Abrupt historical change in the nature of a Ramsar Wetland: Seaford Swamp, greater Melbourne, Australia

Introduction

Wetlands change in response to natural drivers such as climate and hydrology (Reidinger-Whitmore et al. 2023), as well as a result of pressures from humans (Mills et al. 2023). In many instances, the first impacts of people are ancient (Dubois et al. 2018) and so the natural state of the wetland is obscured by time. Long-term data, historical records and palaeoecological information can expand our understanding of the response of ecosystems to natural or anthropogenic impacts (Manzano et al. 2020). The knowledge derived from these approaches provides evidence for the condition pathway through which a wetland has passed. This evidence may be critical for current management because it provides evidence for variability, present transience or stability, and the primary drivers of condition. Although human recollection may provide evidence for past condition, it can be punctuated, subjective and dominated by recent experience (Tibby et al. 2008). Palaeoecological approaches often provide continuous records of change, and while resolution diminishes with sediment depth, can provide evidence of state and transition beyond memory and documentary records.

The extraction of evidence of past condition is available through the exhumation of continuous sediment sequences and the extraction from the sediment layers of biological and chemical indicators of wetland state. These palaeolimnological approaches have now been widely applied to wetlands across south-eastern Australia (e.g. Head 1987; Gell et al. 2005a; MacGregor et al. 2005; Reid et al. 2007; Cook et al. 2016), often showing natural character states that have been obscured with time.

The Ramsar Convention on Wetlands came into effect in 1975, with the aim of mitigating the loss and degradation of the world’s wetlands (Stroud et al. 2022). It was widely recognised that, by then, many wetlands had been lost or changed from their historic condition (Gell et al. 2023). This trend of loss and degradation has continued (Davidson et al. 2023), but it is believed that the Convention has acted to slow its progress. The ecological character of Ramsar sites (wetlands listed under the Convention as internationally important) are described at the time-of-listing, but signatory nations are not obliged to use the ‘time-of-listing’ state as the natural ‘ecological character’ (Pritchard 2022). Where evidence shows a past condition, then restoration to that target may be appropriate, particularly if that represents a ‘positive change’. Importantly, site managers otherwise seek to retain the ecological character, including the biota and ecosystem services, of the wetlands, as well as the biodiversity criteria used to list the site as internationally important (Pittock et al. 2010).

In wetlands that are on a trajectory of degradation, management interventions that reinstate a past condition (or ecological character) and retain at least one criterion for listing may be preferred to options that seek to retain time-of-listing conditions. Such a decision was undertaken in Chilika Lake in India that suffered from high sedimentation and a transition to primarily freshwater (Kumar et al. 2020a). Here palaeoecological evidence (Pandey et al. 2014) showed greater connection to the ocean in the past and management invested in establishing an opening to the sea to recover lost fisheries production as well as species more dependent on brackish–saline conditions, such as mangrove communities (Kumar et al. 2020a, 2020b).

The Seaford and Edithvale Wetlands are located on Bunurong Country and were originally part of a continuous ‘Carrum Carrum Swamp’ that ran parallel to the coast behind the foredune systems, from near the modern-day Melbourne suburbs Mordialloc to Frankston (Department of Sustainability and Environment 2012). From 1866, this wetland system was subjected to considerable deliberate change for the purpose of establishing, first, market gardening and dairy farming, and ultimately, residential communities within the developing Melbourne metropolitan area (Table 1). Today the hydrological system of Seaford Swamp has substantially changed in that it no longer receives water from Carrum Swamp and is now fed from a pump on a drain and via tidal inflows from Kananook Creek via another drain (Lane 2011). Quin et al. (2016) noted that tall marsh vegetation, dominated by common reed (Phragmites australis), has been expanding into other local wetland vegetation types reducing floristic diversity and faunal habitat suitability. This is presently managed by grooming at considerable cost to the local managers.

There are many biological indicator groups that can be extracted from sediments to reconstruct the past condition of wetlands. Diatoms are a form of algae that are sensitive to water quality (Reid et al. 1995). They have a skeleton of silica and so they are often preserved as fossils in sediments. The availability of databases of diatom species and water-quality data from studies across Australia (e.g. Gell 1997; Philibert et al. 2006) and internationally (e.g. Hall and Smol 1992; Gasse et al. 1995; Witkowski et al. 2000) provide a means of interpreting water quality from the fossil diatom record. The extraction of diatoms from swamp and estuarine sediment sequences can, therefore, provide a record of changes in water quality over time (Taffs et al. 2017a). Similarly, pollen can be preserved in wetland sediments and the identification and enumeration of these can yield evidence for past terrestrial and aquatic vegetation (Ellison 2017). Important to understanding ecological change from these fossil assemblages is the establishment of a timeline. The chronology of sediment cores can be estimated using biological (exotic pollen markers) and geochemical (radiometric isotopes) means (Last and Smol 2001).

Materials and methods

Seaford Swamp was visited on 22 July 2022. Three sites were selected for coring on the basis of ease-of-access, and the anecdotal record of permanency of inundation. The sites selected were in the south (SFS), central (SFC) and northern (SFN) sections of the Swamp (Fig. 1). The central points in sites with open water were accessed on foot with a support canoe (Fig. 2). The initial cores, extracted using a d-section hand-corer (Jowsey 1966), were short. Maximum core lengths were 48 cm (SFS), 60 cm (SFC) and 85 cm (SFN). Preliminary analysis of the sediments showed that diatoms were preserved and that broad changes in the assemblages were evident down core. This outcome justified a detailed analysis combining diatom analysis, with analysis of preserved plant pollen (palynology) and radiometric dating (²¹⁰Pb; ¹⁴C). Seaford Swamp was visited again for sediment coring on 2 February 2024. In this instance, a 2.0-m core was extracted from the northern site (SFN; Fig. 3) and cores of 50 and 64 cm were extracted from the central (SFC) and southern locations (SFS) shown on Fig. 1. All cores were wrapped with plastic and stored in a refrigerator until they were subsampled.

Fig. 1.

Map of Seaford Swamp showing core sites.

Fig. 2.

Location of Core SFN showing open water conditions and stands of common reed in the distance.

Fig. 3.

The upper 50 cm of Core SFN showing abrupt change (arrow) in stratigraphy.

The longer core (SFN) was selected as a focus for detailed analysis and dating. In total, 30 samples were taken from the 2.0-m core for pollen and diatom analysis at increasing intervals down core, and 20 samples were taken for supplementary diatom analysis of the SFC and SFS cores.

For diatom analysis, the samples were treated with HCl and H₂O₂ and mounted on slides with Naphrax, following standard procedures (Battarbee et al. 2001). The coverslips were scanned with a Zeiss Axiolab microscope at 1000× magnification under phase contrast. At least 100 diatoms, and as many as 260, were identified from each sample, by reference to Krammer and Lange-Bertalot (1986, 1988, 1991a, 1991b), Foged (1978) and Witkowski et al. (2000), and enumerated. Species counts were expressed as a percentage of the total and presented as a stratigraphic diagram by using the program C2 (ver. 1.8.0, see https://www.staff.ncl.ac.uk/stephen.juggins/software/C2Home.htm; Juggins 2007). Species were classified as marine, mesosaline, fresh–oligosaline and eutrophic on the basis of the above flora and a range of national and international databases. The diatoms were grouped into zones delineated by sustained changes in the main flora.

Samples for pollen analysis were prepared following standard acetolysis approaches following Gosling et al. (2022). Usually at least 300 pollen grains were identified following the Australian pollen database. Pollen counts are expressed as a percentage of total dryland pollen and are presented as a stratigraphic diagram by using C2.

Ten subsamples across the upper 50 cm of Core SFN were sent to the laboratories at the Australian Nuclear Science and Technology Organisation (ANSTO) in Menai (Sydney) for radiometric analysis. Samples were prepared following standard procedures and activated in an alpha (small) counter. Initially, four samples were analysed (0, 10, 20, 30 cm) and the results of these were used to select the other two samples for analysis (5, 15 cm). The basal sediments of Core SFN held many small shells and these were extracted, wrapped in alfoil, and sent to the Chronos Radiocarbon Laboratory at UNSW for AMS ¹⁴C dating.

Results

Sediment stratigraphy and dating

The 200-cm SFN core showed a clear stratigraphic change, with marl-like grey clays from the base to 25 cm (Fig. 3), above which were organic sediments with plant remains. Similar grey clays were evident at the base of the 50-cm core from SFC and the 64-cm core from SFS. Generally, there was a sharp transition to organic surface sediments, although at SFS this varied with alternating grey clay and organic sediments.

The preliminary ²¹⁰Pb data from the initial four samples are provided as Fig. 4. Here, the sample at 30 cm is fully depleted (−2 ± 2 Bq kg⁻¹), suggesting an age of >> 80 years. The unsupported activity at 20 cm suggests an age of 72 or 90 years, depending on whether the Constant Flux Constant Sedimentation (CFCS) or Constant Rate of Supply (CRS) models are applied (see Appleby 2001). This would suggest that the sediments at 20 cm were deposited in c. 1943 CE, yielding an average sedimentation rate of ~2.5 mm year⁻¹. The sample at 15 cm yielded background results (Table 2), confounding the initial age–depth curve attempted from the initial samples. Further, the unsupported ²¹⁰Pb at 5–6 cm was lower than at 10–11 cm. With these additional results, the low values at 5 cm are now considered ‘background’, even though an age was initially inferred for 20–21 cm. ²¹⁰Pb analyses can be affected by low sedimentation rates, variable sources of sediment, and salinity (Gell et al. 2005b). It may also be affected by bioturbation; however, the clear diatom stratigraphy questions that prospect. Here, we tentatively accept the sedimentation rates inferred from the first four samples; however, clearly the timeline requires support from additional analyses.

Fig. 4.

²¹⁰Pb dating output from the initial four samples. The difference between the supported ²¹⁰Pb (a) and the total ²¹⁰Pb (b) provides the unsupported ²¹⁰Pb values (c). The depletion of the unsupported ²¹⁰Pb is modelled to yield the age of the samples (d) relative to the surface.

Table 2.²¹⁰Pb data from alpha analysis.

ANSTO ID	Depth (cm)	Supported 210Pb	Unsupported 210Pb
Z605	0–2	20 ± 2	96 ± 6
Z606	5–6	19 ± 2	8 ± 2
Z607	10–11	16 ± 1	9 ± 2
Z608	15–16	22 ± 2	−11 ± 2
Z609	20–21	17 ± 2	12 ± 2
Z611	30–31	19 ± 2	−2 ± 2

One AMS ¹⁴C age was sought from shells collected from sediments between 190 and 195 cm. The raw result from the analysis was 6330 ± 21 years BP (UNSW-3375), which was converted to 6713 ± 83 years BP (Table 3) after the marine reservoir effect was factored in, following Reimer and Reimer (2001).

Table 3.Results from¹⁴C AMS analysis.

UNSW lab code	Sample label	Date 14C year BP	Date ± 14C year BP	Calibrated age	Age±
UNSW-3375	SFN (190–195 cm)	6330	21	6713	83

The arrival of Pinus pollen at 12 cm confirms a post-European age supported by increases in exotic plantain and daisies at ~30 cm (Fig. 5). The arrival of pine pollen has been used to mark 1850 AD, but Bickford et al. (2008) showed that it first appeared in some sediment records much later.

Fig. 5.

Pollen record from Core SFN.

Diatoms

Most diatom valves were readily identified to species. The identification of some types within the grey clay sediments were identified with less confidence because the valves were often fractured, and in places partly dissolved. Here, depending on species shape, the central area was counted as one, whereas ends were counted as a half. So, the more heavily silicified species were identified more readily and others were mostly recognised where large fragments or entire valves were seen.

Core SFN

The diatom analysis of 30 samples from the SFN core yielded 146 diatom types. The C2 stratigraphic diagram of the diatom assemblages across the 2.0 m of the SFN sediment core is provided as Fig. 6.

Fig. 6.

Diatoms record from Core SFN.

Zone SFN-3 (200–50 cm)

The diatoms of the lower 150 cm are dominated by marine–estuarine taxa, in particular Grammatophora oceanica and Paralia sulcata. Other species typical of coastal waters recorded were Ardissonia robusta, Biddulphia pulchella, Campylodiscus clypeus, Cocconeis scutellum, Cocconeis heteroidea, Diatomella balfouriana, Mastogloia baldjikiana, Navicula (syn. Pinnuavis) yarrensis, Opephora olsenii, Plagiogramma staurophorum, Rhaphoneis amphiceros and Trachyneis australis. Within these sediments, there is a transition from Biddulphia pulchella, Tryblionella panduriformis and Trachyneis australis below 1 m to Campylodiscus clypeus above 1 m. There were very few records of species that prefer fresh to oligosaline waters in those samples, with the proportion of marine taxa ranging from 60 to 100% of valves.

Zone SFN-2 (50–25 cm)

Above 50 cm, there is an abrupt decline in the proportion of marine species from 80% at 50 cm to 12% at 45 cm. The main coastal species that persisted into this zone was Campylodiscus clypeus, a species that has been recorded in inland saline systems (Gell 1997). These species were mostly replaced by mesosaline taxa, principally Aneumastus tuscula and Amphora spp., whereas euryhaline species Cocconeis placentula, Cyclotella meneghiniana and small species within the Fragilariaceae (e.g. Staurosira venter) also were common. The upper part of this zone marked the arrival of freshwater taxa, notably Frustulia rhomboides and the epiphytic Gomphonema spp. No marine species were recorded in the uppermost sample of this zone.

Sone SFN-1 (25–0 cm)

This zone is marked by the decline in mesosaline species and their replacement by fresh to oligosaline, and aerophilous forms. No marine taxa were recorded in this zone. Frustulia rhomboides and Gomphonema spp. continue to be well represented into this zone, with the latter decreasing towards the surface. Most notably, the often-acidophilous genera Eunotia spp. and Pinnularia spp. enter the record towards the top of SFN-2 and dominate in this zone. Other taxa occur more sporadically, with Planothidium delicatulum, Achnanthidium minutissimum and, towards the surface, Tabularia fasciculata becoming common. The latter, along with Nitzschia spp. (incl. N. filiformis) and Cyclotella meneghiniana, mark an increase in eutrophication, particularly above 10 cm.

There are several eurytopic (widely tolerant) species present through the record, including Cocconeis placentula and small species in the genera Staurosira and Staurosirella (ex-Fragilaria). The latter typically predominate in disturbed situations where there is a loss of habitat suitable for other species (Fluin et al. 2007).

Core SFC

The diatom analysis of the 20 samples from Core SFC yielded 114 taxa. The C2 stratigraphic diagram of the diatom assemblages across the 49 cm of the SFC sediment core is provided as Fig. 7.

Fig. 7.

Diatom record from Core SFC.

Zone SFC-4 (49–47 cm)

The lowermost sample was dominated by the marine forms Biddulphia pulchella, Grammatophora oceanica and Paralia sulcata, along with Diatomella balfouriana and Cocconeis scutellum. Ardossonia robusta, Diploneis didyma, Mastogloia baldjikiana, Navicula yarrensis and Rhaphoneis australis were also recorded in this lowermost sample. Marine forms represented over 80% of the total.

Zone SFC-3 (47–27.5 cm)

This zone is marked by the loss of marine forms and the persistence of the estuarine diatoms Cocconeis scutellum, Diatomella balfouriana and then Campylodicsus clypeus. The estuarine forms were largely lost to the record by 30 cm, with only Campylodiscus clypeus persisting to the top of the zone. Estuarine forms were increasingly replaced by mesosaline types, including Amphora spp., Aneumastus tuscula, Mastogloia spp. and Seminavis sp.

Zone SFC-2 (27.5–15 cm)

From 30 to 16 cm, the mesosaline types that entered in Zone SFC-3 dominated the assemblage, along with Aneumastus tuscula, Cocconeis placentula, Cyclotella meneghiniana and Seminavis sp. Frustulia rhomboides entered the record from before this zone. Species common in freshwater, including those in the genera Gomphonema and Pinnularia, increased from 20 cm. Small species in the Fragilariaceae were common from 45 cm, declined through the next zone and were largely absent in SFC-1. The epiphytes Cocconeis placentula and Gomphonema spp. were common.

Zone SFC-1 (15–0 cm)

By 12 cm, Eunotia spp. and Pinnularia spp. dominated. Aerophilous species, including Caloneis sp., Diploneis spp., and Hantzschia amphioxys, increased dramatically from 16 cm to become common from 14 cm. Planothidium delicatulum increased from 8 cm and Hippodonta capitata was common at the surface.

Core SFS

The diatom analysis of the 20 samples from Core SFS yielded 114 taxa. The C2 stratigraphic diagram of the diatom assemblages across the 64 cm of the SFS sediment core is provided as Fig. 8.

Fig. 8.

Diatom record from Core SFS.

Zone SFS-3 (64–37.5 cm)

Like in Cores SFN and SFC, the basal sediments (here 64–40 cm) were dominated by the marine forms Biddulphia pulchella, Grammatophora oceanica and Paralia sulcata, along with Campylodiscus clypeus. Ardissonia robusta, Cocconeis scutellum, Lyrella lyra, Mastogloia baldjikiana and Navicula yarrensis were also recorded in Zone SFS-3. The marine sum was lower than for the other cores, at ~30%. Here, the marine dominance was shared with mesosaline and oligosaline forms, including Aneumastus tuscula, Cyclotella meneghiniana, Craticula spp. and Nitzschia spp.

Zone SFS-2 (37.5–22.5 cm)

The assemblage between 35 and 25 cm was represented by a variable mix of mesosaline and oligosaline taxa. Species that were common in only one or few samples included Bacillaria paxillifer, Craticula spp., Ctenophora pulchella, Nitzschia capitellata, Pinnularia divergentissima and Tabularia fasciculata. Small forms of the Fragilariaceae became common from the top of SFS-3 and declined through this zone.

Zone SFS-1 (22.5–0 cm)

Above 25 cm, the mesosaline Amphora spp. increased, along with indictors of oligosaline to fresh conditions, including Tabularia fasciculata. Other species commonly observed through this zone included Cocconeis placentula, Navicula salinarum, Craticula spp. and Platessa oblongella. Several common taxa recorded in Zone SFS-1 are known as aerophilous types (e.g. Fallacia tenera), thriving in mudflat habitats. Nutrient indicators (e.g. Nitzschia spp.) are common in SFS-1, but several were evident throughout the core (e.g. Cyclotella meneghiniana).

Discussion

Diatom assemblages can be used to infer past water quality and habitat availability on the basis of the known ecology of living species. Because many diatom species have a cosmopolitan distribution, their known ecological preferences can be derived from works published elsewhere (e.g. Witkowski et al. 2000). In Australia, many diatom data bases have been established (e.g. Gell 1997; Philibert et al. 2006; Chessman et al. 2007) that allow for conditions to be deduced from inland waterways. Although there has been work on the modern and fossil diatom assemblages from Australian estuaries (Taffs et al. 2017b), relatively little work has been undertaken on the diatoms of Australian marine environments. Here, the observations of Foged (1978) during a tour of eastern Australia in 1966, combined with the global treatise of Witkowski et al. (2000), form a basis for understanding the ecological preferences of the thalassic (marine) species observed in the lower sediments of the three cores.

The distinct diatom assemblages allow for a confident interpretation of the environmental history to be made. Clearly, Seaford Swamp was, for millennia prior to European settlement, an open tidal system with mesosaline to eusaline waters. The high numbers of the estuarine plankton form Paralia sulcata suggest that the inflows of sea water to Seaford Swamp were substantial. The main marine taxa were observed at the base of all three cores, suggesting that, although being more variable in SFS, tidal conditions existed throughout the Seaford Swamp for millennia. The existence of a saltmarsh environment is inferred from the abundance of saltbush pollen and low numbers of aquatic plants such as grass, sedge and milfoil. Away from the swamp, the vegetation was dominated by sheoak and Banksia.

The ¹⁴C dating suggests that the wetland was established soon after sea levels stabilised at 7000 years ago. The period from 3500 to 2200 years ago was dry across western Victoria, relative to the mid-Holocene (Jones 2001) and probably also dry in the Port Philip Bay region. Dry conditions are implicated in a hiatus, from 2800 to 1000 years BP, in the deposition of delta deposits at the mouth of the Yarra River suggesting low sea levels in Port Phillip Bay (Holdgate and Norvick 2017). The loss of thalassic diatoms by 45 cm may have resulted from the receding Bay at 2800 years BP. However, such a timeline would invoke extremely low sedimentation rates from 45 to 30 cm, spanning the resumption of sea levels at 1000 years to the arrival of Europeans. In lieu of more detailed dating, the abrupt change at 45 cm is considered here to have occurred in early European settlement. So, the stability in the diatom assemblage at SFN between 200 and 45 cm suggests that the Bunurong people witnessed relatively stable wetland conditions leading into European settlement.

In the recent past, the hydrology of the system changed progressively from brackish-saline to mostly fresh. Although this may occur with gradual geomorphic changes at the coast, the abrupt changes recorded here suggest that this was probably due to urban and rural development, land drainage and hydrological changes associated with the construction of the Patterson Cut, a 10-m-wide drainage channel, in 1879 (Ross n.d.). The ²¹⁰Pb dating places this change at before 1943, and possibly as early as the late 19th Century. This has resulted in the blocking of sea water inflow to the wetland. The replacement of marine species with brackish-saline suggests the wetland lost tidal input but continued to be influenced by the residual salts for some time. From the late 19th Century, much of the catchment was cleared for agriculture and there has been substantial urban development. This has increased fresh catchment and urban runoff that has been diverted into the system as part of the development of the region. In Cores SFN and SFC, this has seen the complete replacement of the original diatom flora with fresh-to-oligosaline taxa, many with affinities for organic acids. This pattern is less clear in Core SFS, which was more variable and transitioned to mesosaline, eutrophic and aerophilous species. The pollen record shows a decline in saltbush (albeit temporary) and a rise in aquatic plants such as grass (probably reed), rush, milfoil and sedge that typically prefer fresher conditions. This appears to be most evident in the north of the swamp as reflected in the greater abundance of epiphytic diatoms (Cocconeis spp., Gomphonema spp.) and those preferring acid–neutral conditions (Eunotia spp., Pinnularia spp.) consistent with organic accumulation. The transition was abrupt in the north and more gradual in the south, implicating catchment and urban drainage entering the system from the north as the cause of the change. The changes in the diatom flora also suggest post-European increases in nutrient and sediment flux, again probably owing to changes in land uses in the catchment. Settlement has seen the decline in Banksia and Allocasuarina and increase in Eucalyptus.

The shift post-European settlement from a saline state to fresh is not unusual with coastal lagoons elsewhere in southern and eastern Australia, such as in northern and central New South Wales (Tibby et al. 2008; Leyden et al. 2011), south-eastern South Australia (Haynes et al. 2007) and the lower Barwon River in central Victoria (Reeves et al. 2016; Gell et al. 2022) experiencing considerable freshening.

The terrestrial pollen shows a decline in Allocasuarina and Banksia. Such a decline has been documented elsewhere on the Mornington Peninsula (Calder 1975), owing to their vulnerability to disease and preference as a source of fuel and timber. Eucalyptus has been advantaged by this disturbance, as has Pittosporum, which may have been widely planted or invaded unaided (e.g. P. undulatum). Reed (Phragmites australis) is intolerant of high salinity (Hart et al. 1991) and so has been advantaged by the increase in fresh water. This poses a management problem because it is now dominating other aquatic habitats.

Seaford Swamp has been identified as a system that has been affected by human development and hydrological management over recent centuries. This record attests to that change and identifies the natural state of the system enjoyed by the original inhabitants prior to this intensive disturbance. The long-term state was one strongly influenced by tidal inflows and was probably a saline marsh. Most of this habitat has been lost and the system is now largely fresh-oligosaline, allowing for the colonisation of aquatic plants typical of freshwater such as reed. The growth of reed now poses an issue for the management authority that now invests resources in controlling its growth.

Seaford Swamp was listed as an internationally important wetland under the Ramsar Convention in 2001 where it was broadly recognised as a predominantly freshwater marsh. The identification here, of a prior estuarine state, and the timing and nature of its transition to fresh water, is clear evidence that the ecological character of the wetland had changed before being listed as internationally important. That is, the ecological character (or condition) described when it was listed does not represent its natural ‘ecological character’, but rather it represents a character that was the result of land and water use and development following European settlement of the area.

The extended record of change in the Seaford Swamp identified in this investigation ‘opens up’ the suite of target states that management may consider. This could include the present freshwater state or a more saline one that might reduce the proliferation of reeds and reflect that witnessed by Bunurong People in the past. The pre-European estuarine state identified here also provides a future option under scenarios of inundation under sea-level rise (DEECA 2024). Whichever, it remains that management needs to also consider the listing criteria that underpin its original nomination as well as the obligation to maintain the ecological character, which includes the key features of the biodiversity, the ecological processes that sustain the wetlands, and the ecosystem services that it provides for people.