based sensors has excellent potential to discriminate both hydrodynamically bounded
landscape components and the internal fine-scale spatial distribution of their
metabolically distinct benthic communities (Green et al., 1996; Green et al., 2000;
Maltus and Mumby; 2003; Mazel, 1999). Available mapping sensors may be
categorized according to their appropriate application within the framework for the
proposed reef system E model (Figure 1). Following Green et al. (2000), we divide the
direct remote sensing of reefs into two categories: 1) moderate resolution imaging of
landscape zones with corresponding functional modes (Andréfouët et al., 2001b); and
2) the finer-scale mapping of intra-zone biotopes (Mumby et al., 1997a; Mumby and
Edwards, 2002).
Moderate spatial resolution satellite sensors (>10 m to <100 m) are useful in
creating maps of reef system zones (see Chapter 12). Compared to other digital
sensors, these instruments have the longest history of use in coastal studies. This sensor
class includes the Landsat Multispectral Scanner (MSS), Thematic Mapper (TM), and
Enhanced Thematic Mapper Plus (ETM+), the SPOT High Resolution Visible Scanner
(HRV), and the Indian Remote Sensing Linear Imaging Self-Scanning Sensor (LISS)
(Green et al., 2000). Most studies based on these moderate resolution satellite sensors
have provided geographical information on geomorphological zones, such as forereef,
reef crest, algal rim, lagoon, or emergent atoll (Green et al., 1996).
Khan et al. (1992) used a principal components approach to classify sub-tidal
habitats in the Arabian Gulf on a Landsat TM image. At a Dominican Republic site,
Michalek et al. (1993) examined the use of Landsat TM images for delineating
geomorphology and three basic bottom classes - coral, seagrass, and sand. Ahmad and
Neil (1994) compared the capabilities of Landsat TM and Landsat MSS for mapping
coral reef zonation at Heron Reef, within the Great Barrier Reef of Australia, and found
that TM provides greater geomorphological detail than MSS. More recently,
Matsunaga and Kayanne (1997) used Landsat TM to investigate coarse-scale temporal
habitat change on fringing reefs at Lshigaki Island in the Ryukyu chain. Peddle et al.
(1995) carried out an analysis of radiative transfer in the use of SPOT HRV for
mapping shallow habitats at the Fiji Islands. LeDrew et al. (1995) also studied the coral
reef ecosystem at the Fiji Islands using SPOT HRV, and determined that for depths less
than 10 m, SPOT bands 1 and 2 were useful in the spectral discrimination of coral reef
features. Mumby et al. (1997b) determined that Landsat TM and SPOT HRV have
limited capabilities to map seagrass standing crop at broad spatial scales, provided that
extensive field verification efforts are undertaken.
Bainbridge and Reichelt (1988) concluded that moderate resolution satellite
imagery is more appropriate for studying reef geomorphology than reef biology. The
more recent studies cited above have substantiated this early judgment on the
limitations of intermediate resolution satellite sensors, concluding that moderate spatial/
spectral resolution satellite sensors are useful only for the coarse descriptive mapping of
reef systems (Malthus and Mumby, 2003). In a comparative study of the capabilities of
various sensors for coral reef habitat mapping conducted for reefs of the Turks and
Caicos Islands, Mumby et al. (1997a) found that for all moderate resolution satellite
scanners, map accuracy typically dropped significantly from coarse- to fine-scale
habitat discrimination. Moderate resolution satellite imagery is not sufficient for the
detailed mapping of benthic habitats primarily due to intra-pixel mixing (Andréfouët
et al., 2003a). Additional limitations are posed by the loss of radiometric contrast due
to atmospheric effects (Lubin et al., 2001) and limited spectral resolution (Hochberg
and Atkinson, 2003; Malthus and Mumby, 2003). Intermediate resolution satellite
117
Coral Reef Ecosystem Process Studies & Remote Sensing
landscape components and the internal fine-scale spatial distribution of their
metabolically distinct benthic communities (Green et al., 1996; Green et al., 2000;
Maltus and Mumby; 2003; Mazel, 1999). Available mapping sensors may be
categorized according to their appropriate application within the framework for the
proposed reef system E model (Figure 1). Following Green et al. (2000), we divide the
direct remote sensing of reefs into two categories: 1) moderate resolution imaging of
landscape zones with corresponding functional modes (Andréfouët et al., 2001b); and
2) the finer-scale mapping of intra-zone biotopes (Mumby et al., 1997a; Mumby and
Edwards, 2002).
Moderate spatial resolution satellite sensors (>10 m to <100 m) are useful in
creating maps of reef system zones (see Chapter 12). Compared to other digital
sensors, these instruments have the longest history of use in coastal studies. This sensor
class includes the Landsat Multispectral Scanner (MSS), Thematic Mapper (TM), and
Enhanced Thematic Mapper Plus (ETM+), the SPOT High Resolution Visible Scanner
(HRV), and the Indian Remote Sensing Linear Imaging Self-Scanning Sensor (LISS)
(Green et al., 2000). Most studies based on these moderate resolution satellite sensors
have provided geographical information on geomorphological zones, such as forereef,
reef crest, algal rim, lagoon, or emergent atoll (Green et al., 1996).
Khan et al. (1992) used a principal components approach to classify sub-tidal
habitats in the Arabian Gulf on a Landsat TM image. At a Dominican Republic site,
Michalek et al. (1993) examined the use of Landsat TM images for delineating
geomorphology and three basic bottom classes - coral, seagrass, and sand. Ahmad and
Neil (1994) compared the capabilities of Landsat TM and Landsat MSS for mapping
coral reef zonation at Heron Reef, within the Great Barrier Reef of Australia, and found
that TM provides greater geomorphological detail than MSS. More recently,
Matsunaga and Kayanne (1997) used Landsat TM to investigate coarse-scale temporal
habitat change on fringing reefs at Lshigaki Island in the Ryukyu chain. Peddle et al.
(1995) carried out an analysis of radiative transfer in the use of SPOT HRV for
mapping shallow habitats at the Fiji Islands. LeDrew et al. (1995) also studied the coral
reef ecosystem at the Fiji Islands using SPOT HRV, and determined that for depths less
than 10 m, SPOT bands 1 and 2 were useful in the spectral discrimination of coral reef
features. Mumby et al. (1997b) determined that Landsat TM and SPOT HRV have
limited capabilities to map seagrass standing crop at broad spatial scales, provided that
extensive field verification efforts are undertaken.
Bainbridge and Reichelt (1988) concluded that moderate resolution satellite
imagery is more appropriate for studying reef geomorphology than reef biology. The
more recent studies cited above have substantiated this early judgment on the
limitations of intermediate resolution satellite sensors, concluding that moderate spatial/
spectral resolution satellite sensors are useful only for the coarse descriptive mapping of
reef systems (Malthus and Mumby, 2003). In a comparative study of the capabilities of
various sensors for coral reef habitat mapping conducted for reefs of the Turks and
Caicos Islands, Mumby et al. (1997a) found that for all moderate resolution satellite
scanners, map accuracy typically dropped significantly from coarse- to fine-scale
habitat discrimination. Moderate resolution satellite imagery is not sufficient for the
detailed mapping of benthic habitats primarily due to intra-pixel mixing (Andréfouët
et al., 2003a). Additional limitations are posed by the loss of radiometric contrast due
to atmospheric effects (Lubin et al., 2001) and limited spectral resolution (Hochberg
and Atkinson, 2003; Malthus and Mumby, 2003). Intermediate resolution satellite
117
Coral Reef Ecosystem Process Studies & Remote Sensing