50
Level 2 data sets contain physical parameters calculated
from L1B data (e.g., sea surface temperature). L2 data
require the application of multiple channels, land-seamasks and cloud masks, and usually an atmospheric correction that accounts for the influence of the atmosphere
on the signal.
Sometimes, Levels 3 products are available for specific
locations or a specific gridding. They may also contain temporally merged products, e.g., monthly means, to reduce data
gaps due to clouds or other obstacles. Figure 5a illustrates
daily Chlorophyll a product from 28 July 2017 and a monthly
mean for the month of June 2017 is given in Fig. 5b. Level 4
data may incorporate match-up data of in situ and field measurements. In oceanic applications, the atmospheric correction
for the Level 2 data is mainly applied to convert the TOA measurements into values that would have been measured if the
atmosphere were absent. There is a wide range of applications
and algorithms that can conduct atmospheric corrections.
Applications
All levels are usually provided in a scientific binary data format and are mostly available for free (e.g., MODIS data) or on
request for scientific purposes. Oceanic remote sensing data
can be downloaded via the following selected webpages:
• https://oceancolor.gsfc.nasa.gov/ for SeaWiFS, MODIS,
MERIS, CZCS, and others. They provide Chlorophyllconcentration, sea surface temperature, and a quasi-RGB
image per orbit or on a temporal averaged base.
• https://scihub.copernicus.eu/s3/#/home for OLCI data as
Level 1 or Level 2
EumetView provides a quick access to OLCI data with
orbital RGB images (access via http://eumetview.eumetsat.
int/mapviewer/). The RGB image can be downloaded. The
WorldView page for the MODIS measurements (https://
worldview.earthdata.nasa.gov/) supports RGB images,
reflectance, and several layers (e.g., sea ice and chlorophyll
concentration) for an easy overview of the entire globe. Both
views provide images in near-real time.
The free software SNAP (download via http://step.esa.int/
main/toolboxes/snap/) by ESA is a useful tool for statistical
analyses of satellite data and display results per band or as
RGB image (see Figs. 4 and 6). For some sensors, it provides
atmospheric corrections, conversion to a higher level and
generation of products like chlorophyll concentration.
For example, the two band ratio “blue-green-ratio” gives
a first estimation of the amount of algae and chlorophyll,
respectively, in the ocean in arbitrary units (Martin 2014). In
clear waters, the maximum reflectance is located in the blue
part of the visible spectrum. The peak shifts to the green
regime in the presence of phytoplankton. The blue-greenratio
BG
=
“blue
channel”/“green
channel” = R(440 nm)/R(555 nm) compares these two
extremes. A higher blue-green-ratio indicates a “more blue”
water and we expect low or no algal content. Unfortunately,
this ratio is easily disturbed by additional substances
(CDOM, sediments) that change the shape of a reflectance
spectrum. Therefore, this ratio is only applicable for a first
guess in clear waters with phytoplankton dominance.
Bio-optical Models
Bio-optical models link optical measurements of reflectance
or radiance and biological parameters like chlorophyll a concentration, water quality, euphotic depth, and others. These
biogeochemical variables are a main interest of the end-users
who want to decide or analyze specific issues. Depending on
the observed water, the complexity of a bio-optical model
can vary from a ratio to an extensive non-linear function.
Morel and Prieur (1977) introduced optically simple waters,
which only contain phytoplankton (case-1) and optically
complex waters (case-2). Exemplarily in case-1 waters, several chlorophyll a concentration algorithms are based on
blue-green ratios (BGs). BG is the relation between a “green”
and “blue” measurement band in the VIS providing a qualitative estimate of the relative presence of phytoplankton.
According to Martin (2014), the MODIS Ocean-Color-3band-Algorithm OC3M is an empirical relation between the
maximal ratio of some blue-green ratios from measured
reflectance and chlorophyll a [mg m
−3
] measurements:
R
R
R
L
R S
R S
=
(
)
(
)
(
)
(
)=
log max
,
/
log
.
10
10
443
488
551
0
nm
nm
nm
Chla
2 2424 2 742
1 802
0 002
1 228
2
3
4
−
∗ +
∗ +
−
.
.
.
.
R
R
R
R
L
L
L
L
Where R RS refers to remote sensing reflectance, R L to the
maximal ratio and Chla to chlorophyll a concentration.
There is a wide range of “OC” algorithms depending on
the instrument and the degree of the polynomial. Bio-optical
models also can describe the spectral shape of IOPs based on
only a few measurements, for example,
a
a
S
CDOM
CDOM
=
( )∗
− ∗ −
(
)
(
)
440
440
exp
λ
where a CDOM refers to absorption by CDOM, and S to the
slope of the absorption spectra. CDOM absorption (m
−1
)
exponentially decreases with longer wavelengths and a biooptical model for spectral CDOM is often based on a measurement at approximately 440 nm. The equations for a cdom
and phytoplankton absorption (a ph ) are commonly used models to describe CDOM in case-2 waters and phytoplankton
absorption in case-1 waters (Gilerson et al. 2008; Brewin
et al. 2011; McKee et al. 2014). Usually, the shape factor S
V. Mascarenhas and T. Keck
Level 2 data sets contain physical parameters calculated
from L1B data (e.g., sea surface temperature). L2 data
require the application of multiple channels, land-seamasks and cloud masks, and usually an atmospheric correction that accounts for the influence of the atmosphere
on the signal.
Sometimes, Levels 3 products are available for specific
locations or a specific gridding. They may also contain temporally merged products, e.g., monthly means, to reduce data
gaps due to clouds or other obstacles. Figure 5a illustrates
daily Chlorophyll a product from 28 July 2017 and a monthly
mean for the month of June 2017 is given in Fig. 5b. Level 4
data may incorporate match-up data of in situ and field measurements. In oceanic applications, the atmospheric correction
for the Level 2 data is mainly applied to convert the TOA measurements into values that would have been measured if the
atmosphere were absent. There is a wide range of applications
and algorithms that can conduct atmospheric corrections.
Applications
All levels are usually provided in a scientific binary data format and are mostly available for free (e.g., MODIS data) or on
request for scientific purposes. Oceanic remote sensing data
can be downloaded via the following selected webpages:
• https://oceancolor.gsfc.nasa.gov/ for SeaWiFS, MODIS,
MERIS, CZCS, and others. They provide Chlorophyllconcentration, sea surface temperature, and a quasi-RGB
image per orbit or on a temporal averaged base.
• https://scihub.copernicus.eu/s3/#/home for OLCI data as
Level 1 or Level 2
EumetView provides a quick access to OLCI data with
orbital RGB images (access via http://eumetview.eumetsat.
int/mapviewer/). The RGB image can be downloaded. The
WorldView page for the MODIS measurements (https://
worldview.earthdata.nasa.gov/) supports RGB images,
reflectance, and several layers (e.g., sea ice and chlorophyll
concentration) for an easy overview of the entire globe. Both
views provide images in near-real time.
The free software SNAP (download via http://step.esa.int/
main/toolboxes/snap/) by ESA is a useful tool for statistical
analyses of satellite data and display results per band or as
RGB image (see Figs. 4 and 6). For some sensors, it provides
atmospheric corrections, conversion to a higher level and
generation of products like chlorophyll concentration.
For example, the two band ratio “blue-green-ratio” gives
a first estimation of the amount of algae and chlorophyll,
respectively, in the ocean in arbitrary units (Martin 2014). In
clear waters, the maximum reflectance is located in the blue
part of the visible spectrum. The peak shifts to the green
regime in the presence of phytoplankton. The blue-greenratio
BG
=
“blue
channel”/“green
channel” = R(440 nm)/R(555 nm) compares these two
extremes. A higher blue-green-ratio indicates a “more blue”
water and we expect low or no algal content. Unfortunately,
this ratio is easily disturbed by additional substances
(CDOM, sediments) that change the shape of a reflectance
spectrum. Therefore, this ratio is only applicable for a first
guess in clear waters with phytoplankton dominance.
Bio-optical Models
Bio-optical models link optical measurements of reflectance
or radiance and biological parameters like chlorophyll a concentration, water quality, euphotic depth, and others. These
biogeochemical variables are a main interest of the end-users
who want to decide or analyze specific issues. Depending on
the observed water, the complexity of a bio-optical model
can vary from a ratio to an extensive non-linear function.
Morel and Prieur (1977) introduced optically simple waters,
which only contain phytoplankton (case-1) and optically
complex waters (case-2). Exemplarily in case-1 waters, several chlorophyll a concentration algorithms are based on
blue-green ratios (BGs). BG is the relation between a “green”
and “blue” measurement band in the VIS providing a qualitative estimate of the relative presence of phytoplankton.
According to Martin (2014), the MODIS Ocean-Color-3band-Algorithm OC3M is an empirical relation between the
maximal ratio of some blue-green ratios from measured
reflectance and chlorophyll a [mg m
−3
] measurements:
R
R
R
L
R S
R S
=
(
)
(
)
(
)
(
)=
log max
,
/
log
.
10
10
443
488
551
0
nm
nm
nm
Chla
2 2424 2 742
1 802
0 002
1 228
2
3
4
−
∗ +
∗ +
−
.
.
.
.
R
R
R
R
L
L
L
L
Where R RS refers to remote sensing reflectance, R L to the
maximal ratio and Chla to chlorophyll a concentration.
There is a wide range of “OC” algorithms depending on
the instrument and the degree of the polynomial. Bio-optical
models also can describe the spectral shape of IOPs based on
only a few measurements, for example,
a
a
S
CDOM
CDOM
=
( )∗
− ∗ −
(
)
(
)
440
440
exp
λ
where a CDOM refers to absorption by CDOM, and S to the
slope of the absorption spectra. CDOM absorption (m
−1
)
exponentially decreases with longer wavelengths and a biooptical model for spectral CDOM is often based on a measurement at approximately 440 nm. The equations for a cdom
and phytoplankton absorption (a ph ) are commonly used models to describe CDOM in case-2 waters and phytoplankton
absorption in case-1 waters (Gilerson et al. 2008; Brewin
et al. 2011; McKee et al. 2014). Usually, the shape factor S
V. Mascarenhas and T. Keck