footprint would not necessarily be of the same shape or of equal dimension from
location to location.
For sun-synchronous satellites, the rotation of the Earth beneath the satellite
means that the satellite track is not directly oriented North–South along the Earth’s
surface, and because the orbiting satellite moves as the cross-track scan (e.g., for
AVHRR; Table 11.1) is acquired, neither is the scan East–West oriented
(Fig. 11.3b). In combination with variations in footprint size, this results in raw data
being acquired in irregular grids. However, data users often prefer products aligned
into regular grids with known, and fixed, longitude and latitude bounds, requiring
that orbital data be spatially transformed, which may involve averaging and/or subsampling. Many users may not consider such impacts when using the data. An
understanding of the geometry of satellite retrieval and the potential impact on
interpretation and data resolution is an important consideration for users.
As discussed previously, satellite data resolution is best described by the
instrument’s field-of-view and, as such, the corresponding areal footprint at the
Earth’s surface depends upon the satellite altitude. Orbital period (i.e., the time to
orbit the Earth) increases with orbit radius (i.e., satellite altitude) according to the
gravitational laws, as does the ground-footprint for a given field-of-view angle. For
example, a geo-synchronous satellite (with orbital period 24 h) is further from the
Earth’s surface and has lower spatial resolution than a polar-orbiter (of period
approximately 100 min) with the same field-of-view angle. Thus there is an
application-dependent compromise between frequency of coverage (and sensor
field-of-view characteristics) and spatial resolution. One must therefore take into
consideration how instrument capability (i.e., field-of-view angle), the resulting
spatial resolution, and the return period impact each remote sensing application.
11.2.2 Acquisition Logistics
The most significant factor influencing satellite remote sensing of the ocean surface is the effect of the atmosphere. Of the radiation emitted by the ocean surface,
aerosols in the atmosphere (e.g., water vapor, H 2 O; ozone, O 3 ; oxygen, O 2 ; carbon
dioxide, CO 2 ; nitrous oxide, N 2 O; methane, CH 4 ; nitrogen dioxide, NO 2 ; nitrogen,
N 2 ; dust; and particulates) attenuate particular wavelengths via absorption and
scattering (see Fig. 11.4). Note that the atmosphere is fairly transparent at visible
wavelengths (400–700 nm), which is beneficial for visible remote sensing (e.g.,
ocean color), and also transparent at other specific ‘‘windows’’, which are
employed in instrument channel design (see Table 11.1). For measurement of SST,
the general technique of using information from multiple channels within these
optically transparent windows includes ‘‘dual window’’, ‘‘split window’’ and
‘‘triple window’’ algorithms (see Li et al. 2001 and references therein). These
algorithms rely upon parameters derived from regression of satellite data with
in situ measurements of temperature, and, in some cases, an a priori (reference)
estimate from an external SST dataset.
11 Thermal and Radar Overview
291
location to location.
For sun-synchronous satellites, the rotation of the Earth beneath the satellite
means that the satellite track is not directly oriented North–South along the Earth’s
surface, and because the orbiting satellite moves as the cross-track scan (e.g., for
AVHRR; Table 11.1) is acquired, neither is the scan East–West oriented
(Fig. 11.3b). In combination with variations in footprint size, this results in raw data
being acquired in irregular grids. However, data users often prefer products aligned
into regular grids with known, and fixed, longitude and latitude bounds, requiring
that orbital data be spatially transformed, which may involve averaging and/or subsampling. Many users may not consider such impacts when using the data. An
understanding of the geometry of satellite retrieval and the potential impact on
interpretation and data resolution is an important consideration for users.
As discussed previously, satellite data resolution is best described by the
instrument’s field-of-view and, as such, the corresponding areal footprint at the
Earth’s surface depends upon the satellite altitude. Orbital period (i.e., the time to
orbit the Earth) increases with orbit radius (i.e., satellite altitude) according to the
gravitational laws, as does the ground-footprint for a given field-of-view angle. For
example, a geo-synchronous satellite (with orbital period 24 h) is further from the
Earth’s surface and has lower spatial resolution than a polar-orbiter (of period
approximately 100 min) with the same field-of-view angle. Thus there is an
application-dependent compromise between frequency of coverage (and sensor
field-of-view characteristics) and spatial resolution. One must therefore take into
consideration how instrument capability (i.e., field-of-view angle), the resulting
spatial resolution, and the return period impact each remote sensing application.
11.2.2 Acquisition Logistics
The most significant factor influencing satellite remote sensing of the ocean surface is the effect of the atmosphere. Of the radiation emitted by the ocean surface,
aerosols in the atmosphere (e.g., water vapor, H 2 O; ozone, O 3 ; oxygen, O 2 ; carbon
dioxide, CO 2 ; nitrous oxide, N 2 O; methane, CH 4 ; nitrogen dioxide, NO 2 ; nitrogen,
N 2 ; dust; and particulates) attenuate particular wavelengths via absorption and
scattering (see Fig. 11.4). Note that the atmosphere is fairly transparent at visible
wavelengths (400–700 nm), which is beneficial for visible remote sensing (e.g.,
ocean color), and also transparent at other specific ‘‘windows’’, which are
employed in instrument channel design (see Table 11.1). For measurement of SST,
the general technique of using information from multiple channels within these
optically transparent windows includes ‘‘dual window’’, ‘‘split window’’ and
‘‘triple window’’ algorithms (see Li et al. 2001 and references therein). These
algorithms rely upon parameters derived from regression of satellite data with
in situ measurements of temperature, and, in some cases, an a priori (reference)
estimate from an external SST dataset.
11 Thermal and Radar Overview
291