MULTIBEAM SONAR MAPPING AND SCALLOP STOCK ASSESSMENT: GIS DATA INTEGRATION IN SUPPORT OF SUSTAINABLE FISHERIES MANAGEMENT

(1)Gerry Sutton, (1) Dwyer, N., (2) Tully, O., Hervas, A., Hickey, J., (3) Monteys, X.

(1) Coastal and Marine Resources Centre, (IE)
(2)Irish Sea Fisheries Board (BIM), (IE)
(3) Geological Survey of Ireland, (IE)

Abstract
This study describes the integration of biological and geophysical data within a GIS environment. The research is being undertaken as a principal component in a multidisciplinary approach to the development a strategic plan for the management of scallop stocks (Pecten maximus) off the south east coast of Ireland. A series of GIS tools are used in conjunction with a geodatabase in order to assist in evaluating the relationship between seabed sediment type and scallop stock density. Geophysical data layers including multibeam sonar maps (MBES bathymetry, morphology and acoustic backscatter) and other seabed data layers (sediment samples, sub-sea video imagery, statistical sediment classifications) are overlain and analysed in combination with layers of quantitative biological data showing scallop catch rates. Initial results indicate that high scallop catch rates are strongly correlated with one of two predominant and acoustically distinct sediment types that occur extensively within existing scallop grounds. Seabed imagery acquired during ongoingfield surveys with georeferenced underwater towed video cameras is being integrated within the GIS database in order to further analyse, ground truth and refine inferred sediment classes and their spatial configuration. Catch rate results from stock assessment survey transects positioned on the basis of sediment backscatter imagery have demonstrated the potential for applying integrated digital mapping techniques in order to predict and operationally target areas with a high potential scallop yield. Scopethus realised for improving catch efficiency (CPUE) can be used in concert with closed area and other conservation measures to scientifically underpin future sustainable management policy initiatives for this economically important fishery.

Background
The south east coastal shelf is the location for Ireland's most important national King Scallop (Pecten maximus) fishery. Maintenance of CPUE and controlling total effort are specific management objectives within the current development of a long-term plan to ensure the future viability of this fishery (Tully, 2002). A detailed scientific understanding of the mechanisms governing abundance and distribution of both adult and larval scallop are important in order to ultimately identify and delineate management zones within the fishery e.g. for protection of spawning stock. Existing knowledge of scallop ecology indicates that population distributions are patchy (Robert and Butler, 1998), and that high scallop abundance correlates with coarser sediments such as sands and gravels (Bousfield, 1960; Robert, 1997; Kostylev, 2001). Multibeam sonar systems have become the tools of choice in the mapping of seabed topography, morphology and sediment characteristics (Mitchell & Hughes-Clarke, 1994, Courtney & Shaw 2000). When used in conjunction with optical imagery (still and video recordings) and sediment samples the composite picture thus generated facilitates very detailed spatial characterisation of seabed substrates and habitats. This paper highlights the seabed mapping and GIS aspects of research that began in 2001, and which is due for completion in 2004. The work is being conducted within the framework of a multidisciplinary research program being funded under the Irish National Development Plan (NDP). The project participants are drawn from three academic institutions: Trinity College Dublin and National Universities of Cork and Galway, the Irish Sea Fisheries Board (BIM), and the South East Shellfishermen's Association.

Methods

Biological Stock Assessment
Initial research surveys were carried out in 2001 using scallop dredges towed from a commercial fishing vessel. Since the primary objective at this stage was to map the general distribution of scallop, sample tows were made on a regular grid designed to cover the entire known extent of the fishing grounds. For each thirty-minute tow, the mean number of scallops within specific shell size ranges were recorded along with start and end co-ordinates. All data were stored in MS Excel Tables allowing direct transfer to ArcView GIS via an SQL link. Numbers per tow were then gridded and contour plots generated in order to visualise the spatial distribution of scallop. A follow-up survey was undertaken during September 2002 in which scallop stock sampling was stratified on the basis of the two dominant acoustic facies (identified from backscatter data) and within three broad depth zones. A final survey will be carried out during September 2003 using the same stratified sampling methodology.

Seabed Mapping
MBES sonar data were collected by the project team using the RV Celtic Voyager of the National Marine Institute equipped with a Simrad EM 1002. Coherent overlapping (20-30%) swathes of sonar coverage were generated within discrete blocks whose size and location were prioritised in order to coincide with areas of high scallop density as determined from the results of the initial stock assessments. Samples of surficial seabed sediments were collected during the surveys using a Shipek grab and initially described on the basis of their physical appearance (e.g. clean fine sand with many small shell fragments). Subsequent granulometric analyses were conducted using laboratory standard sieves and laser particle size analysis for finer fractions. All MBES data were managed and post-processed using CARIS™ HIPS (Hydrographic Information Processing System, CARIS, 2003). This software contains a suite of functional modules designed to facilitate complex and time consuming QC and data cleaning procedures, and allow reduction of all sounding data to a common vertical datum (e.g. Mean Sea Level). HIPS also facilitated production and export of the two main data products in GeoTiff format: relative backscatter (five and ten metres grids), and sun illuminated bathymetry (three and five metre grids). Bathymetric data in the form of numerical soundings were also output in ASCII "x,y,z" format at various grid intervals for subsequent use as a fundamental data layer in the GIS. The bathymetry is also being used to generate a digital terrain that will support the development of a high-resolution hydrodynamic numerical model. Preliminary testing of statistically based backscatter classification technique has been undertaken using proprietary (QTC Multiview, Quester Tangent Corporation, 2002) software with the assistance of the Geological Survey of Ireland (GSI). A representative set of acoustically inferred sediment facies/habitat types so far identified will be targeted in the final ground truthing survey to be undertaken during the first week of October 2003.

GIS
GIS is an essential element of study, and ArcView (V3.3) has been used to provide a common platform in which all spatial data are integrated and where analytical operations are undertaken. Tasks range from initial operational planning for survey coverage through to data integration, analysis, presentation and map production. All data are projected from WGS 84 geographic co-ordinates to a common reference frame in UTM. Tabulated point data (sediment samples, photographic locations, scallop sample tow locations) are imported via SQL, whilst MBES data products are imported directly.

Results
The initial broad-scale stock density surveys detected commercial sized scallop at densities of up to 300 individuals per tow. Densities in excess of 80 were encountered over a large semi-continuous area extending seaward to the 80m isobath directly south of the major estuary of the River Suir. The distribution of scallop within the study area is shown in Figure 1.0.
Coherent MBES data covering approximately 65% of the total area of the south coast scallop grounds was generated during two annual field (2001/2) campaigns. The seabed in the study area slopes gradually from approximately the 40m isobath in the north (landward) to +/- 90m water depth in the south, over a distance of 50km. Initial visual and qualitative interpretations reveal, in the north of the area, a narrow band of outcropping rock bisected by a single large incised sediment channel (palaeochannel) with other conspicuous fluvioglacial features. Examples of these facies are shown in Figure 2.0. This distinctive terrain gives way to a rather flat submarine plain dominated by two main acoustically distinct sedimentary facies extending southward to the seaward limit of the scallop grounds. This area is characterised by a presumed gravel lag over which are migrating extensive elongate trains of large low relief arcuate dune-like structures (see Figure 3.0). These dunes and other mobile sedimentary features with distinct topographic expressions are also clearly discernable in the backscatter imagery, typically possessing a distinctive "dark" appearance. Intervening areas characterised by coarser gravely sediments have a much lighter signature. Mean scallop stock density among all samples trawled from the "pale" seabed facies was 115 against a mean of 15 for the "dark" facies. Total catch from the light facies was higher in all cases but one. Figure 4.0 shows a backscatter image of a representative part of the seabed within the study area on which are superimposed location and catch data from a number of stock assessment tows.

Future Program
During the final stages of the project statistical classification techniques will be applied to the whole dataset in order to refine the largely qualitative analyses and correlations so far outlined. This will lead to the production of sediment/biotope maps in which the most potentially productive scallop areas are delineated. These maps will thus provide the spatial basis to underpin future stock management plans. It is also anticipated that the methodologies developed during this work can be validated and geographically extended enabling the identification of new areas with high scallop fishery potential.

Figure 1.0 Map showing the location and general arrangement of the south east coast scallop grounds. Colour coded contours indicate the broad distribution of scallop within the areas under study.

Figure 2.0 Sun illuminated shaded relief image generated from raw multibeam sonar bathymetric data. Obvious artefacts (horizontal along track stripes, shape outlined in black) will be normally removed during subsequent post processing.

Figure 3.0 Sun illuminated shaded relief image generated from raw multibeam sonar bathymetric data. Low relief linearly arranged dune structures are clearly visible.

Figure 4.0 Multibeam backscatter image showing the sharp contrast in acoustic signature derived from the sandier dune structures (dark) and surrounding coarser gravely sediments (light). Scallop catch data are overlain showing numbers per tow along side each track line. Coloured spots indicate start and end points of each sample tow.

References