Characterising the fish habitats of the Recherche Archipelago - Benthic habitats (FRDC (2001/060)

Full description

The Fisheries Research and Development Corporation Project (FRDC) No. 2001/060 characterised and identified the distribution of the different benthic habitats of the Recherche Archipelago and identified the distribution of assemblages of fish, mobile invertebrates, filter feeding communities and seagrasses and the patterns and processes which structure them. Benthic habitat were determined from the interpretation of Landsat imagery, sidescan sonar imagery and video validation data collected within the Recherche region from May 2002 to November 2003.

Lineage

Maintenance and Update Frequency: notPlanned

Statement: Habitat definition - The level to which a habitat can be classified will largely be determined by the resolution of the survey method used, although all types of survey methods should result in the same habitat classes or community definitions to allow cost-effective mapping across large areas. The hierarchical classification scheme developed for the Recherché allow habitats to be both broadly defined using remote sensing methods, such as Landsat or sidescan sonar, or in more detail using methods such as video, to determine factors, such as substrate type, relief and percent cover. For the purposes of this study, broad habitats defined include: low profile reef, high profile reef, dense seagrass, medium seagrass, sparse seagrass, sand, sand with medium vegetation, sand with sparse vegetation, rhodoliths and deep water sponge communities. Habitats have been broadly defined at the 1:100,000 cartographic scale, the minimum mapping unit being set at 100m. Habitat polygons have not been interpreted below this size for the purposes of the broadscale mapping required, although the nature of the data in specific locations allows much more detailed, functional habitats to be mapped (eg. using 1m resolution sidescan sonar or video data). Reliability measures - Within the attribute table, each polygon classified was tagged according to the data sources used to help define it (eg. sidescan, drop video, towed video, Landsat, Kirkman data - see individual metadata for source data accuracy). Reliability weightings can now be assigned to each polygon based on the inputs used to classify it. For example, Landsat imagery would be a more reliable source to classify habitats with, when accompanied by towed video or sidescan data, as opposed to areas that only have isolated drop camera footage. Habitat validation - Spatial datasets, such as sidescan and Landsat imagery, were interpreted to define habitat polygons within a Geographic Information System (GIS). Given the clarity of the water, the extent of large features could often be determined using Landsat at 30-40m in depth, although it was not known what the features were. Sidescan data was often then used to validate features visible on Landsat. For example, sidescan data was able to determine the presence of rhodolith beds, which otherwise were not distinguishable from what was expected to be low profile reef. Similarly, seagrass communities outlined using Landsat, were also further defined according to the degree of cover (eg. sparse, medium or dense seagrass) using both sidescan and video. In many cases, the video validated cover differences that were observable on the sidescan record. Classified video points were used to not only validate polygons but also where possible, extend the boundaries into areas that did not have spatial data inputs.

Notes

Credit
Fisheries Research and Development Corporation (FRDC)

Purpose
OBJECTIVES 1. To identify, classify and map the distribution of, different benthic habitats in the Recherche Archipelago and link their distributions to bottom type and exposure to swells and currents. 2. To provide detailed ecological information to ensure the responsible management of aquaculture fisheries in the region. 3. To increase community awareness of fish habitats through community involvement in the development of baseline studies and ongoing monitoring programs.