Australia: Seagrass Rehabilitation in Adelaide Metropolitan Coastal Waters


A seagrass rehabilitation experiment aimed at determining the best method for facilitating the natural recruitment of Amphibolis seedlings was carried out at two sites along the Adelaide metropolitan coast. Previous seagrass restoration in South Australia has focused on methods developed elsewhere, including the transplantation of mature seagrass units and the culturing and planting out of seagrass seedlings. The use of these methods has proven difficult, and results have generally been poor owing to the exposed nature of the Adelaide metropolitan coast. An alternative approach is to facilitate the natural recruitment of seagrass seedlings in situ. This project was conducted to assess the ability of various artificial substrates to aid in facilitating the natural recruitment of Amphibolis antarctica and A. griffithii seedlings. A range of biodegradable hessian bags, strips and mats in various configurations were deployed at two sites along the Adelaide coast, and the relative effectiveness of each substrate at facilitating recruitment was evaluated. The methods developed in this study offer significant advantages to other methods of restoration, particularly as they are non-destructive, cost effective, and may easily be deployed over large spatial scales.

Quick Facts

Project Location:
Somerton Yacht Club, 26B Esplanade, Somerton Park SA 5044, Australia, -35.00028209999998, 138.4996920072266

Geographic Region:
Australia & New Zealand

Country or Territory:


Coral Reef, Seagrass & Shellfish Beds

Organization Type:
Governmental Body


Project Stage:

Start Date:

End Date:

Primary Causes of Degradation

Urbanization, Transportation & Industry

Degradation Description

Over recent decades, the Adelaide metropolitan coast has lost more than 5,200 ha of seagrass habitat due to increasing anthropogenic pollution (e.g. sewage, stormwater discharges, etc.) and coastal development. Holdfast Bay alone has lost more than 4,000 ha of seagrasses since the 1940s (Figure 1; Hart, 1997). This loss of seagrass has contributed to considerable and ongoing erosion of the Adelaide foreshore and marine environment, and has reduced the amount of habitat available for many local species of flora and fauna.

Reference Ecosystem Description

Holdfast Bay, the site of the current study, is located adjacent to Adelaide (138°35’55″E, 34°55’42″S), South Australia. The Bay contains Posidonia angustifolia, P. australis, P. sinuosa, and Amphibolis antarctica beds interspersed with small patches of A. griffithii (Shepherd and Robertson, 1989). Heterozostera tasmanica and Halophila australis also occur throughout, often as an understorey component to larger species, although Heterozostera may occur more densely in deeper areas (Shepherd and Robertson, 1989). The seagrass community once dominated the soft-bottom coastline to within several hundred metres of the shore.

Project Goals

To assess the ability of various artificial substrates to aid in facilitating the natural recruitment of Amphibolis Antarctica and A. griffithii seedlings.


The project does not have a monitoring plan.


The stakeholders in this project include the Department for Environment and Heritage, the residents of the Adelaide coast, and the community of seagrass researchers and restoration practitioners that will ultimately benefit from the findings of this study.

Description of Project Activities:
Both study sites were located approximately two kilometres from the shore, in unvegetated areas previously dominated by seagrass, and were surrounded by dense seagrass beds consisting largely of Amphibolis. The first site (138º28'03.5"E, 34º54'01.9"S) was situated adjacent to Grange in approximately 8 m of water, while the second was located further north (138º27'47.9"E, 34º52'20.1"S), adjacent to Semaphore Park in approximately 6.5 m depth. Inshore surface water temperature within the area varies seasonally between 11.0 ºC and 26.6 ºC, while total dissolved solids generally vary between 33.0 and 38.8 μgL-1. To determine the spatial distribution of habitats surrounding the two sites chosen for this study, eight transects radiating from the centre of each site were established. Habitat types (sand, Amphibolis, Posidonia, Heterozostera and Halophila) were characterised along each transect, and in most cases, at least fifty metres of habitat was recorded beyond the edge of the sand patch in which the study site was located. The habitats surrounding the two sites selected for this study differed. Site one was characterised by a larger sand patch than site two (radius of 30.9 ± 3.39 m), and was surrounded almost entirely by monospecific meadows of Amphibolis antarctica (97.3%). The seagrass community surrounding site two consisted of dense Amphibolis antarctica beds (53.9%) and Posidonia sinuosa and P. angustifolia beds (44.2%). To quantify the light climate at each site, Odyssey light loggers (Dataflow Systems Pty Ltd, New Zealand) were attached to star pickets approximately 1 m above the sediment surface at the centre of each site in July 2005. The light loggers were programmed to record Photosynthetically Active Radiation every half an hour and were replaced on a fortnightly basis to avoid the build up of sediments and the growth of algae on the sensor element. As expected, the shallower of the two sites (site two) generally received more light during August, September and October. The opposite trend was observed during July, as near-shore, turbid waters associated with storm periods can substantially reduce seabed light availability in these shallower areas (S. Bryars, personal communication). With site data recorded, a range of biodegradable substrates in various forms was deployed at the two study sites in an attempt to enhance the natural recruitment of Amphibolis seedlings. Three different biodegradable materials were used in creating the recruitment facilitation substrates: a fine-weave hessian, a coarse-weave hessian, and mats constructed from interwoven seagrass (see Figure 5 in attached document). Using these materials, ten different recruitment facilitation techniques were trialed (each of varying size and configuration), including six different hessian bags, three different strips and a seagrass mat. The hessian bags and weighted strips were deployed on the September 7, 2004. In the week that followed (September 15-16), divers rearranged the bags and strips in a predetermined, randomised manner. Bags were randomly placed on either side of four transects, at least one metre apart, radiating in north, south, east and west directions from the centre of each site; while strips were randomly positioned along transects radiating in north-west, north-east, south-east and south-westerly directions. During the same period, the hessian strips and the seagrass mats were set out by divers along each transect. The three-metre long hessian strips were planted by digging a trench approximately 30 cm deep and pegging the hessian into the substrate at each end and in the middle. The trench was then filled in, burying the bound edges but leaving approximately 30 cm of the hessian exposed. Strips were laid in two different orientations (east-west and north-south) to assess the potential for long-shore currents to enhance recruitment. A three-metre gap separated each strip. The seagrass mats were anchored to the sediment with four steel tent pegs, one in each corner. Bags filled with sand were sufficiently heavy so as not to need staking down. Following deployment, all bags, strips and mats were labeled. At each site between eight and ten replicates of each method were deployed. To determine the effectiveness of the different recruitment facilitation methods, the number of seedlings on each unit was monitored over time. Monitoring was undertaken shortly after deployment on 12 October 2004 and subsequently in December 2004 and February, April, June and September 2005. In order to compare the effectiveness of the different methods in capturing and maintaining seagrass seedlings, seedling numbers were converted to seedlings per square meter for analysis. The growth of Amphibolis seedlings over time was assessed by randomly collecting 15 seedlings from the hessian units at each site in December 2004 and February and October 2005 (representing seedlings that were approximately 14, 23 and 57 weeks old). Seedlings that had been recently released from the parent plant were also collected in September 2005, and these were assumed to represent age 0. Divers collected the seedlings in plastic bags, which were then transported to the laboratory where morphological measurements--including seedling height, the number of shoots, the number of short shoots, and the number and length of all roots--were undertaken.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
Following the deployment of 196 recruitment facilitation units, over 16,500 Amphibolis seedlings recruited at the two study sites. Seedling recruitment and retention over the 53-week experimental period varied significantly with method and site, and at the end of the experimental period ranged from 6.7 ± 4.44 to 141.8 ± 32.84 seedlings per square meter. Approximately 31.4% of initial seedlings survived the 53-week experimental period, and seedling survival varied significantly with method. The most effective method in terms of seedling density at the end of the experimental period was hessian bags covered with a coarse weave hessian layer. This method, together with large hessian bags, was also among the most cost-effective. The most successful method in terms of initial seedling recruitment was not necessarily the most successful over the longer term. Indeed, seedling recruitment was highest on hessian strips, but this method had the poorest survival rate, with 10.6% and 27.1% on north-south and east-west oriented strips, respectively. At the end of the experimental period, the seagrass mats differed significantly from all other methods. They were the most ineffective method in terms of seedling recruitment and survival, with an average of 15.2 ± 8.48 seedlings per square meter at site one and 6.7 ± 4.44 seedlings per square meter at site two. Notwithstanding this, it is worth noting that no attached seedlings were observed on adjacent sandy substrates. Significant differences in seedling recruitment were observed between the two sites, and are likely to reflect the composition of the surrounding seagrass beds, depth and local hydrodynamic regime. Site one was more effective in recruiting Amphibolis seedlings, with an average 238% greater seedling density than site two (221.6 ± 18.49 and 92.9 ± 8.10 seedlings per square meter, respectively). The double-layered hessian bags supported the highest density of seedlings at both sites (site 1, 141.8 ± 32.84 seedlings per square meter; site 2, 95.6 ± 14.48 seedlings per square meter), although significant differences between sites were apparent. At site one, although mean seedling density was greatest on the double-layered hessian bags, it did not differ significantly from that on many of the other substrates, including the east-west oriented strips, weighted strips, hessian bags, large hessian bags, half buried hessian bags, hessian bags with mats and bags with a flap. Interestingly, at this site the north-south oriented strips retained far fewer seedlings than did the east-west oriented strips. At site two, however, the double-layered hessian bags retained significantly more seedlings than any other method. Above-ground biomass of Amphibolis seedlings increased substantially over time, owing to increased seedling height and an increase in the number of shoots and short shoots. Seedling height almost doubled over the 57 week period, increasing from approximately 104.4 mm to 198.5 mm, while the number of short shoots increased from 0 to 4.2. Secondary shoots were initially observed on seedlings approximately 23 weeks old and at 57 weeks, seedlings generally consisted of either 2 or 3 shoots. Below-ground biomass of Amphibolis seedlings increased almost seven fold over the 57 week period as a result of significant root development. Seedlings recently released from the parent plant generally only had 1 root, approximately 2.2 mm long, however, after 57 weeks seedlings had an average of 10.2 roots with a combined length of 639 mm. At this time, maximum root length had reached 213 mm and averaged 123.6 mm.

Factors limiting recovery of the ecosystem:
Differences in the ability of different methods to recruit and retain Amphibolis seedlings are likely to reflect a number of factors including material type (weave of fabric), profile on the seabed and ability to withstand degradation of the substrate. Seedling recruitment was generally highest on the coarse weave hessian fabrics (including the hessian strips and double-layered hessian bags), probably because the grappling hooks at the base of the seedlings become easily entangled in the hair-like nature of the fabrics' fibres. Fewer seedlings recruited onto the finer weave hessian fabrics, with recruitment particularly poor on the seagrass mats and the mats attached to hessian bags. Poor recruitment onto the seagrass mats may be explained by the fact that the interwoven seagrass making up the mats was coarse and inflexible, making it difficult for the seedlings to become entangled. Furthermore, the mats were situated close to the substrate and may have been more rapidly buried by sediments than other methods, reducing the time available for seedlings to attach. The half buried hessian bags are also likely to have been more rapidly buried, and recruited substantially fewer seedlings than their non-buried counterparts. Seedling retention was poorest on the hessian strips, probably owing to the fact that many of these strips weakened at the sediment level and were lost, along with attached seedlings. Differences in recruitment levels between sites may be explained by a number of factors, including variation in the proximity to surrounding seagrass beds and the reproductive output and density of those beds. Depth, hydrodynamic conditions and sediment dynamics may have also been factors in the differences between sites. Site one was located in about 8 m of water, approximately 1.5 m deeper than site two. Because the Adelaide metropolitan coast is considered to have moderate wave energies (Townsend, 2002), a difference in depth of 1.5 m between sites is likely to have a significant effect on the influence of currents and wave attenuation, both of which are likely to be considerably higher at site two.

Socio-Economic & Community Outcomes Achieved

Economic vitality and local livelihoods:
Seagrass meadows have long been recognised for their ecological and economic importance, and together with coral reefs and mangroves, are thought to represent one of the world's most productive coastal habitats (Short and Wyllie-Echeverria, 1996). Seagrass beds not only play a critical role in primary production (Borum et al., 2006) and nutrient cycling (Hillman et al., 1989; Romero et al., 2006), but they also provide habitat for a diverse array of marine organisms (Bell & Pollard, 1989; Short & Wyllie-Echeverria, 1996; Connolly et al., 1999; Duarte, 2002), and increase the stability of the seafloor through the growth of extensive rhizome mats (Fonseca & Fisher, 1986). Thus, restoration of these seagrass beds promises many long-term benefits for the residents of the Adelaide coast.

Key Lessons Learned

We have shown here that the provision of suitable substrates can facilitate the natural recruitment of Amphibolis antarctica and A. griffithii seedlings, representing a new method of seagrass restoration that may be suitable for deployment over large spatial scales in moderate wave energy environments in temperate Australian waters. Furthermore, it may be valuable to investigate the use of hessian bags in protecting other species of seagrasses in restoration efforts.

Not only does this method hold promise for improving the effectiveness of seagrass restoration efforts, it has also been found to be highly cost-effective. Extrapolating from the costs of the small-scale trial conducted here, it will cost on the order of $10,000 to rehabilitate one hectare of seagrass. This figure compares favourably with that for other methods, and also relative to the estimated economic value of seagrasses ($12,635 to $25,270 per ha per yr).

In spite of the inherent potential in this methodology, the recovery of seagrass beds through the recruitment of seedlings is highly dependent upon survival and growth of the seedlings over time. While recovery as a result of seedling recruitment has been observed (Preen et al., 1995; Kendrick et al., 1999; Whitfield et al., 2004; Olesen et al., 2004), in many cases, seedling survival is poor. Therefore, additional research is needed to explore ways of optimizing the recruitment process and ensuring the long-term survival of seedlings.

Long-Term Management

It is essential to continue monitoring the growth and survival of recruited seedlings to ensure that seedlings survive over a longer period and that the roots of the seedlings grow through the bags. It is also important to ensure that the shoots develop beyond the edge of the bag and begin to recolonise the bare substrates surrounding them. In addition to monitoring the growth of seedlings, it would also be beneficial to monitor the growth of surrounding Amphibolis beds, the reproductive output of these beds and other relevant aspects of their biology and morphology.

Sources and Amounts of Funding

The Department for Environment and Heritage, Coast and Marine Branch (CMB) provided the Aquatic Sciences division of the South Australian Research and Development Institute (SARDI) with financial support for this project.

Other Resources

South Australian Research and Development Institute (SARDI)

Primary Contact

Organizational Contact