Analysis of approaches for monitoring biodiversity in the Commonwealth marine area: Solitary Islands

Thicklip trevally underwater.

IMAGE: Thicklip Trevally (Carangoides orthogrammus), Solitary Islands. Image: Reef Life Survey.

An important part of establishing monitoring protocols is to determine the optimum sampling regime for seabed biodiversity.

This project inventoried shelf rocky reef habitats, benthic communities and demersal fish at the temperate Shelf Rocky Reef Key Ecological Feature (KEF) near the Solitary Islands off New South Wales. Underwater imaging equipment was tested and improved during the process, and protocols were developed to improve the efficiency and rigour of future monitoring.


Non-extractive survey methodologies were developed, deployed and evaluated under a range of conditions. In addition, an experimental, two-day deployment of 30 baited remote underwater video (BRUV) systems located at varying distances apart examined spatial autocorrelation (the degree of clustering) between observations of selected fish assemblages. This evidence will help to determine the minimum distance between BRUVs at which samples are statistically independent and therefore useful for detecting trends. The effect of habitat characteristics (such as reef rugosity) on fish communities was also examined.

Map showing multi-beam sonar tracks

IMAGE: Additional multibeam sonar data acquired in the Solitary Islands KEF. Image: CSIRO.

Key findings

High-resolution multibeam sonar swath mapping and analysis of 40 km² of seafloor habitat identified ancient-coastline reefs and mobile sands on the seafloor south of South Solitary Island. The abundance of fish at every trophic level was strongly related to the three-dimensional structural complexity of reefs: generally, abundance increased with complexity. At small-scales, complexity influenced the abundance and composition of four of the five trophic groups examined.

Data from the experimental BRUVs deployment proved inconclusive, so the project team analysed a much larger data set from Queensland’s Moreton Bay. Initial model results indicated that the spatial correlation between observations of snapper was negligible after about 300 m (indicating that for statistical efficiency in these areas BRUVS should be placed at least 300 m apart).

The project examined how to streamline the manual scoring process of complex, underwater video images, and defined a ‘stopping rule’ after which little further information would be gained by further analysis. With 25 points per image, 20 images were sufficient to characterise benthic biodiversity in the Solitary Islands KEF.

New knowledge and opportunities

The distribution and importance of rocky reefs in this area appear to part of a larger complex of reefs associated with ancient coastlines that occur around Australia.

Fine-scale BRUV data on the distribution of selected fish are being combined with broader-scale (swath acoustic) mapping to predict the occurrence of different species over the broader area of the KEF that will be useful in determining its ecological values and status.

Some 30,000 image labels have been provided to University of Sydney researchers to ‘train’ supervised automatic image classification algorithms.

Outputs and outcomes

This project has generated an improved understanding of the location and morphology of shelf rocky reefs in Commonwealth waters near the Solitary Islands and key determinants of demersal fish community composition in this KEF. It has inventoried demersal fish and benthic communities and identified effective methods of design, collection and analysis of underwater imagery in rocky reef habitat.


Keith Hayes
(03) 6232 5260