6 mangroves and coastal saltmarsh – past and futurechapter 6: mangroves and coastal saltmarsh –...

chapter 6: mangroves and coastal saltmarsh – past and future 277

6 mangroves and coastal saltmarsh – past and future

6.1 Pre-Europeandistributions

introduction

Maps representing vegetation patterns before European colonisation have become increasingly important in Australia, and are generally referred to as ‘pre-1750’ maps. As well as being historically interesting in their own right, they may also provide a reference point to calculate relative degrees of change and depletion and are sometimes used as aspirational targets for restoration of degraded environments (Oliver et al. 2002). In Victoria, the degree of depletion is essential information for determining the conservation status of vegetation and in designing a reserve system that is ‘comprehensive, adequate and representative’ (Commonwealth of Australia 1997). Depletion is usually measured within a bioregion, so that each recognisable vegetation unit can be assigned a Bioregional Conservation Status (BCS) (Natural Resources Management Ministerial Council 2005; see Fitzsimons & Robertson 2005 for an example with freshwater wetlands).

Two main approaches are used to infer long-term changes in vegetation and/or landscapes: i) cultural approaches; and ii) biological approaches (Lunt 2002; Tibby 2003; Egan & Howell 2005). Historical maps and survey data provide valuable cultural material for an analysis of vegetation change. Parish maps or early survey plans, for example, often show vegetation and land-use patterns and may be used to infer changes since European colonisation (Cook & Yugovic 2003). Such maps can often be supplemented with historical oblique (i.e. landscape) and, more recently, aerial photographs (Pickard 2002; Sinclair 2007). In the case of aerial photographs, high-quality images are generally available only for the period after World War 2 and so cannot be used to infer changes that occurred in earlier times (Fensham & Fairfax 2002). Notwithstanding this limitation, historical suites of aerial photographs can be very useful in quantifying changes in wetland vegetation over the past ~40–50 years (e.g. see Boon et al. 2008 for brackish-water wetlands of the Gippsland Lakes).

Other forms of cultural evidence include written records and oral histories. Edmonds (2005) has outlined the problems with historical reconstructions based on written records, and Fogerty (2005) the strengths and weaknesses of relying on oral histories. Oral histories are frequently used to infer changes that have occurred to landscapes over the past ~50 years and thus are of little value in constructing pre-1750 maps of likely vegetation distributions. Moreover, there is often a wide discrepancy in the value attributed to local knowledge and the oral tradition by decision makers and scientists involved in natural resource management.

Biological evidence can be used to gain perspectives extending further back in time than cultural evidence allows, especially in the case of Australia where the written record extends only ~200 years in the past (e.g. see Tibby 2003). In terrestrial studies, analysis of tree rings and fire-scar histories have been useful for determining the ages of trees in closed forests, open forests, woodlands and shrublands (Lunt 2002; Pearson & Searson 2002). Studies of floristic degradation, mostly using degradation sequences or environmental gradients, have found extensive use in grassy woodlands and tussock grasslands (Lunt 2002). Some of these studies use both historical written documents and field data, thus bridging the gap between the cultural and biological approaches. In the case of Victoria, there are scientific records of vegetation extending back to the landing of George Bass at Wilsons Promontory in 1798, James Grant’s visit to Western Port in 1801 and, most importantly, the collections of Robert Brown when he sailed with Matthew Flinders through The Heads in 1802 (Willis 1966; Moxham et al. 2009). These can provide useful background information on floristics, but are obviously limited in spatial extent and by the seasonality of the visit.

mangroves and coastal saltmarsh of victoria: distribution, condition, threats and management278

Palynolimnological studies can be used to document century-scale vegetation changes, particularly in moist sites (such as wetlands) where pollen is preserved. Although such studies tend to quantify changes in floristics rather than changes in wetland boundaries (Lunt 2002), they have been very useful in showing the loss of submerged angiosperm vegetation from floodplain wetlands across much of south-eastern Australia and the current dominance of primary production by algae and, in particular, phytoplankton (Leahy et al. 2005; Reid et al. 2007).

A tremendous difficulty with historical reconstructions of the pre-European Australian landscape is that there are ‘radically different readings of the same landscape’ (Horton 2000, p 72). As examples, controversy rages as to pre-European fires regimes and their ecological impact (e.g. see Benson & Redpath 1997; McLoughlin 2004; Gott 2005), the density of trees in grassy woodlands (Benson & Redpath 1997; Lunt 1997; Horton 2000; Griffith 2002) and the density of regrowth and so-called woody weeds (Benson & Redpath 1997). Against this background of the suite of methods available for analysis – all with limitations – is the possibility of a strongly political or ideological element to many of the interpretations that are made (e.g. see Horton 2000).

Reconstructing the pre-colonisation patterns and distribution of vegetation in Australia is obviously fraught with difficulties. In the case of mangroves and coastal saltmarsh, those difficulties are compounded by a number of factors. In some cases vegetation change is a simple case of subtraction: vegetation has been removed from the landscape without altering the pre-existing patterns, so that current vegetation maps are simply ‘cut outs’ from the pre-1750 maps. Saltmarsh differs significantly from that simple scenario, since it has both expanded and contracted in different parts of Victoria and in some places saltmarsh now occurs in places where it did not in 1750. The complexity results from the great dynamism of coastal saltmarsh, a vegetation type which can respond fairly rapidly (certainly within a century) to changes brought on by coastal engineering.

Second, many patches of saltmarsh are small and intricately shaped, meaning they are not easily mapped when they have been destroyed, and their size can be difficult to estimate without substantial error. Third, saltmarshes do not leave behind easy clues for historical ecologists to unravel. Often they have been utterly destroyed through changes to the whole landscape, such as from land-claims using earthworks, extensive landscaping for saltworks or water treatment, or the removal of whole areas of land for marinas and other types of coastal development. Most other vegetation types are removed without such wholesale alteration of the landscape. When forests or woodlands have been removed, for example, the soil usually remains accessible for interpretation, as do a few relict trees. In contrast, coastal saltmarsh can disappear without a trace.

Pre-1750 vegetation mapping was completed by the State Government for the whole of Victoria over several years in the 1990s. That dataset has a number of limitations which make a new and updated analysis desirable. The most notable limitation is the coarse scale of the original analysis; the mapping was done as part of a very large project, which limited the detail that could be allocated to particular areas or vegetation types. Our study aimed to revise the existing mapping with a finer-scale analysis. We consider only those vegetation types relevant to the map dataset outlined in Chapter 5. The delineation of units is discussed in more detail below.


mapping approach and methods

We used a variety of techniques to produce a statewide, pre-1750 map of the likely distribution of these types of coastal vegetation. In different areas, different historical resources or existing clues were available, so different techniques were employed regionally. Three approaches, however, were used routinely.

The first tool was to use historical maps. In some places, early surveyors’ plans show saltmarsh (often referred to as ‘samphire’) and mangrove patches quite clearly. In other cases, low water, high water and the spring-tide maximum are shown as lines (Figure 6.1). On some plans wetlands are shown and labelled ‘salt’ or ‘fresh’, thus allowing some assumptions to be made as to the EVC that formerly occupied the site. Other maps show unannotated wetlands that are ambiguous. We searched for useful early maps using the catalogues of the National Library of Australia and the State Library of Victoria, along with sections of the Public Records Office microfiche collection. The maps are cited in the text where relevant (see note on citation in the references section).

Figure6.1: Examples of historical maps used to infer pre-European distribution of coastal saltmarsh. Panel A shows a portion of Shallow Inlet mapped by Smythe in the 1840s (undated a). Note the use of the terms ‘Samphire’ (i.e. succulent chenopods) and ‘Mangrove’; the latter no longer occur in Shallow Inlet. Panel B shows a portion of the coast of Port Phillip, mapped by Noone (1880a). Rather than commenting directly on the vegetation, it shows the tide marks. The area shown in panel B is now occupied by the Western Treatment Plant. The maps are oriented for ease of display.


The second tool was on-ground interpretation. Where the field teams noted evidence of land-claim, such evidence was recorded where possible. The field teams attempted also to mark on aerial images the former extent of the saltmarsh that had been claimed. Although that was sometimes a simple task, the interpretation is nevertheless often confusing, particularly where no vestige of the former high-tide boundary remained or where obvious palaeo-coastal features from periods much earlier than 1750 (e.g. from the Holocene maximum: see Pirazzoli & Pluet 1991) could be mistaken for more recent shorelines.

Third, we used remote-interpretation tools. In places where no useful early maps could be located and sites were not visited in the field, aerial photographs were reviewed to check for clues about the prior vegetation. Evidence of channels or levies used to reclaim low-lying coastal plains, for example, were taken as evidence of former saltmarsh vegetation. We also used elevation data to check that our view of pre-1750 saltmarsh only included low-lying areas. As we did not have access to LIDAR elevation data at the time for most of Victoria, we used freely available elevation data from the NASA Endeavour ‘Shuttle Radar Topography Mission’ 2000 (see Farr et al. 2007). Remote-sensing tools were particularly important in the Corner Inlet and Nooramunga areas, where the historic record is not detailed.

We used the information from these diverse sources to compile a map of the likely pre-1750 distribution of coastal marsh across Victoria. The pre-1750 maps are presented in the same GIS format as the contemporary vegetation mapping shown in Chapter 5. Metadata are provided in Appendix H, and the files are held by the Department of Sustainability and Environment. Polygons were created as described in Chapter 5.2. In cases where the early cartography was good and vegetation patterns relatively simple, polygons were captured using transparencies traced from the historic plan placed over the screen. Where the original plans were distorted in relation to our maps, due to cartographic errors or different projections, the historic plans were captured digitally and geo-rectified to align them in space with our map data. This was done by digitally stretching the images against reference points, using the ‘ImageWarp’ software available as an extension to Arcview 3.2 (ImageWarp version 2.0, 1999). The rectified images were then used as templates in the same way as the aerial photographs. In cases where the information was insufficient to geo-rectify the image or where the level of accuracy in the original product was very poor, we interpreted the historic information ‘by eye’. In some cases the different pieces of evidence were at variance. Multiple historic plans, and the evidence visible on the aerial imagery do not always tell precisely the same story. In all such cases a subjective compromise was made.

The different methods for determining pre-1750 vegetation patterns vary in their ability to resolve different vegetation types, but the ability to distinguish types is generally very poor. For clarity, we chose to represent only two units: • Mangroves – This unit corresponds exactly to the EVC Mangrove Shrubland.• Other coastal marsh – This unit includes all the Saltmarsh Aggregate units, Estuarine Wetland, Saline

Aquatic Meadow (where within the marsh, as for the extant map layer), bare ground (where within the marsh, as for the extant map layer) and Seasonally Inundated Sub-saline Herbland.

To reflect the uncertainty around which vegetation types grew where, the general term ‘coastal marsh’ is sometimes used in the discussion below to refer to relevant units where their type is not precisely known.

Calculating accurate depletion statistics is complicated by the different resolution of the current and pre-1750 maps. For the accurate area statements presented in Chapter 5, we excluded areas of bare ground, brine pools and other minor inlying patches of non-saltmarsh vegetation. We cannot predict the distribution of these


inliers on a fine scale in the pre-1750 map. For this reason, we have compared the pre-1750 area of marsh to an increased (compared to Table 5.4) current saltmarsh total that includes these inlying areas (this inflation factor is, on average, only 1.1).

overview of depletion following european colonisation

Table 6.1 shows a numerical summing of estimated depletions in each sector of the coast (coastal sectors were defined in Chapter 5). In the absence of other information, we assumed that the relative proportions of Estuarine Wetland (EVC 10) and Coastal Saltmarsh Aggregate (EVC 9) remained constant during any process of historical depletion.

The Gippsland Lakes area presented significant problems when it came to calculating depletion statistics. As discussed in detail below, in this part of the coast there have been potentially large losses and gains across the years, many of which are difficult to identify using the available evidence. The primary difficulty is with existing areas of saltmarsh, some of which are natural occurrences, some of which are expansions of saltmarsh since colonisation. Given this uncertainty, we provide upper and lower bounds on our depletion estimates for the Gippsland Lakes area, based on two extreme scenarios:

• Scenario 1 – All of the ambiguous saltmarsh areas are natural. • Scenario 2 – All of the ambiguous saltmarsh areas are recent expansions (in which case they are counted

as gains, that offset other losses in any view of net statewide change).

Our pre-1750 map suggests that Victoria once supported approximately 34,000–39,000 ha (i.e. ~340–390 km2) of coastal marsh vegetation, about 80–95% of which remains (Scenarios 2 and 1, respectively).


Table6.1: Estimated depletion of Mangrove Shrubland, Coastal Saltmarsh (see text), and Estuarine Wetland since 1750, by coastal sector (see text for discussion of limitations). All figures are rounded to the nearest 5% or 5 ha. Two scenarios are given for two sectors associated with the Gippsland Lakes, as described in the text.

Sector Net loss of all coastal marsh (ha)

Mangroves% remaining

Saltmarsh% remaining

Estuarine Wetland % remaining

Overall %remaining

Glenelg < 5 - ~100 ~100 ~100

Fawthrop-Belfast 55 - 55 55 55

Western estuaries ~0 - ~100 ~100 ~100

Aire-Gellibrand < 5 - - ~100 ~100

Surf-coast estuaries 5 - 80 80 80

Breamlea 50 - 85 85 85

Connewarre-Barwon 375 ~100 85 85 85

Lonsdale Lakes 225 - 40 40 40

Salt Lagoon < 5 - ~100 ~100 ~100

Swan Bay 85 - 85 85 85

Mud Islands Assumed nil - ~100 - ~100

Port Phillip 1840 ~100 50 50 50

The Inlets 45 ~100 55 55 60

Western Port 300 95 85 85 90

French Island 30 ~100 95 95 ~100

Rhyll Inlet 15 ~100 90 90 95

Lang Lang coast 5 (gain) ~100 ~130 ~130 ~130

Bass River 65 ~100 70 70 75

Powlett-Kilcunda 130 - ~100 40 40

Anderson Inlet 660 > 100 40 40 45

Shallow Inlet 300 0 60 60 40

Wilsons Promontory Assumed nil 100 100 100 100

Corner Inlet 1070 80 35 35 55

Nooramunga Coast 645 ~100 80 80 83

Nooramunga Island Assumed nil 100 100 100 100

Jack Smith Lake 115 - 95 95 95

Lake Reeve 660 - 80 80 80

Lake Wellington (Scenario 1)

310 - 90 90 90

Lake Wellington (Scenario 2)

3000 (gain) - > 100 > 100 > 100

Lakes Victoria and King (Scenario 1)

130 - 95 95 95

Lakes Victoria and King (Scenario 2)

1830 (gain) - > 100 > 100 > 100

East Gippsland Inlets 185 - 85 85 85

Statewide (Scenario 1) ~7250 90 75 75 80

Statewide (Scenario 2) ~2045 90 95 95 95


historical changes along the victorian coastline

Table 6.1 demonstrates that the historical changes to coastal saltmarsh and mangroves have not been evenly spread across the Victorian coast: some sectors have experienced relatively little change since 1750 whereas others have been severely depleted. In this section of the report we discuss the changes in more detail, on a sector-by-sector basis, with reference to the strength of evidence used to estimate the change. Note that the discussion is concerned largely with wetland extent and not with change in ecological condition, as the latter could not be easily or consistently inferred by the methods used in the study. Note also that our statewide perspective inevitably leads to a glossing-over of some losses that were of comparatively minor extent; such a bias should not be seen to trivialise the smaller losses that have taken place since European colonisation.

Glenelg(Brid)

Reports from the field work suggested that relatively little change has occurred to the extent of intertidal marsh in this small sector of the coast. Some Estuarine Wetland (EVC 10; along with Estuarine Reedbed [EVC 952]), particularly between the Glenelg River and the South Australian Border, has apparently been converted to pasture (some areas are recorded as a change from Estuarine Reedbed to Estuarine Wetland, where Phragmites has been selectively removed by continual grazing).

Fawthrop-Belfast(FawthropVVP;BelfastWaPandVVPalthoughnotmappedassuch)

Field work and several early plans indicate that significant change has occurred to the extent of intertidal marsh in this area (Anon c.1843, 1843, 1850; Barrow undated; Myers 1843). Fawthrop Lagoon has suffered very considerable hydrological change through channelling, changes to its catchment, and the conversion of the outlet to an artifical channel. The freshwater-saltwater boundary is now sharply defined by a causeway. Despite these changes, the basic shape of the lagoon probably remains the same in the saltmarsh zone (Stanley 1869a). There is no evidence for a substantial loss of coastal saltmarsh here, but it is likely that the brackish zone was once larger (indicating a likely loss of Estuarine Wetland). Since the mouth of the Moyne River has been permanently open, Belfast Lough has been permanently exposed to the Southern Ocean. It now contains meadows of seagrass in the subtidal zone (Zostera muelleri, Ball & Blake 2009) and large expanses of Wet Saltmarsh Herbland and some Wet Saltmarsh Shrubland. This wetland, however, was shown in 1843 as being ‘fresh during the winter months’ (Myers 1843). Although that evidence is cursory, it suggests that the wetland has become saltier, which probably allowed saltmarsh to increase at the expense of Estuarine Wetland. (We have assumed half the area was once Estuarine Wetland for the purposes of area calculations.) As well, the wetland has been drained around its margins. The small saltmarsh that now occurs opposite Griffith Island occurs in an area that has been altered greatly by engineering works since European colonisation. The marsh was originally larger, and probably located further inland.

WesternEstuaries(WaP)

This sector is located in an agricultural landscape in which native vegetation has been almost entirely removed, including formerly vast areas of freshwater wetland. Extensive field work for this and a previous project (Sinclair & Sutter 2008), however, indicate that intertidal marsh was naturally restricted and its post-European loss has been relatively minor. The direct historic record is sparse (Barrow 1853; Urquart undated; Anon 1854; Watson 1855) and field work suggests saltmarsh was probably restricted to small patches in Rutledge’s Cutting and at the original mouth of the Merri River (near Thunder Point, Warrnambool), at


the mouth of the Fitzroy River, and the lower portions of Lake Yambuk. In all cases, only Wet Saltmarsh Herbland and some Coastal Tussock Saltmarsh is present. Most of these patches remain and have suffered only minor losses since European colonisation.

The loss of Estuarine Wetland is difficult to assess, particularly as sparse stands of Sea-rush Juncus kraussii may remain in grazed pasture after the removal of many other species that may be associated with a range of EVCs other than Estuarine Wetland (e.g. Estuarine Reedbed, Estuarine Scrub). Losses may have occurred in the hind-dune marshes on the Fitzroy River, and possibly also on the lower Merri River, although extensive drainage works between 1859 and 1870 and sand movement have changed the system radically (Anon 1854; Sinclair & Sutter 2008). All the losses in this sector of the coast are related to pasture creation, through drainage works and/or through continual grazing by stock.

Aire-Gellibrand(GellibrandWaP;AireOtP)

The Aire and Gellibrand systems drain steep hills which experience some of the highest rainfall in Victoria. The rivers discharge across large floodplains separated from the sea by natural barriers. Neither estuary currently supports saltmarsh or mangrove vegetation, and there is no evidence that either did in pre-European times. Both river systems, however, support extensive stands of Estuarine Wetland. Field inspection and a recent mapping study of the entire Aire floodplain (Osler et al. 2010) confirm that extensive drainage works have radically altered its hydrological character, and led to a massive loss of native vegetation. Estuarine Wetland seems to have suffered only small declines, although it is difficult to reconstruct its exact pre-1750 distribution (Osler et al. 2010).

Surf-coastestuaries(OtP)

This sector of the coast supports only small patches of coastal marsh vegetation. Nonetheless, it has suffered some significant changes since European colonisation. The Erskine River at Lorne retains much of its Estuarine Wetland. Painkalac Creek at Aireys Inlet also remains largely intact and supports mostly Wet Saltmarsh Herbland and Estuarine Wetland, but some of marsh on the eastern side of the inlet has been filled. On the Anglesea River, a section of intertidal marsh (Wet Saltmarsh Herbland and Coastal Tussock Saltmarsh) and Estuarine Wetland on the western side of the inlet was filled and converted to an urban park and roadway (Osler et al. 2010). Interestingly, the Anglesea River has recorded an increase in Estuarine Wetland, which has invaded areas of former Estuarine Scrub which were burnt out and bulldozed in the 1983 fires (Osler et al. 2010). The tiny areas of Wet Saltmarsh Herbland, Coastal Tussock Saltmarsh and Estuarine Wetland on the Spring Creek at Torquay have been fragmented and reduced in area by incremental works on the shoreline, plantings and weed invasion.

Breamlea(OtPandVVP)

Early plans (Smythe 1847; Daintree et al. 1863) show that the area of coastal saltmarsh across most of this sector has changed little since the early days of pastoral settlement. That has occurred despite a history of varied and intensive land use. The eastern end of the marsh, however, abuts a wastewater treatment plant, which has destroyed a portion of the marsh. The other large loss of coastal marsh occurred as a result of the landfill located at the south-western end of the sector. A lagoon which was once located between Salt Swamp (Connewarre-Barwon sector) and Breamlea has also been destroyed. Other incremental losses have occurred as a result of drainage for pasture, particularly in the north and centre of the sector.


Connewarre-Barwon(OtPandVVPalthoughnotmappedassuch)

The lower Barwon River system remains a formidable and complex area of saltmarsh and brackish-water marshes (containing all the EVCs relevant to this report in the one sector), despite its proximity to large cities such as Melbourne and Geelong and the long history of European land use. The historical record is detailed (Smythe 1847; Byerley 1855; Daintree & Ross 1862a) and shows that significant losses of marsh have occurred as a result of drainage for pasture in the south-western corner, at the southern boundary of Salt Swamp, along the course of the Barwon River (notably the southern side), and on the eastern side of Lake Connewarre. Comparatively minor losses have occurred elsewhere through dumping of material, development and weed invasion. Interestingly, it is possible that some small (< 10 ha) gains in saltmarsh extent have also occurred since colonisation, with areas on Paceys Island which were once wooded with Estuarine Scrub (probably Melaleuca lanceolata) having been converted to ‘saltmarsh’ through loss of their overstorey (Byerley 1855).

LonsdaleLakes(OtP)

This section of the coast has suffered significant losses of a diverse range of saltmarsh types (probably including Wet Saltmarsh Herbland, Wet Saltmarsh Shrubland, Coastal Hypersaline Saltmarsh, Coastal Dry Saltmarsh or Coastal Tussock Saltmarsh) (Hoddle 1843, Byerley 1855; Daintree & Shepherd 1863a). The losses result from the accumulation of many local incremental changes involving drainage, conversion to pasture, and the removal of natural hydrological regimes through the construction of roadways. The Lonsdale Lakes were once more-or-less joined to Swan Bay by an expanse of marsh vegetation. Only a small section of this on the margin of Swan Bay is recorded as ‘samphire’ by Hoddle (1843) (here corresponding to Wet Saltmarsh Shrubland and Wet Saltmarsh Herbland). The remainder was presumably made up of other units such as Coastal Tussock Saltmarsh. The loss of this area, apparently partly filled in to make way for suburban Point Lonsdale, is recorded in this sector, not Swan Bay.

SaltLagoon(OtP)

Despite being surrounded by semi-urban development, agricultural land and roads, Salt Lagoon has largely retained its former shape and area (Daintree & Shepherd 1863b). All losses are of very minor extent.

SwanBay(OtP)

Sand movement has altered the entrance to Swan Bay markedly since European colonisation, partly as a result of natural processes, and partly due to coastal engineering works. It is difficult to tell from the primary historical resource (Hoddle 1843; Ross 1859; Cox & Stanley 1874; numerous maps of Port Phillip Bay) how much change in marsh extent these developments have caused; we have assumed no change in extent at the entrance. Saltmarsh has certainly been lost at the south-western end of the Bay because of the construction of roads and railways. (As noted above, the large area of marsh lost between Swan Bay and the existing marsh at Lonsdale Lakes has been counted under the Londsdale Lakes sector.) Some small losses have occurred to the north and north-west as a result of drainage for pasture, but the saltmarsh area was evidently always narrow in this region, as the hinterland rises quickly. The Edwards Point area retains essentially all its pre-European area of marsh.


MudIslands(nobioregionassigned)

Due to their isolation in Port Phillip Bay, Mud Islands have been largely free from the development pressures that have led to the destruction of so many other coastal marshes in Victoria. But this is not to say this area has been static: historical records collated by Yugovic (1998) demonstrate that natural geomorphic processes have changed the shape of the islands and the marsh greatly since the early 19th century. Even so, we have assumed that there has been no net loss of saltmarsh vegetation since pre-1750.

PortPhillip(VVPandOtP)

This sector has suffered massive losses of intertidal marsh, brought about by intensive land use, particularly ponds for water treatment and salt production. Fortunately the historic record is extensive and detailed (Russell 1837; Anon 1841; Garrard & Shaw 1850; Christie 1853; Ross 1859; Daintree & Ross 1862b; Cox 1864; Noone 1880a, b; Goldsmith & McGauran 1896; Proeschel undated). The marshes which originally occurred in inner Melbourne and the port area have been mostly swallowed by the growing city. The marshes of Altona retain much of their area, but have been corralled in a largely suburban landscape. The Cheetham wetland at the mouth of Skeleton Creek has been almost entirely converted to evaporation ponds, and the upper marsh has been replaced by the Sanctuary Lakes housing development. The once vast string of marshes which stretched between the Werribee River and Point Wilson has been swallowed by the Western Treatment Plant. Similarly, the large marshes between Point Wilson and Limeburners Bay, along with the marshes of Point Henry, have been almost fully converted to salt ponds. Only the RAAF Lake and The Spit (Point Wilson) remain as large, relatively intact marshes in this part of the Victorian coast.

TheInlets(GipP)

The Inlets today present an unusual superimposition of the past and present. The once vast Koo Wee Rup swamp now discharges via large drains. However, between the Inlets the imprint of the original tidal channels and portions of marsh and other vegetation remains intact. The largest losses have been west of Lyall Inlet, while most of the marsh east of Lyall Inlet apparently remains intact, (Smythe 1842a; Yugovic & Mitchell 2006). This is not to say, however, that vast areas of other types of wetland (e.g. Swamp Scrub) have not been lost following the drainage of Koo-Wee-Rup Swamp (see East 1935 for a historical overview).

WesternPortcoast(GipP)

The Western Port coast retains a massive area of intact saltmarsh, but it has also suffered historical losses as a result of land-claims. The primary historic resource is good for Western Port, as it consists of several early survey plans and aerial photography taken prior to some land building (Smythe 1842a, 1842b, 1842c, 1842d; Bailliere 1866; Air Photo Westernport [868] A2 Central Plans Office neg. No. 20898 [1958]). The construction of the Hastings foreshore and marina resulted in the destruction of approximately 33 ha of saltmarsh, and the industrial developments at nearby Long Point resulted in the destruction of a further 23 ha. HMAS Cerberus near Sandy Point also required minor land-claims into former saltmarsh. Apart from these major developments, land-claims for pasture have also destroyed large expanses of marsh, particularly from Watson’s Inlet around to Yallock Creek; and near Stockyard Point. That destruction of saltmarsh and mangroves has not ceased, as evidenced by the example shown in Figure 6.2. In contrast, the marshes on the Phillip Island and the Gurdies-Grantville coast remain largely intact, despite being bordered by private


land, and have suffered only slight losses through drainage and grazing. That retention of coastal wetland is presumably a function of the slightly steeper terrain at the inland border of the marshes.

Figure6.2. Recent saltmarsh destruction in Western Port resulting from a land-claim. Note the new fenceline surrounding an area of recently imported soil deposited on top of the marsh. In the background, the original upper-limit of saltmarsh is visible, although heavily grazed.

FrenchIsland(GipP)

The vast stands of pre-European saltmarsh and mangroves on French Island remain almost entirely intact (Smythe 1842b; Bailliere 1866). Some losses have occurred as a result of clearing of Mangrove Shrubland in some places, and some Estuarine Wetland has been converted to pasture in the upper Redbill Creek; extensive saltpans have modified but not removed much of the saltmarsh on the north-eastern coast.

RhyllInlet(GipP)

Although it retains large areas of marsh, Rhyll Inlet has suffered losses since European colonisation. The rubbish tip on Cowes-Rhyll Road has obliterated 5 ha of intertidal marsh (Wet Saltmarsh Shrubland and Wet Saltmarsh herbland). Extensive drainage works on the eastern end near Cowes have destroyed several more coastal wetlands. A detailed early map (Smythe 1842b) suggests that the spit sheltering Rhyll Inlet has lengthened since colonisation, and there may have been some modest expansion of saltmarsh at the eastern end of the inlet.

LangLangCoast(GipP)

The Lang Lang coast is unusual among all the defined sectors of coast in that it shows an increase in the extent of coastal saltmarsh since European colonisation (see also Lake Wellington and Lakes Victoria and King). At settlement, this section of the Victorian coast contained virtually no saltmarsh or mangroves, and was fringed instead with a dense stand of Swamp Paperbark Melaleuca ericifolia where the Tobin Yallock Swamp directly met the sea (Smythe 1842c; Yugovic & Mitchell 2006). Fresh water spilled from this swamp,


recorded by Smythe (1842c) as ‘numerous rills of freshwater continually running’. The coast at this time was probably cliffed and eroding (Yugovic & Mitchell 2006). Drainage, however, has left the former landscape unrecognisable. A series of bund walls now line the coast, and fresh water is channelled to the sea. A band of saltmarsh has formed on and in front of the bund walls, above the intertidal zone, and presumably receives salty water from ocean spray. This marsh is mostly species-poor Wet Saltmarsh Herbland. Small areas of remnant marsh occur only at the abandoned mouth of the old Yallock Creek.

BassRiver(GipP)

Pasture-creation has destroyed a large proportion of the marsh on the Bass River estuary, with almost the entire margin being bounded by walls, drains or fill. *Spartina grows in extensive mats in channels and on mudflats throughout the marsh. Interestingly, an early plan of Western Port (Smythe 1842c) does not show any mangroves at the mouth of the Bass River, despite showing them elsewhere, including nearby at Settlement Point. Whether the extensive stands which now exist are a recent invasion is doubtful, and it is conceivable that the survey plan was incomplete.

Powlett-Kilcunda(GipP)

The small portion of marsh at Kilcunda, supporting Wet Saltmarsh Herbland, Coastal Tussock Saltmarsh and Estuarine Wetland, retains most of its former extent. The Powlett River also retains most of its saltmarsh (again Wet Saltmarsh Herbland and Coastal Tussock Saltmarsh), but very large areas of Estuarine Wetland have been lost to pasture on the large plain to the north east of the mouth of the river.

AndersonInlet(GipP)

Anderson Inlet has experienced some of the largest losses of coastal marsh anywhere in Victoria, since the first surveys were undertaken in the 19th century (Smythe c.1847; Stanley 1869a,b; Anon undated a). Much of the saltmarsh and mangrove areas have been destroyed following the extensive drainage works for conversion to pasture. Losses have been very substantial at the eastern end of the inlet, and around the mouth of the Tarwin River. Moving sands also have had a minor impact on saltmarsh extent near Point Smythe. The spread of *Spartina in the Inlet ensures that changes continue (Chapter 5.5).

ShallowInlet(GipP)

This section of the coast has undergone substantive changes since European colonisation, representing some of the largest losses in Victoria. Land-building works have destroyed vast areas of saltmarsh, particularly on the southern side of the inlet. Bird (1993, Figure 167) shows the loss of wetland that has occurred since European colonisation (see the reproduction of his map in Chapter 1). Interestingly, both the early plans consulted for this part of the coast show extensive areas of both ‘mangroves’ and ‘samphire’ (i.e. saltmarsh) (Smythe 1848, Anon undated a; see Figure 6.1 panel A). There are currently no mangroves in Shallow Inlet, and unless we assume both early plans were in error, it would seem that the prior stands (of perhaps 250 ha in total) have been entirely lost.

WilsonsPromontorycoast(WPro)

This small sector is entirely within a National Park and is bordered by intact native vegetation. We have assumed that coastal saltmarsh or mangroves have suffered no measurable decline in extent.


CornerInletcoast(GipP)

Some of Victoria’s largest losses of saltmarsh have occurred in this sector. Most of the existing saltmarsh is bordered by channels and walls/levees, and the upper marsh has been drained and converted to pasture. Unfortunately the historic record seems to lack detail at a fine scale (Smythe 1848, undated a; Wright 1849; Anon undated b), however field observations, aerial photographs and elevation data come together to provide strong and consistent evidence of former marsh extent. We have attempted to mark in a pre-1750 boundary between mangroves and saltmarsh (Wet Saltmarsh Shrubland in this area), using intact areas to guide us in the relative widths of these bands.

NooramungaCoast(GipP)

Although the historic record is detailed around Port Albert (Anon 1842; Smythe 1843; Anon undated c), it is poorly resolved elsewhere in the sector. Like Corner Inlet, however, much of the marsh is now bounded inland by sea walls and channels, which provide direct evidence of the extent to which the landscape has changed.

NooramungaIslands(GipP)

The Noormunga Islands remain almost entirely uncleared and undeveloped and, as such, probably represent the least disturbed large sector of the Victorian coast. Although we are aware that some geomorphic changes have taken place in the area since settlement, the historic record does not allow an accurate reconstruction to be made of past patterns (Smythe 1848; Wright 1849; Smythe undated b; Anon undated b); thus we have assumed that the extent of marshes in this area has remain unchanged since European colonisation.

JackSmithLake(GipP)

The historic record for Jack Smith Lake is poor (Smythe undated d), and field inspection and aerial photograph interpretation suggest that much of the marsh area remains, but with small areas lost to land-claim and grazing.

LakeReeve(GipP)

Lake Reeve itself retains most of its former area of marsh, with the exception of one large area claimed for pasture at the western end. The historic record is detailed and it shows the vegetation has otherwise changed very little in its extent since the early 19th century (Smythe undated c, undated d, undated e; Anon undated d). Perhaps the biggest losses have been outside Lake Reeve itself, but included in this sector: Smythe (undated c) shows extensive patches of ‘saltmarsh’ between Lake Reeve and Lake Wellington (> 750 ha in total; distinguishable from areas labelled ‘swamp’ and ‘tea-tree’). Such areas were clearly not intertidal, but presumably interacted with saline groundwater. Similar patches of saltmarsh lacking tidal connection exist elsewhere (e.g. Lonsdale Lakes). The examples near Lake Reeve have largely been converted to pasture or, in one case, a large dam.

LakesVictoriaandKing(GipP)

The ever-shifting natural entrance to the Gippsland Lakes from Bass Strait was replaced by a permanent opening as a result of engineering works between 1870 and 1889 (Bird 1965; Bird & Lennon 1989). Bird (1965, 1966) marshalled evidence to indicate that the creation of the permanent opening to the Southern


Ocean had caused an increase in salinity in the Gippsland Lakes system, leading to an increase in saltmarsh at the expense of freshwater wetlands. He suggested that Lakes Victoria and King were mostly brackish in pre-European times. His conclusions are broadly supported by our field work, which showed strong evidence of dieback of the fringing freshwater and brackish-water wetland systems (e.g see Figure 6.3 below). Nonetheless, the historic record shows clearly that saltmarsh was present and extensive around Lakes Victoria and King in the early 19th century. Saltmarsh or salt lakes apparently occurred naturally right around the lakes: existing maps show them in Victoria Lagoon and the surrounding depressions, in Jones Bay (Wilkinson 1849), around Loch Sport, Beacon Swamp, Blond Bay, Point Wilson, near Paynesville, on the Mitchell Silt Jetties (Smythe undated c, undated d; Wilkinson 1849), on Raymond Island (Anon undated e) and on the Boole Poole Peninsula.

It would appear that saltmarsh has expanded in some places around the Gippsland Lakes, remained static in others, and perhaps decreased in others. Direct losses of coastal saltmarsh, caused by local land-claims, have in comparison been minor. Given the uncertainty as to whether some large expanses of marsh are remnant or adventive, we have presented two extreme scenarios for this sector (as described above).

LakeWellington(GipP)

Figure6.3: Loss of former Common Reed Phragmites australis and Swamp Paperbark Melaleuca ericifolia vegetation, Lake Coleman, southern shore of Lake Wellington. The inset shows the intensity of the secondary salinisation; the white soil surface is salt crystals.


Like Lakes Victoria and King, Lake Wellington has been influenced by the opening of the artificial entrance to the ocean at Lakes Entrance. Bird (1965, 1966) suggested that Lake Wellington was once essentially a freshwater lake, and that the saltmarsh on its margins has invaded recently, replacing the prior freshwater systems. The historical record partially supports this view. Myers (c.1840) shows Lake Wellington as ‘fresh’ (Lake Reeve is shown ‘Salt’; Lakes Victoria and King are unannotated). In contrast, other early sources (Smythe undated c; Anon undated e) show very extensive areas of saltmarsh interlaced with ‘tea-tree’ and ‘swamp’ on the southern margins of Lake Wellington (and Lake Coleman), and in many depressions inland between Lake Wellington and Lake Reeve (see discussion under the Lake Reeve sector, above). In his more-detailed 1965 study used as a basis for the 1966 summary, Bird acknowledges the natural (more inland) occurrence of these saltmarsh areas.

Again, it would seem that saltmarsh has a complex history in the area. There is, however, no evidence for an extensive pre-1750 occurrence of saltmarsh on the western and northern shores of Lake Wellington, which suggests that it has expanded its overall area on this part of the coast. As for Lakes Victoria and King, we have provided two scenarios for Lake Wellington, to account for several large patches of uncertain history. Figure 6.3 shows an example of the effects of secondary salinisation along the southern shores of Lake Wellington, where prior Common Reed Phragmites australis and Swamp Paperbark Melaleuca ericifolia vegetation is in the process of being replaced by saltmarsh vegetation.

EastGippslandInlets(EGL)

Although little direct evidence of saltmarsh loss could be found for this sector of the coast in the historical record, an inspection of aerial photographs reveals land-claims on the lower Snowy River, which have lead to large, local losses of intertidal marsh (Wet Saltmarsh Herbland) and Estuarine Wetland (Lees 1855; Anon undated d, undated f ).


assigning a bioregional conservation status to each evc

As outlined in the introduction to this chapter, one of the primary reasons for creating pre-1750 maps is to calculate depletion levels for various vegetation types across the state. Such information is important for assigning Bioregional Conservation Status to different vegetation types, which in turn has a direct impact on planning decisions. Table 6.2 shows the criteria used to assign Bioregional Conservation Status in Victoria.

Table6.2: Criteria for assigning Bioregional Conservation Status to vegetation types (from www.dse.vic.gov.au).

Status Criteria

Presumed Extinct

Probably no longer present in the bioregion.

Endangered Contracted to less than 10% of former range; OR

Less than 10% pre-European extent remains; OR

Combination of depletion, degradation, current threats and rarity is comparable overall to the above:

• 10 to 30% pre-European extent remains and severely degraded over a majority of this area; or

• naturally restricted EVC reduced to 30% or less of former range and moderately degraded over a majority of this area; or

• rare EVC cleared and/or moderately degraded over a majority of former area.

Vulnerable 10 to 30% pre-European extent remains; OR

Combination of depletion, degradation, current threats and rarity is comparable overall to the above:

• greater than 30% and up to 50% pre-European extent remains and moderately degraded over a majority of this area; or

• greater than 50% pre-European extent remains and severely degraded over a majority of this area; or

• naturally restricted EVC where greater than 30% pre-European extent remains and moderately degraded over a majority of this area; or

• rare EVC cleared and/or moderately degraded over a minority of former area.

Depleted Greater than 30% and up to 50% pre-European extent remains; OR

Combination of depletion, degradation and current threats is comparable overall to the above and greater than 50% pre-European extent remains and moderately degraded over a majority of this area.

Rare Rare EVC (as defined by geographic occurrence) but neither depleted, degraded nor currently threatened to an extent that would qualify as Endangered, Vulnerable or Depleted.

Least Concern

Greater than 50% pre-European extent remains and subject to little to no degradation over a majority of this area.

Our map datasets provide a partial basis for assigning depletion levels and Bioregional Conservation Status to each of the units mapped in the study. Although they provide an accurate representation of the current extent of each unit, they do not allow a comprehensive assessment of how extensive each unit was pre-1750, because the historical record is not detailed enough. Given that it cannot indicate changes in ecological condition, nor does our dataset allow us to judge the level of degradation for each EVC. Accordingly, the assignment of Bioregional Conservation Status to saltmarsh EVCs is based on a combination of data and expert knowledge. Table 6.3 shows the recommended Bioregional Conservation Status for each unit mapped in the study. It shows also the currently assigned Bioregional Conservation Status, where applicable. We do not recommend an updated Bioregional Conservation Status to EVC 538 Brackish Herbland, because our study did not examine the full range of these EVC in Victoria (see Chapter 5).


Table6.3: Recommended Bioregional Conservation Status (BCS) ratings for each Ecological Vegetation Class (EVC) described and mapped in this report. The bioregion abbreviations are those used by the Department of Sustainability and Environment (DSE). BCS ratings shown in round brackets are those previously assigned by DSE, where applicable. Abbreviated BCS ratings are as per Table 6.2.

EVC Bioregion

Brid WaP OtP VVP GipP WPro EGL

Wet Saltmarsh Herbland R R R E D R R

Wet Saltmarsh Shrubland V D E D R R

Coastal Tussock Saltmarsh V E R E V R R

Coastal Saline Grassland E R E R R

Saltmarsh-grass Swamp E X?

Coastal Dry Saltmarsh V E E R R

Coastal Hypersaline Saltmarsh V E V

Coastal Saltmarsh (Aggregate) V(V) V V(E) E(V) V(LC) R(LC) D(V)

EVC 947 Brackish Lignum Shrubland

V X?

EVC 10 Estuarine Wetland E(E) E(D) R(E) E(E) D(LC) R(R) V(V)

EVC 140 Mangrove Shrubland R(V) V(V) LC(LC) R(R)

EVC 196 Seasonally inundated Sub-saline Herbland

D(R) E R

6.2 Thefutureforsaltmarshunderrisingsealevel–WesternPortcasestudy

introduction

Chapter 1.13 provided a detailed overview of the likely impacts of climate change on Victorian mangroves and coastal saltmarsh. In brief, both vegetation types typically occupy a restricted elevation range largely controlled by the tides, generally extending from or around the local mean sea level to the local extent of the highest spring tides. Given this intertidal position, it is highly likely that substantial impacts to the distribution, extent and condition of saltmarsh and mangrove will occur if there are significant changes to mean sea level in the future. When sea levels rise and/or tidal amplitude is increased, intertidal plant communities may persist by migrating inland, or where there is sufficient sediment supply they may ‘adapt’ in situ where vertical accretion of the marsh surface keeps pace with sea-level rise (Reed 1990, 1995; Callaway et al. 1996; Day et al. 1999). Where such adaptive responses are impossible, saltmarshes and mangroves may be lost to deeper or more permanent inundation by the sea. Therefore, the rate of sea-level rise and the shape of the coastline and its hinterland (including human structures and developments) will interact to determine the fate of intertidal vegetation at any one location.

The purpose of this section of the investigation is to explore some of the local consequences for intertidal vascular plant communities that may result from likely sea-level rise along the Victorian coast. The mainland coast of Western Port was chosen as the case-study site, chiefly because of the current availability for that part of the Victorian coast of high-resolution terrestrial elevation data obtained using LIDAR.


The specific objective was to use existing bio-physical models of terrain and elevation to explore potential impacts on intertidal marshes assuming a sea-level rise of 0.80 m. That value was chosen in order to be consistent with regional coastal planning in Victoria, which assumes a 0.80 m rise over the coming century (Victorian Coastal Council 2008). Predictions vary as to the magnitude of future sea-level rise, and many authorities have modelled substantially higher rises by the end of the 21st century (see Chapter 1.13 for a detailed review). According to projections from the Intergovernmental Panel on Climate Change, the rate of sea-level rise will likely increase over the coming century as a consequence of human-induced global warming. Values for predicted sea-level rise over the course of the next century typically range from 90 mm to 0.88 m, based solely on the thermal expansion of the oceans. Contributions to sea-level rise from the melting of ice and snow captured in glaciers and polar-ice sheets are more difficult to predict and incorporate into models.

It is generally accepted that the global sea level has been rising over the last century at around 1.7 ± 0.3 mm year–1 (IPCC 2007; see Chapter 1.13 for information specific to the Australian coast). Recently an acceleration in sea-level rise of around 0.013 ± 0.006 mm year–1 has been detected over the period 1870 to 2004 (Church & White 2006; Wöppelmann 2009). Rises in the relative mean sea level around the Australian coastline for the period 1920 to 2000 has been estimated to be around 1.2 mm year–1 (Church et al. 2006).

the dynamics of sea-level rise

When the IPCC and other authorities speak of sea-level rise associated with human-induced climate change, they are usually referring to elevated world-wide sea levels, often referred to as the eustatic sea level. As outlined in Chapter 1.13, eustatic sea-level changes may not correspond closely to changes experienced at specific locations. The latter are greatly influenced by local factors and processes, and are often very difficult to predict. Some of the pertinent factors are listed below; it must be stressed that they are dynamic and profoundly interactive:• Hydraulic capacity and arrangement of tidal channels and embayment entrances• Climate and frequency of extreme weather events• Shape, arrangement, nature and movement of tidal barriers (such as sandbars)• Local ocean currents• Local geology• Anticipated regime of extreme or storm amplified tidal events• Local land subsidence or uplift• Vegetation composition, structure and productivity• Accretionary rates (rates of sedimentation)• Inflows from streams and estuaries• Rate or regime of eustatic sea-level rise. (Both coastal erosion and marsh accretion may play out very

differently under slow or rapid sea-level rise scenarios).

Further to these considerations, the response of humans in coastal regions to climate change will invariably involve changes to land use within local catchments and the widespread construction of groins and sea walls, along with ongoing channel dredging and/or widening. Future change in these variables cannot be accurately anticipated even for a given section of coast.


Given these factors, it is not surprising that published studies that have addressed likely changes in the area of intertidal wetlands under sea-level rise have been limited to the analysis of a limited number of variables, and are generally unable to make predictions that take account of all the important variables and their temporal interactions. Some studies have chosen to look at marsh dynamics in detail but without any spatial aspect (e.g. Morris et al. 2002); others have focussed on aspects of primary production and sea level without considering bathymetry or geomorphology in detail (e.g. Callaway et al. 1996; Day et al. 1999); and finally some have been restricted to conceptual models (e.g. Semeniuk 1994; Michener et al. 1997). Our study is similarly limited in scope, as described below.

case-study site – western port

Western Port presents a study area where many of the factors noted above apply and interact, which makes it an interesting – and by no means simple – case-study area. It is a large, land-locked marine embayment with estuarine features (Figure 6.4: see also Bird 1976). Two islands (French Island, Phillip Island) dominate much of the bay and protect the internal coastlines from wind and wave energy. The modern bay was formed some 10,000 years ago when low-lying areas were inundated by rising sea levels (Marsden et al. 1979). The bay is generally very shallow, and supports extensive areas of intertidal mudflats and seagrass beds. For more detailed accounts of the environmental and geomorphological setting of Western Port, see Jenkin (1974), Bird & Barson (1975), Marsden & Mallett (1975), Ministry for Conservation (1975), Shapiro (1975a,b), Spencer-Jones et al. (1975), Marsden (1979), Marsden et al. (1979), Sternberg & Marsden (1979), and May & Stephens (1996).

Tidal waters from Bass Strait enter the bay via two systems, one to the south and east of French Island and the other to the west and north of French Island (Miles 1974, 1976). The flood and ebb tides coalesce and separate respectively along a tidal divide in the vicinity of Lang Lang (Figure 6.5). The tide is enhanced as it passes into the bay, such that the tidal range increases from about 2 m at the entrances to more than 3 m to the north of French Island (Bird 2008).

Western Port supports extensive areas of mangrove and saltmarsh vegetation (see Chapter 5.1). Mangroves and saltmarsh occupy a narrow elevation range, generally extending from around 0.3 above mean sea level (MSL) to 1.8 m above MSL, a range of around 1.5 m. The bulk of this range from around 0.3 above MSL to 1.2 m above MSL some 0.9 m is occupied by mangroves. Figure 6.6 shows the current distribution of mangroves and saltmarsh in Western Port, as mapped in this report.

As outlined in Chapter 1.11, substantial areas of mangrove around Western Port were harvested in the early 19th century to produce barilla for the production of glass and soap. Considering this historical loss and the subsequent loss of mangrove and associated vegetation to intensive coastal development (see Chapter 6.1 above), the distribution of mangroves today is surprisingly similar to that before European colonisation. Rogers (2004) documents the recent expansion in the extent of mangrove at Rhyll Inlet and Quail Island. That expansion has largely been on the landward side of the mangrove stands and has resulted in a corresponding decline in the area of saltmarsh. The encroachment of mangroves into saltmarsh has been observed widely in south-eastern Australia by a number of authors (see Saintilan & Williams 2000). Rogers (2004) also examined sedimentation rates at Rhyll Inlet and Quail Island and estimates vertical accretion rates of around 2.5 mm + 0.2 mm per year at Rhyll and 1.4 mm + 0.2 mm per year at Quail Island. These rates of


vertical accretion were found to be consistent over the last ~100 years. Given these data, there is reasonable evidence that sedimentary processes in Western Port have kept pace with, or have exceeded, the background rate of sea-level rise over at least the past century.

Figure6.4: The case-study area: Western Port and its immediate hinterland. Green areas denote intertidal regions peripheral to mangrove and saltmarsh. These areas typically support mudflats and seagrass beds.

Figure6.5: The intricate networks of tidal drainage channels in the north east of Western Port revealed in an infrared image. In the foreground is the Lang Lang coast, which is characterised by a steep 0.5–2 m earth cliff. This section of coast in the northern part of Westernport Bay is notable as it has historically never supported mangroves. Source: Marsden et al. (1979).

approach and methods

The objective of this component of the project was to model the extent of intertidal vegetation following a rise of 0.80 m in eustatic sea levels. It is apparent that to build such a model for Western Port requires the incorporation of locally relevant data on climate, marine and terrestrial hydrological processes, marine and terrestrial geomorphological processes and feedbacks. These data are currently not available. Notwithstanding the intellectual merits of attempting to construct such a model, given the substantial degree of uncertainties in generalised global models of climate change, it is doubtful that any usefully predictive local model could be created in the near future. Moreover, the parameters determining the distribution of intertidal vegetation types are numerous and interactive and cannot be predicted into the future with useful certainty. As argued in Chapter 1.13, there are considerable uncertainties at every level of attempting to predict climate change impacts on vegetation and ecological processes.

In the absence of a holistic dynamic model, we decided to explore only two variables important in determining the distribution of saltmarsh and mangrove: i) tidal regime; and ii) wave energy. The idea was to define the tidal and wave energy constraints on intertidal vegetation from their currently mapped extent within the Bay, then attempt to define the environmental space in a future with a 0.80 m rise in sea level while maintaining constant all the other relevant variables. The approach was selected on the basis of the data we have access


to (e.g. LIDAR, accurate mapping) and that which we did not have access to (e.g. local rates of primary production for dominant plant species, responses to increased CO2 concentrations, responses to changes in frost frequency etc.).

Kilometres

Figure6.6: Current extent of mangrove (shown in green) and saltmarsh (shown in red-brown on the landward side).

Topographyoftheseafloorandlandsurface

A Digital Terrain Model for Western Port was constructed from several spatial datasets. The Department of Sustainability and Environment has recently completed a LIDAR survey of the entire Victorian coastline and much of the hinterland (see Figure 6.7, showing the Port Phillip and Western Port section of the survey). At the time of this study (November 2008), only the Western Port section of the Coastal DEM project had been processed to obtain a bare earth surface dataset. The LIDAR data were supplied in 190 tiles at 1 m pixel resolution in ASCII grid format. These were combined and resampled to 5 m pixel resolution using the Bilinear Interpolation routine within the ArcGIS ArcView 8.1 GIS. Crude sea-floor data were created from 1:50,000 bathymetric contour data obtained from the geospatial data library of the Department of Sustainability and Environment, using the seaward edge of the LIDAR surface as an additional contour. The contours were then transformed to a gridded and interpolated surface for the sea floor of Western Port using the ‘Topo to Raster’ routine within the ArcGIS ArcView 8.1 GIS (see Hutchinson & Dowling 1991). It is an interpolation method specifically designed for the creation of hydrologically correct digital elevation models. These two datasets – the Western Port LIDAR derived Coastal DEM, and Western Port bathymetry – were merged with the Department’s existing 20 m resolution digital elevation model for Victoria, clipped to the Western Port region to form a continuous terrestrial and bathymetric terrain model (see Figures 6.8 and 6.9).


Figure6.7: Coastal digital elevation model shown in the Port Phillip and Western Port region. (Note the absence of LIDAR-generated elevation data for French and Phillip Islands.)

Vegetationtypemodelling

The map dataset generated for this report was used to delineate saltmarsh and mangrove vegetation in Western Port. For clarity, only the aggregate saltmarsh unit was used for the analysis, along with Mangrove Shrubland. Within these two broad zones, bay-wide statistics from the corresponding regions of the continuous terrestrial and bathymetric terrain model were extracted. The mean and standard deviations for each zone were first used to predict the distribution of intertidal vegetation within the bay and then used to naively predict (assuming fixed bathymetry, shoreline and hinterland) into a future with sea levels elevated beyond the current mean sea level by 0.80 m. Areas of vegetation in man-made environments (e.g. salinised pasture behind sea walls, etc.) were excluded from the base dataset.


Wave-exposuremodelling

Wave exposure was modelled using WEMo (Wave Exposure Model) 3.1, which is a freeware extension to ArcGIS 8.1. WEMo is a simple hydrodynamic model that calculates the wind wave exposure of a site (Malhotra & Fonseca 2007, 2008). It can implement two methods for calculating wind wave exposure: i) Representative Wave Energy (RWE); and ii) Relative Wave Exposure Index (REI). The REI alone was calculated as the data required to parameterise a RWE analysis were not available to the project. The RWE analysis uses an empirical approach based on inverse-distance-weighting function of bottom depth applied to wave rays (Malhotra & Fonseca 2007). In contrast, the REI mode computes an internally consistent relative index value calculated for a site and provides a context for evaluating how exposed a site is to wind-generated waves in comparison to any other site. As a higher REI value at a given site implies higher wave energy, REI values were used for comparisons among different sites under seemingly similar conditions.

To compute the REI at a given location, the WEMo program must be supplied with bathymetry, a shoreline (i.e. mean sea level) and wind data for each of the major eight compass headings. Wind speed and frequency of wind from each major compass direction are combined with fetch to calculate a relative exposure index. Fetch is defined as the distance from the site to land along a given compass heading. Wind speed and frequency were obtained from Bureau of Meteorology data for Rhyll (see Figure 6.10).

Figure6.8: Continuous terrestrial and bathymetric terrain model for the Western Port embayment floor to 10 m above mean sea level.

Figure6.9: Digital elevation model from the sea floor to 30 m above mean sea level. Lighter coloured areas are more elevated. The high degree of precision and accuracy can be seen in the LIDAR data on the mainland coast in contrast with the less accurate and far less precise elevation data for the islands and the sea floor.


Figure6.10: Wind roses for Rhyll. These plots show wind direction and speed at 9:00 am (left) and 3:00 pm (right), averaged for the period 1 December 1991 to 31 December 2006. Source: Bureau of Meteorology.

Eight hundred points within Western Port were selected at random from a 100 m x 100 m grid. The REI was calculated for these points using the 2009 shoreline and the putative ‘2090 shoreline’ supplemented by 200 additional points reflecting the expanded size of Western Port when the eustatic sea-level rise prediction of 0.80 m was applied naively. After the REI for all points was calculated, the sites were imported into the GIS (ArcGIS ArcView 8.1) and a continuous surface was interpolated and applied across open water using the Inverse Distance Weighting (IDW) function.

results and discussion of case study

Before further discussion it is worth making the salient point that the results of this investigation, while interesting, are not useful for coastal planning in either a nature-conservation or coastal-development context. Notwithstanding the uncertainties around the rate and amplitude of anticipated sea-level rise due to global warming, many of the fundamental data that would support a locally informative predictive model of inundation are either currently unavailable or may not be usefully predicted given all the anticipated geomorphological and ecological feedbacks and interactions. That does not, of course, preclude planning bodies and responsible agencies from responding to the threat of sea-level rise. Rather, it heightens the need to be radically precautionary in near-coastal regions. What is axiomatic is that if predicted rates of sea-level rise are realised, much of the Victorian public lands which currently support intertidal vegetation will be inundated and the conservation of saltmarsh and mangrove will require substantial areas of what is currently freehold land to be set aside for their landward migration and reassembly.

Wave-exposuremodelling

The results of the wave-exposure modelling were limited by the spatial resolution of the analysis and the generalised, often inaccurate, nature of the bathymetry that informed the analysis. Figure 6.11 shows the results of the REI model for Western Port. The model is generally indicative of regions where low wave energy – shown in pink to white shading – may facilitate the development of either mangrove or saltmarsh. In particular locations such as Sandy Point, the REI values are clearly an anomaly which presumably arises from errors in bathymetry. Further, the model does not directly consider the contribution of ocean swells to wave energy. Such swells are common within and around the western entrance to the bay. Despite these limitations,


the model renders the current near-coastal wave energy reasonably well and could be considered broadly accurate if not precise. For example, wave energy created by fetch is comparatively high along the Lang Lang coast, where intertidal vegetation has apparently – at least, since European colonisation – never supported extensive mangrove or saltmarsh (see Figure 6.6).

Lang Lang Coast

Figure6.11: Relative wave exposure (REI) index modelled within Western Port and along the existing coastline. The region of Bass Straight in the south-west corner of the figure denoted by crosses was not modelled. Glaring errors in the modelling – particularly along the coastlines stretching from Point Leo to Sandy Point and on the northern coast of Phillip Island – will be evident to those familiar with the bay. They are a consequence of errors in the bathymetry.

Figure 6.12 shows the results of the REI model for Western Port determined for a bay expanded by the imposition of a 0.80 m rise in sea levels. The model is not particularly informative and is shown here only for the purpose of highlighting the problems with this approach. The REI calculations are particularly sensitive to ocean depth and the shape and position of the coastline. If sea-level rise is realised, both the sea floor and the coastline will respond morphologically. Having no way to anticipate such responses, we have merely used the contemporary bathymetry and have naively expanded the shoreline by the requisite 0.80 m vertical height. As a consequence, we have realised an unlikely future scenario where very extensive regions on the periphery of the northern section of Western Port are uniformly covered by exceedingly shallow sheet of water at high tide. Very shallow stretches of water typically support negligible wave energy and our model reflects this. Further gross limitations of the future REI modelling relate to the potential for local and regional wind strength and directional bias to change under a rapidly changing climate.

Figure6.12: Modelled relative wave energy in 2090. This schematic is the result of a very simplistic and potentially misleading analysis of ‘Relative wave exposure’ (REI) within Western Port employing a shoreline determined by adding 0.80 m vertical height to the existing shore. The region of Bass Straight in the south-west corner of the figure denoted by crosses was not modelled.


Vegetationtypemodelling

Elevation parameters extracted from the saltmarsh and mangrove mapping were used to define the local tidal niche of these vegetation types. The results of this GIS analysis are shown in Table 6.4.

Table6.4: Current elevational range for saltmarsh and mangroves in Western Port.

Vegetation type Mean sea level (m) Standard deviation (m)

Range + 1 standard deviation (m)

Mangrove 0.757 0.466 0.291 – 1.223

Saltmarsh 1.524 0.265 1.259 – 1.789

By applying these intervals as rules to the continuous terrestrial and bathymetric terrain model, we were able to make useful predictions regarding the distribution of these vegetation types in 2009. If we compare Figures 6.6 and 6.13, it is clear there is considerable congruence between the mapped distribution and the modelled current distribution, particularly in areas where there is existing saltmarsh vegetation (ignoring the predictions of intertidal vegetation on the islands where the coastal digital elevation model is not LIDAR-derived). A higher-resolution comparison is provided in Figure 6.14. A consequence of the deterministic nature of the model is that it erroneously predicts the occurrence of saltmarsh across the low-lying coastal plains that are currently protected from tidal inundation by constructed sea walls and levees. If these structures were removed, tidal channels would extend and form across these plains and saltmarsh and other coastal wetlands would certainly begin to establish. The ‘problem’ is particularly noticeable across the coastal plains behind Lang Lang and Tooradin.

The set of elevation rules for saltmarsh and mangrove was then applied to the terrestrial and bathymetric terrain model using a mean sea level elevated by 0.80 m above the present level. The results of this process are shown in Figure 6.15. In applying these rules to some future sea-level scenario, we have necessarily assumed that all the dynamic, geomorphological and hydrological processes are maintained as they operate at present including tidal regime (frequency, duration of inundation, seasonality, etc.) and the land and sea surface remain fixed are shown. Given the severe limitations of such assumptions and the crude determinism of the elevation rules, the results should be interpreted with great caution and not assumed to represent an accurate prediction.

Our simple model demonstrates the broad areas where saltmarsh migration is conceivable, such as the coastal plains around Tooradin, against those where it is impossible due to steep hinterland terrain, such as the San Remo coast. However, such propositions must be interpreted with great caution. We have considered only basic topography, the tidal range of saltmarsh and mangroves, and a crude measure of relative wave energy in our model. We have not considered any of the other factors (listed previously) that influence the destruction and re-establishment of estuarine vegetation under changed sea levels. Consequently our model may be both an over- or under-estimate of the extent of saltmarsh in various places, given the multitude of potential emergent scenarios. It is possible that future studies will be able to incorporate more variables to gather further insights, but we suggest that the level of complexity, interactivity and profound stochasticity among the variables will make accurate predictions of the distribution of intertidal vegetation unlikely in the foreseeable future. Certainly no currently published studies have succeeded in accounting for all relevant variables.


Figure6.14: Comparison of the observed (A) and predicted (B) contemporary distribution of saltmarsh (brown) and mangroves (dark green) when the elevation range rules are applied to the contemporary shoreline. The area shown is near Tooradin, on the northern end of Western Port. Note that this is a static model and the impediment posed to tidal inundation by sea walls and levees is in the plains behind Tooradin are not taken into account (right hand side of map). This demonstrates the local fidelity of saltmarsh and mangrove to sea level and tidal regime.

Figure6.13: Application of elevation range rules for mangrove (dark green) and saltmarsh (red-brown) to the contemporary shoreline of Western Port. Compare with distributions shown in Figure 6.6. Due to the static nature of the model it erroneously predicts the occurrence of saltmarsh across extensive low-lying coastal plains that are protected from tidal inundation by sea walls and levees. This is particularly noticeable across the Lang Lang and Tooradin coastal plains. Note also that the predictions for the islands and Sandy Point are crude and reflect the low quality of the DEM and bathymetry respectively.

Figure6.15: Application of elevation range rules (see Table 6.4) for mangrove (dark green) and saltmarsh (red-brown) to a possible future late 21st century shoreline of Western Port (0.80 m rise in eustatic sea level). Predictions for the islands and Sandy Point are not shown due to the low quality of the DEM and bathymetry. The hatched region of the map indicates the Lang Lang region where low 0.5–2 m peaty ‘earth-cliffs’ indicate that the current coastline is in retreat (see Gell 1974). Such processes may be reasonably expected to continue and as such, the development of extensive saltmarsh in this region under future sea levels is considered unlikely.

6 mangroves and coastal saltmarsh – past and futurechapter 6: mangroves and coastal saltmarsh –...

Documents