Monitoring coral reef bleaching from space : a feasibility study using a physics-based remote sensing approach
There is a requirement for effective, worldwide monitoring of the health of coral reef systems, in particular the coral bleaching phenomenon. The transient nature of bleaching events, the extent and remoteness of tropical reefs, and the complex nature of the light field in submerged coral reef environments, requires novel and sophisticated remote sensing-based detection and monitoring approaches. An objective, physics-based approach (SAMBUCA) was developed for retrieving water column concentrations, bathymetry, and bottom substrate composition from remote sensing data on a per-pixel basis. MERIS images, offering a suitable spatial and temporal coverage for monitoring coral bleaching, were acquired before and during a minor bleaching event at Heron reef (Australia) in 2004. Benthic field surveys, conducted during and after the bleaching event, were analysed, and the optical properties of Heron reef substrates and waters were investigated. The latter represents the most comprehensive optical characterisation of tropical coral waters. SAMBUCA was parameterised with the optical properties of Heron coral reef waters and benthos, and applied to the FR MERIS data. Bathymetry was mapped to a high degree of accuracy, optically deep water was identified, and changes in benthic substrate composition were inferred pointing towards a minor bleaching event. Changes in per-pixel percent cover of live coral were used as a proxy indicator of coral bleaching. This study showed that environmental information on coral bleaching may be objectively retrieved at the appropriate scales for climate change assessment. Further development and validation of applying the SAMBUCA physics-based approach to global remote sensing data is recommended. Ultimately, this may lead to the implementation of regional or global tropical coral health monitoring systems. The approach presented in this thesis is unique in its application of radiative transfer theory to coral reef environments. It is furthermore applicable to other areas such as seagrass beds, and this thesis therefore represents a significant contribution to the remote sensing of (optically shallow) aquatic environments.