SATELLITE MEASUREMENTS
149
performed on them, and whether they are suitable for the purpose. For
example some data products may contain “cosmetic” filling of values in
locations where there would otherwise be gaps due to cloud cover or other
obstructions to the remote sensing process. Ideally only true observations,
and not the artefacts of data processing, should be presented to the numerical
model.
This chapter is therefore written for those engaged in developing
operational oceanography systems, to give them a basic background in the
methods of ocean remote sensing so that they can appreciate what issues to
consider as they evaluate the quality of satellite data. It is split into three
main sections. The first is a generic overview of the subject. The second
introduces the basic remote sensing methodology for some of the key
variables used by models. These are sea surface height from altimetry,
ocean colour, sea surface temperature (SST) and ocean waves. The third
main section uses the example of SST to explore how measurements
retrieved from several different sensor systems and supplied by different
agencies can be most effectively combined to serve the needs of ocean
forecasting models. It presents the methods adopted by an international
programme established by the Global Ocean Data Assimilation Experiment
(GODAE) for this purpose.
2.
Methods of satellite oceanography: An outline
2.1
Using the electromagnetic spectrum
All satellite remote sensing sensors use electromagnetic (e.m.) radiation
to view the sea. The ability of particular sensors to measure certain
properties of the ocean and how well they can view through the atmosphere
or penetrate clouds depends critically on which part of the e.m. spectrum
they use. Figure 1 shows the section of the electromagnetic spectrum that is
of relevance to remote sensing, and the four broad classes of sensors that are
used. The diagram also shows how the transmittance of the atmosphere
varies with e.m. wavelength, which accounts for why sensors are found only
in certain wavebands. A much fuller account is given by Robinson (2004).
For much of the e.m. spectrum the atmosphere is opaque and therefore
unusable for remote sensing of the ocean. However in a number of
“window” regions of the spectrum most of the radiation is transmitted
although it may be attenuated to some extent. These windows provide the
opportunities for ocean remote sensing.
One of the windows extends from the visible part of the spectrum
(between 400 nm and 700 nm, used by the human eye) into the near infrared
(NIR). This is used by “ocean colour” radiometers that observe sunlight
reflected from the ocean, both from the surface and from within the upper
149
performed on them, and whether they are suitable for the purpose. For
example some data products may contain “cosmetic” filling of values in
locations where there would otherwise be gaps due to cloud cover or other
obstructions to the remote sensing process. Ideally only true observations,
and not the artefacts of data processing, should be presented to the numerical
model.
This chapter is therefore written for those engaged in developing
operational oceanography systems, to give them a basic background in the
methods of ocean remote sensing so that they can appreciate what issues to
consider as they evaluate the quality of satellite data. It is split into three
main sections. The first is a generic overview of the subject. The second
introduces the basic remote sensing methodology for some of the key
variables used by models. These are sea surface height from altimetry,
ocean colour, sea surface temperature (SST) and ocean waves. The third
main section uses the example of SST to explore how measurements
retrieved from several different sensor systems and supplied by different
agencies can be most effectively combined to serve the needs of ocean
forecasting models. It presents the methods adopted by an international
programme established by the Global Ocean Data Assimilation Experiment
(GODAE) for this purpose.
2.
Methods of satellite oceanography: An outline
2.1
Using the electromagnetic spectrum
All satellite remote sensing sensors use electromagnetic (e.m.) radiation
to view the sea. The ability of particular sensors to measure certain
properties of the ocean and how well they can view through the atmosphere
or penetrate clouds depends critically on which part of the e.m. spectrum
they use. Figure 1 shows the section of the electromagnetic spectrum that is
of relevance to remote sensing, and the four broad classes of sensors that are
used. The diagram also shows how the transmittance of the atmosphere
varies with e.m. wavelength, which accounts for why sensors are found only
in certain wavebands. A much fuller account is given by Robinson (2004).
For much of the e.m. spectrum the atmosphere is opaque and therefore
unusable for remote sensing of the ocean. However in a number of
“window” regions of the spectrum most of the radiation is transmitted
although it may be attenuated to some extent. These windows provide the
opportunities for ocean remote sensing.
One of the windows extends from the visible part of the spectrum
(between 400 nm and 700 nm, used by the human eye) into the near infrared
(NIR). This is used by “ocean colour” radiometers that observe sunlight
reflected from the ocean, both from the surface and from within the upper