Australia: Salt Marsh Restoration at the Olympics 2000 Site (Homebush Bay)


As part of a large-scale remediation strategy for the Olympics 2000 site, the New South Wales Olympic Coordination Authority commissioned the restoration of two saline wetlands in Homebush Bay near Sydney, Australia–North Newington and Haslams Creek. Before beginning the full restoration of both sites, however, project planners first conducted experimental transplantation at a pilot site in the Haslams Creek wetlands. Modifications were made to the site’s hydrology and sedimentation dynamics, and then cuttings of six species from nearby donor sites were transplanted along a gradient. During the subsequent 2-year monitoring period, growth and survival rates were documented in an effort to better understand the effects of salinity, sedimentation, and tidal influence on the species’ survival, and to predict patterns of colonisation and zonation in the recreated saltmarsh systems.

Quick Facts

Project Location:
Sydney Olympic Park NSW, Australia, -33.8466458, 151.0580539

Geographic Region:
Australia & New Zealand

Country or Territory:


Estuaries, Marshes & Mangroves

Area being restored:
2.5 hectares

Project Lead:
New South Wales Olympic Coordination Authority

Organization Type:
Governmental Body


Project Stage:

Start Date:

End Date:

Primary Causes of Degradation

Urbanization, Transportation & Industry

Degradation Description

The wetlands of the Parramatta River basin were part of the first river catchment on the Australian continent to experience urban development, having been visited and mapped within 10 days of the arrival of the First Fleet in 1788. The highest wetland losses in the catchment have been of saltmarsh, partly because they are drier and therefore more easily `reclaimed’ than mangroves, and partly because mangroves tend to regenerate by colonising new mud flats created by increased sedimentation from works upstream. The Homebush Bay wetlands are estimated to have been reduced to less than one-sixth their original size over the last 200 years (Katchka, 1992).

The Newington North wetlands were used for salt-farming and cattle grazing in the early nineteenth century, and were then enclosed in the Royal Australian Navy Armaments Depot (RANAD) Newington, occupying approximately one-quarter of its area, in the north-east of the site adjacent to the River. The Naval Depot operated for approx. 100 years until its recent transfer to the OCA. The presence of the Navy, while it protected the area from many impacts of the later urbanisation of Sydney, led to further trenching, waste dumping and associated disturbances by the Navy itself. Nevertheless, the plant community appears to be a remnant of the original vegetation (Clarke and Benson, 1988). The most recent major impact on the Newington North wetlands has been the loss of tidal flushing. This has been partly the result of dredge-fill dumping from the river, which has elevated the area somewhat relative to the shoreline, and partly because, over the last decade, channels dug by the Navy to enhance tidal flushing have become almost entirely choked by sedimentation and mangrove colonisation.

The second area of saline wetlands is represented by remnant patches of what was once a continuous border along the tidal length of Haslams Creek. Abattoirs and brickpits were constructed adjacent to the Creek earlier this century, leading to polluted storm-water run-off, and a large area of wetlands and was used as a waste dump for domestic and industrial purposes.

Reference Ecosystem Description

The estuarine intertidal ecosystem of New South Wales consists of mosaics or zonal bands of mangroves and saltmarsh. The mangroves normally occupy the area from just above the low tide mark to mean high tide levels (the area occupied by `lower marsh’ in colder climates), and the saltmarsh extends from mean high to high high tide levels (that is, only the `upper marsh’ zone in colder latitudes).

The saltmarsh communities of Homebush Bay in particular are regarded as the most important in the Sydney region, because of their size and species composition (Adam, 1996). They are also valued as a waterbird habitat, and have been recognised by the Australian Government via inclusion on the register of the National Estate, in bilateral protection agreements with China and Japan, and through the government’s assent to the Convention on Wetlands of International Importance, Especially as Waterfowl Habitat (RAMSAR) (OCA, 1995; Straw, 1996).

Project Goals

The OCA’s goal for wetland rehabilitation is the restoration of full ecosystem functioning, if possible, by means of rehabilitation of the plant communities (OCA, 1996). It is hoped that the rehabilitation of habitat will lead to the natural recovery of the faunal complement of the system, which is at present very depauperate (Berents, 1993). Particular attention will be given to the re-establishment of three saltmarsh species which are rare in Port Jackson, and whose biology is largely unknown. These species are Lampranthus tegens (F. Muell.) N. E. Br., Wilsonia backhousei Hook. f., and Halosarcia pergranulata (J. Black) Paul G. Wilson. If possible, the area covered by these three species will be increased, and the maintenance of local biodiversity will thus be ensured.


The project does not have a monitoring plan.

Description of Project Activities:
Site modeling and baseline ecological studies were the first project activities undertaken. Data was collected to ascertain the presence and distribution of different saltmarsh species (especially those of the three rarer species), the extent of tidal influence, and sediment conditions at the project site. These data were compared to corresponding data from reference sites in Port Jackson and the wider Sydney region in order to compare species composition, zonal distribution and substrate conditions supporting mangrove and saltmarsh species. Restoration then began at an experimental plot on a tidal inlet of Haslams Creek that had been used over the previous 30 years or more as an industrial waste dump. Between the waste dump and the edge of the creek was a band approx. 15 m deep of mangroves, Avicennia marina var australasica. (Forsk) Vierh. The banked up area behind the mangroves was excavated to a depth of 3-4 m to remove the waste, and re-filled to a lower level using clean fill from a building site in the city. The experimental plot was approximately 70 m wide by 80 m deep, at right angles to the Creek. This area included a band about 20 m deep immediately behind the mangroves, and at the same level (i.e. mean high tide mark) so that the transplants would not be shaded by the mangroves. The remaining 50 m length was graded to form a ramp, at a slope of approximately 1 in 100 m, i.e. to a total elevation of about 50 cm above the level of the mangroves, before curving up into a fairly steep bank. Levels were checked by laser-aided survey, and at each successive 10 cm elevation, 20 planting stakes were set out across the ramp. The plot was then left for tidal inundation and salination to take place over a 3-month period before the first transplantations were carried out. Two series of trial cuttings were taken, in November 1993 (late spring) and in May 1994 (late Autumn). In each trial six species were used, the three dominants S. quinqueflora, S. australis and S. virginicus, and the three species rare in the area, W. backhousei, L. tegens and H. pergranulata. For each species, new season tip-and-stem cuttings were taken from other nearby saltmarsh stands. The cuttings were taken from as many plants or clumps as possible, both to avoid damage to individual plants, and to maximise genetic diversity of the transplant stock. The cuttings were placed between sheets of wet horticultural capillary matting for transportation back to the glasshouse. The cuttings were first washed in tap water to remove any salt build-up, and then dipped for 5 s into 1% sodium hypochlorite solution for surface sterilisation, and rinsed in distilled water. Except for S. virginicus, for which internode segments were used, tip segments were used, and sometimes two or three single-node segments from immediately behind the tips. On the basis of a pilot test of rates of root production, the cut ends of all but L. tegens and S. quinqueflora were then dipped into root-promoting hormone gel, `Clonex Purple', Growth Corporation, Fremantle, WA. The cuttings were set out in germination trays containing a 1:1 sand: peat mixture. The trays were then placed in a fogging tent in the glasshouse with bottom heat under a sand base (25°C±5°C). When root and shoot growth were established, the plantlets were hardened off, first by lifting the tent flaps and lowering the level of fog, and then on the glasshouse bench. They were then potted up into small pots usually containing a 3:2 sand:peat mixture. After approximately 10 days to acclimatise, the plantlets were gradually salinated at 2-day intervals over a 2-week period, to a final watering concentration of half-strength seawater, before being planted out at the experimental plot. At each of the 20 stakes, at each of the five successive 10-cm elevation levels, two plantlets of each of the six species were planted in a radius of approximately 0.5 m around the stake, i.e. a total of 40 cuttings of each of the six species at each elevation. The plantlets were hand watered 2-3 times per week with tap water for 6 weeks after planting out, to help them survive the summer heat. For the follow-up study the experimental ramp was divided into a grid consisting of seven columns (i.e. at right angles to the Creek) by seven bands (across the site). Three of the columns were then randomly chosen for sampling plants and substrate. Up each of the three ramp columns, three 0.25 m2 quadrats were placed randomly within each of the seven bands, to measure numbers and sizes of coloniser plants. Substrate samples to a depth of 10 cm were also collected from every band of each of the sampled columns, and pH, water content, conductivity and organic matter (loss on ignition) were estimated. Rehabilitation activities were later conducted at the Newington wetlands, although not in association with the pilot project discussed here. At the Newington site, weed clearing, re-levelling at putative saltmarsh zonal levels, and transplantation of S. quinqueflora, L. tegens, W. back- housei were carried out during 1996 in the first of four segments of the wetlands to be inundated. The transplantations in this area were not of cuttings, such as were used in the experimental trials at Haslams Creek, but clumps or mats of plant material which had to be removed from the Haslams Creek area in order to make way for the major reconstruction of the wetlands there as part of the restoration exercise. In February 1997, tidal flushing was restored to the Newington site, by the opening of a large canal (approximately 10 m width, 4 m depth) to the river.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
Seasonal measurements were taken over a 2-year period at the Haslams site. Colonisation by both mangrove seedlings and saltmarsh species took place on the experimental ramp, as well as along unplanted corridors (approximately 2 m wide) that ran along either side of the ramp to provide access for heavy machinery used in the construction phase. The survival and rooting rates for all species in the spring trial were higher than 90%, whereas in the autumn trial the rates ranged from 60 to 80%. In addition, the spring/summer transplants had much better field survival and growth rates than the autumn/winter series, which were very low. Of the three common dominants, S. quinqueflora showed significantly higher survival rates up the whole tidal gradient than the other two species, indicating its tolerance of a wide range of substrate conditions. With S. australis and S. virginicus, it appeared that they may have needed a higher level of soil water than was consistently available in the middle levels of the ramp, but which was supplied by more frequent tidal inundation at the lower elevations, and by a slightly greater freshwater input from the bank behind, at the upper edges of the site. With respect to the three rare species, it was the other chenopod, H. pergranulata which, like S. quinqueflora, showed the highest survival rates up the entire tidal gradient. This was an interesting result, since in our ecological studies it was found that the H. pergranulata most commonly occurred either in supra-marsh elevations, above S. virginicus, or on salt scalds. The results indicate that in this species, although growth can be sustained over a wide range of substrate conditions, recruitment and establishment may depend on other factors which more strictly limit its distribution in the wild. L. tegens survived and flourished only at the uppermost elevations of the ramp, which is consistent with the levels found in the ecological study. W. backhousei had very low field survival rates, but they occurred at both the lowest and highest elevations, again where it can be inferred there would be more consistent substrate water contents. The growth of S. quinqueflora was greatest at the lowest elevations, which is consistent with the ecological findings. As noted above, S. australis had similar survival rates at the bottom and top of the gradient; however, growth was ten times as great at the upper elevations of the ramp, which had some intermittent freshwater seepage from the bank behind. For reasons which are unclear, S. virginicus, which normally grows on the upper edges of the marsh, in this experiment grew a little better at the lower levels (P < 0.05), though survival rates were similar in both zones. Although there were reasonable to high survival rates for H. pergranulata at every elevation, plant growth was two to three times greater towards the top of the gradient. It was found that over the period from the first transplanting, the three commonest Sydney wetland species had colonised much of the site, A. marina at the lowest elevation, with S. quinqueflora and S. australis at higher levels. The water levels on the ramp, as expected, declined with elevation, although from field observations it was known that the moisture levels at the top of the gradient, under the bank at the back of the plot, appeared higher from time to time than at the middle levels which were only inundated during spring tides each month. Conductivities increased up the ramp to the highest elevation sampled, where it declined once more, indicating the balance varying with elevation between relative frequencies of inundation that would increase salinity, and the leaching effects of freshwater input. Organic matter levels were twice as high at intermediate elevations than at either the top or bottom of the ramp. This may have been due to an accretion effect from the tides at these levels, since the total plant cover of transplants and colonisers was sparser there than at higher or lower levels. Colonising S. quinqueflora and S. australis were present at every level sampled. The conductivity levels at this site, which is on a fully flushed tidal creek, were much lower than those recorded in the ecological studies of the Newington North wetlands. However, elements of the normal zonation bands of the wetlands had been established within the 3 years. At the Newington North wetlands, tidal flushing was restored to the site via the canal constructed in 1997. As expected, the canal has caused the draining of a previously more or less permanent stagnant lake, which had existed since the cessation of tidal influence over 10 years ago. Consequently, the whole segment is somewhat drier than it had been before, excepting the zone flooded during neap tides. This new hydrology should favour an increase in the area of saltmarsh at the expense of mangroves over the next 3-5 years (Burchett and Pulkownik, 1996; Mamouney, 1996; Burchett et al., 1998).

Factors limiting recovery of the ecosystem:
At the Haslams site, S. quinqueflora showed significantly higher survival rates up the whole tidal gradient than the other two dominant species. This success was likely due to S. quinqueflora's tolerance for a wide range of substrate conditions. S. australis and S. virginicus, on the other hand, may have needed a higher level of soil water than was consistently available in the middle levels of the ramp. These species met with more success at the upper and lower elevations, where more frequent tidal inundation (lower) and a slightly greater freshwater input (upper) created more favourable conditions. None of the rare species was among the colonizers observed at Haslams, either on the ramp or in the corridors. This suggests that their limited distribution and abundance in the area are related to biological factors limiting establishment--i.e. low seed production or fertility, inefficient dispersal mechanisms and/or low seedling viability. At the Newington wetlands site, the renewed tidal inundation did not follow completely the pattern predicted by modelling and therefore, did not reach the transplants. Consequently, transplants are being irrigated manually until they can be transferred to other patches in Newington or back to a restored Haslams Creek wetland. Almost all of the S. quinqueflora planted at Newington died within 3 months, probably as a result of dehydration, but both of the rare species are so far surviving.

Socio-Economic & Community Outcomes Achieved

Economic vitality and local livelihoods:
Over the last decade there has been an increased community awareness in Australia of the value of wetlands and the need for their conservation to ensure plant biodiversity, protect fisheries nurseries, improve water quality in the estuary, and assist in the survival of migratory bird species. There is a community interest in the restoration and preservation of these wetlands, as evidenced by frequent reminders in the public media of the environmental obligations of the OCA. Community concerns have been expressed to the State Government regarding water quality improvement, pollution reduction, and fisheries restoration, as a result of which special funding has recently been made available for cleaning up Sydney. It is to be hoped that the example of coastal ecosystem rehabilitation offered at the Olympics 2000 site will assist in the evolution of future policies and practices in this region, as well as in a further public appreciation of the ecological and resource values of these areas.

Key Lessons Learned

The transplantation trials reported here had a number of shortcomings. They were confined to one site; two sets of cuttings only were assayed; and the tests were conducted on only six of a suite of about 15 species regularly found in the Sydney saltmarsh. In addition, the fill used as the substrate for the experimental plot was unlike the natural sediments in which saltmarsh is found in this region. Nevertheless, the results have been of use in providing a further understanding of the biology of the species concerned, and practical pointers to the development of rehabilitation strategies for the site. First, they show that the six species used can all be propagated readily from cuttings, using standard horticultural methods, with no added salt. It has also been shown that survival and growth from cuttings taken in spring/summer (temp. range 18-28°C) are much more successful than from those taken in autumn/winter (temp. range 1117°C).

This project has also shown that if a new area is constructed in which the appropriate conditions of hydrology, salinity and tidal flushing are provided, at least some of the common wetland species will colonise and cover a significant proportion of the area within 3 years. Thus, transplantation is not necessarily required, at least for a number of the common species. All that appears to be needed for natural rehabilitation of these species is the provision of an intertidal area with appropriate environmental conditions.

Long-Term Management

The rehabilitated natural areas of the Olympics 2000 site, which include two stands of estuarine wetlands, (named the Newington North and Haslams Creek stands), as well as freshwater wetlands, a stand of Cumberland Plain (eucalypt-dominated) woodland and a native grassland, will become part of a new Millennium Parklands. The Parklands will be maintained in perpetuity as an ecological amenity for the State of New South Wales and its visitors.

Sources and Amounts of Funding

$137,000,000 USD The pilot project undertaken at the Haslams wetlands site was just one component of a larger rehabilitation and remediation effort. An estimated $137 million was invested by the New South Wales Government’s Olympic Coordination Authority (OCA) on remediation measures at the Olympics 2000 site in Homebush Bay. The remediation resulted in extensive parklands providing recreational opportunities and conservation areas, rejuvenated waterways, world-class recreation and entertainment venues plus residential and commercial districts. Some of the savings were also spent on an “Enhanced Remediation Strategy” which was set up to address community concerns about perceived risks and long-term management of the site after the Games.

Other Resources

Sydney Olympic Park

Primary Contact

Organizational Contact