Mapping Overview

To better manage our oceans, we must first have a clear picture of the condition of the ocean floor. Using the latest hydro-acoustic equipment we are able to measure the depths and intensity of the back-scattered sound to measure the bottom roughness or texture to efficiently produce high-resolution maps of the ocean bottom. These combined values provide an effective and efficient means of habitat identification, e.g. sand, rock, reef, etc. Extensive underwater tow video surveys can then also be undertaken to ground-truth these interpretations and classify the areas in to habitat maps. In broad terms the benthic habitat mapping process falls under the following three main stages:

Habitat Characterisation, Habitat Classification and Habitat Mapping

The primary goal for habitat mapping is to accurately identify the spatial location, extent and characteristics of differing habitats on the seafloor. Traditional approaches of diving, still camera or towed video surveys provide the most direct mechanism to observe the characteristics of habitats. Unfortunately, these fall short on the spatial location and extents of the habitat in question. When costed on a coverage basis, these traditional techniques prove to be prohibitively expensive, prone to omissions and relatively unsafe survey techniques. A more cost effective and rigorous approach is required to identify all habitats at a broad scale as efficiently as possible. This would then be followed up with targeted towed video camera and stills photography in identified areas of interest. Hydro acoustic surveys provide the most cost effect broad-scale means of acquiring the seafloor data that can be used in conjunction with well planned follow up fine scale surveys for the actual benthic habitat classification and spatial mapping of classified habitats.

It should be noted that hydro acoustic survey techniques do not directly observe marine habitats, but the data they collect can be used as a surrogate to identify differing habitats.

For the purposes of benthic habitat mapping, broad scale hydro acoustic surveys generally fall into the realm of Multibeam bathymetry and sidescan sonar. Whilst it is also possible to include single beam echo sounding in this category, it is fast becoming out dated as not being a cost effective solution where full coverage of the seafloor is expected.

Back to top of page

Habitat characterisation

Spatial characterisation representation is carried out by means of broad scale acoustic and electro-acoustic mapping tools, such as depth measurements, and satellite and aerial photography and fine scale characterisation using geotechnical sampling methods such as grabs and coring. See technical section for details of mapping survey tools.

Habitat classification

The process of classification reduces the seabed characterisations in to a set of habitat classification types. See how to describe habitats.

Back to top of page

Habitat Mapping

The ultimate goal of a habitat mapping project is to generate a classified map of the various habitats in the survey area. A long standing goal is to create a rigorous automated classification system. The success of such a system requires good quality data inputs. At present the bathymetric surfaces consistently provide such quality surfaces, as seen in the example below, for further information see the case study on comparison of acoustic systems, Ref (Figure 43) in the Marmion Case Study.

The above imagery is an ISO unsupervised classification based on the bathymetric surface and its derivatives. The region has been classified into areas of low reef, high reef, deep sand and sand inundated. Correlation to the bathymetry and backscatter surfaces can clearly be seen.

Back to top of page

Important basic considerations in using acoustic systems regarding benthic habitat mapping

The following information is taken from the CWHM Report "Acoustic Techniques for Seabed Classification" (Download 11 MB PDF) by J D Penrose, P J W Siwabessy, A Gavrilov, I Parnum, L J Hamilton, A Bickers, B Brooke, D A Ryan and P Kennedy, and to see the rest of this report refer to the technical content of the broad-scale acoustic sensors section.

1. Habitat and Habitat Surrogates

One of the many definitions of habitat is given by the Shorter Oxford English Dictionary and cited in Harden-Jones(1994) as; "The locality in which a plant or animal naturally grows or lives; habitation. Applied (a) to the geographical area over which it extends; (b) to the particular station in which a specimen is found; (c) but chiefly used to indicate the kind of locality, as the sea-shore, chalk hills, or the like" . In some usages, the term habitat is extended to include the biological communities associated with a given locality. In such cases, description of a marine habitat may involve consideration of the total biomass in an area and its biodiversity. Some commentators make use of the term "biotope" to represent the seabed physical habitat and its associated biological communities. " Habitat " in the context of the marine environment is not the same as used in context of land environments where it is related to " cover". Habitat in context of the marine environment is generally described as the biological and physical characteristics of where an organism lives. The goal is to classify and map the seabed in a meaningful way in terms of its physical and biological characteristics. The processes involved in benthic habitat mapping require a habitat characterisation that describes the habitat based on geological, biological, chemical and oceanographic observations. The habitat classification process produces a set of habitat types based on a suite of standard descriptors of topographical, geological, biological, and natural and anthropogenic features and processes. The habitat map gives rise to a spatial representation that describes and classifies the various habitat units based on assumptions that organisms distribute themselves along environmental gradients and that their clusters define distinct sets of environmental factors.

Benthic habitat map of Parks Victoria - click on image below to go to case studies section on of Parks Victoria

Acoustic assessment of the seabed has commonly been employed essentially as a sediment classification technique. Considerable attention has been given (see Sternlicht (1999) for a review of this issue) to the linked parameters of sediment grain size, density, porosity, compressional and shear sound speeds and absorptions and surface roughness. This work has informed the acoustic system discussion in the toolkit and in part underlies the method of operation of the commercially available systems which use acoustics for seabed classification. As discussed in the technical section concerning single beam sounders, the concepts of seabed acoustic "roughness" and "hardness" are used as seabed descriptors. Here and later in this toolkit "acoustic hardness" is used as a descriptor of the acoustic impedance of the substrate type and hence of the impedance contrast offered to an acoustic wave by the water-seabed interface. The physical roughness of a surface influences the amount of sound backscattered to a receiver, so that measured backscatter or "acoustic roughness" are often proxies for physical sediment roughness. (Backscattered sound is that part of the total scattered sound that goes back towards the source) For sedimentary seabeds "hardness" in particular can sometimes be linked to sediment density and compressional sound speed, which in turn link to other sediment parameters. A difficulty is that these acoustic descriptors "hardness" and "roughness" need not necessarily describe the actual physical hardness or geometrical roughness properties of the seabed. The roughness of sedimentary seabeds is seen as a consequence of the sources and sinks of sediment and of the kinetic energy delivered to the seabed by waves, tides and currents. However, high reflectivity and high backscatter from even a small shell content in sediments can degrade the use of acoustic "roughness" and "hardness" parameters for inference of geometrical seabed roughness. Additionally, epibenthic biota, particularly seagrasses and macroalgae of sufficient density can dominate acoustic returns and mask signals from the seabed itself.

A recent UK Government study discusses an approach to marine and coastal environments, listing seven areas of coherent action which are in turn linked to 12 guiding principles of the Ecosystem Approach as defined by the Convention on Biological Diversity. One such area of coherent action is concerned with the environment and includes a priority action item "Using surrogate information sources" . This appears to be a response to the need to access information on benthic habitat structure and function, amongst other factors, which can guide management and usage of the marine environment under conditions of limited understanding of ecosystem dynamics and usually, limited spatial and temporal data coverage.

As currently employed acoustic techniques provide several surrogate descriptors of habitat. These descriptors are usually linked to more direct habitat parameters by a variety of techniques, including photography and spot sampling of sediments and biota.

In recent decades, a number of studies have linked sediment grain size with infaunal invertebrate distributions. This issue has been reviewed by Snelgrove and Butman (1994) who conclude however that "the complexity of soft-sediment communities may defy any simple paradigm relating to a single factor, and we propose a shift in focus towards understanding relationships between organism distributions and the dynamic sedimentary and hydrodynamic environment" . In the context however of a broader range of seabed types, including harder sediments and exposed reef structures, some generalisations appear to be justified. Thus Siwabessy (2001), as discussed in Results from North West Shelf and Southeast Fisheries Regions, of the technical section concerning single beam sounders, has shown that certain groupings of near bottom fish species can be related to acoustically determined seabed type.

2. Vertical Extent of Bottom Habitat

Most discussion of acoustic interaction with the seabed does not consider that benthos above the sediment or reef surface contributes to measurable backscatter. Thus most commonly, biota such as macroalgae and sea grass are treated as "invisible" to acoustic sensing. This simplification appears to be a usable first approximation where low frequencies are used, and where only the dominant seabed surface return is sought. However, as shown by Hundley, Zaboudil and Norall (undated), seabed plant assemblies may be detected by appropriate acoustic techniques. Such detection may well provide a more direct surrogate measure of relevance to habitat classification than the measures currently employed, at least where substantial plant biomass areal density exists. A system known as SAVEWS (Submerged Aquatic Vegetation Early Warning System) has been developed by the U.S. Army Engineer Waterways Experiment Station to characterise vegetation in shallow water environments. SAVEWS uses a BioSonics DT4000 digital hydroacoustic sounder with a narrow-beam transducer (Sabol and Burczynski 1998). The system records the depths of the tops of vegetation, usually appearing as "a jagged pattern" . The pattern is interpreted visually or automatically. Koniwinski et al. (1999) have used this system.

Discussion of sub-surface contributions to acoustic backscatter are much more common. At the high frequencies relevant to the present CWHM work, such sub-surface scattering can be expected from targets such as shell material, sediment grains, biological matter such as rhizomes, and gas bubbles. Sub-surface scattering necessarily follows in time what is usually a dominant seabed reflection/scattering signal and is mixed with off-axis acoustic returns from the major interface. Thus, for most acoustic geometries information about sub-surface scatterers is not retrievable from signals such as echosounder returns, although sub-surface scattering contributes to the received echo return. Sternlicht (1999) reviews this issue, which has received considerable attention at a somewhat deeper depth scale than is of importance for habitat description, in the context of sensing for buried munitions. However, the acoustic systems are usually very good detectors of shell beds, as these have high acoustic reflectivity and backscatter e.g. Smith et al. (2001), and organisms such as horseshoe crabs and brittle stars (Magorrian et al. , 1995). Tseng et al. (2005) have described a technique for distinguishing several classes of seabed biotic cover using a Genetic programming technique for processing single beam return signals.

In discussing acoustic techniques applied to seabed habitat description, it is important to note that the "seabed" extends both above and below the sediment or reef interface with the water column. Biotic material above the direct interface is susceptible to acoustic detection, while that below is difficult to distinguish from signals derived from the direct interface itself.

3. Sampling Statistics and Coverage

Many, if not most, oceanographic measuring and sampling tasks must be carried out under costs and effort constraints which allow for only under-sampled data sets to be gained. The task of adequately assessing and classifying seabed regimes in the many areas of interest highlights this issue. Acoustic techniques are attracting considerable interest because they offer the potential to provide comparatively rapid assessment of some seabed properties. Nonetheless, the techniques described in the body of this toolkit, involving as they do the use of boats as operating platforms, are also subject to their own cost and effort constraints on coverage. This is particularly so in the case of the earliest developed classification techniques, which use single beam echosounder technology. Such systems allow for data acquisition at slow to medium vessel speeds and from a strip of the seabed of cross-track width which depends on sounder beamwidth and water depth. Such track widths are a function of beamwidth, and may approach the water depth for beam widths of 50-60 degrees common in fish finding sounders, but are usually somewhat less than this figure. The sampled area provided from echosounder systems thus provides good along-track coverage and across-track coverage linked to the spacing between tracks. Interpolation between tracks, needed if a full 2D representation of the area surveyed is to be attempted, thus calls for some assumptions concerning the spatial variability of the seabed and benthic types. Siwabessy (2001) has approached this problem on the North West Shelf of Western Australia by firstly assessing, from a variety of track directions, whether or not spatial variability shows a clear distinction between tracks at constant depth, i.e. from tracks which follow depth contours, and tracks which follow maximum gradients in the area. In both cases an estimate of the autocorrelation characteristic length of roughness/hardness, i.e. lengths in which attributes of roughness/hardness data are self correlated, is made from the time/space sounder records. In his example, no evident contrast was seen between the along-track and cross- track runs. These autocorrelation characteristic lengths are then used as a basis for interpolating between tracks.

A further statistical consideration concerns the "patch size" provided by an acoustic system incident on the sea floor. An example of patch size is provided by the seabed area insonified by an echosounder. Commonly this is approximated as the projection of the central beam on the seabed interface at the depth of interest; the angular limits of the beam being represented by the 3dB points of the transmit beam pattern. In shallow water such a patch size is necessarily small and may be less than the horizontal roughness scale lengths of the surface. A succession of returns gained as the sounder is translated across such an environment will provide a distribution of reflected/scattered signal amplitudes. At greater depths a larger patch size is insonified and may now exceed the roughness length scales. This will provide a different form of signal amplitude distribution. This issue has been addressed by, amongst others, Dugelay et al. (2000) who also report on the effect of grazing angle on echo ensemble distribution form.

Back to top of page