49
sensors at top at the atmosphere. Mathematically, the measured signal is a function of all conditions and constituents of
the atmosphere and the water. We can access these parameters by inversion (exemplary, see inversion techniques
described in Rodgers 2000). Generally, the retrieved TOA
signal contains approximately 90% atmospheric and 10%
oceanic information.
Additionally, the TOA measurement changes with the
spectrum and the viewing geometry. As outlined beforehand,
in the near infrared we expect a very low or no signal above
(clear) water bodies due to the strong absorptive properties
of waters beyond 750 nm but a high response in the blue visible range. Most of the optical instruments have a nadir viewing geometry where the measurement sensor looks directly
downwards. An off-nadir measurement with a viewing zenith
angle (VZA) greater than 0° from the normal axis between
satellite and surface increases the path between the location
of the water-leaving radiation and the sensor. The signal can
increase by diffuse scattering in the atmosphere and/or attenuates due to more opportunities for absorption and scattering
by molecules and particles.
The main natural obstacles (along others) in optical
remote sensing are clouds, sunglint, and the interfering
atmosphere. Clouds appear thick and white to the human
eyes and also to optical sensors. In different wavelengths
regimes, for example for microwave measurements, clouds
are transparent and the sensors can measure the underlying
surface. Usually, microwave instruments are used for the
detection of sea surface temperature, surface height or land
applications. For optical measurements, certain algorithms
(“cloud masks”) exclude pixels with expected cloud coverage. Sunglint occurs at smooth and highly reflective surfaces
such as water or fresh snow if the solar light is directly
reflected into the sensor. The bright reflection usually oversaturates a sensor’s measurement capability and also contains very low or no information about the water body.
However, a change in the viewing geometry reduces or even
avoids the measurement of sunglint.
The measured TOA radiation has to pass the atmosphere,
which highly changes the received signal that leaves the
water. The measurement can be “back-calculated” to a BOAmeasurement, which is ideally equal to the water-leaving
signal. Therefore, it is necessary to estimate the influence of
the atmosphere on the TOA-signal by proxies and additional
measurements. Using the estimation, the signal can be corrected for the atmosphere (“Atmospheric Correction”).
Selected Sensors for Water Remote Sensing
Historically, scientific Earth observation started in the late
1950s to support weather forecasts and to analyze weather
phenomena. In the following, we present some selected sensors that have or had the main mission to observe water bodies. Therefore, each sensor’s bands were carefully chosen for
water applications. However, they are also used above land
and most of the introduce sensors also have land and atmospheric missions.
The Coastal Zone Color Scanner (CZCS) onboard the
US-platform NIMBUS-7, operational from 1978 to 1987,
was one of the first satellite sensors mainly designed to
observe the oceans. The CZCS measurements were a first
step towards global mapping of chlorophyll a concentration
and the impact of the oceans on the carbon cycle. In 1996,
the Sea-viewing Wide Field-of-view Sensor (SeaWiFS)
onboard Seastar began sensing the ocean in eight channels
within 400 nm to 900 nm. SeaWiFS operated until 2010
and was slightly tilted to avoid sun glint. MODIS, introduced in section “Instruments”, is mounted on board the
satellites Aqua and Terra operating from 1999 and 2002,
respectively, until present time. The Medium Resolution
Imaging Spectrometer MERIS was one of 11 instruments
onboard the Environmental Satellite Envisat that operated
from 2002 until a technical platform failure in 2012.
MERIS supported the chlorophyll a fluorescence investigation with a band at 681 nm nearly to 683 nm where the fluorescence peaks and some bands usable for chlorophyll a
algorithms (ESA 2006).
The Ocean and Land Color Imager (OLCI) on board
Sentinel-3 continues the heritage of MERIS with 6 additional bands (ESA 2013). Sentinel-3A was launched in 2016
and Sentinel-3B is planned for 2018 (https://earth.esa.int/
web/guest/missions/esa-eo-missions/sentinel-3, 29 July
2017). Hyperspectral imagers usable for water measurements are the Hyperspectral Imager for the Coastal Ocean
(HICO) installed on the International Space Station (ISS)
and the Hyperspectral Imager (HSI) onboard EnMAP. HICO
operated from 2008 to 2014 (http://hico.coas.oregonstate.
edu/, 29 July 2017-07-29) and EnMAP is planned for launch
in 2019 (http://www.enmap.org/).
Using Remote Sensing Measurements
Preprocessing
Before the space-borne measurements are available for the
user, they are usually preprocessed. The state of processing
is defined by its level. The processing is mostly done by the
operating space agency and, hence, the expressions may
sometimes vary slightly and the agencies may not provide all
levels for all sensors. Referring to Martin (2014) the levels
(L) are briefly introduced:
Level 0 data sets contain the raw measurements without any
correction besides measurement or transfer artifacts.
Level 1 data sets contain temporal and spatial information.
Level 1B data provide measurements converted to a
radiometric unit (e.g., radiance).
Marine Optics and Ocean Color Remote Sensing
sensors at top at the atmosphere. Mathematically, the measured signal is a function of all conditions and constituents of
the atmosphere and the water. We can access these parameters by inversion (exemplary, see inversion techniques
described in Rodgers 2000). Generally, the retrieved TOA
signal contains approximately 90% atmospheric and 10%
oceanic information.
Additionally, the TOA measurement changes with the
spectrum and the viewing geometry. As outlined beforehand,
in the near infrared we expect a very low or no signal above
(clear) water bodies due to the strong absorptive properties
of waters beyond 750 nm but a high response in the blue visible range. Most of the optical instruments have a nadir viewing geometry where the measurement sensor looks directly
downwards. An off-nadir measurement with a viewing zenith
angle (VZA) greater than 0° from the normal axis between
satellite and surface increases the path between the location
of the water-leaving radiation and the sensor. The signal can
increase by diffuse scattering in the atmosphere and/or attenuates due to more opportunities for absorption and scattering
by molecules and particles.
The main natural obstacles (along others) in optical
remote sensing are clouds, sunglint, and the interfering
atmosphere. Clouds appear thick and white to the human
eyes and also to optical sensors. In different wavelengths
regimes, for example for microwave measurements, clouds
are transparent and the sensors can measure the underlying
surface. Usually, microwave instruments are used for the
detection of sea surface temperature, surface height or land
applications. For optical measurements, certain algorithms
(“cloud masks”) exclude pixels with expected cloud coverage. Sunglint occurs at smooth and highly reflective surfaces
such as water or fresh snow if the solar light is directly
reflected into the sensor. The bright reflection usually oversaturates a sensor’s measurement capability and also contains very low or no information about the water body.
However, a change in the viewing geometry reduces or even
avoids the measurement of sunglint.
The measured TOA radiation has to pass the atmosphere,
which highly changes the received signal that leaves the
water. The measurement can be “back-calculated” to a BOAmeasurement, which is ideally equal to the water-leaving
signal. Therefore, it is necessary to estimate the influence of
the atmosphere on the TOA-signal by proxies and additional
measurements. Using the estimation, the signal can be corrected for the atmosphere (“Atmospheric Correction”).
Selected Sensors for Water Remote Sensing
Historically, scientific Earth observation started in the late
1950s to support weather forecasts and to analyze weather
phenomena. In the following, we present some selected sensors that have or had the main mission to observe water bodies. Therefore, each sensor’s bands were carefully chosen for
water applications. However, they are also used above land
and most of the introduce sensors also have land and atmospheric missions.
The Coastal Zone Color Scanner (CZCS) onboard the
US-platform NIMBUS-7, operational from 1978 to 1987,
was one of the first satellite sensors mainly designed to
observe the oceans. The CZCS measurements were a first
step towards global mapping of chlorophyll a concentration
and the impact of the oceans on the carbon cycle. In 1996,
the Sea-viewing Wide Field-of-view Sensor (SeaWiFS)
onboard Seastar began sensing the ocean in eight channels
within 400 nm to 900 nm. SeaWiFS operated until 2010
and was slightly tilted to avoid sun glint. MODIS, introduced in section “Instruments”, is mounted on board the
satellites Aqua and Terra operating from 1999 and 2002,
respectively, until present time. The Medium Resolution
Imaging Spectrometer MERIS was one of 11 instruments
onboard the Environmental Satellite Envisat that operated
from 2002 until a technical platform failure in 2012.
MERIS supported the chlorophyll a fluorescence investigation with a band at 681 nm nearly to 683 nm where the fluorescence peaks and some bands usable for chlorophyll a
algorithms (ESA 2006).
The Ocean and Land Color Imager (OLCI) on board
Sentinel-3 continues the heritage of MERIS with 6 additional bands (ESA 2013). Sentinel-3A was launched in 2016
and Sentinel-3B is planned for 2018 (https://earth.esa.int/
web/guest/missions/esa-eo-missions/sentinel-3, 29 July
2017). Hyperspectral imagers usable for water measurements are the Hyperspectral Imager for the Coastal Ocean
(HICO) installed on the International Space Station (ISS)
and the Hyperspectral Imager (HSI) onboard EnMAP. HICO
operated from 2008 to 2014 (http://hico.coas.oregonstate.
edu/, 29 July 2017-07-29) and EnMAP is planned for launch
in 2019 (http://www.enmap.org/).
Using Remote Sensing Measurements
Preprocessing
Before the space-borne measurements are available for the
user, they are usually preprocessed. The state of processing
is defined by its level. The processing is mostly done by the
operating space agency and, hence, the expressions may
sometimes vary slightly and the agencies may not provide all
levels for all sensors. Referring to Martin (2014) the levels
(L) are briefly introduced:
Level 0 data sets contain the raw measurements without any
correction besides measurement or transfer artifacts.
Level 1 data sets contain temporal and spatial information.
Level 1B data provide measurements converted to a
radiometric unit (e.g., radiance).
Marine Optics and Ocean Color Remote Sensing