Benthic Habitat Mapping in Cockburn Sound 2004

Brief description

Detailed surveys were undertaken to accurately map the benthic coverage and assemblages of the Eastern Shelf, Cockburn Sound for summer 2004, using sidescan sonar and extensive groundtruthing. The priorities were to obtain additional information on the following: - Species assemblage (for example, distinguish between seagrass species, reef areas, wrack material, mussel beds); - Vegetation percentage cover; - Presence of epiphyte material; and - Habitat information for areas deeper than 10 m or with poor water clarity. These results were presented in Report No. 321/1

Lineage

Maintenance and Update Frequency: notPlanned

Statement: - Habitat mapping - This project for habitat mapping in the Eastern Shelf of Cockburn Sound made use of the following data: - Visual examination of the 2002 aerial imagery to define the SSS survey area and provide indicative habitat information for areas not covered by the SSS; - Extensive Sidescan Sonar (SSS) survey across the Eastern Shelf region; - Towed underwater video to provide ground truth data and assist in the interpretation of the SSS data; and - Spot dives in selected areas which were identified as having high habitat complexity. Survey sites are shown in thumbnail. Aerial photography The 2002 imagery was used as an initial guide to the extents and location of the vegetated and unvegetated areas within Cockburn Sound. Scrutiny of the imagery also allowed the identification of those areas which could not be accurately mapped from the aerial photographs (due to high water column turbidity or water depth in excess of 10 m). This information was used to define the region of the SSS survey. Sidescan Sonar High frequency Sidescan Sonar (SSS) has been popular for the mapping and visualisation of the seabed since the 1980s and provides a time and cost effective method for the delineation of habitat types in relation to their acoustic properties. Sidescan Sonars utilize two transducers which produce a fan-shaped pulse (ping) of sound energy. A towfish containing the transducers is towed behind a boat at a constant depth below the water surface, transmitting pings through the water column. As the ping interacts with the seabed, a small proportion is reflected back to sensors on the towfish (the 'return' signal). The energy of the return signal is amplified and recorded Towed video groundtruthing Despite the wealth of data provide by the SSS imagery on the hardness, texture and relief of the substrate, groundtruthing is still essential in enabling benthic habitats to be confidently and accurately defined. The first component of the ground truth surveys was completed using towed underwater video. An underwater video camera was towed behind the survey vessel at speeds of 1.5-2.5 knots and the resultant digital signal was supplied to an onboard video recorder and screen. The depth of the camera was controlled manually to ensure that it remained an optimal distance (<0.5 m) from the seabed for identification of the benthic habitats. The camera, at this height above the seabed, records benthic habitat across an approximately 1m wide transect. The underwater video was used to examine features of interest identified from the SSS imagery and to identify habitat types which corresponded to the different tones and textures obtained in the SSS imagery. This was done by recording the position of boundaries between the different habitat types identified as well as recording habitat characteristics at fixed intervals through the video footage within each habitat type. This allowed the classification of the SSS imagery. Spot dives Spot dives were used to examine areas of high habitat complexity following the initial analysis of the SSS imagery and video footage. These spot dives assisted in the determination of the composition and small-scale variability present in these complex areas. The dives also allowed more detailed descriptions of each habitat type to be made. Dives were carried out at five separate locations, chosen to cover a range of habitat types as determined by the analysis of the SSS imagery. The location of each dive site was determined using a DGPS and the divers annotated laminated copies of the SSS imagery with information on habitat composition.

Statement: - Mapping methods - The boundaries of the habitat types were defined visually from the SSS data at a scale of 1:5,000. In areas where SSS coverage was not obtained, the 2002 aerial photography was employed. This mapping was undertaken using ArcView to create a single coverage of benthic habitat types. The habitat polygons were identified based on visual interpretation of the SSS imagery (and aerial photography) using predefined mapping control rules. The previous (2002) habitat mapping from aerial photography distinguished habitat classes: bare sand, vegetated and land. In the absence of groundtruth survey data for 2002, it was not possible to distinguish between seagrass and reef, and both were classified as vegetated. Consistency in the mapping of these habitat types was maintained through the use of the following control rules (Kendrick et al., 2000) which were implemented using the semi-automated Spann-Wilson segmentation method: 1. Vegetated patches, that were isolated and less than 30 m2, were not mapped. 2. Vegetated patches, that were greater than 30 m2 and less than 100 m2, were mapped as separate patches when the distance between one patch and another was greater than the diameter of the patch. 3. Vegetated patches, that were greater than 30 m2 and less than 100 m2, were mapped as a single meadow when the distance between one patch and another was less than the diameter of the patch. 4. Vegetated patches greater than 100 m2 were mapped separately. Unvegetated regions within these patches, with areas greater than 100 m2, were mapped as bare sand. For the present study, the mapping control rules were modified to more closely reflect the resolution of the groundtruthing methods employed as follows: 1. Isolated habitat features of less than 200 m2 were not mapped. 2. Seagrass patches mapped individually if clearly distinguishable from surrounding patches on SSS imagery (at scale 1:5,000). 3. All habitat features greater than 200 m2 were mapped from the SSS imagery at a scale of 1:5,000. 4. Areas covered by a combination of seagrass and reef structures, in which the two habitats were indistinguishable in the SSS imagery (at scale of 1:5,000), were classified as mixed habitats. 5. In areas for which detailed habitat data from spot dives was available, separate mapping using greater detail was carried out. For the present study, the habitat classes were expanded to enable more detailed mapping of the broad range of benthic habitats identified. The habitat classification of each polygon was determined from an examination of the underwater video and spot dive information. Whilst the soft sediment habitat corresponds directly to the old unvegetated habitat classification, additional habitat types have been added to more accurately describe the range of vegetated habitats present. The resolution of the mapping depended upon the types of survey coverage achieved within each area. For those areas solely covered by the 2002 aerial imagery, mapping resolution was low, with only bare substrate areas and vegetated areas (areas consisting of seagrass, wrack or reef) being mappable. For those areas covered by SSS, the extents of the habitats could be more accurately defined, and more detailed classification of habitat types made. Areas covered by SSS and towed video could be mapped at a higher resolution still, with habitat boundaries and habitat types more accurately mapped. Those areas covered by SSS and spot dives (and sometimes towed video) could be mapped at an even higher resolution, with the production of highly detailed habitat maps possible.

Notes

Credit
This study was funded by the following members of the Cockburn Sound Management Council: Department of Industry and Resources, LandCorp and Fremantle Ports.

Purpose
Accurately map the benthic coverage and assemblages of the Eastern Shelf, Cockburn Sound for summer 2004, using sidescan sonar and extensive groundtruthing.