because of the long history of data acquisition, global coverage of coral reef areas,
and relatively low cost. Multispectral satellite sensors are still evolving. Most
notably, spatial resolution of the sensors is getting higher (i.e., smaller pixel sizes),
contributing to higher habitat classification accuracy and more detailed mapping of
spatial features on coral reefs (Mumby and Edwards 2002; Andréfouët et al. 2003).
Multispectral satellite remote sensing of coral reefs has approximately 40 years
of history, dating back to the launch of the first Landsat satellite in 1972. The first
application using multispectral imagery for coral reef mapping was achieved in the
Great Barrier Reef of Australia using Landsat MSS data with spatial resolution of
80 m (Smith et al. 1975; Jupp et al. 1985), where analysis focused on products for
approximating bathymetry and reef geomorphologic zones. Deployment of SPOT
HRV and Landsat TM in the 1980s gave the scientific community access to data
with improved spatial resolution of 20–30 m. Though the data contributed significantly to improved regional land cover mapping, mapping habitats on coral
reefs was limited to several basic classes in the 1980s (Vercelli et al. 1988). During
this period the potential for characterization and quantification of coastal tropical
environments for management and planning was suggested (Loubersac and Populus 1986), but nonetheless Kuchler et al. (1988) concluded ‘‘that current technologies are unable to fulfill this goal because of their lack of refinement for coral
reef environments.’’ In the early 1990s, some studies demonstrated mapping and
change detection using Landsat TM (e.g., Zainal et al. 1993; Ahmad and Neil
1994), but they did not include quantitative classification accuracy and the use of
remote sensing approaches to management of coral reefs remained scarce.
Significant advances occurred in the late 1990s from both application and
technology perspectives. Green et al. (1996) reviewed applications of remote
sensing for the assessment of tropical coastal resources and pointed out the
importance of comparing the capabilities of different sensors in order to aid
managers’ decisions in selecting a remote sensing technique. Quantitative evaluation of cost (inclusive of both image and processing costs), benefit and classification accuracy of several remote sensors was assessed by Mumby et al. (1997). A
systematic classification scheme of benthic features for habitat mapping of coral
reefs, which is widely used today, was also developed (Mumby and Harborne
1999). A series of studies serving as guidelines were summarized and included in a
remote sensing handbook for tropical coastal management (Green et al. 2000).
Significant advances in technology were made by the Landsat 7 Enhanced
Thematic Mapper (ETM+) and IKONOS sensors, which were both launched in
1999. Though the spatial resolution and spectral arrangement of Landsat ETM+
was the same as Landsat TM, the acquisition schedule specifically targeted coral
reefs throughout the world (Arvidson et al. 2001). This meant that mapping all the
coral reefs was now possible by satellite data (Andréfouët et al. 2006). IKONOS
provided the first publicly-available commercial multispectral data with high
spatial resolution (4 m), which delivered high classification accuracy and a scale
suitable for mapping details within geomorphic zones (Mumby and Edwards 2002;
Andréfouët et al. 2003).
52
H. Yamano
and relatively low cost. Multispectral satellite sensors are still evolving. Most
notably, spatial resolution of the sensors is getting higher (i.e., smaller pixel sizes),
contributing to higher habitat classification accuracy and more detailed mapping of
spatial features on coral reefs (Mumby and Edwards 2002; Andréfouët et al. 2003).
Multispectral satellite remote sensing of coral reefs has approximately 40 years
of history, dating back to the launch of the first Landsat satellite in 1972. The first
application using multispectral imagery for coral reef mapping was achieved in the
Great Barrier Reef of Australia using Landsat MSS data with spatial resolution of
80 m (Smith et al. 1975; Jupp et al. 1985), where analysis focused on products for
approximating bathymetry and reef geomorphologic zones. Deployment of SPOT
HRV and Landsat TM in the 1980s gave the scientific community access to data
with improved spatial resolution of 20–30 m. Though the data contributed significantly to improved regional land cover mapping, mapping habitats on coral
reefs was limited to several basic classes in the 1980s (Vercelli et al. 1988). During
this period the potential for characterization and quantification of coastal tropical
environments for management and planning was suggested (Loubersac and Populus 1986), but nonetheless Kuchler et al. (1988) concluded ‘‘that current technologies are unable to fulfill this goal because of their lack of refinement for coral
reef environments.’’ In the early 1990s, some studies demonstrated mapping and
change detection using Landsat TM (e.g., Zainal et al. 1993; Ahmad and Neil
1994), but they did not include quantitative classification accuracy and the use of
remote sensing approaches to management of coral reefs remained scarce.
Significant advances occurred in the late 1990s from both application and
technology perspectives. Green et al. (1996) reviewed applications of remote
sensing for the assessment of tropical coastal resources and pointed out the
importance of comparing the capabilities of different sensors in order to aid
managers’ decisions in selecting a remote sensing technique. Quantitative evaluation of cost (inclusive of both image and processing costs), benefit and classification accuracy of several remote sensors was assessed by Mumby et al. (1997). A
systematic classification scheme of benthic features for habitat mapping of coral
reefs, which is widely used today, was also developed (Mumby and Harborne
1999). A series of studies serving as guidelines were summarized and included in a
remote sensing handbook for tropical coastal management (Green et al. 2000).
Significant advances in technology were made by the Landsat 7 Enhanced
Thematic Mapper (ETM+) and IKONOS sensors, which were both launched in
1999. Though the spatial resolution and spectral arrangement of Landsat ETM+
was the same as Landsat TM, the acquisition schedule specifically targeted coral
reefs throughout the world (Arvidson et al. 2001). This meant that mapping all the
coral reefs was now possible by satellite data (Andréfouët et al. 2006). IKONOS
provided the first publicly-available commercial multispectral data with high
spatial resolution (4 m), which delivered high classification accuracy and a scale
suitable for mapping details within geomorphic zones (Mumby and Edwards 2002;
Andréfouët et al. 2003).
52
H. Yamano