350
A. Dekker, V. Brando, J. Anstee, S. Fyfe, T. Malthus and E. Karpouzli
such as Landsat TM and SPOT (e.g. Mumby et al.,
1997a, 1998; Jakubauskas et al., 2000) and is a
geometrically accurate, cost-effective alternative to
aerial photography (Mumby et al., 1997a).
II. Principles of Remote Sensing
of Seagrass Ecosystems
The remote sensing of seagrasses and related seabed
habitats is based on the principle that a remote sensor
can ‘see’ the substratum and the vegetation growing
on or in (microphytobenthos) that substratum. Seagrasses are covered by a water column that attenuates the light reaching, interacting with (see Chapter
13), and being reflected from the benthos. While
the remote sensing of terrestrial plants makes significant use of the red edge (i.e. the steep slope between strong red wavelength absorption and strong
near-infrared reflectance characteristic of the spectral signatures of healthy plant leaves), aquatic plants
cannot be recognized by this feature since wavelengths beyond 680 nm are significantly attenuated
by pure water (Kirk, 1994), beyond a water column
depth of 1–2 m. In coastal waters, spectral scattering and absorption by phytoplankton, suspended organic and inorganic matter, and dissolved organic
substances further restricts the light passing to the
benthos (Dekker et al., 2001). Zimmerman (Chapter 13) discusses the effects of higher concentrations
of optically active materials in the water column on
seagrass photosynthesis. Spectral discrimination between aquatic plant species must therefore concentrate on pigment related spectral features within the
visible wavelengths, where light penetrates the water column and can be reflected back to the sensor
(Fyfe, 2003). Absolute reflectances from submerged
species are generally low, often lower than the reflectance from a deep water column, and from a
visual perspective seagrass areas generally appear
darker than non-vegetated areas. One of the reasons
for seagrasses generally appearing darker in remote
sensing images and aerial photography is the shading that takes place within the canopy (see Chapter
13 for details).
In all physics-based applications where the substratum is mapped through a water column, a
bathymetry estimate (i.e. a water column depth estimate) is implicitly or explicitly involved. Optically
shallow waters are a special case in the remote sensing of aquatic systems. Under these conditions a
measurable signal will be detected from the substratum or plants, through the water column and through
the air–water interface. The water is considered to
be optically deep if there is no measurable influence
of the benthos or substratum on the remotely sensed
reflectance (Dekker et al., 2001).
The bathymetry and the water column optical
properties are thus important to remote sensing in
aquatic environments, in particular for seagrass and
macro-algal mapping (Dekker et al., 2001) since optimal results will be obtained from remote sensing if
the signal from the seagrass canopy is maximized
by compensating for the attenuating influence of
the water column. Whether a benthic feature such
as seagrass can actually be discriminated depends
on the spectral optical depth of the water column
(and the atmosphere between the sensor and the
water surface), the brightness and density of the
vegetation, and the spectral contrast between it and
the substratum, as well as on the spectral, spatial,
and radiometric sensitivity of the remote sensing
instrument.
Remote sensing of aquatic environments (seagrass, sand, macro-algae, muds, and coral reefs)
requires sensors with greater sensor signal-to-noise
ratio than those applied in terrestrial environments.
Coupled with this factor is the number of quantization levels to which the sensor can record, referred
to as the radiometric resolution of the sensor. This
must be high enough to allow a range of brightness
levels over which a classification can be performed
and sensitive enough to be able to detect the lower reflectance of the deeper seagrass beds (Dekker et al.,
2001). Seagrasses may grow with sparse cover and
can be spectrally confused with other benthic features such as areas of macro-algae, detritus, and
corals (Mumby et al., 1997a,b). The small size and/or
linear shape and patchy nature of many seagrass
meadows means that in many cases high spatial resolution is also required to accurately determine their
distribution and abundance. Another limiting factor
in accurately detecting change in seagrass meadows
may be the temporal cover of remote sensing data.
The spatial extent of meadows may decline in response to human impacts or natural dynamics and
may change rapidly, within weeks to months. For
example, monitoring seasonal change in the extent
and density of certain seagrass species would require
a series of temporal cover data sets. The issue of
temporal cover is being increasingly addressed with
the availability of more satellite sensors, some with
A. Dekker, V. Brando, J. Anstee, S. Fyfe, T. Malthus and E. Karpouzli
such as Landsat TM and SPOT (e.g. Mumby et al.,
1997a, 1998; Jakubauskas et al., 2000) and is a
geometrically accurate, cost-effective alternative to
aerial photography (Mumby et al., 1997a).
II. Principles of Remote Sensing
of Seagrass Ecosystems
The remote sensing of seagrasses and related seabed
habitats is based on the principle that a remote sensor
can ‘see’ the substratum and the vegetation growing
on or in (microphytobenthos) that substratum. Seagrasses are covered by a water column that attenuates the light reaching, interacting with (see Chapter
13), and being reflected from the benthos. While
the remote sensing of terrestrial plants makes significant use of the red edge (i.e. the steep slope between strong red wavelength absorption and strong
near-infrared reflectance characteristic of the spectral signatures of healthy plant leaves), aquatic plants
cannot be recognized by this feature since wavelengths beyond 680 nm are significantly attenuated
by pure water (Kirk, 1994), beyond a water column
depth of 1–2 m. In coastal waters, spectral scattering and absorption by phytoplankton, suspended organic and inorganic matter, and dissolved organic
substances further restricts the light passing to the
benthos (Dekker et al., 2001). Zimmerman (Chapter 13) discusses the effects of higher concentrations
of optically active materials in the water column on
seagrass photosynthesis. Spectral discrimination between aquatic plant species must therefore concentrate on pigment related spectral features within the
visible wavelengths, where light penetrates the water column and can be reflected back to the sensor
(Fyfe, 2003). Absolute reflectances from submerged
species are generally low, often lower than the reflectance from a deep water column, and from a
visual perspective seagrass areas generally appear
darker than non-vegetated areas. One of the reasons
for seagrasses generally appearing darker in remote
sensing images and aerial photography is the shading that takes place within the canopy (see Chapter
13 for details).
In all physics-based applications where the substratum is mapped through a water column, a
bathymetry estimate (i.e. a water column depth estimate) is implicitly or explicitly involved. Optically
shallow waters are a special case in the remote sensing of aquatic systems. Under these conditions a
measurable signal will be detected from the substratum or plants, through the water column and through
the air–water interface. The water is considered to
be optically deep if there is no measurable influence
of the benthos or substratum on the remotely sensed
reflectance (Dekker et al., 2001).
The bathymetry and the water column optical
properties are thus important to remote sensing in
aquatic environments, in particular for seagrass and
macro-algal mapping (Dekker et al., 2001) since optimal results will be obtained from remote sensing if
the signal from the seagrass canopy is maximized
by compensating for the attenuating influence of
the water column. Whether a benthic feature such
as seagrass can actually be discriminated depends
on the spectral optical depth of the water column
(and the atmosphere between the sensor and the
water surface), the brightness and density of the
vegetation, and the spectral contrast between it and
the substratum, as well as on the spectral, spatial,
and radiometric sensitivity of the remote sensing
instrument.
Remote sensing of aquatic environments (seagrass, sand, macro-algae, muds, and coral reefs)
requires sensors with greater sensor signal-to-noise
ratio than those applied in terrestrial environments.
Coupled with this factor is the number of quantization levels to which the sensor can record, referred
to as the radiometric resolution of the sensor. This
must be high enough to allow a range of brightness
levels over which a classification can be performed
and sensitive enough to be able to detect the lower reflectance of the deeper seagrass beds (Dekker et al.,
2001). Seagrasses may grow with sparse cover and
can be spectrally confused with other benthic features such as areas of macro-algae, detritus, and
corals (Mumby et al., 1997a,b). The small size and/or
linear shape and patchy nature of many seagrass
meadows means that in many cases high spatial resolution is also required to accurately determine their
distribution and abundance. Another limiting factor
in accurately detecting change in seagrass meadows
may be the temporal cover of remote sensing data.
The spatial extent of meadows may decline in response to human impacts or natural dynamics and
may change rapidly, within weeks to months. For
example, monitoring seasonal change in the extent
and density of certain seagrass species would require
a series of temporal cover data sets. The issue of
temporal cover is being increasingly addressed with
the availability of more satellite sensors, some with