Linking seafloor characteristics to biological communities

The effective management of our marine ecosystems requires the capability to identify, characterise and predict the distribution of benthic biological communities within the overall seascape architecture. In order to achieve this, detailed knowledge of resources relevant to the scale of resource exploitation is required. The QUEST decision tree classifier was used to predict benthic biological community distributions in the Anglesea site, a 54 square kilometre area off the central coast of Victoria from depths of 7 to 56 metres. This paper presents a method to integrate bathymetry and backscatter derivative data from high resolution multibeam hydroacoustics with acoustically positioned towed video data. A set of 11 derived predictor variables were integrated with video observation data to classify 7 dominant benthic biological communities. QUEST runs with a combination of bathymetry and backscatter predictor variables produced significantly better results than other methods employed. Predictor variables influencing the distribution of biological communities were found to vary with depth. This paper demonstrates that decision tree classifiers are capable of integrating variable data types and are highly adaptable for mapping benthic biological communities, critical to maintain biodiversity and other system services in the marine environment. Examples of potential applications integrating seascape data for resource and biodiversity assessment at arrange of spatial scales will be discussed.