Since beds of horse mussels have acoustic reflectivity characteristics that often differ markedly from the surrounding seabed, it is possible to locate and map the broad areas occupied by the beds remotely. This has been done in Strangford Lough and off the Lleyn Peninsula using RoxAnnÔ acoustic seabed discrimination equipment and off the northern tip of the Isle of Man using side-scan sonar. We would recommend that a combination of RoxAnnÔ and side-scan be used. It is essential that Differential GPS is used for position fixing and would recommend the use of a side-scan system that automatically logs the positions and makes allowance for the lay back of the sonar fish. The best practice for post-survey processing of such data is still evolving, but it is likely to involve the use of Geographic Information Systems for bringing displays to common scales, formats and displays.

Ground-truthing is probably best carried out by drop-down video. Since the absolute values given by acoustic systems are seldom 100% stable over time and under different circumstances, we would recommend that ground truthing is carried out each time a monitoring survey of a particular bed is done. Special attention will need to be made to determining an agreed standard or standards for what is to be regarded as the edge of the bed. This will particularly be a problem if there is fragmentation at the edge of the bed or if it continues as a wider spread of isolated clumps. If there is adequate standardisation, and using facilities either within the GIS software or using an area measurement package such as Sigma-Scan, it should be possible to compare areas determined on different surveys over time.

With RoxAnnÔ the precise extent of the footprint of each recorded reading is at present not known so it is impossible to indicate the minimum patch size detectable. Indeed, even if there is a pronounced difference in the acoustic properties of the horse mussel bed and the underlying stony seabed, it will not be clear whether intermediate acoustic readings come from footprints that fall partly on the two types of ground or whether they come from deposits of dead shell that give intermediate acoustic readings. Because of these complications, special efforts should be made to make ground truth video observations at the margins of the bed. Consideration should be given to undertaking video transects across the edge of the bed along tracks that are in repeatable directions. Special attention might also need to be given to the boundaries of the bed where they coincide with local inshore fishery, ie the 6 mile limit of the jurisdiction of a Sea Fisheries Committee where byelaws limit the size of vessels or types of gear.

RoxAnnÔ makes measurements just along the track taken by the survey vessel, so interpolation is needed between survey lines. Surfer and DGM3 are amongst the interpolation packages routinely used with RoxAnnÔ data. Particularly where the beds are somewhat patchy, there are currently concerns about the extent to which the interpolated outputs really represent the spread of the patches. More work is needed to see how differences in the spacing of survey lines and or their orientation affects the displayed output from the interpolations. However, if side-scan is used in conjunction with RoxAnnÔ it should be possible within the GIS to make more intelligent adjustments to the end result.

On side-scan sonar displays horse mussel beds show up both as having a different texture from the surrounding seabed and, particularly where mounds or waves form, in a different pattern of relief. From acoustic shadows it is possible to measure the height of features in the bed. Some modern side-scan sonar systems have built in software to do this.

Next section                     References