Integrating dual frequency side scan sonar and high spatial resolution satellite imagery for monitoring coral reef benthic communities

In this study, the characteristics of both optical and acoustic data types were compared to determine if synergistic use of both methods improved the accuracy of classification of benthic reefs and associated habitats. IKONOS imagery and dual frequency side scan sonar data were acquired in the western Caribbean encompassing coral, seagrass, algal and sediment habitats. Both data types were analysed in isolation and in combination at both habitat and community classification levels. The accuracies achieved with the combined dataset (61% and 52% for the coarse and medium classification levels, respectively) were significantly higher than what was achieved on the basis of the two datasets used in isolation, demonstrating the synergy of the acoustic and optical datasets. The greater accuracy improvements were attributed to the higher spatial resolution of the sonar data, its greater depth penetration, and information contained in the sonograms regarding structural organisation of the habitats. Misclassifications using the combined datasets could be attributed to spectral and textural similarities between different classes, quantity and classification of the ground truth sites, and uncertainties in the co-registration of the two datasets.