Investigating the effects of higher spatial resolution on benthic classification accuracy at Midway Atoll

Abstract

Effective monitoring of coral reefs is important for ecological and economic reasons, and satellite remote sensing has been shown to be useful for mapping and monitoring these ecosystems. This thesis will compare 2 multispectral systems and investigate the effects of increased spatial resolution on benthic classifications in the highly heterogeneous coral reef environment of Midway Atoll. It will evaluate the utility of QuickBird's increased spatial resolution compared to IKONOS imagery in the same study area at multiple scales. Previous studies (e.g.

Mumby and Edwards, 2002; Capolsini et al.

2003; Wang et al.

2004; Benefeild et al.

2007) comparing various satellite sensors suggest that greater spatial resolution should lead to more accurate classifications, but a direct comparison of QuickBird and IKONOS sensors has not been carried out in marine environments. Light interactions in marine environments are complex and add difficulty to spectral discrimination, producing more variable results in classification accuracy than in terrestrial environments. This research does not find any significant improvements in thematic mapping accuracy of benthic environment from QuickBird's higher spatial resolution satellite imagery. Additionally, a cost benefit analysis did not show a decisive advantage in choosing either imagery type for the application of monitoring the extent, biodiversity, and health of coral reef habitats.