Franklin and Zeehan Marine Park Multibeam mapping and drop-camera validation for Parks Australia

This report provides a comprehensive overview of the fine-scale substrata multibeam bathymetry and habitat mapping efforts conducted in the Franklin and Zeehan Australian Marine Parks (AMPs) located off the west coast of Tasmania, Australia. The primary objective of this mapping exercise was to map the bathymetry of the shelf areas of these AMPs (designated as Multiple Use Zones – MUZ’s) for the first time and to understand the distribution of habitats and key sessile invertebrate cover and species within these areas. These habitats support a variety of species including commercially valuable ones such as jackass morwong (Nemadactylus macropterus), striped trumpeter (Latris lineata) and, potentially, southern rock lobster (Jasus edwardsii). Characterisation of the seafloor was undertaken in two phases. An initial stage to comprehensively map these areas by multibeam sonar, followed by extensive ground validation undertaken by a drop-camera system that also recorded the presence of dominant sessile invertebrate fauna as well as benthic/demersal fish species.
Details of the multibeam mapping component of this study are presented in an associated CSIRO data report (Davey et al. 2022) and some additional mapping generated from this data is presented here, including examples of characteristic features of each of the parks surveyed. This survey was undertaken in appropriate weather conditions, with 30% swathe overlap and suitable vessel velocity to ensure all acquired data could be gridded at 2 m x 2 m scale (i.e. as high a resolution as possible for mapping at depths in excess of 100 m) in all shelf waters to enable fine feature detection and allow for optimal subsequent habitat distribution modelling and validation by drop-camera. This provided key insights into the range and extent of seabed features and an initial understanding of the overall bathymetry of shelf waters in both parks. Notable features include the shallow and relatively high profile mesophotic reefs in the northern section of the Franklin MP as well as the extensive low profile pavement reef found in the Zeehan MP, a feature also present in the southern region of the Franklin MP which underpins the shallower complex reef in the north of this park. Closer inspection of this area (in more finely gridded data) revealed clear geomorphological features that suggest the shallower reefs are likely of volcanic origin (likely basalt), reflecting the adjacent coastal and island geology of the region.
The location of the 300 camera drops in each park was based on a spatially balanced design (informed by the prior multibeam mapping). A camera system with near 360-degree views was deployed on the seabed for 5 minutes using a standard approach so fish assemblages and sessile invertebrates could also be recorded. The habitat data derived from the imagery was subsequently used in conjunction with the multibeam data to model and classify habitat distribution within the parks, providing the first validated habitat maps of the shelf MUZ’s of these parks for use in spatial planning and management. This habitat modelling component, coupled with a comprehensive catalogue of direct imagery of the seabed, revealed key insights about the seabed structure and substrata composition within the Franklin and Zeehan MPs. In the Franklin Marine Park, the dominant substratum types were shelf unvegetated sediments, mixed shelf reefs, and rocky reefs. In the south of the park, lower-profile reef systems (presumably of limestone origin) outcropped in approximately 80 m but were often sand-inundated except at distinct and often linear step-features in the bedrock, whereas, at the northern margin, a more complex high-profile system was present,
extending as shallow as 35 m. Overall, mesophotic rocky reefs and rariphotic shelf reefs represented only a small portion of the mapped area. The soft sediment features that dominate this park were usually observed to be markedly rippled, even at depths below 100 m, showing the significant influence of oceanic swells in this region, that likely regularly impact this substrate preventing any significant colonisation by invertebrate cover. The automated habitat mapping prediction accuracy (assessed by drop-camera validation) for habitat maps generated for the Franklin MP was moderate, with an overall accuracy of 71% and a kappa value of 0.55.
In the Zeehan Marine Park, specifically the continental shelf region of the Multiple Use Zone (MUZ), mapped features exhibited a similar pattern, with shelf unvegetated sediments being the most prevalent substrata type, followed by mixed rariphotic shelf reefs and shelf unvegetated sediments. Typically, the inner third of the shelf region in the park was sand dominated, with increasing flat pavement rariphotic rocky reef in the mid-third, with notable long step-features (often multiple km in length), while in the outer third the pavement (presumed to be limestone, similar to that found in the Franklin MP) became significantly more fractured into distinctively elevated blocks, with more step-features and less sand-inundation. The step-features (ledges) were rarely undercut (based on multiple camera observations), limiting habitat available for crevice-dwelling species like rock lobsters. Much of the shelf area in the park ranged from 100 to 120 m, yet despite this depth, the soft-sediment habitats were often distinctly rippled, showing evidence of significant swell-related disturbance, reflecting the high wave energy of the western King Island region. The automated habitat mapping accuracy for the Zeehan MP based on drop-camera validation of modelled habitat distribution was also moderate, with an overall accuracy of 71% and a kappa value of 0.47.
Accompanying the mapping results, the cover of sessile biota associated with different substrata classes was quantitatively examined. The results describe a diverse range of marine life, from bare rippled sediments to high-profile reefs covered in sponges, gorgonians, hard bryozoans, and tube worms. The diversity of the biota was particularly high on the volcanic reef features mapped at the northern end and mid-eastern margin of the Franklin MP. Additionally, tall Ecklonia radiata kelp forests were identified as a prominent feature in the upper mesophotic rocky reefs of the Franklin MP.
The use of drop camera habitat validation also allowed initial quantitative counts of the fish species observed within the Franklin and Zeehan MPs. The Franklin MP exhibited a higher fish species count, with butterfly perch (Caesioperca lepidoptera) being the most abundant species, which was typically found in association with the complex mesophotic reef systems. Other species like ocean perch (Helicolenus percoides) and jackass morwong (Nemadactylus macropterus) were present in the park but in lower numbers. In comparison, the Zeehan MP had a slightly lower fish species count, with butterfly perch and cosmopolitan leatherjackets (Meuschenia scaber) being the most abundant species. Similarly, other species such as jackass morwong and red gurnard perch (Neosebastes scorpaenoides) were present but in smaller quantities. Striped trumpeter (Latris lineata), a highly targeted species, was observed in small numbers in the deeper regions of the mapped area along the shelf break of the Zeehan MP. An initial assessment would indicate that the more fractured nature of the reef systems of the outer shelf area of this park probably provide more suitable habitat for both striped trumpeter and jackass morwong than the inner regions. Notably, no southern rock lobsters (Jasus edwardsii) were sighted in the imagery, presumably due to the absence of any suitable crevice-like habitat.
Overall, the use of panoramic drop cameras, when undertaken at high frequency utilising spatially balanced designs, provides a cost-effective and robust tool for initial exploration of MPs at shelf-wide scales, enabling robust validation of multibeam mapped features where such mapping is available. With sufficient replication (as undertaken here) it can also provide accurate substrata mapping (even in the absence of prior multibeam mapping). Importantly, this method of sampling also allows for a robust initial assessment and understanding of habitat relationships of associated demersal fish and dominant sessile invertebrate assemblages in these otherwise unknown regions. Thus, providing valuable information to inform initial management planning and subsequent biological assessments. However, despite the relatively high sampling density utilised here (300 drops per park), this approach only provides pilot-study scale information on the biological assemblages present. Additional sampling effort would be required to gain adequate sample sizes for reliable mapping/monitoring of specific biotic habitat features, as well as the overall distribution of fish assemblages within these parks. Ideally this information would be subsequently obtained by methods more targeted to these tasks.
While only 50% of the Franklin MP was surveyed using multibeam sonar due to bad weather, the subsequent drop camera survey was able to sample multiple sites throughout this unmapped region, suggesting much of this region is likely dominated by soft sediments with the notable exception of a complex mesophotic reef extending into the park midway down the eastern boundary of the MP. It is recommended that completing this multibeam mapping is a priority step to underpin future monitoring, particularly along the eastern margin.
It is also recommended that baited remote underwater video and autonomous underwater vehicle surveys are completed to provide a more comprehensive assessment of the demersal fish assemblages and sessile biota biodiversity of these parks.
In summary, this report, in conjunction with Davey et al. (2022), describes bathymetric features and distribution of subst