Department for Environment and Heritage Eyre Peninsula Natural Resources Management Board
Marine Habitats within bays of the Eyre Peninsula NRM Region
Final Report to the Eyre Peninsula Natural Resources Management Board for the program: Establishing Marine Baselines
By David Miller, Grant Westphalen, Ann Marie Jolley and Ben Brayford
Marine Habitats within the bays of the Eyre Peninsula NRM Region
Final Report to the Eyre Peninsula Natural Resources Management Board for the program:
Establishing Marine Baselines
Coast and Marine Conservation Branch Department for Environment and Heritage
David Miller, Grant Westphalen, Ann Marie Jolley and Ben Brayford.
For further information, please contact:
Coast and Marine Conservation Branch Department for Environment and Heritage GPO Box 1047
Adelaide SA 5001
This publication may be cited as:
Miller, D. 1, Westphalen, G.2, Jolley, A. M. 1 and Brayford, B. 3. 2009. Marine Habitats within the Sheltered Bays of the Eyre Peninsula NRM Region. Final Report to the Eyre Peninsula Natural Resources Management Board for the program:
Establishing Marine Baselines. Coast and Marine Conservation Branch, Department for Environment and Heritage, Adelaide, SA.
1 Department for Environment and Heritage, Coast and Marine Conservation Branch, Adelaide, SA.
2 Westphalen Consulting, Adelaide SA.
3 SKM Consulting, Perth WA.
Yvette Eglinton, Victoria Hendry, Henry Rutherford, Dimitri Colella Shane Holland, Dennis Gonzalez, Alison Wright, Amanda Spezialli, Neva Perry, Fab Graziano, Shelley Harrison, Charles Maddison, Andy Burnell, Bryan McDonald and Peter Fairweather . Also, thanks go to Dave Armstrong and other DEH staff, as well as Matt Guidera (Streaky Bay and Sceale Bay Bluewater Charters) for assistance with field work.
© Department for Environment and Heritage
This work is copyright. Except as permitted under the Copyright Act 1968 (Cmth), no part of this publication may be reproduced by any process, electronic or otherwise, without the specific written permission of the copyright owners. Neither may
information be stored electronically in any form whatsoever without such permission.
Every attempt has been made to provide accurate information in this document.
However, no liability attaches to the Department, its employees or collaborators or any other organisation or individual concerned with the supply of information or preparation of this document for any consequences of using the information contained in the document.
Table of contents
TABLE OF CONTENTS ...1
1 EXECUTIVE OVERVIEW ...2
2 BACKGROUND ...4
3 MARINE HABITAT MAPPING AND BROAD SCALE SURVEYS IN THE EYRE PENINSULA NRM REGION ...7
3.1 MARINE MANAGEMENT REGIONS...8
3.1.1 Bioregions... 8
3.1.2 Biounits... 9
3.1.3 Marine parks... 10
3.2 HABITAT MAPPING...11
3.2.1 Fisheries habitat areas ... 12
3.2.2 Other marine benthic habitat mapping... 13
3.2.3 Coastal vegetation mapping ... 13
3.2.4 Satellite imagery ... 14
3.2.5 Other potential data sources and GIS layers... 14
3.2.6 Aquaculture monitoring... 15
3.2.7 Beach profiling ... 16
3.3 REEF SYSTEMS...16
3.5 SOFT BOTTOM HABITATS...18
3.6 THREATS TO MARINE SYSTEMS IN THE EPNRM REGION...19
4 REMOTE SENSING AND MARINE HABITAT MAPPING – DEVELOPMENT OF A STANDARDISED APPROACH ...22
5 BENTHIC HABITAT MAPPING AND GROUND TRUTHING METHODS WITHIN THE EP NRM REGION...24
5.2 DIGITISATION OF AERIAL IMAGERY...24
5.3 VIDEO GROUND TRUTHING...25
5.4 ACOUSTIC GROUND TRUTHING...25
5.4.1 Classification of habitats/production of maps ... 26
5.4.2 Data and map limitations ... 29
6 BENTHIC MAPPING OBSERVATIONS...31
6.1 PROPORTIONAL COVER OF BROADER HABITAT GROUPS...33
6.2 AREAS OF HIGH HABITAT DIVERSITY...41
6.3 CONCLUSIONS AND RECOMMENDATIONS...43
APPENDIX A - LIST OF BRYARS (2003) FISHERIES HABITAT AREAS (FHA) THAT OCCUR IN THE EP NRM REGION...53
1 Executive overview
Under the Natural Resources Management (NRM) Act 2004, NRM boundaries include all State waters. Therefore, NRM planning and programming must provide for the ecologically sustainable use of marine environments.
Measuring the effects of human activities in marine environments requires the
establishment of baseline habitat mapping against which specific threats and condition targets can be measured and assessed. Habitat mapping currently available is at a scale of 1:100,000, which does not provide adequately for the management needs of NRM Boards.
The Eyre Peninsula NRM (EPNRM) Board has commenced a program to address this critical knowledge need by engaging the Department for Environment and Heritage (DEH) in a program of broad scale marine habitat mapping at a resolution that is more suited to local management needs.
The mapping and underlying GIS data outlined in this report form an invaluable resource for managers within the EP NRM region, providing a critical baseline against which future changes can be measured and include recommendations for future
monitoring and research.
Detailed spatial mapping of seafloor habitats was conducted across eight large embayments within the EP NRM region. The embayments mapped, each of high economic, social and environmental value to Eyre Peninsula communities, were Denial Bay, Smoky Bay, Streaky Bay, Baird Bay, Venus Bay, Coffin Bay, Franklin Harbor and False Bay.
This summary document forms part of a set of information which also includes:
- a detailed map book;
- an interactive Arc Reader DVD (which will serve as a basis for identifying monitoring and management requirements as well as a driver of basic research and an educational tool); and
- an addendum to the report which includes a summary of baseline biodiversity information for specific habitats within the eight bays.
The total area mapped in this exercise (~ 1200 km2) is relatively small when
compared with the total marine environment encompassed by the EP NRM region (~
4.15%), but includes high resolution (1:20,000) baseline data on important coast and marine habitats adjacent to areas of population where threats are most likely to arise.
Large areas of seagrass dominate each of the bays considered, although Smoky Bay and False Bay had a high proportion of patchy-sparse seagrass cover, which may indicate systems under stress. While some species of seagrass undergo substantial, natural inter and intra-annual biomass changes, targeted future monitoring of seagrass cover and water quality within these bays will enable the EPNRM Board to determine to what degree the seagrass dynamics in this area are a product of human influences.
Similarly, large continuous seagrass beds dominate Denial Bay, Streaky Bay and Coffin Bay. Ongoing, periodic assessment of the condition of seagrasses in these bays
(again in relation to water quality) is recommended, coupled with terrestrial and marine management practices targeted towards the maintenance of seagrass health.
Areas of high habitat diversity were observed within Baird Bay and Coffin Bay. It is recommended that these should be considered in greater detail to determine the nature of this variability and whether these areas are in fact localised biodiversity hotspots which may be of regional importance.
It is widely accepted that sustainable management of natural assets should be approached at a holistic systems-level rather than that of individual species. This approach recognises the interconnectivity within and between habitats such that factors which may affect only one species will have flow on effects to the rest of the system (e.g. Fairweather 1999, GESAMP 2001, Allee et al. 2000, Flaherty and Sampson 2005). Management at broader ecosystem scales has a number of advantages (Fairweather 1999, GESAMP 2001, Flaherty and Sampson 2005) including (amongst others):
- Recognition that many environmental stress factors are non-specific, - Broader understanding of the ecosystem effects that may result from
exploitation of a resource, with concomitant realignment of what might constitute “sustainability”,
- Management and monitoring strategies are more efficient,
- Ecosystems scale data will present the integrated impact of a number of anthropogenic and natural stress factors,
- A greater understanding of the natural dynamics and processes of systems, particularly at larger scales,
- Understanding that environmental threats are now recognised as operating at very large spatial scales including regional (i.e. urbanisation and habitat fragmentation), national (i.e. catchment degradation) and global levels (i.e.
- Local scale issues (e.g. water pollution, etc) may be placed within a broader biogeographic context (see Connell and Irving 2008), and
- Providing a more effective, cohesive and consistent basis for engagements with all stakeholders that have interests in the system(s) concerned.
Note that a systems level approach to environmental resource management does not preclude or discount the targeted strategies required for rare, threatened and
endangered species, or indeed the specific approaches required for high priority pests.
Within the framework of large scale monitoring, there is a concomitant need to increase our understanding of the physical and biological factors that structure ecosystems and to identify areas of high biodiversity. Understanding spatiotemporal variability and biodiversity differences within systems across a range of scales leads to:
- Increased understanding of the ecosystem services provided by the resource, which may lead to improved engagements with stakeholders.
- A capacity to prioritise monitoring and management interventions in areas of high biodiversity.
- More efficient application of conservation/multiple use strategies.
- Identification of specific threats.
- Development of a notion of ecosystem “health” within the context of the broader habitat type (i.e. subtidal reef systems see Turner et al. 2007).
Following on from the Australian Government’s Natural Heritage Trust (NHT) funded mapping of the upper Gulf St Vincent and Spencer Gulf areas in 2005, and the Adelaide Mount Lofty region between 2006 and 2008, in late 2006 the Eyre Peninsula Natural Resources Management (EP NRM) region (in partnership with the NHT) developed a project with the Department for Environment and Heritage (DEH) to produce a detailed spatial Geographic Information System (GIS) layer of seafloor habitats within six sheltered bays within the EP NRM region. These included Denial, Baird, Venus and Coffin Bays on the west coast and False Bay and Franklin Harbour on the east coast. Together it was envisaged that these bays would provide a
representative example of the range of habitats likely to be present in the EP NRM region. In 2008 a further two bays, Streaky and Smoky Bays were included in the survey.
Work associated with this project included an update of previously available broad scale (southern Australia) marine benthic habitat maps produced by CSIRO, covering the inshore waters of the bays at a spatial scale relevant to regional management issues. In addition, the survey protocol and marine habitat definitions were aligned with those being developed elsewhere in Australia with the aim of developing habitat maps that will fit within a broader national framework.
Effective large-scale marine management requires a capacity to obtain data on changes in systems at large spatial scales. Marine benthic habitat mapping offers a cost effective approach to obtaining data on shallow (< 20 m) nearshore systems.
Furthermore, the development of a hierarchical approach to habitat differentiation has resulted in a framework for mapping that is readily repeatable, consistent at the national scale and encompasses the capacity to incorporate additional data.
Within the EP NRM region, the need for large scale marine habitat assessment capability is a major motivating factor in the development of Monitoring, Evaluation and Reporting Frameworks (see AMLR NRM 2008). However, while there is a need for large scale baselines, there is also a need to identify, monitor and manage smaller scale biodiversity and conservation “hotspots” as well as understanding
spatiotemporal variability and identifying the physical environmental drivers that structure marine systems across a range of spatial and temporal scales. This knowledge allows for ready identification of threats and appropriately targeted management responses. However, the EP NRM region poses a number of challenges to development of finer scale marine mapping, not least of which is the extent of the coastline, which is the longest of any NRM region in South Australia.
The following describes relevant background to the marine habitat mapping process for the Eyre Peninsula (EP) NRM region, including three main aspects:
- Marine management regions, broadscale marine observations and mapping in the EP NRM region, including what is understood with respect to risks to nearshore systems.
- A brief summary of the results of recent marine habitat mapping within eight large embayments within the region.
- Links between results of mapping relative to earlier benthic surveys as well as risks.
This document is analogous to similar reports related to marine habitat mapping developed for the Adelaide and Mount Lofty Ranges NRM Board (DEH 2009a), Northern and Yorke NRM Board (Miller et al 2009a) and the South East NRM Board (Miller et al 2009b). The structure of these documents and portions of the text related to marine management areas and habitat mapping are therefore similar (if not
identical), dealing with the same source material in many instances. While it is certainly feasible to reference this material to the companion documents in such instances, it was felt by the authors that every effort should be undertaken to ensure each report formed a “stand alone” entity.
Aims of this study were thus to:
- Establish baselines for coast, marine and estuarine biodiversity that will enable monitoring of change in resource condition within the EP NRM region.
- Develop marine habitat mapping at scales relevant to management for eight large embayments within the EP NRM region.
- Generate map books at a scale of 5 × 5 km and an interactive DVD of benthic habitat maps and other relevant GIS information layers.
This document summarises the management frameworks, approaches and history of habitat mapping for the purposes of natural resource management in the EP NRM region. The summary will cover four areas related to marine environmental management including:
- Current and planned marine management regions within the EP NRM region, - The history of habitat mapping within the region,
- Large scale habitat characterisation and comparison studies in reef, seagrass and soft bottom systems that might support habitat mapping.
- Risks to coast, estuarine and marine systems within the EP NRM region.
From a mapping perspective this document includes:
- A brief summary of the mapping methodology, including ground truthing approaches.
- Some summary statistics of the results of the mapping, including areas that may be of further interest for marine managers.
In addition, an addendum to this report will outline the methodology and summary results relating to the collection of baseline marine and estuarine biodiversity data that will form part of a monitoring baseline for the detection of change in resource
condition within the bays surveyed.
3 Marine habitat mapping and broad scale surveys in the Eyre Peninsula NRM region
Southern Australian nearshore marine systems are widely regarded for their high complexity, diversity and levels of endemism (e.g. Keough and Butler 1995, Edyvane 1999a, Connell 2007). Development of sustainable management strategies for these systems therefore presents a particular challenge (Turner et al. 2007), particularly in light of the broad range of potential or actual threats and given (Edyvane 1996, FAO 2003, Baker 2004, Flaherty and Sampson 2005, NY NRM 2008):
- A lack of historical/baseline data on marine systems in most instances,
- A diverse array of stakeholders competing for access to a range of overlapping resources and
- The physical difficulties and logistics of obtaining data in the marine environment at scales relevant to managers across a vast and often isolated coastline.
Broad scale habitat mapping has been a key feature of NRM in terrestrial systems, but has increasingly been applied to coast, estuarine and marine environments - although there is a concomitant need to develop a unified classification system (DEH 2007a, Mount et al. 2007). Baker (2004) describes a diverse group of marine benthic habitats from southern Australia:
- Freshwater outputs (overlaps with estuaries), - Tidal flats,
- Saltmarsh and samphire, - Mangroves,
- Seagrass meadows, - Reefs,
- Benthic sand habitats,
- Shallow and deep water sponge “gardens”, - Benthic mud habitats,
- Island habitats and
- Mixed assemblages and gradients between broader habitat groups.
All of the above occur to some extent within the EP NRM region. However, the EP NRM coast and marine systems present a particular challenge to marine managers relative to other South Australian NRM regions. This region has the largest length of coast (~ 1600 km; EP NRM 2008) versus the next highest at around 1380 km for the NY NRM (NY NRM 2008), but it is worth noting that the linear distances between locations on either side of Yorke Peninsula are relatively short. In addition, the EP NRM would appear to have retained a far larger proportion of its original terrestrial habitat relative to other regions (43% of original native vegetation versus 10-20% for
other regions; AMLR NRM 2007, EP NRM 2008, NY NRM 2008, SE NRM http://www.senrm.sa.gov.au/Home/tabid/243/Default.aspx, Accessed March 2009).
Modification of terrestrial landscapes is a major contributor to water quality decline and habitat degradation within nearshore marine environments (e.g. Bryars 2003, AMLR NRM 2007, Fox et al. 2007, Turner et al. 2007). Coast, estuarine and marine systems within the EP NRM are thus large, remote and, at this point in time,
potentially relatively undisturbed by altered terrigenous inputs. However, this should not be interpreted as suggesting that marine systems within the EP NRM region are not exposed to threats.
The following describes marine management regions, broadscale marine observations and mapping in the EP NRM region, including what is understood with respect to risks to nearshore systems.
3.1 Marine management regions
Marine habitat management regions within the EP NRM region comprise:
- Marine Planning/Ecosystem-based Management Guidelines (based on IMCRA Bioregions),
- Edyvane (1999a, b) Biounits, - Marine Parks.
It is worth noting that Australian NRM zones are largely based on terrestrial catchments, bioregions or State Government management boundaries (Australian Government, http://www.nrm.gov.au/nrm/region.html, Accessed April 2009, Planning South Australia, http://www.planning.sa.gov.au/go/SAGovernmentRegions, Accessed April 2009). The marine borders for NRM regions have no relationship to IMCRA bioregions and similar. For this reason, bioregions and biounits often overlap NRM marine boundaries.
The Integrated Marine and Coastal Regionalisation of Australia (IMCRA Version 4.0;
Commonwealth of Australia 2006) classification places three coastal and two offshore provincial regions that occur to some extent within South Australia, with the EP NRM region including areas from the Spencer Gulf IMCRA Province and the Great
Australian Bight IMCRA Transition (Commonwealth of Australia 2006). Mesoscale bioregions (that include the coastal regions defined under IMCRA Version 3.0) include eight coastal areas either wholly or partly within South Australia, five of which occur to some degree within the EP NRM region (Figure 1), including:
- Eucla – transitional warm to cold temperate rocky cliff coast, - Murat – transitional warm to cold temperate rocky crenulate coast, - Eyre – transitional warm to cold temperate rocky coast,
- Northern Spencer Gulf – confined inverse estuary on tidal coastal plain and - Spencer Gulf – semi-confined inverse estuary on tidal coastal plain.
For full descriptions of these areas, including information on climate, oceanography, geology and geomorphology, biota and estuaries, see IMCRA Technical Group (1998).
Figure 1 - Map of the EP NRM region showing Bioregions as well as the areas covered in the current benthic habitat mapping.
Marine biounits, based on CSIRO habitat mapping (1:100,000 scale) and the work undertaken by Edyvane (1999a, b) comprise 35 areas along the South Australia coast
to a depth of around 50 m. There are 15 biounits that occur wholly or partly within the EP NRM region with summary information on each included as part of the draft State of the Region report (EP NRM 2008). For full descriptions of each biounit, see Edyvane (1999b), including information relative to (amongst others): biogeography, conservation values and status, fisheries, recreation and tourism, science, research and education as well as cultural and historical aspects.
IMCRA bioregions and/or Edyvane (1999a, b) biounits may be used to as the first layer in defining areas/natural assets that may be of particular interest as well as the broader targeting of management activity (IMCRA Technical Group 1998, Baker 2004). Indeed, the IMCRA bioregions have played a role in the determination of MPAs (DEH 2009b; see below). Similarly, biounits are employed as descriptive components of State of the Region reporting (AMLR NRM 2007, EP NRM 2008, NY NRM 2008). However, both regional classifications are based on integrated
biogeographic data from a range of species groups as well as related
geomorphological and physical environmental factors. These regions are therefore difficult to relate to specific areas/habitat types that may require targeted management intervention. Furthermore, most of the stress factors (or threats – see discussion) identified for marine systems relate to habitat destruction and water quality issues that are generally concentrated to the near shore fringe (Bryars 2003, AMLR NRM 2007) at smaller scales than either classification can readily resolve.
3.1.3 Marine parks
Marine Protected Areas (MPAs) are a major marine environmental management and conservation initiative within South Australia. Designation of MPA areas was based on 14 design principles that include biological, social and cultural aspects (DEH 2009b). The system of 19 MPAs spread across the South Australia coast will form a key element for the protection and conservation of marine biodiversity as well as cultural and historical values within a framework that will allow for ecologically sustainable development of marine resources. The associated management and monitoring strategies thus have important implications for NRM throughout the state.
There are 11 proposed MPAs that occur wholly or partly within the EP NRM region including (DEH 2009b):
- Far West Coast Marine Park, - Nuyts Archipelago Marine Park, - West Coast Bays Marine Park, - Investigator Marine Park, - Thorny Passage Marine Park,
- Sir Joseph Banks Group Marine Park, - Neptune Islands Group Marine Park, - Gambier Islands Group Marine Park, - Franklin Harbour Marine Park, - Upper Spencer Gulf Marine Park and
- Western Kangaroo Island Marine Park (small portion thereof).
Although MPA boundaries have been defined, each requires further development in terms of internal multiple-use zoning, associated management plans and development of Performance Management Systems that will likely include some level of physical environmental and/or biological monitoring (NY NRM 2008, DEH 2009b). Zoning for Marine Parks in SA will include four types of internal zones plus provision for establishing special purpose areas (Marine Parks Act 2007;
http://www.legislation.sa.gov.au/LZ/C/A/MARINE%20PARKS%20ACT%202007/CURRENT/2007.6 0.UN.PDF). These zones/areas are defined as follows:
- General managed use zones - zones established so that an area may be managed to provide protection for habitats and biodiversity within a marine park, while allowing ecologically sustainable development and use.
- Habitat protection zones – zones established so that an area may be managed to provide protection for habitats and biodiversity with a marine park, while allowing activities and uses that do not harm habitats or the functioning of ecosystems.
- Sanctuary zones - zones established so that an area may be managed to provide protection and conservation for habitats and biodiversity within a marine park, especially by prohibiting the removal or harm of plants, animals or marine products.
- Restricted access zones - zones established so that and area may be managed by limiting access to the area.
- Special purpose areas - areas within a marine park with boundaries defined by the management plan, in which specified activities, that would otherwise be prohibited or restricted as a consequence of the zoning of the area, will be permitted under the terms of the management plan.
In addition to MPAs, there is a range of existing conservation, recreation parks and reserves within the EP NRM region. See the draft State of the Region report (EP NRM 2008) for a summary.
3.2 Habitat mapping
Relative to elsewhere in the state, the EP NRM region has limited historical data on benthic habitats. As with the rest of the South Australian coast, on the broadest scale, there is the CSIRO 1:100,000 benthic habitat maps that were used by Edyvane
(1999a, b) to develop biounit designations. Shepherd and Womersley (1981) used diver and boat observations to map the benthic community within Waterloo Bay using six reef and five seagrass (including bare sand) habitat types. Results of the survey suggest that the arrangement of habitats is spatially and temporally dynamic relative to water movement and depth.
The adjacent NY NRM Board in collaboration with the Department for Environment and Heritage undertook a fine scale habitat (1:10,000) mapping exercise between 2005 and 2007 in the upper reaches of Gulf St Vincent and Spencer Gulf to a depth of 15 m (DEH 2007a, b, c). These areas encompass the largest areas of seagrass in South Australia as well as other unique environmental values (NY NRM 2008, see Winnonowie and Clinton Biounit information Appendix A, Edyvane 1999a, b). A focus on habitat mapping and development of an understanding of both natural changes (see Seddon 2000) and anthropogenic sources of change is critical to
appropriate management. As reverse estuaries (e.g. Edyvane 1999a), the biological systems within the upper reaches of both gulfs may be particularly sensitive to factors that may further increase water temperature and salinity such as proposed desalination operations as well as global warming. Within Spencer Gulf, the mapped area
encompassed the coast from the Munyaroo Conservation Park on the east coast of Eyre Peninsula to Port Broughton on the west coast of Yorke Peninsula and therefore included a substantial area within the EP NRM region (DEH 2007a, c).
Importantly, these observations were undertaken based on cover assessments of a hierarchy of physical and/or biological characteristics along similar lines to the framework developed by Allee et al. (2000) and the Tasmanian Aquaculture and Fisheries Institute (SEAMAP 2008) including:
- Geomorphic type (hard/soft bottom),
- Biogeomorphic type (vegetated or unvegetated), - Substratum/ecotype (seagrass, algae, sand/silt or reef), - Structure (habit and density of cover) and
- Cover (extent % of the substratum coverage).
The resultant mapping was verified with extensive video ground truthing (DEH 2007a).
3.2.1 Fisheries habitat areas
An inventory of benthic habitats that are important for fisheries was undertaken by Bryars (2003) through an assessment of coastal near shore assets across South
Australia (up to 20 m depth or 3 km offshore – whichever came first). This summary classified benthic communities relative to 13 basic habitat types (that included the associated overlying pelagic component):
- Reef, - Surf beach,
- Seagrass meadow, - Unvegetated soft bottom, - Sheltered beach,
- Tidal flat, - Tidal creek, - Estuarine river, - Coastal lagoon, - Mangrove forest, - Saltmarsh,
- Freshwater spring and - Artificial habitats.
Habitat areas were only included if they were relatively large and/or significant to local fisheries. The depth/distance limit employed in this survey was based on a lack of data on deepwater systems as well as the view that shallow near shore areas were most threatened. The Bryars (2003) inventory was used to define 62 Fisheries Habitat Areas (FHAs) across the South Australian coast, including 30 within the EP NRM region that variously included all of the above habitat types except Coastal Lagoon and Freshwater Spring (Appendix A).
Sustainable management of commercial and recreational fisheries is a critical element of marine NRM. However, the consideration of habitats in terms of their importance to fisheries may discount other values. For example, a large area of reef may support a number of fisheries relative to small, isolated outcrops, but the latter may be
critically important in terms of biodiversity/conservation at local scales. In addition, the resolution of habitats within this assessment would appear to be too coarse to determine anything other than major changes through time. This issue may be compounded by the overlapping of some of the habitat types (Appendix A).
3.2.2 Other marine benthic habitat mapping
Alternative sources of information on benthic habitats might be obtained from environmental impact assessments and monitoring associated with current and proposed coastal developments including (amongst others):
- Jetty and port facilities (e.g. SANTOS Limited 1981), - Aquaculture zoning,
- Housing developments,
- Stormwater and wastewater outfalls,
- Desalination plants (notably a proposed desalination facility at Point Lowly) and
- Specific “one off” events such as the 1992 Era oil spill at Port Bonython (Wardrop et al. 1992, Connolly 1994).
There is a diverse array of ‘grey’ literature associated with the above, the availability of which and relevance in support of benthic habitat mapping is variable. The Draft Environmental Impact Statement for Port and Terminal Facilities at Stony Point, South Australia, describes five distinct intertidal and subtidal habitat types in the vicinity of the (then) proposed development (SANTOS Limited 1981). However, while the habitat types employed in the SANTOS Limited (1981) summary have some resemblance to those identified elsewhere (notably the deeper water group), the potential for alignment with habitat groups at the larger scale is perhaps limited.
3.2.3 Coastal vegetation mapping
The “Biological Survey of South Australia” database (DEH, http://www.environment.
sa.gov.au/biodiversity/ecological-communities/biosurveys.html#surveys, Accessed April 2009) provides a nationally consistent approach to vegetation classification called the National Vegetation Information System (NVIS) with more than 9000 distinct habitat types based on the vegetation and physical environmental data (DEH 2006, DEWR 2007). Part of the South Australian biological survey includes a state-
wide investigation into coastal, dune and cliff-top vegetation that employed 22 broad vegetation types (Opperman 1999). A similar survey of saltmarsh and mangrove habitats was completed by Canty and Hille (2002) and included 69 habitat codes based on a five-tiered classification system using landform, estuarine influence, degree of inundation, vegetation cover and integrity.
There are 16 recognised estuaries within the EP NRM region, with most being classified as tide-dominated, with the Tod River being the only permanently flowing water course (DEH 2007d). Detailed descriptions of each estuary relative to physical environment (catchment area, flows, etc.), habitats, bird and fish species, protection arrangements, cultural assets, economic importance, activities and pressures are
presented in the Estuaries Information Package for the EP NRM region (DEH 2007d).
3.2.4 Satellite imagery
Much of the following is based on a summary developed for Gulf St Vincent (see Petrusevics 2008) but should nonetheless be valid for most, if not all, of the South Australian coast.
Satellite remote sensing provides almost daily data (cloud permitting) on
oceanographic, meteorological and hydrodynamic data at a resolution of ~ 1 km2 since the 1970s (Petrusevics 2008). A range of observational datasets is available from a succession of satellites, with varying degrees of emphasis on either sea surface temperature or visible light imagery including:
- Very High Resolution Radiometer (VHRR, 1972 – 1978), - Coastal Zone Color Scanner (CZCS, late 1970s),
- Advanced Very High Resolution Radiometer (AVHRR, 1978 – 1984), - Sea-viewing Wide Field-of-view Sensor (SeaWiFS, 1979 – 2004) and
- Moderate Resolution Imaging Spectrometer (MODIS, Aqua and Terra – from 2000).
3.2.5 Other potential data sources and GIS layers
Analysis and interpretation of GIS-based habitat mapping would benefit from access to a range of additional information and/or layers related to a range of features including (among others):
- infrastructure (shipping channels, jetties, breakwaters, etc), - coastal inputs (outfalls, rivers and stream),
- tourist attractions (recreational beaches, boating/fishing or SCUBA diving areas, etc.),
- aquatic and coastal reserves,
- local and State Government planning regions and - hydrodynamic modelling.
There are a variety of sources available for this type of information, generally at the state level, including (amongst others):
- The extensive list of GIS layers summarised by Caton et al. (2007) as part of
“Conservation Assessment of the Northern and Yorke Coast”, many of which have relevance across the state,
- Atlas of South Australia (http://www.atlas.sa.gov.au/ - Coastal Management Area, accessed May 2008),
- South Australian Waters: an Atlas and Guide (Boating Industry Association of South Australia 2008),
- A number of management strategies developed by the Coastal Protection Board related to acid sulphate soils, coastal weeds, coastal erosion and beach monitoring (see http://www.environment.sa.gov.au/coasts/management.html, accessed March 2009),
- Fisheries stock assessments,
- Aquaculture monitoring (see below) and
- Non-mapping environmental monitoring and research.
3.2.6 Aquaculture monitoring
All marine-based aquaculture in South Australia is required to maintain a level of environmental monitoring as part of licensing (Aquaculture Regulations 2005).
With the expansion of the aquaculture industry in the Eyre Peninsula (notably southern bluefin tuna farming off Pt Lincoln; EP NRM 2008) there has been a growing demand for suitable locations as well as an increased monitoring and
research capability at least at the scale of specific operations. Aquaculture operations within the EP NRM region are relatively extensive compared to other regions. The associated water quality and environmental monitoring may thus form a regular input of data that may support the interpretation of habitat mapping.
From 2001, tuna farming operations were monitored under a compliance-based report card approach (called TEMP) that includes annual monitoring of seacages in their present location within lease areas outside Boston Bay
bluefin_tuna_environmental_monitoring, Accessed March 2009). These observations are based around video and benthic infauna surveys (macro-invertebrate species living within surface sediments), which in 2007 included 12 compliance (on farm) and 18 control locations (Loo et al. 2008). Earlier assessments of tuna farming operations were conducted within Boston Bay in the mid 1990s (prior to all operations being moved offshore) and included observations of the soft bottom benthic community as well as infauna (Cheshire et al. 1996a, b).
The South Australia Shellfish Quality Assurance Program (SASQAP) has operated since the early 1990s and ensures that farmed shellfish within 18 regions across the State are fit for consumption through an ongoing program of water quality monitoring (SASQAP 2004). Currently the EP NRM region includes nine SASQAP
(http://www.pir.sa.gov.au/aquaculture/monitoring__and__assessment/sasqap, Accessed April 2009). However, it needs to be realised that the primary focus of SASQAP monitoring relates to microbial, phytoplankton and biotoxin monitoring for the purposes of food safety. Nonetheless, information on the pattern of restrictions
placed on SASQAP monitoring areas may form a useful indicator for more targeted investigation.
3.2.7 Beach profiling
The Coastal Protection Branch (DEH) has undertaken annual sand profile
observations of up to 44 locations along the Eyre Peninsula coast since 1986 (Eaton et al. 2001). These observations are targeted to monitoring sand movements (both accumulations and losses) across beaches including dunes and the subtidal nearshore zone. Substantial changes are apparent at most beaches over the sampling period, although most are considered to be of natural origin. However, beaches at North Shields, Lucky Bay, Arno Bay, Smoky Bay and Venus Bay warrant further attention, if not management intervention (Eaton et al. 2001).
3.3 Reef systems
The EP NRM coastline has protracted stretches of rocky/cliff coastline and consequently supports large and diverse reef communities (Edyvane 1999b).
Great Australian Bight waters have been reported to support macroalgal communities to a depth of 70 m (Shepherd and Womersley 1971, 1976), which suggests that water clarity within this stretch of coast is very high. However, investigations of reef systems within the EP NRM region are temporally and spatially limited. Notably there have been no reef health surveys along the lines of Turner et al. (2007) within the EP NRM region, but in 2007 and 2009 investigations of the lower Eyre and far west coast reef communities were done as part of a collaboration between the University of Tasmania and DEH using the Edgar and Barrett (1997, 1999) methodology (yet to be reported at the time of writing).
Shepherd and Womersley (1971, 1976) describe the composition and structure of macroalgal communities at Pearson Island and St Francis Island respectively. More recent surveys under the general banner of “SA Offshore Island Expeditions” have repeated and expanded on these earlier surveys of offshore islands. These surveys entail a collective effort from a range of organisations (including SARDI, DEH and the Universities) spanning a range of disciplines. The investigations include;
- Isles of St Francis in 2002
Benthic community investigations were mostly focussed on observations of the reef habitats relative to composition and biogeography (Womersley and Baldock 2003), community composition and productivity (Turner and Cheshire 2003) and zonation (Baker and Edyvane 2003).
- Investigator Group in 2006
Investigations included (amongst others) surveys of seagrass systems (Bryars and Wear 2008), benthic habitats (Miller and Wright 2008) and the macroalgal community (Baker et al. 2008).
Another expedition to the Sir Joseph Banks Group has just been completed (May 2009).
Connell and Irving (2008) undertook an investigation into the composition and structure of reef systems across different spatial scales (1-10 km, > 100 km and >
1000 km) across the whole of southern Australia (Cape Leeuwin in Western Australia to Mooloolaba in southern Queensland),which included observations at Port Lincoln on southern Eyre Peninsula. This study showed that differences between reefs at all scales could largely be explained by biogeography (latitude and longitude of each site).
Observations by the community-based monitoring program “Reef Watch” within the EP NRM region are limited, comprising only “Feral or in Peril” observations (17 spread across six locations) that include sightings of a selection of species that are readily recognised marine pests (Feral) or species that may be of conservation concern/public interest (in Peril). These observations include some information on locations, but no real data of benthic community composition
(http://www.reefwatch.asn.au/, Accessed March 2009).
Mapping, site comparison or monitoring of seagrasses on the EP NRM coasts are rather limited. Shepherd and Robertson (1989) summarise a number of targeted seagrass investigations/mapping exercises or point observations within the EP NRM region. This summary suggests that seagrasses beds are patchy on the exposed southern Australian coasts such as those within the Great Australian Bight, generally occurring in sheltered areas or at depths where wave energy is reduced but there is still sufficient light. Posidonia angustifolia and Posidonia coriacea are reported to occur to 25 - 30 m depth at the base of cliffs off western Eyre Peninsula (Shepherd and Robertson 1989).
Shepherd (1975) mapped benthic communities in the vicinity of two outfalls (one for domestic wastewater, one for a fish cannery) at Proper Bay, Port Lincoln. This survey described a number of habitat types relative to each outfall. At Billy Lights Point these community types included:
- Bare sand,
- Posidonia australis, - Pinna/holothurian, - Rubble bottom,
- Heterozostera tasmanica and - Granitic reef.
Within Proper Bay a slightly altered set of community types was considered:
- Ulva lactuca and Posidonia, - Posidonia australis (b1), - Posidonia australis (n1), - Bare sand.
Note that the two forms of Posidonia australis (b1 and n1) are probably separate species with n1 being Posidonia sinuosa and/or Posidonia angustifolia. The Shepherd (1975) survey describes areas of seagrass loss or decline in the vicinity of both outfalls, in particular the Proper Bay site.
Aerial photography has been used to map seagrass distribution and change over a wide variety of coastal environments in South Australia at resolutions superior to Landsat imagery. Within the EP NRM region this includes:
- Boston Bay, Port Lincoln (Hart 1999) and - False Bay, Whyalla (Cameron 2002).
Hart (1999) considered changes in seagrass cover within the entirety of Proper Bay and strips along the coast of Boston Bay and Boston Island to 10 m depth based on differences between orthorectified aerial images from the mid 1970s and 1996.
Although this investigation produced maps of seagrass distribution within the total area considered (~98 km2), there were no distinctions in terms of seagrass species composition or density (Hart 1999). Rather, a cutoff coverage of 50% was used to differentiate between seagrass cover and bare substrate. Results were thus considered relative to four categories:
- Seagrass no change, - Substrate no change, - Seagrass loss and - Substrate loss.
It is worth noting that the areas of seagrass loss or decline observed by Shepherd (1975) within Proper Bay would appear to be consistent with Hart (1999)
observations. From the mid 1970s to 1996 there was a net increase in substrate (or loss of seagrass) of approximately 1.685 km2 across the entire area considered.
benthic mapping and that differences in seagrass cover (either loss or gain) may be confounded by drifting macroalgae. Nonetheless, results of this mapping can be employed to look for historical changes in seagrass cover within the areas considered, if only in terms of all seagrasses with coverage of 50% or more.
The Cameron (2002) investigation at False Bay, Whyalla is not publically available at the time of writing (D. Hart Pers. Comm. 2009).
3.5 Soft bottom habitats
Soft bottom systems form the largest marine environment within the EP NRM region (EP NRM 2008), although mapping, surveys and research are spatially limited. Most observation has occurred in the Pt Lincoln region in relation to southern bluefin tuna seacage aquaculture with initial surveys and reporting conducted in 1996 when the bulk of farming was conducted inside Boston Bay (Cheshire et al. 1996a, b). These surveys included observations (mostly video) at fixed distances from cages as well as controls at least 1 km or more distant. Results of these surveys may serve to ground truth habitat maps, although at more than a decade old, the validity of such a
comparison is open to question. There are also the results of seacage monitoring in the current farming locations outside Boston Bay, although the data for mapping may be of limited use owing to the area and depths involved (seacages currently operate in 20 – 24 m). There may also be commercial/confidentiality constraints on data access.
Commercial prawn trawling within Spencer Gulf began in 1967 (PIRSA 2003), but unfortunately there is no analogous investigation of benthic systems along the lines of Shepherd and Sprigg (1976) against which the long term effects of trawling might be measured. Svane et al. (2009) undertook a range of benthic observations from five areas within current prawn fishing grounds (21 – 23 m deep) that had a varied, but known, level of accumulated prawn trawling history. The benthic community at these sites was found to be dominated by sandy sediment and some fine gravel in some areas, with varied but overall low macro fauna/flora cover that negatively correlated with the accumulated prawn trawling effort (Svane et al. 2009). Less trawled areas were characterised by a mixture of bearded mussel (Trichomya hirsutus), southern hammer oyster (Malleus meridianus) and razor clam (Pinna bicolor) that may be analogous to the Malleus-Pinna assemblage identified in Gulf St Vincent by Shepherd and Sprigg (1976). It was concluded that, similar to Gulf St Vincent (see Tanner 2005), prawn trawling was likely to have a strong negative influence on the structure of benthic communities in Spencer Gulf, although there is a north-south
environmental gradient that may explain some of the differences between sampling areas (Svane et al. 2009). Both Tanner (2005) and Svane et al. (2009) noted a lack of eelgrass (Heterozostera tasmanica) in their observations, although this species was considered to be abundant over large areas of Gulf St Vincent (Shepherd and Sprigg 1976) and probably within Spencer Gulf to a depth of around 30 m, although in the absence of historical data for the latter this inference cannot be confirmed. It is important to note that, while prawn fishing grounds cover less then 15% of Spencer Gulf waters, they actually include a large proportion of the deeper areas (> 15 m;
Svane et al. 2009).
3.6 Threats to marine systems in the EP NRM region
There are a diverse range of threats to coast, estuarine and marine systems in South Australia derived from an equally variable array of activities and stakeholders (Edyvane 1996).
In a risk assessment of coast, estuarine and marine assets within the EP NRM region, Cheshire et al. (2008) reported a group of 18 assets that were juxtaposed against 23
“issues” (or threats; Table 1). Coastal pest plants and animals, coastal development (both construction and operational phases) and coastal access (off-road vehicles, trail bikes, beach camping, etc) were noted as imposing extreme threats to a range of coastal assets, in particular intertidal and subtidal seagrasses, coastal vegetation, estuarine environments and enclosed soft sediment systems. However, the majority of threats (19 out of 23) imposed at least a high level of risk to some form of asset.
Similarly, all assets were associated with some type of high risk (Cheshire et al.
2008). An attempt to offer some spatial component to asset-threat combinations across the region was only partially successful owing to a lack of data on the distribution of the asset and/or the associated threat.
Note that although Cheshire et al. (2008) acknowledged climate change as a threat to coast, estuarine and marine assets it was considered that this subject required a separate risk assessment. Similarly, invasive marine pests as well as threats with respect to rare, endangered or threatened species were also recommended as subjects for targeted assessments.
The threats identified by Cheshire et al. (2008) in the Coast and Marine Prioritisation Workshop for the EP NRM Region broadly align with those described in the State of the Region reporting for the EP NRM region (EP NRM 2008) including:
- Climate change,
- Development – marinas, residential, holiday shacks, - Flooding and erosion,
- Desalination plants, - Ferry operations,
- Tourism visitor facilities and coastal infrastructure, - Aquaculture,
- Commercial fisheries,
- Other industries (e.g. mining),
- Coastal Acid Sulphate Soils (CASS) and - Visitor use and general human impacts.
Successful management of coast, marine and estuarine assets requires mechanisms to address threats, minimise impacts and ameliorate damage. These approaches include the need for accurate and repeatable benthic habitat mapping at relevant scales.
Table 1 – Assets and issues (threats) identified within the Eyre Peninsula NRM Region.
Coastal/veg - Dunes/unvegetated Coastal/veg - Dunes/vegetated
Coastal/veg - Mangroves/intertidal mudflat Coastal/veg - Rocky Cliff
Coastal/veg - Samphire/gahnia/salt marsh Estuary - Tidal and river dominated Saline lakes
Pelagic - Deep Water (> 40 m) Pelagic - Inshore (< 40 m) Reef - Intertidal
Reef - Subtidal
Sand/Soft Sediment - Bays/Sandy Beaches (Open coast) Sand/Soft Sediment - Bays/Sandy Beaches (Enclosed) Sand/Soft Sediment - Shallow subtidal
Sand/Soft Sediment - Deep water (> 40 m) Seagrass - Intertidal
Seagrass - Subtidal
Water quality (clarity, nutrient status, etc) Issues
Point Source - industrial discharge Point Source - waste water Point Source - thermal
Point Source - stormwater pipes, drains Diffuse source - nutrients
Diffuse source - chemical contaminants Diffuse source - sediment inputs
Marine vessel - leakages, hydrocarbons, antifoulant, airborne pollution
Desalination plant impacts
Coastal development - operational Coastal development - construction Dredging
Acid Sulphate Soils
Domestic animals and livestock - grazing, disturbance Mining impacts
Marine invasive species
Pest plants and animals (coastal) Litter, rubbish dumping, marine debris
Water extraction (ground water and loss of surface water flow from catchments)
Access/marine (swimming with sea lions/sharks/tuna, jet skis, etc.) Access/coastal (off-road vehicles, trail bikes, bush camping, beach combing)
Modification of benthic habitat
4 Remote sensing and marine habitat mapping – development of a standardised approach
A key element to the development and implementation of resource condition targets for Natural Resource Management is to establish accurate baselines from which future changes in ecosystem structure (or health) can be compared. Sustainable management of natural resources and the development of conservation strategies at ecosystems levels require a greater understanding of the distribution and status of the supporting habitats (DEH 2007a, Mount et al. 2007). Broad-scale habitat mapping, coupled with geographic information system (GIS) capability is a powerful tool for large-scale environmental management (GESAMP 2001, Flaherty and Sampson 2005, Mount et al. 2007). However, this approach is reliant upon a capacity to consistently
differentiate and map habitat types and therefore presents a particular challenge when dealing with subtidal marine systems wherein traditional remote sensing techniques may be of restricted value (DEH 2007a, Mount et al. 2007). Current marine habitat mapping criteria are targeted at regional scales (Allee et al. 2000, Mount et al. 2007) and there is thus a need to develop standardised national criteria for marine habitat mapping (Allee et al. 2000, DEH 2007a, Mount et al. 2007).
National scale habitat mapping definitions have been established for terrestrial
systems in Australia (see the National Vegetation Information System (NVIS) DEWR 2007), but marine systems are yet to be comprehensively unified (DEH 2007a, Mount et al. 2007). Allee et al. (2000) identified several requirements for a national marine habitat classification system including:
- Universal and consistent coverage that is spatiotemporally sensitive,
- An additive structure such that classification can be taken to finer scales that fit within broader classifications as data become available,
- Combines physical, geomorphic and biotic data, - Compatibility with a GIS framework,
- Amenable to currently available data and technology and
- Provides a basis for identifying functional linkages wherein the observed patterns can be related to ecological processes.
The approach developed by Allee et al. (2000) for the USA employs a hierarchical system of 13 levels, most of which relate to broader scale geomorphic features. A hierarchical approach to habitat mapping has the advantage of flexibility in
development of summaries as well as improving the resolution within more broadly classified regions as data become available (Allee et al. 2000, Mount et al. 2007).
Within Australia, one of the best examples of a large-scale marine habitat mapping program is SEAMAP in Tasmania, which has been in operation since around 2001 (Barrett et al. 2001). More recently major mapping programs have been undertaken in other states (including those by Marine Parks in NSW, Dept for Primary Industry and Deakin University in Victoria, and the Marine Futures program in WA). In South Australia, there is the recently completed benthic mapping of the upper Spencer Gulf (DEH 2007c) as well as the entire AMLR NRM region (DEH 2009a). The
methodologies employed by the SEAMAP and DEH (2007a, c, 2009a) mapping
programs are based on that of Allee et al. (2000), although the hierarchy includes only four levels; geomorphic type, substratum/ecotype, substrate eco-type and a series of modifiers (see Benthic Mapping and ground truthing methods below).
Aerial and satellite imagery have frequently been employed in understanding shallow marine environments, although most historical aerial/satellite imagery has been obtained with a view to terrestrial objectives (Mount et al. 2007) and the analysis of historical images from a marine habitat mapping perspective is frequently restricted (see Hart 1999). The limitations to detecting habitat differences in aquatic systems from aerial images include (Mount 2003, DEH 2007a, Mount et al. 2007):
- Water depth, - Water clarity,
- Sun angle and reflection and - Water surface state.
In spite of these restrictions, remote sensing has proved to be a useful tool in
identifying habitat modification in shallow marine systems (Allee et al. 2000, Mount 2003, Mount et al. 2007). Even so, acoustic technologies and processing techniques are increasingly capable of covering large areas of substrate with substantial accuracy, largely independently of factors that limit more traditional approaches. However, it is important to realise that habitat mapping is never an exact science with sacrifices being made relative to the competing needs for habitat type resolution versus spatial coverage. Further, it needs to be realised that the boundaries between habitat types are often broad transition zones rather than rigidly constrained and that these zones may shift according to seasonal fluctuations in vegetative cover (DEH 2007a).
Regardless of the approach to broader habitat classification, finer scale investigation requires varying levels of ground truthing, generally in the form of video or SCUBA operations (DEH 2007a, Mount et al. 2007).
The following describes a program of marine habitat mapping in the EP NRM region, building on recent developments in subtidal mapping. The aim is to develop a system of reliable, repeatable and relevant habitat mapping capability for near-shore
environments that can be employed as a basis for natural resources monitoring and management.
5 Benthic habitat mapping and ground truthing methods within the EP NRM region
Mapping of marine habitats included eight large embayments across the EP NRM region from False Bay in upper Spencer Gulf in the east to Denial Bay in the west (Figure 1). Mapping covered the area from median high water out to the sheltered extent of each bay, with the exception of Streaky Bay where mapping was limited by the availability of suitable aerial imagery.
Full coverage mapping of the Eyre Peninsula coastline, while desirable as a long- term goal, was not practical within the constraints of this study. The embayments chosen encompassed the major habitats likely to be impacted by shore-based
activities, in particular reef and seagrass systems. Information on the distribution of benthic habitats was collected using a combination of techniques that collected data across increasingly finer scales, including:
- Aerial imagery was used to assess the spatial extent of habitats at the broadest level. Boundaries between habitats such as seagrass, bare substrate and reef are often evident on aerial images and have previously been used to map habitats out to 15 m in South Australia (DEH 2007a provides a simple overview of this process and habitat mapping in general).
- Acoustic data (from a single beam sounder) to further define the extent of habitats in deeper water where light penetration is limited and provide confirmation of habitat extent in areas mapped from imagery.
- Habitat identification and verification carried out using towed underwater video.
All information collected was compiled as spatial layers within a Geographic Information System (GIS) and used to produce hardcopy map books and an interactive ARC reader DVD. The latter enables users to access spatial layers for habitat and video ground truthing as well as underwater images.
The following sections describe this process in detail.
5.2 Digitisation of aerial imagery
Orthorectified aerial imagery used for digitisation of habitat boundaries for the Eyre Peninsula region was collected by DEH in 2004 at a resolution of 1 m per pixel. In 2004, imagery was not available for False Bay and as a result imagery from 2001 collected at a resolution of 2 m was used.
Habitat boundaries were identified on imagery and digitised (digitally traced) based on varying patterns, tones and textures on the orthorectified aerial imagery using GIS (Figure 2).
Figure 2 - Example of habitat delineation on an aerial image.
5.3 Video ground truthing
Extensive video ground truthing was carried out to validate mapped polygons derived from the aerial imagery. Proposed video sampling points were selected from imagery of each bay with the aim of maximising evenness of coverage of each bay (access was limited in some of the shallower bays) with a balanced representation of different habitat types. Video footage was collected at each of the pre-determined sampling points using one of two high-resolution, towed underwater video camera systems Morphcam by Morphvision, connected to a Sony GVD1000e digital video recording deck or a Scielex underwater video camera linked to an Archos portable digital hard drive recorder. Each video sample consisted of a 30 second drift. Differential GPS data was simultaneously encoded on the audio track of the videotape to provide position information relative to video footage.
Benthic habitat data was extracted from video footage using a purpose-built Visual Basic program. The program allows the operator to view videotapes and assign habitat types, which are stored along with the corresponding GPS location from the audio channel. Data were then compiled in a database from which GIS spatial layers were produced. Around 4100 video observations were collected and analysed in the EP NRM region.
5.4 Acoustic ground truthing
Interpretation of aerial imagery is subject to uncertainty due to the water clarity/light penetration and sun reflection on the sea surface and becomes less reliable with depth (Mount 2003, DEH 2007a, Mount et al. 2007). While the majority of mapping for the eight embayments was derived from aerial imagery, in the deeper margins of Coffin Bay and Denial Bay acoustic survey methods were also used.
Five acoustic (echo sounding) transects were undertaken in both Coffin Bay and Denial Bay to increase the confidence of habitat delineation from aerial images and to extend mapping beyond what is normally achievable from imagery in this region (i.e.
10 – 15 m). These surveys used a pole mounted Simrad EQ60 38/200 kHz transducer across a series of parallel transects spaced ~ 1000 m apart run through the deeper central parts of each bay from between 10 and 15 - 20 m depth. All surveys were conducted at a speed of around 3.5 knots. Acoustic data was collected and stored on
the surface control unit hard drive along with differential GPS information. Several types of information were extracted from acoustic data, including;
- Bathymetry (depth), - Substrate composition, - Substrate relief and - Presence of vegetation.
Acoustic data was classified based on data for two frequencies (38 and 200 KHz) from the logged raw sounder files in Echoview software (by Sonar Data Version 3.50). Classification of different habitats was based on the thickness and intensity of acoustic returns and differences between the two frequencies (Figure 3). Harder substrates tend to reflect acoustic energy more strongly thus producing a stronger second echo, while rougher (higher relief) substrates tend to scatter acoustic returns resulting in longer tail on the first echo. Acoustic reflectance above the sounder- detected bottom for the lower frequency (38 kHz) can often signal the presence of vegetation (Lucieer et al. 2007), particularly dense seagrass, although consistent differences in sounder-detected bottom between the two frequencies are also a strong indicator for the presence of seagrass (Figure 3) while regular inconsistencies suggest rough hard bottom (typically reef). Sounder-detected bottoms for the two frequencies tend to be the same in areas dominated by bare sand.
Figure 3 - Example of acoustic echogram for 38 khz (with 38 and 200 khz bottom detection lines overlaid) showing signals for sand, seagrass and reef.
Classified seafloor types based on acoustic data along with spatial geo-referencing information from a differential GPS were used to create a GIS spatial layer of substrate/habitat types.
5.4.1 Classification of habitats/production of maps
The approach used for classification of benthic habitats for marine habitat mapping in the Northern and Yorke NRM region for the upper Spencer Gulf and Gulf St Vincent (see above; DEH 2007a) was modified to include new habitat types encountered in the EP NRM region (and others) comprising four levels (Figure 4; DEH 2009a) in line with approaches used elsewhere in Australia and internationally.