bleaching Degree Heating Week produced by the U.S. National Oceanographic
and Atmospheric Administration (NOAA) Coral Reef Watch Bleaching Outlook
program were only released after successful early warning predictions of bleaching
events were first established using experimental products (Liu et al. 2003).
15.2.3 Data Requirements and Limitations
A summary of the most common research and management applications for coral
reef remote sensing are displayed in Table 15.2 along with some of the corresponding sensor options for producing the required data products. The gaps in the
table also indicate some of the current spatial and temporal limitations of remote
sensing to meet user needs. Some of these limitations are described below.
In most cases, it is not possible to discriminate benthic, sessile organisms to the
species level. For example, because corals all contain similar pigments and
because colonies typically occur at the scale of less than 0.5 m
2 , it is infeasible to
provide comprehensive coral species diversity maps. For the same reasons, it is not
possible to consistently distinguish spectrally similar, but functionally distinct
categories, such as hard and soft corals. This may have implications if researchers
or managers are trying to scale-up properties or processes (e.g., carbonate production) derived from reef building hard corals that are spectrally indistinguishable
for non-calcifying organisms. These problems can partially be addressed by
integrating data from multiple sensor systems, incorporating a combination of
spatial and spectral analysis techniques, and linking field data describing dominant
components of community structure to habitat classes (Mumby 2001).
Mapping and monitoring the extent of coral bleaching has become a strong
priority given the increasing frequency and severity of events in the past few
decades. As corals expel their pigmented zooxanthellae during a bleaching event,
the resulting strong change in color would suggest that remote sensing at visible
wavelengths would be useful for quantifying extent of coral mortality (Clarke et al.
2000; Andréfouët et al. 2002). However, its application has been limited given the
patchiness of events, the varying intensities and susceptibilities of coral to
bleaching, the inability to detect changes deeper than just shallow depths, and the
need for imagery to be acquired within specific timeframes (Andréfouët et al.
2002). Moreover, it is often difficult to acquire imagery over bleached reefs prior
to algal colonization of dead coral surfaces, making differentiation of dead versus
live coral difficult (Clarke et al. 2000). These temporal limitations are particularly
challenging when relying on data from satellite sensors with low revisit frequencies, which makes the likelihood of acquiring cloud-free images from the
tropics directly following a disturbance event fairly low.
410
S. Jupiter et al.
and Atmospheric Administration (NOAA) Coral Reef Watch Bleaching Outlook
program were only released after successful early warning predictions of bleaching
events were first established using experimental products (Liu et al. 2003).
15.2.3 Data Requirements and Limitations
A summary of the most common research and management applications for coral
reef remote sensing are displayed in Table 15.2 along with some of the corresponding sensor options for producing the required data products. The gaps in the
table also indicate some of the current spatial and temporal limitations of remote
sensing to meet user needs. Some of these limitations are described below.
In most cases, it is not possible to discriminate benthic, sessile organisms to the
species level. For example, because corals all contain similar pigments and
because colonies typically occur at the scale of less than 0.5 m
2 , it is infeasible to
provide comprehensive coral species diversity maps. For the same reasons, it is not
possible to consistently distinguish spectrally similar, but functionally distinct
categories, such as hard and soft corals. This may have implications if researchers
or managers are trying to scale-up properties or processes (e.g., carbonate production) derived from reef building hard corals that are spectrally indistinguishable
for non-calcifying organisms. These problems can partially be addressed by
integrating data from multiple sensor systems, incorporating a combination of
spatial and spectral analysis techniques, and linking field data describing dominant
components of community structure to habitat classes (Mumby 2001).
Mapping and monitoring the extent of coral bleaching has become a strong
priority given the increasing frequency and severity of events in the past few
decades. As corals expel their pigmented zooxanthellae during a bleaching event,
the resulting strong change in color would suggest that remote sensing at visible
wavelengths would be useful for quantifying extent of coral mortality (Clarke et al.
2000; Andréfouët et al. 2002). However, its application has been limited given the
patchiness of events, the varying intensities and susceptibilities of coral to
bleaching, the inability to detect changes deeper than just shallow depths, and the
need for imagery to be acquired within specific timeframes (Andréfouët et al.
2002). Moreover, it is often difficult to acquire imagery over bleached reefs prior
to algal colonization of dead coral surfaces, making differentiation of dead versus
live coral difficult (Clarke et al. 2000). These temporal limitations are particularly
challenging when relying on data from satellite sensors with low revisit frequencies, which makes the likelihood of acquiring cloud-free images from the
tropics directly following a disturbance event fairly low.
410
S. Jupiter et al.