Simulation of satellite chlorophyll-a measures using in situ data

Lilian Anne Krug

1


1
Centre of Excellence in Observational Oceanography Nippon Foundation and Partnership
for Observations of the Global Ocean (NF-POGO), Bermuda Institute of Ocean Sciences.
17, Biological Station, Ferry Reach, St. George's GE 01, Bermuda
[email protected]

Abstract. Ocean colour orbital sensors can collect information of the first quarter of the photic zone depth. This
study proposed to calculate the chlorophyll-a concentration for this penetration depth and to compare it with the
satellite product. The photic and penetration depths at the Bermuda Atlantic Time Series (BATS) station were
obtained from the diffuse light attenuation profiles. These, in turn, were based on the inherent optical properties,
absorption and scattering, estimated from in situ chlorophyll-a measurements. The chlorophyll-a concentrations
were calculated for the penetration depths of 440 nm, 550 nm and of the entire photosynthetic active region
(PAR 400-700 nm). The results show good linear correlations between the satellite product and penetration
depth chlorophyll-a concentrations (R
2
> 0.6). Both monthly and weekly SeaWiFS products underestimate in situ
chlorophyll-a when its concentration is higher than 0.15 mg m
-3
, probably due to the algorithm scheme of
chlorophyll-a retrieval. Ten-year time series were constructed to analyze the seasonal patterns. The photic and
penetration depths, as the subsurface chlorophyll-a maximum, tend to become shallower at the beginning of
spring season (February to April) and deepen in summer season (June to July). This pattern might be associated
with phytoplankton blooms and the mixed layer depth (MLD) seasonality.

Palavras-chave: ocean colour, inherent optical properties, penetration depth, cor do oceano, propriedades
pticas inerentes, profundidade de penetrao.

1. Introduction
In the radiometric data received by an orbital optical scanner, the amount of information
on water column relevant to ocean colour studies is impressively low. This occurs because the
sunlight beam is reduced in both directions, down and upwards, both while crossing the
atmosphere and within the ocean.
The attenuation of the light in the water column depends on its inherent optical properties
(IOPs), the absorption and scattering of the water molecules, particulate and dissolved
constituents (Kampel and Novo, 2005). In open waters the phytoplankton may be considerate
the main player on variations of the optical properties, the IOPs, and therefore the light
attenuation, can be expressed as a function of the concentration of its main pigment, the
chlorophyll-a (Sathyendranath et al., 2001).
Due the scarce presence of particulate and dissolved materials in oceanic waters the
photic zone depth the depth where the light is attenuated to 1% of its intensity at the surface
is larger than in coastal waters. Still, satellites can see only the first portion of the photic
zone, the penetration depth, and returns a weighted average value over this layer
(Sathyendranath and Platt, 1989).
The objective of this study is to compare chlorophyll-a concentration estimated by the
satellite with a simulated satellite data, where the concentration at an inferred penetration
depth was estimated based on in situ chlorophyll-a profiles. The seasonal variability of the
chlorophyll-a and the photic and penetration depths are analyzed in a ten-year time series.

2. Methodology
2.1. Photic and penetration depths
In order to obtain the chlorophyll-a concentration at the penetration depth (z
pe
) first is
necessary to estimate the photic depth (z
p
). The photic depth is defined as that depth where
irradiance (E) is reduced to 1% of its value at the surface (Platt et al., 1994), i.e.
) , ( ) , (
01 . 0
surface z
E E
p

= . The irradiance loss at each depth can be determined as:
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7106
7106

) (
) , ( ) , (
1 ) , (
1

=
n n
n
z
n n
z z K
z z
e E E



(1)
where K
(,Zn)
is the light attenuation coefficient at depth n and wavelength .

The attenuation coefficient (K) has the dimensions of length (typically, m
-1
), and the
quantity K
-1
is called the optical depth of the water. The photic zone layer has approximately
4.6 optical depths (Platt and Sathyendranath, 2010):

K
K
E
E
K
z
e
z
surface
e p
6 . 4
100 log
1
log
1
) , (
) , (
~ =
|
|
.
|

\
|
=

(2)

According to Gordon and McCluney (1975), 90% of the upwelling water radiance at sea
surface belongs to the first optical depth, denominated penetration depth. Based on Equation
2, this study assumes that the penetration depth is equivalent to the first 25% of the photic
depth (
4
p
pe
z
z = ).

2.2. Light attenuation, absorption and backscattering coefficients
The light attenuation coefficient can be stated as a function of the inherent optical
properties of the water column, the absorption and backscattering (Sathyendranath et al.,
2001):

d
b
z
n
z
n
z
n
b a
K

) (
) , ( ) , (
) , (
+
=

(3)
where a
(,zn)
is the absorption coefficient for wavelength at depth n, b
b(,zn)
is the
backscattering coefficient for wavelength at depth n and
d
is the mean cosine for
downwelling light.

The absorption coefficient is the joined effect of the absorption of the different elements
present at the water column (Sathyendranath et al., 2001):

( ) ( )
p y w z
a a a a
n
+ + =
) , (
(4)
where a
w
is the absorption coefficient of pure water at wavelength , a
y
is the absorption
coefficient of the yellow substance and a
p
is the particulate material absorption coefficient.
Since this study area is located in an open ocean region, a
p
is assumed to be dominated by
phytoplankton absorption.

While a
w
() is a single wavelength-dependent value (Table 3 in Pope and Fry, 1997), the
other equation components are calculated as (Sathyendranath et al., 2001; Devred et al.,
2006):

) 440 ( 014 . 0
) 440 ( ) (
exp

=

nm y y
a a

(4.1)
where a
y
(440 nm) is the absorption coefficient of the yellow absorption at 440 nm and co-
varies with the phytoplankton absorption at the same wavelength (a
p(440 nm)
):

) 440 ( ) 440 (
3 . 0
nm p nm y
a a =
(4.1.1)

The phytoplankton absorption coefficient was modelled by Sathyendranath et al. (2001)
and improved in Devred et al. (2006). It considers specific absorption coefficients for two
different populations of phytoplankton, distinguished by size and harvest efficiency. It is
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7107
7107
assumed that the total chlorophyll-a concentration (C) is dominated by population 1
chlorophyll-a (C
1
) until C reaches a certain threshold. Higher than this concentration,
population 2 chlorophyll-a turns the dominant component, such that:

| | C a C
C
a a C a
m
m
p
*
) ( 2
1
*
) ( 2
*
) ( 1 1 ) (
)
1
exp( 1

+
(

=

(4.2)
where,
m
C
1
is the maximum value of C
1
,
*
) ( 1
a
and
*
) ( 2
a are specific absorption coefficient
of populations 1 and 2 at wavelength and C is total chlorophyll. From in situ data collected
in several cruises, Devred et al. (2006) computed the parameters
*
) ( 1 1
,

a C
m
and
*
) ( 2
a (Tables 2
and 3 in Devred et al., 2006).

For the backscattering coefficient at wavelength and depth n, Sathyendranath et al.
(2001) present:

( ) ) (
~
) , ( b

p
bp
bw z
b b b b
n
+ =

(5)
where ( )
bw
b is the backscattering coefficient of pure water at wavelength (Equation 5.1),
bp b
~
is the ratio of backscattering and total scattering coefficient by particles (
) (
) (
~

p
bp
bp
b
b
b =
)
and b
p
() is the scattering coefficients of particles (Equation 5.2).

Based on Morel (1974), due to molecules symmetry the backscattering coefficient of pure
water (b
bw
) is 50% of total scattering (b
w
) and is calculated as:
( )
3 . 4
500 ) ( bw
500
5 . 0

|
.
|

\
|
=

nm w
b b

(5.1)
where b
w(500 nm)
is the scattering coefficient of pure water at 500 nm: 0.00288 m
-1
.

The scattering coefficient of particles is also a function of wavelength, defined as (Loisel
and Morel, 1998; Sathyendranath et al., 2001):

C
nm p
b b
10
log
) 660 (
) ( p
660

|
.
|

\
|
=

(5.2)
where C stands for chlorophyll-a concentration and b
p(660nm)
is the scattering coefficient of
particles at 660 nm:
795 . 0
) 660 (
407 . 0 C b
p
=

(5.2.1)

Ulloa et al. (1994) and Sathyendranath et al. (2001) define the backscattering/scattering
ratio for particles as:

) log 42 . 0 78 . 0 ( 01 . 0
10
) ( b
~
C b =
(5.3)

2.3. Chlorophyll-a at the penetration depth
With the light attenuation coefficient (K) profiles calculated, the photic and penetration
depths are estimated (Equations 1 and 2) allowing the calculation of the chlorophyll-a
concentration at the penetration depth (C
pe
). The equation considers that shallower the water
parcel, higher its contribution to the radiance at surface:
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7108
7108

( )
( )
}
}
=
) (
0
) (
0
) (

pe
pe
z
z
z
pe
dz z f
dz z f C
C

(6)
where z
pe
() is the penetration depth (m), C
z
the chlorophyll concentration at depth z and f(z)
is:

}

=
pe
z
dz z K z f
0
2
' ). ' ( exp ) (

(6.1)

The chlorophyll-a concentrations were calculated for the penetration depths of 440 nm, 550
nm and of the entire photosynthetic active region (PAR 400-700 nm).

2.4. In situ chlorophyll-a concentration
The Bermuda Atlantic Time Series (BATS) is located in the western North Atlantic
subtropical gyre (Michaels et al., 1994; Steinberg et al., 2001). Since 1988, the Bermuda
Institute of Ocean Sciences (BIOS) conduct cruises on a biweekly to monthly basis to sample
phytoplankton pigments among others physical, chemical and biological data at BATS station
(3110 N - 6410 W). Sampling methodology is described in Steinberg et al. (2001).
A total of 111 profiles (0 m 250 m) of in situ chlorophyll-a concentration was obtained
from the BATS database (http://www.bats.bios.edu) constituting a ten-years time series
(1997-2007). Typically, 12 samples at different depths cover the first 250 m of the water
column. Here, these values were interpolated to reconstruct profiles with 1 m interval.

2.5. Satellite chlorophyll-a concentration
Averaged concentrations of chlorophyll-a at BATS location were obtained from monthly
and weekly 9 km-resolution products of SeaWiFS for the correspondent period of the in situ
chlorophyll-a time series. The data was acquired from the GES-DISC Interactive Online
Visualization And aNalysis Infrastructure (Giovanni) as part of the NASAs Goddard Earth
Sciences (GES) Data and Information Services Centre
(http://reason.gsfc.nasa.gov/OPS/Giovanni/ocean.seawifs.shtml)
The SeaWiFS chlorophyll-a concentrations are estimated with a fourth order polynomial
equation, where the algorithm OC4v4 makes use of different ratios among blue and green
bands (OReilly et al., 2000). This algorithm was based on empirical relations of in situ water
leaving radiance and chlorophyll-a concentrations in a variety of ocean provinces which have
different optical properties. As older versions, OC4v4 has a considerate good performance in
open ocean waters with chlorophyll-a concentrations between 0.03 mg m
-3
to 1 mg m
-3

(OReilly et al., 2000).

3. Results and discussion
3.1. In situ chlorophyll-a concentration at BATS
For the 10-years time series, BATS average chlorophyll-a concentration at the first 250 m
is 0.1 mg m
-3
with a maximum of 0.7 mg m
-3
(Figure 1), characteristic of low chlorophyll
waters.
Elevated concentrations reach the surface near April and vary from 60 m to 110 m
throughout the rest of the year (Figure 1 and 2). According to Longhurst and Harrison (1989),
the maximum chlorophyll-a concentration occurs near surface during phytoplankton bloom
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7109
7109
and near the photic depth in stable conditions. At this elevated depth, light intensity is low and
the phytoplankton has to increase its cellular chlorophyll content in order to maintain
efficiency. This subsurface chlorophyll-a maximum (SCM) presented an average
concentration of 0.3 mg m
-3
, with minimum and maximum of 0.03 mg m
-3
and 0.7 mg m
-3
,
respectively.

Figure 3 - Chlorophyll-a concentration at Bermuda Atlantic Time Series

3.2. Photic and penetration depths at BATS
The photic depths presented higher depth for the lower wavelength, 440 nm, and the
shallower depths for the green wavelength (550 nm), approximately 160 m and 60 m
respectively (Figure 4). The photic depth for the photosynthetic active region (PAR), which
constitutes the entire visible and near infrared spectra (400 to 700 nm) had an intermediary
average of 110 m. The approximate average for the penetration depth (z
pe
) were 40 m, 15 m
and 27 m for 440 nm, 550 nm and PAR respectively.
The time series showed an oscillating behaviour for z
p
and z
pe
at 440 nm and PAR
wavelengths similar to the subsurface chlorophyll-a maximum (Figure 4). The shoaling at z
p
,
z
pe
at 440 nm and PAR and SCM occurs at the beginning of spring season (February to April)
and deepening in summer (June, July). This pattern might be associated with the mixed layer
depth (MLD) seasonality. In winter, the low temperatures and strong wind conditions deepen
the vertical mixing, bringing nutrient-rich waters from deep layers to the photic zone. In
beginning of spring, with both nutrients and light available, the rate of phytoplankton growth
increases causing the spring blooms. An increase in phytoplankton elevates the particulate and
dissolved matter (from decomposing material) concentrations, increasing the attenuation of
light at surface and consequently reducing the photic depth.
In summer the higher light availability and the increased surface temperature generate a
pronounced, shallower thermocline turning the MLD thinner and ceasing the renewal of
nutrients from deep waters. This situation inhibits the phytoplankton growth, reducing the
particulate and dissolved matter.

Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7110
7110

Figure 4 - Time series of photic depth (Zp), penetration depth (Zpe) and deep maximum chlorophyll-a at BATS
station for different wavelengths.

3.3. Comparison between in situ and satellite chlorophyll-a
The temporal variation of in situ and satellite chlorophyll-a concentrations was very
similar, but with different magnitudes (Figure 5). From October 1997 to July 2007,
chlorophyll-a concentrations extracted of the monthly and weekly SeaWiFS products vary
between 0.04 mg m
-3
to 0.21 mg m
-3
, while chlorophyll-a at the penetration depth for the 440
nm and 550 nm wavelengths and PAR minimum and maximum were 0.01 and 0.59 mg m
-3
,
respectively.

Figure 5 Chlorophyll-a concentrations at 440 nm, 550 nm and PAR penetration depths (C
pe(440 nm)
, C
pe(550 nm)

and C
pe(PAR)
) and monthly (m) and weekly (w) SeaWiFS products.

Although z
pe(440 nm)
and z
pe(550 nm)
had an average difference of 25 m (Figure 4) their
chlorophyll-a content was practically identical: an average of 0.11 mg m
-3
for both (Figure 5).
SeaWiFS products also were very similar with average concentration values of 0.09 mg m
-3

(monthly) and 0.10 mg m
-3
(weekly).
The difference between SeaWiFS and in situ chlorophyll-a concentrations may be a result
of the algorithm that generates the satellite product. OReilly et al. (2000) describe that the
relation between in situ upwelling radiance at surface and chlorophyll-a concentration is the
basis for the polynomial equation OC4v4 uses to prepare the data.
The disparity between both acquisitions (satellite and in situ) was more discernible in
higher concentrations (> 0.15 mg m
-3
Figure 5). The closeness of SCM and penetration
depth for this time periods (Figure 4) may indicate an elevation on chlorophyll-a
concentration in depths neglected by the algorithm. When plotted, the correlation between
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7111
7111
these datasets evidenced good linear relations (Table 1) and this underestimation of
concentrations above ~0.15 mg m
-3
by the SeaWiFS products, here represented by Monthly
SeaWiFS data and z
pe(440 nm)
(Figure 6).

Table 1 Correlation coefficients (R
2
) between the chlorophyll-a concentrations at the penetration depths and
obtained from SeaWiFS products.
SeaWiFS product
In situ Monthly Weekly
C
pe(440 nm)
0.76 0.64
C
pe(550 nm)
0.69 0.69
C
pe(PAR)
0.66 0.63


Figure 6 Relation between chlorophyll-a concentration for the estimated penetration depth at 440 nm and
SeaWiFS monthly average. Trend and identity lines in black and red, respectively.

The results presented here still are object of study. The next steps include detailed study
of the uncertainties originated by the assumptions and analysis of the interannual variations
observed. The relation to environmental variables, such as sea surface temperatures and
eddies, and influence of major events, like hurricanes and ENSO will be investigated.

4. Conclusions
In theory the depth satellites are able to sense is fairly variable depending on wavelength
and time of the year. The depth where chlorophyll-a has its maximum concentration is much
higher than this penetration depth throughout most part of the year, excepting periods of
bloom, where chlorophyll-a increases near surface.
Satellite-derived chlorophyll-a concentrations were well correlated with the chlorophyll-a
present at the penetration depth, but underestimate it when the concentrations are high. This is
probably a result of the algorithm used for the product generation.
It is important to highlight that the usefulness of remote sensing lies in the synoptic
coverage it provides, the accuracy depends on careful algorithm design and the interpretation
is enhanced with in situ observations, reinforcing the value of long time series such as BATS.

5. Acknowledgments
The author wishes to express her gratitude to Dr. Trevor Platt and Dr. Shubha Sathyendranath for the exciting
and challenging exercise and the generous guidance to accomplish it. To Dr. George White III, Dr. Li Zhai and
Dr. Heather Bouman for sharing their knowledge. To the Bermuda Institute of Ocean Sciences (BIOS), the
Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7112
7112
Nippon Foundation (NF) and the Partnership for Observations of the Global Ocean (POGO) for support the
Observational Oceanography training at BIOS.

6. References
Devred, E.; Sathyendranath, S.; Stuart, V.; Maass, H.; Ulloa, O.; Platt, T. A two-component model of
phytoplankton absorption in the open ocean: Theory and applications. Journal of Geophysical Research, v.
111, n. C03011, p.1-11, 2006.

Gordon, H. R; McCluney, W. R. Estimation of the depth of sunlight penetration in the sea for remote sensing.
Applied Optics, v. 14, n. 2, p. 413-416, 1975.

Kampel, M. e Novo, E. M. L. O sensoriamento remoto da cor da gua. In: Souza, R.B. (org.). Oceanografia
por satlites. So Paulo: Oficina de Textos, p. 179-196, 2005.

Loisel, H.; Morel, A. Light scattering and chlorophyll concentration in case 1 waters: a reexamination.
Limnology and Oceanography, v. 43, n. 5, p. 847858, 1998.

Longhurst, A. R.; Harrison, W. G. The biological pump: Profiles of plankton production and consumption in the
upper ocean. Progress in Oceanography, v. 22, n. 1, p. 47 - 123, 1989..

Michaels, A.F.; Knap, A.H.; Dow, R.L.; Gundersen, K.; Johnson, R.J.; Sorensen, J.; Close, A.; Knauer, G.A.;
Lohrenz, S.E.; Asper, V.A.; Tuel, M.; Bidigare, R.R.Seasonal patterns of biogeochemistry at the US JGOFS
Bermuda Atlantic Time-series Study side. Deep-Sea Research I, v. 41, n. 7, p. 10131038, 1994.

Morel, A. Optical properties of pure seawater. In: Jerlov, N. G.; Nielsen, E. S. (org) Optical Aspects of
Oceanography, New York: Academic, p. 1-24, 1974.

OReilly, J. E.; Maritorena, S.; Siegel, D.; O'Brien, M. C.; Toole, D.; Mitchell, B. G.; Kahru, M.; Chavez, F. P.;
Strutton, P.; Cota, G.; Hooker, S. B.; McClain, C. R.; Carder, K. L.; Mller-Karger, F.; Harding, L.; Magnuson,
A.; Phinney, D.; Moore, G. F.; Aiken, J.; Arrigo, K. R.; Letelier, R.; Culver, M. Ocean color chlorophyll a
algorithms for SeaWiFS, OC2, and OC4: Version 4. In: Hooker, S.B.; Firestone, E.R. (org). SeaWiFS
Postlaunch Technical Report Series, Volume 11, SeaWiFS Postlaunch Calibration and Validation
Analyses, Part 3. Maryland: NASA, Goddard Space Flight Center, 2000, p. 9-23.

Platt, T.; Sathyendranath, S. Modelling primary production IV. Daily primary production of the water column:
General introduction. p. 27 35. Available at: < http://www.geosafari.org/kochi/Forms/PPMOMNI.pdf>.
Accessed in: 19.nov.2010.

Platt, T.; Sathyendranath, S.; White, G.; Ravindran, P. Attenuation of visible light by phytoplankton in a
vertically structured ocean: solutions and applications. Journal of Plankton Research, v. 16, n. 11, p. 1461-
1487, 1994.

Pope, R. M.; Fry, E. S., Absorption spectrum (380700 nm) of pure water: II. Integrating cavity measurements.
Applied Optics, v. 36, n. 33, p. 87108723, 1997.

Sathyendranath, S.; Platt, T. Remote sensing of ocean chlorophyll: consequence of non uniform pigment profile.
Applied Optics, v. 28, n. 3, p. 490-495, 1989.

Sathyendranath, S.; Cota, G.; Stuart, V.; Maass, H.; Platt, T. Remote sensing of phytoplankton pigments: a
comparison of empirical and theoretical approaches. International Journal of Remote Sensing, v. 22, n. 2-3, p.
249273, 2001.

Steinberg, D.K.; Carlson, C.A.; Bates, N.R.; Johnson, R.J.; Michaels, A.F.; Knap, A. Overview of the US
JGOFS Bermuda Atlantic Time-series Study (BATS): a decade-scale look at ocean biology and
biogeochemistry. Deep-Sea Research II, v. 48, p. 14051448, 2001.

Ulloa, O.; Sathyendranath, S.; Platt, T. Effect of the particlesize distribution on the backscattering ratio in
seawater. Applied Optics, v. 33, n. 30, p. 70707077, 1994.

Anais XV Simpsio Brasileiro de Sensoriamento Remoto - SBSR, Curitiba, PR, Brasil, 30 de abril a 05 de maio de 2011, INPE p.7113
7113