (Szeto, 2011) and Mitchell and Kahru (2009) (SPGANT-GLIv3) and the
third-polynomial for MODIS-Aqua (their Eq. 5) of Johnson et al. (2013) (Johson2013) are also shown. The maximum band ratio in our study ranged between 1 and 3 (excluding the
outlying data points of 3.19 and 3.44 collected in NWWHT), thus the Mitchell and Kahru (2009)’s algorithm blending was not applied (see Kahru and Mitchell, 2010).

studies have shown that fluorescence-derived chlorophyll-a can be
overestimated up to 2-fold in the Southern Ocean (e.g., Marrari et al.,
2006) when compared to HPLC. Thus, it is important to note that all
studies considered for comparison here (Fig. 9), excluding Johnson
et al. (2013), used fluorescence-derived chlorophyll-a. The slightly de-
part of our data and Garcia et al. (2005)’s equation could be associated
with differences in the phytoplankton communities. Indeed, the two
lowest [Chl a] values (0.11 and 0.13 mg m−3) (Fig. 1, NWWHT) de-
parting from the main trend (Figs. 9a and 9b), contained a very distinct
phytoplankton community (composed by Phaeocystis antarctica) and
spectral absorption (Ferreira et al., in press). These results reveal that,
despite the proximity of those two stations and the Antarctic Peninsula
(Fig. 1), the bio-optical properties of the northern Weddell Sea differ
considerably and the usage of distinct algorithms among regions in the
proximities of Antarctic Peninsula is likely necessary.
Our Rrs(λ) dataset is not large enough to have sufficient re-
presentation of the northern Antarctic Peninsula (n=51), but it is no-
teworthy the agreement of our data to previous studies in the Southern
Ocean and, specifically, in the Antarctic Peninsula. Also, from our
knowledge, this is the first report of Rrs(λ) in Antarctic waters using
Satlantic SAS sensors. In fact, Zeng et al. (2016) measured Rrs(λ) above
water around the Antarctica Peninsula using an ASD Fieldspec Hand-
Held 2, showing less satisfactory agreements among Rrs(λ) ratios and
[TChl a] than those observed here.
Despite being out the scope of our study, a complementary ex-
planation for the lack of a well-defined function between bbp(λ) and
Rrs(λ) is the influence of inelastic radiative processes, such as Raman
scattering by water molecules and fluorescence by colored dissolved
organic matter and chlorophyll-a, which is not generally considered in
the equation Rrs(λ) ~ bbp(λ)/a(λ), but shown to have an influence on
ocean color inverse models (e.g., Westberry et al., 2013; McKinna et al.,
2016). Further efforts to elucidate their role in the optical properties are
recommended (e.g., Lee and Huot, 2014; Li et al., 2016).
4. Conclusions
Phytoplankton contribute to more than 60% of the total non-water
absorption in the blue wavelengths in the northern Antarctic Peninsula,
which may be classified as case 1 waters. The cdom absorption can be
equal to that of phytoplankton in the blue wavelengths, detritus has
very low spectral absorption and none covaries with chlorophyll-a
concentration. Even with very low absorption, detritus was the com-
ponent that better correlated with particulate backscattering, while
phytoplankton cells (dominant in the particles pool) mostly covaried
with particulate scattering. This is an interesting result that contributes
to the comprehension of the backscattering sources not only in
Antarctica, but also other regions of the global ocean.
The very low values of backscattering as previously found in
Antarctica waters was confirmed and corroborated using independent
measurements of the diffuse attenuation coefficient, which corre-
sponded almost 1 to 1 to the absorption coefficient. The shape of the
spectral remote sensing reflectance varied mainly to variability in [TChl
a], while [TChl a] vs. maximum band ratios dependence deviated from
global trends in a very similar fashion as in other studies of the
Southern Ocean, likely due to very low bbp(λ). The present study con-
stitutes the most comprehensive description of concurrent apparent and
inherent optical properties measured in the Antarctic Peninsula and
contributes for further bio-optical studies and ocean color applications.
Acknowledgments
This work is part of a multidisciplinary program under the GOAL
(Brazilian High Latitudes Oceanography Group) activities in the
Brazilian Antarctic Program (PROANTAR). Financial support was pro-
vided by CNPq (National Council for Research and Development) for
the projects POLARCANION (CNPq Grant 556848/2009-8) and PRO-
OASIS (CNPq Grant 565040/2010-3). Amabile Ferreira was supported
by FAPESP (Research Support Foundation of the State of São Paulo,
Grants 2013/16957-2 and 2016/01601-6). Áurea M. Ciotti (PQ
310985/2013-7) and Carlos A. E. Garcia (PQ 311943/2015-2) thank
CNPq for the research fellowship. The analysis of samples for algal
pigments was performed at FURG. The authors thank the crew of the
Polar Vessel “Almirante Maximiano” of the Brazilian Navy and several
investigations on the cruises for assistance. We acknowledge the help
provided by Bianca Tocci and Maria Fernanda C. Giannini during the
field work. The discussions with David Doxaran and Vincenzo Vellucci
on the processing of AC-9 and backscattering data is much appreciated.
Special thanks to Inia Soto that provided the MATLAB code for the SAS
data processing, developed at FURG. This research is a contribution of
the NP-BioMar, University of São Paulo. We acknowledge the two
anonymous reviewers for their valuable contributions that have greatly
improved the manuscript.
References
Antoine, D., Siegel, D.A., Kostadinov, T., Maritorena, S., Nelson, N.B., Gentili, B., Vellucci,

10
A. Ferreira et al.
V., Guillocheau, N., 2011. Variability in optical particle backscattering in contrasting
bio-optical oceanic regimes. Limnol. Oceanogr. 56 (3), 955–973.
Babin, M., Stramski, D., 2004. Variations in the mass-specific absorption coefficient of
mineral particles suspended in water. Limnol. Oceanogr. 49, 756–767.
Babin, M., Stramski, D., Ferrari, G.M., Claustre, H., Bricaud, A., Obolensky, G., Hoepffner,
N., 2003. Variations in the light absorption coefficients of phytoplankton, non-algal
particles, and dissolved organic matter in coastal waters around Europe. J. Geophys.
Res. 108 (3211), 3210 (1029/2001JC000882).
Bartz, R., Zaneveld, Z.R.V., Park, H., 1978. A transmissometer for profiling and moored
observations in water. Proc. SPIE 160, 102–108 in Ocean Optics V, M. B. White, ed.
Behrenfeld, M.J., Boss, E., 2006. Beam attenuation and chlorophyll concentration as al-
ternative optical indices of phytoplankton biomass. J. Mar. Res. 64, 431–451.
Behrenfeld, M.J., Boss, E., Siegel, D.A., Shea, D.M., 2005. Carbon-based ocean pro-
ductivity and phytoplankton physiology from space. Glob. Biogeochem. Cycles 19,
GB1006. http://dx.doi.org/10.1029/2004GB002299.
Bernard, S., Probyn, T.A., Quirantes, A., 2009. Simulating the optical properties of phy-
toplankton cells using a two-layered spherical geometry. Biogeosci. Discuss. 6,
1497–1563.
Boss, E., Collier, R., Larson, G., Fennel, K., Pegau, W.S., 2007. Measurements of spectral
optical properties and their relation to biogeochemical variables and processes in
Crater Lake National Park, OR. Hydrobiologia 574, 149–159.
Boss, E., Pegau, W.S., Lee, M., Twardowski, M.S., Shybanov, E., Korotaev, G., Baratange,
F., 2004. The particulate backscattering ratio at LEO-15 and its use to study particle
composition and distribution. J. Geophys. Res. 109, C0101410 (1029/
2002JC001514).
Bowers, D.G., Harker, G.E.L., Stephan, B., 1996. Absorption spectra of inorganic particles
in the Irish Sea and their relevance to remote sensing of chlorophyll. Int. J. Remote
Sens. 17 (12), 2449–2460.
Bricaud, A., Babin, M., Claustre, H., Ras, J., Tièche, F., 2010. Light absorption properties
and absorption budget of Southeast Pacific waters. J. Geophys. Res. 115, C08009.
http://dx.doi.org/10.1029/2009JC005517.
Bricaud, A., Claustre, H., Ras, J., Oubelkheir, K., 2004. Natural variability of phyto-
plankton absorption in oceanic waters: influence of the size structure of algal po-
pulations. J. Geophys. Res. 109, C11010. http://dx.doi.org/10.1029/2004JC002419.
Bricaud, A., Ciotti, A.M., Gentili, B., 2012. Spatial-temporal variations in phytoplankton
size and colored detrital matter absorption at global and regional scales, as derived
from twelve years of SeaWiFS data (1998–2009). Glob. Biogeochem. Cycles 26,
GB1010.
Bricaud, A., Morel, A., Babin, M., Allali, K., Claustre, H., 1998. Variations of light ab-
sorption by suspended particles with chlorophyll a concentration in oceanic (case 1)
waters: analysis and implications for bio-optical models. J. Geophys. Res. 103,
31,033–31,044.
Bricaud, A., Babin, M., Morel, A., Claustre, H., 1995. Variability in the chlorophyll spe-
cific absorption coefficients of natural phytoplankton: analysis and parameterization.
J. Geophys. Res. 100, 13,321–13,332.
Bricaud, A., Morel, A., Prieur, L., 1981. Absorption by dissolved organic matter of the sea
(yellow substance) in the UV and visible domains. Limnol. Oceanogr. 26, 43–53.
Brown, C.A., Huot, Y., Werdell, P.J., Gentili, B., Claustre, H., 2008. The origin and global
distribution of second order variability in satellite ocean color and its potential ap-
plications to algorithm development. Rem. Sens. Environ. 112, 4186–4203.
Bukata, R.P., Jerome, J.H., Kondratyev, K.Y., Pozdnyakov, D.V., 1995. Optical Properties
and Remote Sensing of Inland and Coastal Waters. CRC Press, London.
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 de-
gradation products. J. Geophys. Res. 96 (C11), 20,599–20,611.
Cetinić, I., Perry, M.J., D'Asaro, E., Briggs, N., Poulton, N., Sieracki, M.E., Lee, C.M., 2015.
A simple optical index shows spatial and temporal heterogeneity in phytoplankton
community composition during the 2008 North Atlantic Bloom Experiment.
Biogeosciences 12, 2179–2194.
Ciotti, A.M., Cullen, J.J., Lewis, M.R., 1999. A semi-analytical model of the influence of
phytoplankton community structure on the relationship between light attenuation
and ocean color. J. Geophys. Res. 104 (C1), 1559–1578.
Dall’Olmo, G., Westberry, T.K., Behrenfeld, M.J., Boss, E., Slade, W.H., 2009. Significant
contribution of large particles to optical backscattering in the open ocean.
Biogeosciences 6, 947–967.
Del Vecchio, R., Blough, N.V., 2004. On the origin of the optical properties of humic
substances. Environ. Sci. Technol. 38, 3885–3891.
Delille, D., 2004. Abundance and function of bacteria in the Southern Ocean. Cell. Mol.
Biol. 50, 543–551.
Dierssen, H., Randoph, K., Garaba, S., 2017. Interpreting backscattering in the Southern
Ocean, International Ocean Colour Science Meeting 2017, Lisbon, Portugal.
Dierssen, H.M., 2010. Perspectives on empirical approaches for ocean color remote sen-
sing of chlorophyll in a changing climate. Proc. Nat. Acad. Sci. 107, 17073–17078.
Dierssen, H., Smith, R., Vernet, M., 2002. Glacial meltwater dynamics in coastal waters
west of the Antarctic Peninsula. Proc. Natl. Acad. Sci. USA 99, 1790–1795.
Dierssen, H.M., Smith, R.C., 2000. Bio-optical properties and remote sensing ocean color
algorithms for Antarctic Peninsula waters. J. Geophys. Res. 105, 26301–26312.
Doxaran, D., Leymarie, E., Nechad, B., Dogliotti, A., Ruddick, K., Gernez, P., Knaeps, E.,
2016. Improved correction methods for field measurements of particulate light
backscattering in turbid waters. Opt. Express 24 (4), 3615–3637.
Estapa, M.L., Boss, E., Mayer, L.M., Roesler, C.S., 2012. Role of iron and organic carbon in
mass-specific light absorption by particulate matter from Louisiana coastal waters.
Limnol. Oceanogr. 57, 97–112.
Ferreira, A., Ciotti, Á.M., Mendes, C.R.B., Uitz, J., Bricaud, A., 2017. Phytoplankton light
absorption and the package effect in relation to photosynthetic and photoprotective
pigments in the northern tip of Antarctic Peninsula. J. Geophys. Res. 122. http://dx.
11
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
doi.org/10.1002/2017JC012964.
Ferreira, A., Ciotti, Á.M., Giannini, M.F.C., 2014. Variability in the light absorption
coefficients pf phytoplankton, non-algal particles, and colored dissolved organic
matter in a subtropical bay (Brazil). Estuar. Coast Shelf Sci. 139, 127–136.
Figueroa, F.L., 2002. Bio-optical characteristics of Gerlache and Bransfield strait waters
during an Antarctic cruise. Deep Sea Res. Part II 49, 675–691.
Garcia, C.A.E., Garcia, V.M.T., McClain, C., 2005. Evaluation of SeaWiFS chlorophyll
algorithms in the Southwestern Atlantic and Southern Oceans. Remote Sens. Environ.
95, 125–137.
Goldstone, J.V., Del Vecchio, R., Blough, N.V., Voelker, B.M., 2004. A multicomponent
model of chromophoric dissolved organic matter photobleaching. J. Photochem.
Photobiol. 80, 52–60.
Gordon, H.R., Morel, A., 1983. In: Barber, R.T., Mooers, C.N.K., Bowman, M.J.,
Zeitzschel, B. (Eds.), Remote Assessment of Ocean Color for Interpretation of Satellite
Visible Imagery: A Review. Springer- Verlag, New York.
Graff, J.R., Westberry, T.K., Milligan, A.J., Brown, M.B., Dall’Olmo, G., Reifel, K.M.,
Behrenfeld, M.J., 2016. Photoacclimation of natural phytoplankton communities.
Mar. Ecol. Prog. Ser. 542, 51–62.
Graff, J.R., Westberry, T.K., Milligan, A.J., Brown, M.B., et al., 2015. Analytical phyto-
plankton carbon measurements spanning diverse ecosystems. Deep Sea Res. Part I
102, 16–25.
Huot, Y., Morel, A., Twardowski, M.S., Stramski, D., Reynolds, R.A., 2008. Particle optical
backscattering along a chlorophyll gradient in the upper layer of the eastern South
Pacific Ocean. Biogeosciences 5, 495–507.
IOCCG - Internacional Ocean-Color Coordinating Group, 2015. Ocean colour remote
sensing in polar seas. In: Babin, M., Arrigo, K., Belanger, S., Forget, M.-H. (Eds.),
IOCCG Report Series, No. 16. IOCCG, Dartmouth, Canada.
IOCCG - Internacional Ocean-Color Coordinating Group, 2006. Remote sensing of ocean
colour in Coastal, and other optically-complex, waters. In: Lee, Z.-P. (Ed.), Reports of
the International Ocean-Colour Coordinating Group, No. 5. IOCCG, Dartmouth,
Canada.
IOCCG - Internacional Ocean-Color Coordinating Group, 2000. Remote sensing of in-
herent optical properties: fundamentals, tests of algorithms, and applications. In:
Sathyendranath, S. (Ed.), Reports of the International Ocean-Colour Coordinating
Group, No. 3. IOCCG, Dartmouth, Canada.
Iturriaga, R., Siegel, D.A., 1989. Microphotometric characterization of phytoplankton and
detrital absorption properties in the Sargasso Sea. Limnol. Oceanogr. 34, 1706–1726.
Johnson, K.S., Berelson, W.M., Boss, E., Chase, Z., Claustre, H., Emerson, S.R., Gruber, N.,
Körtzinger, A., Perry, M.J., Riser, S.C., 2009. Observing biogeochemical cycles at
global scales with profiling floats and gliders: prospects for a global array.
Oceanography 22, 216–225.
Kahru, M., Mitchell, B.G., 2010. Blending of ocean colour algorithms applied to the
Southern Ocean. Remote Sens. Lett. 1, 119–124.
Karl, D.M., Christian, J.R., Dore, J.E., Letelier, R.M., 1996. Microbiological oceanography
in the region west of the Antarctic Peninsula: microbial dynamics, nitrogen cycle and
carbon flux. In: Ross, R.M., Hofmann, E.E., Quetin, L.B. (Eds.), Foundations for
Ecological Research West of the Antarctic Peninsula. American Geophysical Union,
Washington, D. C.
Kirk, J.T.O., 1994. Light and Photosynthesis in Aquatic Ecosystems (509 pp.). Cambridge
University Press, New York.
Kishino, M., Takahashi, M., Okami, N., Ichimura, S., 1985. Estimation of the spectral
absorption coefficients of phytoplankton in the sea. Bull. Mar. Sci. 37 (2), 634–642.
Kostadinov, T.S., Siegel, D.A., Maritorena, S., 2009. Retrieval of the particle size dis-
tribution from satellite ocean color observations. J. Geophys. Res. 114, CO9015.
Lee, Z., Huot, Y., 2014. On the non-closure of particle backscattering coefficient in oli-
gotrophic oceans. Opt. Express 22, 29223–29233.
Li, L., Stramski, D., Reynolds, R.A., 2016. Effects of inelastic radiative processes on the
determination of water-leaving spectral radiance from extrapolation of underwater
near-surface measurements. Appl. Opt. 55, 7050–7067.
Loisel, H., Nicolas, J.M., Sciandra, A., Stramski, D., Poteau, A., 2006. Spectral dependency
of optical backscattering by marine particles from satellite remote sensing of the
global ocean. J. Geophys. Res. 111, C09024.
Loisel, H., Bosc, E., Stramski, D., Oubelkheir, K., Deschamps, P.Y., 2001. Seasonal
variability of the backscattering coefficient in the Mediterranean Sea based on sa-
tellite SeaWiFS imagery. Geophys. Res. Lett. 28, 4203–4206.
Loisel, H., Stramski, D., 2000. Estimation of the inherent optical properties of natural
waters from the irradiance attenuation coefficient and reflectance in the presence of
Raman scattering. Appl. Opt. 39, 3001–3011.
Loisel, H., Morel, A., 1998. Light scattering and chlorophyll concentration in Case 1
waters: a reexamination. Limnol. Oceanogr. 43, 847–858.
Lubac, B., Loisel, H., 2007. Variability and classification of remote sensing reflectance
spectra in the eastern English Channel and southern North Sea. Remote Sens.
Environ. 110, 45–58.
McKinna, L.I.W., Werdell, P.J., Proctor, C.W., 2016. Implementation of an analytical
Raman scattering correction for satellite ocean-color processing. Opt. Express 24
(14), A1123–A1137.
Maffione, R.A., Dana, D.R., 1997. Instruments and methods for measuring the backward-
scattering coefficient of ocean waters. Appl. Opt. 36, 6057–6067.
Marrari, M., Hu, C., Daly, K., 2006. Validation of SeaWiFS chlorophyll a concentrations in
the Southern Ocean: a revisit. Remote Sens. Environ. 105 (4), 367–375.
Matsuoka, A., Hill, V., Huot, Y., Babin, M., Bricaud, A., 2011. Seasonal variability in the
light absorption properties of western Arctic waters: parameterization of the in-
dividual components of absorption for ocean color applications. J. Geophys. Res. 116,
C02007.
Meler, J., Kowalczuk, P., Ostrowska, M., Ficek, D., Zabłocka, M., Zdun, A., 2016.
Parameterization of the light absorption properties of chromophoric dissolved
A. Ferreira et al.
organic matter in the Baltic Sea and Pomeranian lakes. Ocean Sci. 12, 1013–1032.
Mendes, C.R.B., de Souza, M.S., Garcia, V.M.T., Leal, M.C., Brotas, V., Garcia, C.A.E.,
2012. Dynamics of phytoplankton communities during late summer around the tip of
the Antarctic Peninsula. Deep-Sea Res., I 65, 1–14.
Mendes, C.R., Cartaxana, P., Brotas, V., 2007. HPLC determination of phytoplankton and
microphytobenthos pigments: comparing resolution and sensitivity of a C18 and a C8
method. Limnol. Oceanogr. Methods 5, 363–370.
Mitchell, B., Holm-Hansen, O., 1991. Bio-optical properties of Antarctic Peninsula waters:
differentiation from temperate ocean models. Deep Sea Res., Part A 38 (8–9),
1009–1028.
Mitchell, B.G., Kahru, M., 2009. Bio-optical algorithms for ADEOS-2 GLI. J. Rem. Sens.
Soc. Jpn. 29, 80–85.
Morel, A., Claustre, H., Antoine, D., Gentili, B., 2007. Natural variability of bio‐optical
properties in case 1 waters: attenuation and reflectance within the visible and
near‐UV spectral domains, as observed in South Pacific and Mediterranean waters.
Biogeosciences 4, 913–925.
Morel, A., Maritorena, S., 2001. Bio-optical properties of oceanic waters: a reappraisal. J.
Geophys. Res. 106, 7163–7180.
Morel, A., Ahn, Y.-H., 1991. Optics of heterotrophic nanoflagellates and ciliates: a ten-
tative assessment of their scattering role in oceanic waters compared to those of
bacterial and algal cells. J. Mar. Res. 49, 177–202.
Morel, A., 1988. Optical modeling of the upper ocean in relation to its biogenous matter
content (case 1 water). J. Geophys. Res. 93, 10,749–10,768.
Morel, A., Prieur, L., 1977. Analysis of variations in ocean color. Limnol. Oceanogr. 22,
709–722.
Nelson, N.B., Siegel, D.A., 2013. The global distribution and dynamics of chromophoric
dissolved organic matter. Annu. Rev. Mar. Sci. 5, 447–476.
Nelson, N.B., Carlson, C.A., Steinberg, D.K., 2004. Production of chromophoric dissolved
organic matter by Sargasso Sea microbes. Mar. Chem. 89, 273–287.
Nelson. N.B., Siegel, D.A. 2002. Chromophoric DOM in the open ocean. Hansell & Carlson
2002, pp. 547–578.
Nelson, N.B., Siegel, D.A., Michaels, A.F., 1998. Seasonal dynamics of colored dissolved
material in the Sargasso Sea. Deep-Sea Res. I 45, 931–957.
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 chlorophyll algorithm for SeaWiFS. J.
Geophys. Res. 103, 24937–24953.
Organelli, E., Claustre, H., Bricaud, A., Barbieux, M., Uitz, J., D’Ortenzio, F., Dall’Olmo,
G., 2017. Bio-optical anomalies in the world's oceans: an investigation on the diffuse
attenuation coefficients for downward irradiance derived from Biogeochemical Argo
float measurements. J. Geophys. Res. 122.
Preisendorfer, R.W., 1961. Application of Radiative Transfer Theory to Light
Measurements in the Sea. 10. Institut géographique national, pp. 11–30.
Prieur, L., Sathyendranath, S., 1981. An optical classification of coastal and oceanic
waters based on the specific spectral absorption curves of phytoplankton pigments,
dissolved organic matter, and other particulate materials. Limnol. Oceanogr. 26,
671–689.
Reynolds, R.A., Stramski, D., Neukermans, G., 2016. Optical backscattering by particles in
Arctic seawater and relationships to particle mass concentration, size distribution,
and bulk composition. Limnol. Oceanogr. 61, 1869–1890.
Reynolds, R.A., Stramski, D., Mitchell, B.G., 2001. A chlorophyll dependent semi-
analytical reflectance model derived from field measurements of absorption and
backscattering coefficients within the Southern Ocean. J. Geophys. Res. 106,
7125–7138.
Sathyendranath, S., Platt, T., 1989. Remote sensing of ocean chlorophyll: consequence of
nonuniform pigment profile. Appl. Opt. 28 (3), 490–495.
Slade, W.H., Boss, E., 2015. Spectral attenuation and backscattering as indicators of
average particle size. Appl. Opt. 54 (24), 7264–7277.
Stramski, D., Reynolds, R., Babin, M., Kaczmarek, S., Lewis, M., Roettgers, R., Sciandra,
A., Stramska, M., Twardowski, M., Franz, B., et al., 2008. Relationships between the
surface concentration of particulate organic carbon and optical properties in the
eastern South Pacific and eastern Atlantic Oceans. Biogeosciences 5 (1), 171–201.
Stramska, M., Stramski, D., 2005. Variability of particulate organic carbon concentration
in the north polar Atlantic based on ocean color observations with Sea-viewing Wide
Field-of-view Sensor (SeaWiFS). J. Geophys. Res. 110, 1–16.
12
Deep-Sea Research Part II xxx (xxxx) xxx–xxx
Stramski, D., Boss, E., Bogucki, D., Voss, K.J., 2004. The role of seawater constituents in
light backscattering in the ocean. Prog. Oceanogr. 61, 27–56.
Stramski, D., Reynolds, R.A., Kahru, M., Mitchell, B.G., 1999. Estimation of particulate
organic carbon in the ocean from satellite remote sensing. Science 285, 239–242.
Stramski, D., Kiefer, D.A., 1998. Can heterotrophic bacteria be important to marine light
absorption? J. Plankton Res. 20, 1489–1500.
Stramski, D., Kiefer, D.A., 1991. Light scattering by microorganisms in the open ocean.
Prog. Oceanogr 28, 343–383.
Swan, C.M., Nelson, N.B., Siegel, D.A., Kostadinov, T.S., 2012. The effect of surface ir-
radiance on the absorption spectrum of chromophoric dissolved organic matter in the
global ocean. Deep-Sea Res. I 63, 52–64.
Szeto, M., Werdell, P.J., Moore, T.S., Campbell, J.W., 2011. Are the worlds oceans opti-
cally different? J. Geophys. Res. 116, C00H04.
Tassan, S., Ferrari, G.M., Bricaud, A., Babin, M., 2000. Variability of the amplification
factor of light absorption by filter-retained aquatic particles in the coastal environ-
ment. J. Plankton Res. 22, 659–668.
Tassan, S., Ferrari, G.M., 1995. An alternative approach to absorption measurements of
aquatic particles retained on filters. Limnol. Oceanogr. 40, 1358–1368.
Thomson, P.G., Davidson, A.T., vandenEnden, R., Pearce, I., Seuront, L., Paterson, J.S.,
Williams, G.D., 2010. Distribution and abundance of marine microbes in the Southern
Ocean between 30 and 80°E. Deep Sea Res. II 57, 815–827.
Twardowski, M.S., Claustre, H., Freeman, S.A., Stramski, D., Huot, Y., 2007. Optical
backscattering properties of the ‘clearest’ natural waters. Biogeosciences 4,
1041–1058.
Twardowski, M.S., Sullivan, J.S., Dierssen, H., Zaneveld, J.R.V., 2006. Sources of back-
scattering in marine waters, Ocean Sciences Meeting, Honolulu, HI, February 20-24,
Abstract OS32M-02.
Twardowski, M.S., Boss, E., Sullivan, J.M., Donaghay, P.L., 2004. Modeling the spectral
shape of absorption by chromophoric dissolved organic matter. Mar. Chem. 89,
69–88.
Twardowski, M., Boss, E., Macdonald, J.B., Pegau, W.S., Barnard, A.H., Zaneveld, J.R.V.,
2001. A model for estimating bulk refractive index from the optical backscattering
ratio and the implications for understanding particle composition in case I and case II
waters. J. Geophys. Res. 106, 14,129–14,142.
Vaillancourt, R.D., Brown, C.W., Guillard, R.R.L., Balch, W.M., 2004. Light backscattering
properties of marine phytoplankton: relationships to cell size, chemical composition
and taxonomy. J. Plank. Res. 26 (2), 191–212.
Vodacek, A., Blough, N.V., DeGrandpre, M.D., Peltzer, E.T., Nelson, R.K., 1997. Seasonal
variation of CDOM and DOC in the Middle Atlantic Bight: terrestrial inputs and
photooxidation. Limnol. Oceanogr. 42, 674–686.
Werdell, P., Bailey, S., 2005. An improved in‐situ bio‐optical data set for ocean color
algorithm development and satellite data product validation. Remote Sens. Environ.
98 (1), 122–140.
Westberry, T.K., Boss, E., Lee, Z., 2013. The influence of Raman scattering on ocean color
inversion models. Appl. Opt. 52, 55525561.
Westberry, T., Dall’Olmo, G., Behrenfeld, M.J., Moutin, T., 2010. Coherence of particulate
beam attenuation and backscattering coefficients in diverse open ocean environ-
ments. Opt. Express 18 (15), 15,419–15,425.
Whitmire, A.L., Boss, E., Cowles, T.J., Pegau, W.S., 2007. Spectral variability of the
particulate backscattering ratio. Opt. Express 15, 7019–7031.
Whitmire, A.L., Pegau, W.S., Karp-Boss, L., Boss, E., Cowles, T.J., 2010. Spectral back-
scattering properties of marine phytoplankton cultures. Opt. Express 18,
15073–15093.
Yentsch, C.S., Reichert, C.A., 1962. The interrelationship between water-soluble yellow
substances and chloroplastic pigments in marine algae. Bot. Mar. 3, 65–74.
Zaneveld, J.R.V., Kitchen, J.C., Moore, C.M., 1994. The scattering error correction of
reflecting-tube absorption meters. Proc. SPIE 2258, 44–55 Ocean optics XII.
Zeng, C., Xu, H., Fischer, A.M., 2016. Chlorophyll-a estimation around the antarctica
peninsula using satellite algorithms: hintshints from field water leaving reflectance.
Sensors 16, 2075.
Zhang, X., Hu, L., He, M., 2009. Scattering by pure seawater: effect of salinity. Opt.
Express 17 (7), 5698–5710.
Zhang, X., Lewis, M., Lee, M.E.-G., Johnson, B., Korotaev, G.K., 2002. The volume scat-
tering function of natural bubble populations. Limnol. Oceanogr. 47, 1273–1282.