Spartina Spp.) in South-Eastern Australia Induces Island Formation, Salt Marsh Development, and Carbon Storage

Invasive cordgrass (Spartina spp.) in south-eastern Australia induces island formation, salt marsh development, and carbon storage

David M. Kennedy,1* Teresa Konlechner,1 Elisa Zavadil,1 Michela Mariani,1 Vanessa Wong,2 Daniel Ierodiaconou3 and Peter Macreadie4 1School of Geography and National Centre for Coasts and Climate, The University of Melbourne, Parkville, Victoria, Australia 2School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria, Australia 3School of Life and Environmental Sciences, Faculty of Science, Engineering and Built Environment, Victoria, Australia 4Deakin University, School of Life and Environmental Sciences, Centre for Integrative Ecology, Faculty of Science, Engineering and Built Environment, Burwood, Australia *Corresponding author. Email: [email protected]

Received 2 August 2017 • Revised 21 September 2017 • Accepted 23 September 2017

Abstract

Invasive vegetation species can lead to major changes in the geomorphology of coastal systems. Within temperate estuaries in the southern hemisphere, especially Australia and New Zealand, the cordgrass Spartina spp. has become established. These species are highly invasive, and their prolific growth leads to the development of supratidal environments in formerly intertidal and subtidal environments. Here, we quantified the impact of Spartina invasion on the geomorphology and sequestration capacity of carbon in the sediments of Anderson Inlet, Victoria, Australia. Spartina was first introduced to the area in the 1930s to aid in land reclamation and control coastal erosion associated with coastal development. We found that Spartina now dominates the intertidal areas of the Inlet and promotes accretion (18 mm/year) causing the formation of over 108 ha of supratidal islands over the past 100 years. These newly formed islands are calculated to potentially contain over 5.5 million tonnes of CO2 equivalent carbon. Future management of the inlet and other Spartina-dominated environments within Australian presents a dilemma for resource managers; on the one hand, Spartina is highly invasive and can outcompete native tidal marshes, thereby warranting its eradication, but on the other hand it is likely more resilient to rising sea levels and has the potential for carbon sequestration. Whether or not the potential advantages outweigh the significant habitat change that is antici- pated, any management strategies will likely require additional research into costs and benefits of all ecosystem services provided by Spartina including in relation to nutrient cycling, shoreline stabilisation, and biodiversity as well as in response to the longevity of carbon found within the sediments.

Keywords ecosystem services; blue carbon; Spartina; salt marsh; sea level rise; invasive species

Geographical Research • February 2018 • 56(1): 80–91 80 doi:10.1111/1745-5871.12265 D.M. Kennedy et al., Cordgrass salt marsh 81

Introduction Estuary, Tasmania, Australia, led to accretion of a supratidal marsh over previously bare subtidal Salt marshes are commonly found on the margins mud flats (Hedge & Kriwoken, 2000; Kriwoken of estuaries and are characterised by halophytic & Hedge, 2000). In this study, we explore the vegetation that is inundated by the highest tides impact of invasive grass species (Spartina)within (Allen, 2000). Their evolution is closely related Venus Bay contained within a shallow barrier- to sea level and tidal inundation; however, the estuary in Victoria, Australia. Through subsurface precise elevation of a salt marsh within the tidal coring and aerial photo analysis, we assess the prism is the product of a complex interaction of positives and negative environmental impacts of biophysical factors related to allochthonous and invasive-species driven habitat change. autochthonous sedimentation, vegetative communities, and ground water (Rogers, Saintilan & Regional setting Woodroffe, 2014). Salt marshes are therefore not passive environments, and they interact with wider Venus Bay has an approximate surface area of estuarine environments through both vertical and 3.4 km2 located in the south eastern part of lateral accretion. It is their ability to accrete which Anderson Inlet in West Gippsland, Victoria, is most important for their continued presence Australia (Figure 1). Anderson Inlet is predomi- under rising sea level, rather than simply their nantly submarine, with sediment movement asso- elevation in relation to sea level (Lovelock et al., ciated with migrating tidal channels and 2014; Reef et al., 2017). movement of the tidal deltas at the estuary The main external controls on salt marsh mouth. The estuary is 10 km long, being the wid- development are tidal regimes as well as sediment est (2–3 km) in its central regions. It is separated supplies (Allen, 2000). There must also be from the open ocean by a beach-barrier system sufficient accommodation space for marshes to estimated to be 4,500 years old (Li, Gallagher accrete, with the highest tides being the principle & Finlayson, 2000). The estuary is therefore boundary condition. Rates of accretion on the classified as a partially infilled barrier estuary other hand are determined by rates of sediment according to the scheme of Roy (1984). The supply, which can be derived from both organic coastal plain surrounding the inlet is characterised and inorganic sources (Allen, 2000). Vegetation by three terraces ranging up to 6 m elevation. The is a key element as it can both enhance deposition highest corresponds to the last interglacial period of suspended sediment as well as directly contrib- (c. 125 ka) and the lower two, both <2meleva- ute biomass (Reed, 1995). In fact, tidal marshes are tion, relating to early–mid Holocene higher sea among the most efficient ecological systems for levels (Li et al., 2000). the storage of organic carbon (Duarte et al., The coast of Victoria is microtidal with a semi- 2013; McLeod et al., 2011; Pidgeon, 2009), with diurnal spring range of 1.1 m at Port Philip Heads salt marsh ecosystems ranking the highest in (Port of Melbourne, 2013). The tidal prism with organic carbon storage potential among all coastal Anderson Inlet is 22.66 × 106 m3 (McSweeney, wetland and forested terrestrial ecosystems Kennedy, & Rutherfurd, 2017). The mean signifi- (Ouyang & Lee, 2014). The efficiency of carbon cant wave height for the Victorian coast is 2.4 m accumulation is related to the burial of organic with a period of 8.4 s (Hughes & Heap, 2010). material from accreting sediments and the Modelling of open-ocean waves in Victoria presence of anaerobic conditions, which decrease indicates the mean annual wave height for Cape rates of organic matter decomposition (Hedges & Paterson is 1.8 m (WaterTech, 2004). Wave data Keil, 1995; Kristensen, Ahmed & Devol, 1995; within Anderson Inlet are not recorded. Mean Kristensen et al., 2008). annual maximum and minimum air temperatures As marshes’ evolution is dependent on their for Cape Paterson are 9.6–18.8°C with a mean vegetative ecosystems, changes within these habi- annual rainfall of 939 mm (BoM, 2012). tats can have major geomorphic implications. For Spartina (cordgrass) is a genus of approxi- example in Australia, Western Port, Victoria mately 17 rhizomatous perennial grass species (Rogers, Saintilan & Heijnis, 2005) and Tweed found in temperate estuarine and salt marsh envi- River, New South Wales (Rogers et al., 2014), ronments. Spartina spp. are efficient colonisers of changes in tidal prism associated with sea level rise marine habitats renowned for their ability to stabi- have led to mangrove colonisation of salt marsh lise estuarine surfaces, promote accretion, and the habitat. At lower elevations within the tidal prism, conversion of intertidal flats into supratidal mono- the introduction of Spartina grass in the Tamar cultures. Spartina was widely planted for erosion

Figure 1 Venus Bay is located within Anderson Inlet in southern Victoria. (a) Anderson Inlet is one of the most southernmost barrier estuaries on the Australian mainland, occurring on the southern coast of Victoria. (b) Venus Bay (white box) is the area where Spartina grass has the most impact and is the focus of this study control and reclamation in many temperate regions Methods because of these attributes (An et al., 2007; Hacker et al., 2001). The impact of Spartina was evaluated within a Two species of Spartina, Spartina x townsendii 280 ha section of Venus Bay (Figure 1). Airborne and Spartina anglica were introduced to coastal Light Detection and Ranging (LiDAR) data were Victoria (Williamson, 1996). S. x townsendii,a collectedin2007bytheVictorianGovernment’s sterile hybrid of American and English cordgrass Department of Environment and Primary Indus- species, was first introduced to the study site in tries. The surveying was conducted using a LADS the 1930s (Williamson, 1996). The second species, Mk II system coupled with a GEC-Marconi S. anglica, is the fertile product of chromozonal FIN3110 inertial motion sensing system and a dual doubling of S. x townsendii (Hacker et al., 2001). frequency kinematic global positioning system S. anglica was introduced to the study site in the (kGPS). This dataset was processed to produce a 1960s (Williamson, 1996). S. anglica is probably seamless terrestrial–marine mosaic from eleva- the dominant species today, however, the two spe- tions of +10 m to depths of À25 m with a final ras- cies are difficult to distinguish and there are few ter grid of 2.5 m resolution for the entire Victorian data available on the relative distributions of coast (Quadros and Rigby, 2010). Shoreline S. anglica and S. x townsendii within the study site. change on the islands was defined on the basis of

© 2017 Institute of Australian Geographers D.M. Kennedy et al., Cordgrass salt marsh 83 the seaward boundary of the vegetation in aerial according to the criteria of Leeder (1982) and size and satellite photographs from 1950, 1981, 2010, statistics calculated using the graphical procedures and 2015 (Table 1). The 1950s imagery was of Folk and Ward (1957). Analysis of total organic overexposed and composed of three images that carbon and carbonate composition was undertaken had been combined manually prior to digitisation. using the loss on ignition method (Kennedy & Ground truthing of the most recent aerial photo- Woods, 2013). Dating was undertaken at the Wai- graph indicated this boundary is either mangrove kato Radiocarbon Dating Laboratory, New or Spartina grass. Images were georeferenced in Zealand. ArcGIS 10.4.1 using control points obtained from Pollen analysis, using standard procedures the 2015 imagery. Images were transformed using (including hot 10% KOH, 40% HF and acetolysis) a first order polynomial with root mean square er- as outlined in Faegri and Iversen (1989), was rors ranging between 0.42 and 12.2 m. Vegetation undertaken at 10 cm intervals in the uppermost lines were digitized manually by a single operator. part of the core with pretreatment of 10 per cent Positional uncertainty in defining the shoreline HCl prior to addition of HF. Organics were sepa- was calculated based on the pixel, rectification, rated with KOH and acetolysis, with samples and digitization errors (Ford, 2013). Pixel and rec- sieved through a 7 micron mesh. Slides were tification errors are represented by the resolution of counted until at least 300 terrestrial pollen grains the original image and the root mean square error were identified based on species reference collec- of the dereferencing process (Del Rio & Gracia, tions held by the School of Geography at The 2013). Digitization error was calculated as the University of Melbourne. Pollen belonging to standard deviation of the shoreline position from Spartina was not distinguishable from native repeated digitization of the largest mud island by Poaceae spp. Counts were processed and graphed a single operator (Ford, 2013). Total shoreline er- using the computer programme Tilia 2.0.37 ror is calculated as the root sum of all shoreline po- (Grimm, 1999). sition errors and ranged between 4.4 and 28.3 m (Table 1). Island volume and height above mean Results sea level (MSL) was calculated in ArcGIS from the 2007 LiDAR data. MSL was defined as the Islands zero elevation contour line relative to Australian Height Datum. Islands in the aerial photos were defined as being A vibrocore core (76 mm diameter) was taken vegetated communities, separated from the hinter- on the mud island closest to the boat ramp in land by channels, which have a proportion of their Venus Bay (38.67056°S, 145.79892°E) in 2015 surface exposed at or above mean high water (Figure 2). The location was selected as it repre- spring tide elevation. Three vegetated mud islands sents an intertidal environment of the mud island could be clearly identified in 1950 and ranged in region, the area where the impact of Spartina area from 0.14–20.14 ha, with lengths and widths was most likely to be significant and where the of up to 654 and 435 m, respectively (Figure 2a). aerial photo analysis indicated the mud flats to be These islands grew in the 65 years between aerial particularly dynamic. images by progressive expansion of shorelines fol- In the laboratory, sediment grain size was lowing vegetation colonisation of the intertidal analysed using a Beckman Coulter LP13320 laser mud flats and through the formation and coales- particle sizer with grain texture classified cence of new islands. Two islands, identified in Figure 2, exemplify the processes of colonisation and island growth Table 1 Uncertainty in shoreline position on islands within between 1950 and 2015. Island A was not present Venus Bay in 1950 but in 1981 comprised small vegetation patches (<38 m2) indicative of the recent estab- Image Pixel Rectification Digitization Total lishment of plants on the previously unvegetated size (m) error (RMS) error (m) error (m) mudflats.By2010,IslandAhadgrownto comprise an island 1.4 ha in area with a close-to 1950 3.2 12.2 28.2 28.4 100 per cent vegetation cover. Island B was the 1981 0.3 11.6 4.4 4.4 smallest of the three islands present in 1950. By 2010 1 0.42 24.7 27.3 2015 1 21.9 25.1 1981, Island B had expanded in area by 97 per cent to comprise an island 10.4 ha. By 2010, expansion RMS: root mean square of Island B and neighbouring mud islands had

Figure 2 Aerial imagery of Venus Bay from 1950 to 2015 showing the growth of the islands in the past 65 years. Subsurface sediments were samples from the small island closest to shore which accreted during this period (arrowed yellow circle). cumulated in the coalescence of at least three between 2010 and 2015; however, extensive zones islands forming an island measuring 22 ha in area. of scattered vegetation along the margins of sev- Sixteen islands were identified in the 1981 eral islands as well as isolated patches of new veg- imagery: the three islands present in 1950 and an etation are clearly identifiable in the 2015 imagery additional 13 islands that had formed where previ- indicating that island expansion and colonisation ously there had been bare tidal flats (Figure 2b). of the intertidal mud flats remains ongoing. Over- All three original islands had increased in area all, the total area of mud island within the study and the vegetation around the island margins site has increased steadily from 34.4 ha in 1950 comprised scattered colonies indicative of ongoing to 108.3 ha in 2015, a net increase of 214 per cent island expansion. These 16 islands expanded and (Table 2). While no elevation data are available merged to form eight large islands by 2010 and for the study area prior to 2007, the extensive tidal new islands continued to form (Figure 2c). The flats evident in 1950 probably lay at close to MSL. mud islands continued to expand in area between Average island height in 2007 was 0.67 above 2010 and 2015. By 2015, the study site comprised MSL with a combined volume of about ten islands with an average area of 10.8 ha and 688,711 m3 above MSL. lengths and widths of up to 1,100 and 750 m, re- The contemporary mud islands are colonised by spectively (Figure 2d). No new islands formed amixofSpartina and mangrove (Avicennia

Table 2 Changes in island area in the past 65 years however, island expansion appears to be the result of increasing mangrove densities, not Spartina. Year Island area (ha) Rate of increase (ha/year)

1950 34.4 ± 7.9 Stratigraphy 1981 60.9 ± 2.4 0.9 2010 103.6 ± 18.5 1.5 A vibrocore was collected in the centre of a small 2015 108.3 ± 18.1 0.9 island which was a bare sediment ﬂat in 1950 (Figure 2d). A total depth of 3.41 m was reached through coring, and a coarse shelly-sand layer at marina var. australasica) (Figure 3). The initial the base of the core prevented further penetration. increase in island area between 1950 and 1981 A core length of 2.65 m was recovered and, with was largely because of the spread of Spartina since no loss of sediment occurring, resulted in a total 1962 (Boston, 1981), and can be clearly distin- compaction of 22 per cent. guished from the indigenous mangrove/saltmarsh The basal unit of the core is dominated by communities dominating the original mud islands medium–coarse sand and characterised by mud in the 1981 aerial imagery. While we have no lenses and abundant shell material both in the form quantitative data on the ﬂoristic composition of of whole valves of Tellina deltoidalis and broken the mud islands since 1981, examination of the ae- shell hash layers (Figure 4). Total organic carbon rial imagery suggests that the ongoing expansion content is low (<2%) while carbonate grains of Spartina contributed to much of the increase accounts for between 3 and 7 per cent (Figure 5). in island area between 1981 and 2010. Since 2010, The whole shell valves are disarticulated and

Figure 3 (a) the edge of the cored island looking toward the Venus Bay township. (b) Spartina and mangrove vegetation in the centre of the island where the coring was undertaken

Figure 4 Sedimentology and stratigraphic log of the vibrocore. The uppermost parts of the core are fine grained dominated by silt. The sediment progressively coarsens with depth with a coarse sand with shell material found at the base of the core. broken indicative of a high energy environment of therefore representing a maximum age for the base deposition. The mean and medium grain sizes are of the core. similar (1–2 φ,mediumsand),reflecting the Overlying the lower unit is a medium—fine, moderately—moderately well-sorted character of moderately—moderately-well sorted, sand with a the sediments (Figure 4). A single T. deltoidalis minor (<15%) component of mud. Small mud valve from the base of this unit was radiocarbon lenses (millimetre thick) were found throughout dated at 2,747–2,480 years CalBP (Table 3). This this unit, being particularly common from shell was disarticulated and allocthonous, 1.8–2.2 m (2.31–2.83 m compaction corrected)

Figure 5 Total organic carbon (TOC) and carbonate composition through the core. There is a distinct difference between the upper 75 cm and lower parts of the core, with the surface sediment being high in TOC and low in carbonate depth (Figure 4). There was an absence of large, present throughout the core, with a peak at 0.65 m visible shell material, although peaks in carbonate core depth (0.83 m compaction corrected) content to over 12 per cent suggest some shell (Figure 6). Specimens of Rumex spp. are found material is present as sand-size grains (Figure 5). throughout the mud unit. These species are consid- Overlying the sands to 0.8 m (1.2 m compaction ered to be representative of the post-European corrected) depth is a muddy sand unit (mean grain period. Indigenous species Eucalyptus and size 2.5 and 4 φ (very fine to fine sand), which is Myrtaceae spp. are more abundant at the base of very poorly sorted to poorly sorted. Carbonate the mud unit (10%) compared with the top content peaks at 9 per cent, and there is a marked (<5%). Casuarina sp. also are in greater abun- increase in the total organic carbon content, rising dance at the base (Figure 6). During analysis, soot to >6 per cent at the top of the unit. Live mangrove particles, the byproduct of industrial activity, were roots were found within this unit (Figures 4 and 5). found throughout the uppermost unit. Soot was The top of the core from 0.8 m core depth to the classified on the basis of its round organic clast surface is dominated by very poorly sorted, or- appearance, which contrasts to angular charcoal ganic rich (>10% total organic content, TOC, with particles. a peak of 17%) fine mud, which contained large (c. 5 mm width) live mangrove roots but lacked Discussion visually observable shell material (Figure 4). The mean grain size was between 5 and 6 φ (medium Spartina spp. was first introduced to Australia in silt) in the upper 0.5 m and coarsened to 4 φ (coarse the late 1920s and early 1930s and was recorded silt—very fine sand) in the lower parts of this up- in Anderson Inlet by 1932 (Williamson, 1996). permost unit (Figure 3). This unit consistently has At the time of the earliest aerial photos 30 years a low carbonate content (<3%) (Figure 5). later, the grass had already established in the southern section of the estuary. This original Palynology population is likely to comprise the infertile S. x townsendii. A rapid expansion in the area occupied Well preserved pollen was found in the uppermost by Spartina spp. was recorded after the introduc- mud unit. The non-native tree Pinus radiata is tion of S. anglica in 1962 (Williamson, 1996). particularly abundant at the surface, accounting Today, Venus Bay is still dominated by for 20 per cent of pollen, with Plantago lanceolata supratidal islands, which have increased their

Table 3 Radiocarbon age from the tidal channel at the base of the core

Lab Number Material Depth (cm) Conventional age Calibrated age (years BP) (years BP ± 1σ) (68.2% Prob)

Wk41853 Tellina deltoidalis 242 2,938 ± 20 2,747–2,480

Figure 6 Pollen composition of the surface fine organic unit. The abundance of exotic Pinus pollen indicates the post-European settlement age of this material. aerial coverage by 73.8 ha. At present, these and this can be observed on the older marsh marshes are dominated by Spartina, with a fringe surfaces in Anderson Inlet that were present over of mangroves on their seaward edges. The surface 50 years ago (Figure 2). Channel dimensions are sediments extend to over 1.0 m depth at the south not, however, an accurate proxy of marsh age as western part of the island complex. These their evolution can be affected by a range of organic-rich sediments are characteristic of marsh conditions often working independently of climate, environments, particularly those composed of tides, or sedimentological conditions (Perillo & Spartina as described in a global review of 143 Iribarne, 2003). For example, in Argentina (Rio sites by Ouyang and Lee (2014). Underlying this de la Plata and Loyola Bay), channel development organic-rich layer are coarser fine sands, with is related to hydrodynamic processes, while in low TOC more characteristic of the bare sediment Bahia Blanca Estuary biology was critical in flats surrounding the islands. channel initiation (Perillo & Iribarne, 2003). The presence of P.radiata pollen throughout the The marsh-dominated islands in Venus Bay are organic-rich salt marsh unit, combined with an therefore likely the result of the introduction of abundance of soot particles, allows us to infer that Spartina due to the coincidence of plant colonisa- the marsh has developed since European colonisation and island development, all of which occurred tion. Poor preservation of pollen grains below the in the mid-20th century postdating catchment salt marsh unit, as well as possible contamination clearance by a century. The island formation has related to mangrove roots, means that inferring led to a major shift in the geomorphology of Venus the age of the non-marsh sediments is difficult. Bay. In an Australian review, Macreadie et al. The inter-sub tidal flats on which the marsh grew, (2017) found that Spartina-dominated marshes however, must be younger than 2,747–2,480 years had the highest levels of carbon accumulation rates CalBP, based on the radiocarbon age of a tidal among five halophyte genera (incl. Disthichlis, channel unit at the base of the core (Figure 7). Halimione, Juncus,andPhragmites). In this study, The marsh islands, therefore, have primarily the total organic carbon within the Spartina portion grown vertically on top of subtidal muddy sand of the marsh was >10 per cent of the dry sediment flats at rates of 18.5 mm/year. Existing channels weight. Sampling of neighbouring salt marshes for on the bare sediment flats in the 1950s aerial photos organic carbon yields a carbon concentration of À3 appear to delineate the edge of the newly emergent 2.16+/À1.05 g Corg cm (Macreadie et al., marshes suggesting primarily vertical accretion: 2017). Given the values of TOC are similar this is similar to the Beeftink and Rozema (1988) between the two marsh environments, we can model of salt marsh development developed from assume a similar degree of organic carbon storage barrier and estuarine systems in the Netherlands. beneath the mud islands. For the studied islands As the marshes vertically accrete, tidal channels (688,711 m3 aboveMSLbasedontheLiDAR tend to become more entrenched (Allen, 2000), data), this yields a total historical carbon store of

Figure 7 Stratigraphic interpretation of the sedimentary history of the island from 3 ka to present at least 1.5+/À0.7 million tonnes, giving a total invasive, and has caused major geomorphic CO2 equivalent of 5.5 million tonnes. Based on change within the estuary and the creation of inter- the data available, this store is approximate but tidal mud islands. Island accretion has been linear does represent the potential of such systems for since 1950 and therefore will most likely continue carbon storage. The longer term potential of the into the future, especially as Spartina is wide- newly established marshes to sequester carbon spread within Anderson Inlet. Major changes in will, however, depend on the dynamics of the benthic habitats can therefore be expected to marsh-biosedimentary system and the structural include associated impacts on navigation and ame- form of carbon present (that is, labile or not labile). nity use within Venus Bay. Actively accreting The establishment of Spartina therefore poses a marsh environments on the other hand have management dilemma. The species is highly greater potential to provide shoreline protection

© 2017 Institute of Australian Geographers 90 Geographical Research • February 2018 • 56(1): 80–91 to sea level rise than do bare sediment flats. In ad- habitats may provide opportunities for other dition, their high carbon content means there is po- colonisers such as mangrove. A critical unknown tential for carbon storage, assuming that the carbon is the ability of mangroves to outcompete Spartina is primarily in situ. as sea level rises. An additional complication for managers is the impact of sea level rise on vegetative communities. Acknowledgements In Victoria, higher high tides are causing shifts in community compositions with mangroves This project was kindly funded by the West Gippsland colonising salt marsh environments in Western Catchment Management Authority, with Simon Sharp and fi Port Bay (Rogers et al., 2005; Saintilan et al., Darren Hocking providing valuable eld assistance. The insightful comments of two anonymous reviewers were 2014). This vegetative change, in turn, will have appreciated. This project is also supported through funding from implications for the storage of below-ground the Australian Government’s National Environmental Science carbon as rates of sequestration are dependent on Programme. the wetland community structure (Lovelock et al., 2014). A significant unknown is the ability of mangroves to displace Spartina-dominated References marshes when compared with endemic communi- Allen, J.R.L., 2000. Morphodynamics of Holocene salt ties. In Anderson Inlet, the Spartina-dominated marshes: a review sketch from the Atlantic and southern marshes are fringed by mangroves (Avicennia North Sea coasts of Europe. Quaternary Science Reviews, marina var. australasica). In Victoria and eastern 19(12), pp.1155–1231. Australia (specifically the Tweed River), the An, S.Q., Gu, B.H., Zhou, C.F., Wang, Z.S., Deng, Z.F., Zhi, Y.B., Li,H.L.,Chen,L.,Yu,D.H.andLiu,Y.H.,2007.Spartina almost universal shift from saltmarsh to mangrove invasion in China: implications for invasive species manage- communities (Rogers et al., 2005; Rogers et al., ment and future research. Weed Research, 47(3), pp.183–191. 2014; Saintilan et al., 2014) may mean that the Beeftink, W.G. and Rozema, J., 1988. The nature and function- Venus Bay Spartina marshes could be replaced ing of salt marshes. In: W. Salomans, B. L. Bayne, E. K. Duursma and U. Forstner, eds. Pollution of the North Sea: by mangrove in the future. An Assesssment. Berlin: Springer. pp.59–87. Del Rio, L. and Gracia, F.J., 2013. Error determination in the Conclusions photogrammetric assessment of shoreline changes. Natural Hazards, 65(3), pp.2385–2397. The colonisation of the invasive species Spartina Duarte, C.M., Losada, I.J., Hendriks, I.E., Mazarrasa, I. and Marba, N., 2013. The role of coastal plant communities for has led to the development of supratidal marshes climate change mitigation and adaptation. Nature Climate in Anderson Inlet in south eastern Australia. These Change, 3(11), pp.961–968. marshes have developed in the past century on what Faegri, K., and Iversen, J., 1989. Textbook of Pollen Analysis. were initially intertidal sediment flats. Based on ae- 4th edition by Knut Faegri, Peter Emil Kaland, and Knut Krzywinski. Chichester: Wiley. rial photograph analysis, Spartina colonisation has Folk, R.L. and Ward, W.C., 1957. Brazos River bar: a study in been rapid and appears to continue to drive island the significance of grain size parameters. Journal of Sedimen- evolution, with the islands doubling in size since tary Petrology, 27(1), pp.3–26. 1950. The rapid expansion of the Spartina islands Ford, M., 2013. Shoreline changes interpreted from multi- is most likely to continue into the next century. temporal aerial photographs and high resolution satellite images: Wotje Atoll, Marshall Islands. Remote Sensing of The islands that have been created are very high Environment, 135, pp.130–140. in total organic carbon, suggesting their high Hacker, S.D., Heimer, D., Hellquist, C.E., Reeder, T.G., potential as a carbon sink. This potential will, Reeves, B., Riordan, T.J. and Dethier, M.N., 2001. A marine however, be dependent on the type of carbon that plant (Spartina anglica) invades widely varying habitats: potential mechanisms of invasion and control. Biological is present as well as on the long term stability of Invasions, 3(2), pp.211–217. the newly accreted islands and their vegetative Hedge, P. and Kriwoken, L.K., 2000. Evidence for effects of communities. Management of Spartina-dominated Spartina anglica invasion on benthic macrofauna in Little environment results in two major environmental Swanport estuary, Tasmania. Austral Ecology, 25(2), – trade-offs. On the one hand, Spartina is a highly pp.150 159. Hedges, J.I. and Keil, R.G., 1995. Sedimentary organic matter invasive species the presence of which leads to preservation: an assessment and speculative synthesis. major change in the habitats it colonises; this is Marine Chemistry,49(2–3), pp.81–115. because it forms supratidal surfaces in formerly Kennedy, D.M. and Woods, J.L.D., 2013. Determining organic sub-intertidal environments. On the other hand, and carbonate content in sediments. In: A. D. Switzer and D. M. Kennedy, eds. Methods in Geomorphology.SanDiego: Spartina marshes are high in organic content and Academic Press. pp.262–273. can accrete at high rates and provide protection to Kristensen, E., Ahmed, S.I. and Devol, A.H., 1995. Aerobic and the subaerial shoreline. The generation of supratidal anaerobic decomposition of organic matter in marine

sediment: which is fastest? Limnology and Oceanography, Perillo, G.M.E. and Iribarne, O.O., 2003. Processes of tidal 40(8), pp.1430–1437. channel development in salt and freshwater marshes. Earth Kristensen, E., Bouillon, S., Dittmar, T. and Marchand, C., Surface Processes and Landforms, 28(13), pp.1473–1482. 2008. Organic carbon dynamics in mangrove ecosystems: a Pidgeon, E., 2009. Carbon sequestration by coastal marine review. Aquatic Botany, 89(2), pp.201–219. habitats: important missing sinks. In: D. Laffoley and G. Kriwoken, L.K. and Hedge, P., 2000. Exotic species and estuar- Grimsditch, eds. The Management of National Coastal ies: managing Spartina anglica in Tasmania, Australia. Carbon Sinks. Switzerland: IUCN, Gland. pp.47–51. Ocean and Coastal Management, 43(7), pp.573–584. Reed, D., 1995. The response of coastal marshes to sea-level Leeder, M.R., 1982. Sedimentology, Processes and Product. rise: survival or submergence. Earth Surface Processes and London: HarperCollins Academic. Landforms, 20(1), pp.39–48. Li, L., Gallagher, S. and Finlayson, B., 2000. Foraminiferal Reef, R., Spencer, T., Möller, I., Lovelock, C.E., Christie, E.K., response to Holocene environmental changes of a tidal McIvor, A.L., Evans, B.R. and Tempest, J.A., 2017. The ef- estuary in Victoria, southeastern Australia. Marine Micropa- fects of elevated CO2 and eutrophication on surface elevation leontology, 38(3–4), pp.229–246. gain in a European salt marsh. Global Change Biology,23(2), Lovelock, C.E., Adame, M.F., Bennion, V., Hayes, M., pp.881–890. O’Mara, J., Reef, R. and Santini, N.S., 2014. Contemporary Rogers, K., Saintilan, N. and Heijnis, H., 2005. Mangrove rates of carbon sequestration through vertical accretion of encroachment of salt marsh in Western Port Bay, Victoria: sediments in mangrove forests and saltmarshes of South East the role of sedimentation, subsidence, and sea level rise. Queensland, Australia. Estuaries and Coasts, 37(3), Estuaries, 28(4), pp.551–559. pp.763–771. Rogers, K., Saintilan, N. and Woodroffe, C.D., 2014. Surface Macreadie, P.I., Ollivier, Q.R., Kelleway, J.J., Serrano, O., elevation change and vegetation distribution dynamics in a Carnell, P.E., Lewis, C.E., Atwood, T.B., Sanderman, J., subtropical coastal wetland: implications for coastal wetland Baldock, J., Connolly, R.M. and Duarte, C.M., 2017. response to climate change. Estuarine, Coastal and Shelf Carbon sequestration by Australian tidal marshes. Scientiﬁc Science,149,pp.46–56. Reports, 7(7), pp.1–10. Roy, P.S., 1984. New South Wales estuaries: their origin and Mcleod, E., Chmura, G.L., Bouillon, S., Salm, R., Björk, M., evolution. In: B. G. Thom, ed. Coastal Geomorphology in Duarte, C.M., Lovelock, C.E., Schlesinger, W.H. and Australia. North Ryde: Academic Press. pp.99–121. Silliman, B.R., 2011. A blueprint for blue carbon: toward Saintilan, N., Wilson, N.C., Rogers, K., Rajkaran, A. and an improved understanding of the role of vegetated coastal Krauss, K.W., 2014. Mangrove expansion and salt marsh habitats in sequestering CO2. Frontiers in Ecology and the decline at mangrove poleward limits. Global Change Environment, 9(10), pp.552–560. Biology, 20(1), pp.147–157. McSweeney, S.L., Kennedy, D.M. and Rutherfurd, I.D., 2017. A Williamson, R. 1996. Spartina in Victoria—an overview. In: J. geomorphic classiﬁcation of intermittently open/closed A.E. Rash, R.C. Williamson and S.J. Taylor, eds. Proceed- estuaries (IOCE) derived from estuaries in Victoria, Australia. ings of the Australasian Conference on Spartina Control Progress in Physical Geography, 41(4), pp.421–449. (10–12 May 1995). Victoria: Department of Conservation Ouyang, X. and Lee, S.Y., 2014. Updated estaimates of carbon and Natural Resources. pp.26–29. accumulation rates in coastal marsh sediments. Biogeosci- ences, 11, pp.5057–5071.