variables (Fig. 15.5; Knudby et al. 2010a), while random forest tree-models were
the best predictors of Fijian fish assemblages from both QuickBird and IKONOSderived habitat measures (Knudby et al. 2011). Most studies to date have shown
the strongest relationships between fish species richness and remote sensingderived structural complexity exists at relatively small spatial scales (2.5–42.5 m;
Kuffner et al. 2007; Purkis et al. 2008; Wedding et al. 2008; Pittman et al. 2009;
Knudby et al. 2010a); however, there is some evidence to suggest that coarser
scale (200–225 m) rugosity may also contribute to total fish biomass (Pittman
et al. 2009; Knudby et al. 2010a).
At present, the Wildlife Conservation Society in Fiji is using outputs of predictive models of fish species richness and food fish biomass, in concert with data
on reef resilience and opportunity costs to fishers, to optimize MPA network
design. The results are being used to provide recommendations to community
managers on options for reconfiguring the existing MPA network to reduce conflict and improve food security for the future.
15.3.3 Threat and Damage Assessments
Remote sensing data have become rapidly integrated into many local and global
scale monitoring efforts to assess threats and damage to coral reefs. In this section,
we present four brief case studies highlighting different applications for coral reef
monitoring and management: (1) water quality assessments in the Great Barrier
Reef lagoon; (2) predictions of crown-of-thorns (Acanthaster planci) outbreaks in
the Pacific; (3) oil spill threat mapping in the Gulf of Mexico; and (4) damage
assessments following a ship grounding in Australia.
Water quality. The Queensland Government in Australia has set an ambitious
goal ‘‘to ensure that by 2020 the quality of water entering the Reef from adjacent
catchments has no detrimental impact on the health and resilience of the Great
Barrier Reef’’ (State of Queensland 2009). In order to achieve this goal, the
management partners need to: assess changes to sediment and nutrient loads
entering the Great Barrier Reef lagoon; identify reefs at risk from flood plumes;
and assess reef ecosystem condition over time. Catchment models such as SedNet
and ANNEX, which incorporate digital elevation models and land use and land
cover maps derived from remote sensing data, can be used for estimating river end
loads of sediments and nutrients (McKergow et al. 2005a, b). Such maps can be
periodically updated to monitor temporal change based on land cover change in
response to management initiatives. New techniques have been recently used in
the Wet Tropics region to identify reefs at risk based on frequency of exposure to
flood plumes mapped and classified from remote sensing data (Devlin and
Schaffelke 2009). The boundaries of primary, secondary and tertiary plumes were
defined for each flood level based on threshold concentrations of chlorophyll-a and
colored dissolved organic matter (CDOM) derived from freely available MODIS
Aqua and Terra data (oceancolor.gsfc.nasa.gov). Results were combined in a GIS
418
S. Jupiter et al.
the best predictors of Fijian fish assemblages from both QuickBird and IKONOSderived habitat measures (Knudby et al. 2011). Most studies to date have shown
the strongest relationships between fish species richness and remote sensingderived structural complexity exists at relatively small spatial scales (2.5–42.5 m;
Kuffner et al. 2007; Purkis et al. 2008; Wedding et al. 2008; Pittman et al. 2009;
Knudby et al. 2010a); however, there is some evidence to suggest that coarser
scale (200–225 m) rugosity may also contribute to total fish biomass (Pittman
et al. 2009; Knudby et al. 2010a).
At present, the Wildlife Conservation Society in Fiji is using outputs of predictive models of fish species richness and food fish biomass, in concert with data
on reef resilience and opportunity costs to fishers, to optimize MPA network
design. The results are being used to provide recommendations to community
managers on options for reconfiguring the existing MPA network to reduce conflict and improve food security for the future.
15.3.3 Threat and Damage Assessments
Remote sensing data have become rapidly integrated into many local and global
scale monitoring efforts to assess threats and damage to coral reefs. In this section,
we present four brief case studies highlighting different applications for coral reef
monitoring and management: (1) water quality assessments in the Great Barrier
Reef lagoon; (2) predictions of crown-of-thorns (Acanthaster planci) outbreaks in
the Pacific; (3) oil spill threat mapping in the Gulf of Mexico; and (4) damage
assessments following a ship grounding in Australia.
Water quality. The Queensland Government in Australia has set an ambitious
goal ‘‘to ensure that by 2020 the quality of water entering the Reef from adjacent
catchments has no detrimental impact on the health and resilience of the Great
Barrier Reef’’ (State of Queensland 2009). In order to achieve this goal, the
management partners need to: assess changes to sediment and nutrient loads
entering the Great Barrier Reef lagoon; identify reefs at risk from flood plumes;
and assess reef ecosystem condition over time. Catchment models such as SedNet
and ANNEX, which incorporate digital elevation models and land use and land
cover maps derived from remote sensing data, can be used for estimating river end
loads of sediments and nutrients (McKergow et al. 2005a, b). Such maps can be
periodically updated to monitor temporal change based on land cover change in
response to management initiatives. New techniques have been recently used in
the Wet Tropics region to identify reefs at risk based on frequency of exposure to
flood plumes mapped and classified from remote sensing data (Devlin and
Schaffelke 2009). The boundaries of primary, secondary and tertiary plumes were
defined for each flood level based on threshold concentrations of chlorophyll-a and
colored dissolved organic matter (CDOM) derived from freely available MODIS
Aqua and Terra data (oceancolor.gsfc.nasa.gov). Results were combined in a GIS
418
S. Jupiter et al.