Geochemical investigation of the South Wellesley Island wetlands: Insight into wetland development during the Holocene in tropical northern Australia

Abstract

The South Wellesley Islands in the Gulf of Carpentaria, northern Australia, were the recent focus of a palynological investigation which found vegetation change during the Holocene was driven by coastal progradation and regional climate. Here, we present new elemental data from x-ray fluorescence core scanning which provides non-destructive, continuous and high resolution analysis from three wetlands across Bentinck Island, the largest of the South Wellesley Islands. Elemental data and grain size analyses are combined with lead-210 (²¹⁰Pb) and accelerator mass spectrometry (AMS) carbon-14 (¹⁴C) dates. An open coastal environment was present 1250 cal. a BP on the south east coast of Bentinck Island, with sediment supply incorporating fluvial deposition and detrital input of titanium and iron from eroding lateritic bedrock. Prograding shorelines, dune development and river diversion formed a series of swales parallel to the coast by ~800 cal. a BP, forming the Marralda wetlands. Wetlands developed at sites on the north and west coasts ~500 and ~450 cal. a BP, respectively. Geochemical and grain size analyses indicate that wetlands formed as accreting tidal mudflats or within inter-dune swales that intercepted groundwater draining to the coastal margins. The timing of wetland initiation indicates localised late-Holocene sea level regression, stabilisation and coastal plain development in the Gulf of Carpentaria. Elemental data provide new records of wetland development across Bentinck Island, highlighting the value of a multi-proxy approach to understanding environmental change during the Holocene in tropical northern Australia.

Keywords

laser particle sizer lead-210 palaeoproxy tropical Australia wetlands XRF

Introduction

A growing body of research suggests stabilising sea levels and regional climate drove wetland development across tropical northern Australia during the late-Holocene. Freshwater swamps developed by 2600 cal. a BP on Mua and Badu islands in the Torres Strait (Rowe, 2007a, 2007b) and 2700 yr BP in central Cape York Peninsula (Butler, 1998; Stephens and Head, 1995). Freshwater collected behind cheniers, coastal swales and in palaeochannels creating freshwater resources across the northern coastal plains by 2000 yr BP (Chappell, 1988; Woodroffe, 1988). Groote Eylandt in the Gulf of Carpentaria records increasing wetness in the last 1000 cal. a BP after a period of dune building (Shulmeister, 1992), while Big Willum Swamp in Cape York expanded to its greatest extent in the last 600–400 cal. a BP (Stevenson et al., 2015). These studies suggest a region wide driver of wetland development during the late-Holocene. However, degradation of traditional proxies in tropical environments and a lack of spatially diverse records limit our understanding of the role climate and sea level change played.

The coastline of tropical northern Australia has significantly changed throughout the Holocene. Rapid postglacial sea level rise reached modern levels 7400 ± 200 cal. a BP in the South Alligator River, van Diemen Gulf (Lewis et al., 2013; Woodroffe et al., 1985, 1987, 1989), and 6400 cal. a BP in the southern Gulf of Carpentaria (Chappell et al., 1982; Lewis et al., 2013; Rhodes, 1982; Rhodes et al., 1980). A mid-Holocene highstand between 1 and 2 m higher than present mean sea level (pmsl) inundated low lying coastal regions and mangrove communities expanded. As sea level stabilised, mangroves flourished during a ‘Big Swamp’ phase with the rapid vertical accretion of mangrove muds evident in sedimentary archives throughout northern Australia and the Torres Strait (Crowley, 1996; Grindrod et al., 1999; Grindrod and Rhodes, 1984; Rowe, 2007a, 2007b; Woodroffe et al., 1985, 1989). Falling sea levels in the late-Holocene caused prograding coastlines associated with the contraction of mangrove communities and succession to freshwater swamps (Crowley et al., 1990; Crowley and Gagan, 1995; Luly et al., 2006; Proske et al., 2014; Rowe, 2007a, 2007b).

Variability in late-Holocene records from tropical northern Australia is often explained by increased seasonality due to the coupled El Niño–Southern Oscillation (ENSO) (Luly et al., 2006; Prebble et al., 2005; Rowe, 2007a; Shulmeister and Lees, 1995). ENSO drives interannual climatic variability across tropical northern Australia by reducing monsoon intensity and causing both droughts and extensive fires (Gagan et al., 2004; Lo et al., 2007; McBride and Nicholls, 1983; Sturman and Tapper, 1996). The alternate La Niña phase increases precipitation, cyclone activity and flood events in the region (Nicholls, 1992). During the mid-Holocene, a general drying trend was identified across northern Australia (Abram et al., 2009; Haug et al., 2001; Reeves et al., 2013; Stott et al., 2004). ENSO intensified between 3700 and 2000 cal. a BP (Reeves et al., 2013; Shulmeister, 1999; Shulmeister and Lees, 1992; Turney and Hobbs, 2006; Williams et al., 2008) before strong La Niña–like conditions increased effective precipitation between 1500 and 1000 cal. a BP across parts of northern Australia (Markgraf and Díaz, 2000; Moss et al., 2011; Shulmeister, 1999; Williams et al., 2010). The role that climate has played in the development and persistence of freshwater resources during the late-Holocene is difficult to quantify without high resolution multi-proxy studies.

Geochemical analysis of wetland sediments provides a tool to explore past environments across a range of sites where other proxies are absent or compromised. Micro x-ray fluorescence (XRF) core scanning rapidly records continuous, high-resolution and non-destructive measurements of relative geochemical variations downcore (Croudace et al., 2006; Francus et al., 2009). XRF scanning of lake sediments records local and regional environmental change (Davies et al., 2015). However, XRF scanning techniques are yet to be applied to palustrine environments in tropical northern Australia. This study presents µXRF geochemical data from three coastal wetlands across Bentinck Island, in the Gulf of Carpentaria, northern Australia. Variations in elemental profiles correspond to changes in catchment weathering, sedimentation, hydrological regimes and post depositional processes (Boyle, 2002; Croudace et al., 2006). We aim to identify relationships between elements in a tropical coastal setting including titanium (Ti), iron (Fe), manganese (Mn), calcium (Ca), rubidium (Rb), strontium (Sr), silicon (Si), bromine (Br), potassium (K) and a ratio of molybdenum (Mo) incoherent (Compton) to coherent (Rayleigh). Changes through time are related to detrital input, anaerobic conditions and local hydrology. The organic content and grain size analyses of sediments identify changes in the wetlands which may affect the elemental data within the profile. Reconstruction of the catchment development, sediment input and wetland stability are related to wider changes in tropical northern Australia during the late-Holocene. This study finds µXRF is a valuable tool to examine palaeoenvironments in regions where other proxies may be absent or compromised.

Methodology

Sampling site

Bentinck Island (143.67 km², 22 m a.s.l.) is the largest of the South Wellesley Islands in the Gulf of Carpentaria (Figure 1). The Island is low lying (<5 m a.s.l.) with coastal depositional environments including beaches, supratidal mudflats, cheniers and aeolian dunes. Shallow sandy soils overlie clay or lateritic bedrock along the coastline (Grimes and Sweet, 1979) with wetlands and swales supporting organic loam/clay rich soils. Inland the lateritic bedrock produces well drained, clay rich soils with poor nutrient availability. The Wellesley Islands formed between 8000 and 6500 BP as sea levels rose with archaeological evidence of human occupation from 3500 years ago (Memmott et al., 2016; Ulm et al., 2010). Limited European presence and disturbance of the Islands have preserved palaeoecological records and archaeological sites.

Figure 1.

Map of Bentinck Island and surrounding South Wellesley Islands (Albinia, Fowler and Sweers Island). Sites discussed in this paper are labelled.

The tropical Australasian region experiences significant seasonal variation in prevailing winds and rainfall with the southward migration of the Inter Tropical Convergence Zone bringing the Australian summer monsoon. Northwesterly winds and 92% of the annual rainfall occur during the short wet season (November to March), while drier south east trade winds dominate during the long dry season (April to October). Tropical summer cyclones contribute additional rainfall to the region and are common in the area with 37 passing within 100 km of the Wellesley Islands between 1906 and 2006 (Bureau of Meteorology, 2014a). January and February are the wettest months with a mean rainfall of 326 mm and 307 mm, respectively (Bureau of Meteorology, 2014b). The mean maximum temperature is 33°C in November with a mean minimum in July of 16°C (Bureau of Meteorology, 2014b).

Three wetland sites across Bentinck Island were sampled using a D-section corer. Marralda Swamp (MARR04) runs parallel to the south east coast, consisting of several freshwater lakes and swamps. A pollen and micro-charcoal record from MARR01 located at the north east end of the chain is published in Moss et al. (2015). MARR04 (S 17.09631, E 139.54265, length = 50 cm) is ~100 m south of MARR01 and contains permanent freshwater to a depth of 50 cm. Open coastal woodland with Pandanus spiralis, Melaleuca acacioides and a mixed grassland including Spinifex longifolius surrounds the site with Eleocharis dulcis and Typha sp. growing in the wetland.

Well Swamp (WS01) (S 16.9846, E 139.49504; length = 50 cm) is an ephemeral site established behind a mangrove fringe including back beach species Ceriops tagal, Bruguiera exaristata and Avicennia marina. The site is surrounded by open grassland with P. spiralis and Casuarina equisetifolia subsp. equisetifolia also present in the area. A 50–90 cm core was sampled using a sand auger which contained increasing shell hash, sand and mottling towards the lateritic bedrock (90 cm). The unconsolidated material was not analysed as part of this study. West Coast Swamp (WCS01) (S 17.0948, E 139.43964; length = 50 cm) developed between dunes, containing open water to a maximum depth of 1 m with Typha sp. around the swamp fringe. Woodland surrounds the back dune site including Eucalyptus pruinosa, Corymbia setosa and Acacia sp. Four sediment cores were collected across a transect, and the longest core was selected for analysis.

Stratigraphy and chronology

Cores were described in the field noting variations in sedimentary characteristics, including pisoliths and shell hash. Loss on ignition (LOI) analysis determined the total organic content at continuous 1 cm intervals downcore with subsamples dried at 65°C and then heated to 450°C. Grain size distribution was measured at 1 cm intervals using a Malvern Mastersizer 2000 laser diffraction analyser situated at the Australian Nuclear Science and Technology Organisation (ANSTO), Sydney, NSW. All samples underwent pre-treatment prior to analysis. First, organic material was destroyed using 30% H₂O₂ heated to 65°C before disaggregating samples in sodium pyrophosphate (Na₄P₂O₇). Samples were dispersed using ultrasound (Ryzak and Bieganowski, 2011) and grain size readings taken three times. The mean of each sample determined the percent of clay (<2 µm), silt (2–63 µm) and sand (>63 µm) present in each fraction. Grain size parameters including the graphical median, standard deviation, skewness and kurtosis after Folk and Ward (1957) were calculated using GRADISTAT (Blott and Pye, 2001).

A total of 15 lead-210 (²¹⁰Pb) and 4 accelerator mass spectrometry (AMS) radiocarbon (¹⁴C) bulk sediment samples were analysed at ANSTO. The age chronology for ²¹⁰Pb was determined by measuring radioisotopes (Po²¹⁰ and Ra²²⁶) for the top 20 cm of each core using alpha spectrometry following Harrison et al. (2003). The sedimentation rate and age was calculated using both the constant rate of supply (CRS; Appleby and Oldfield, 1978) and the constant initial concentration (CIC) models (supplementary material, available online; Goldberg et al., 1977; Robbins and Edgington, 1975). The CIC chronology was incorporated into age–depth models as the error margin was smaller and the sedimentation rate was relatively consistent across sites. Bulk sediment samples were selected from the basal unit of each core for radiocarbon dating. Organic fragments were removed and samples treated with acid and alkali to remove humic acids and carbonates before analysis (see Table 1). Radiocarbon dates were calibrated using the ShCal13 curve (Hogg et al., 2013). A smooth spline interpolation incorporating ²¹⁰Pb and ¹⁴C dates produced age–depth models using Clam (Blaauw, 2010) in RStudio (Chazdon et al., 2011). The stratigraphy and chronology for each site are shown in Figures 2 –4.

Table 1.

AMS ¹⁴C ages of bulk sediment samples from Marralda Swamp (MARR04), Well Swamp (WS01) and West Coast Swamp (WCS01). Dates were analysed at ANSTO and calibrated using Clam (Blaauw, 2010) and the ShCal13 calibration curve (Hogg et al., 2013). All calibrated ages are reported with a 95.4% probability distribution.

Site	Laboratory no.	Sample depth (cm)	Radiocarbon age BP	δ¹³C per mil	Calibrated age yr BP (95.4%)
MARR04	OZS567	35–36	570 ± 20	−23.7	554 (536) 515
MARR04	OZR035	46–47	1240 ± 30	−24	1257 (1114) 986
WS01	OZR037	48–49	420 ± 25	−19.4	502 (461) 328
WCS01	OZR036	48–49	575 ± 25	−16.9	560 (538) 510

Figure 2.

Sediment profile of MARR04 including high-resolution optical image and x-radiograph of the core provided by ITRAX. Age–depth models were constructed using Clam in RStudio (RStudio Team, 2015) with a smooth spline age–depth model and include CIC-modelled ²¹⁰Pb dates which have a lower error than CRS modelled dates. Grain size, median, sorting, skewness and kurtosis after Folk and Ward (1957) were calculated using GRADISTAT (Blott and Pye, 2001). Zones discussed throughout the text are identified based on changes in the grain characteristics and images.

Figure 3.

Sediment profile of WS01. Smooth spline age–depth models constructed using Clam in RStudio (RStudio Team, 2015) including CIC-modelled ²¹⁰Pb dates. Grain size, median, sorting, skewness and kurtosis after Folk and Ward (1957) were calculated using GRADISTAT (Blott and Pye, 2001). Zones discussed throughout the text are identified based on changes in the grain characteristics and images.

Figure 4.

Sediment profile of WCS01. Smooth spline age–depth models constructed using Clam in RStudio (RStudio Team, 2015) including CIC-modelled ²¹⁰Pb dates. Grain size, median, sorting, skewness and kurtosis after Folk and Ward (1957) were calculated using GRADISTAT (Blott and Pye, 2001). Zones discussed throughout the text are identified based on changes in the grain characteristics and images.

Geochemical analysis

All cores were scanned using the ITRAX µXRF scanner for sedimentological evaluation at ANSTO. ITRAX produces high resolution RGB and X-radiographic images which aid in identifying subtle changes in stratigraphy. Micro-XRF elemental profiles from aluminium to uranium were measured using a molybdenum (Mo) tube in 1 mm increments with an exposure time of 10 s per step. Data with significant variations in the total counts generated at the ends of core scans indicated gaps or unconsolidated material and were deleted presenting unnormalised data may primarily reflect changes in water, carbonates and organic content rather than environmental change due to the closed sum effect (Lowemark et al., 2011; Van der Weijden, 2002). Elemental data are normalised by the sum of incoherent and coherent scattering to compensate for grain size changes, water content and sediment density (Burn and Palmer, 2014; Kylander et al., 2011) before being smoothed using a five-point running mean. A correlation matrix of normalised data identifies the strength of association between elements within cores (supplementary material, available online). An unconstrained principal component analysis (PCA) explores relationships between elements within sites and was executed in RStudio (RStudio Team, 2015). Samples at 10 mm intervals were included in the PCA bi-plot to illustrate relationships between elements throughout the cores (Figure 5). Single elements (Mn, Ti, Fe and Ca) and ratios detecting changes in redox sensitive elements (Mn/Ti and Fe/Ti), local hydrology (Ca/Ti, Sr/Ti and Mn/Ti) and organic content (Mo incoherent/coherent) are shown in Figures 6 –8.

Figure 5.

Principle component analysis of µXRF data by depth. Data were analysed and graphed in RStudio (RStudio Team, 2015). The first two axes explain 70.5% of the variance in MARR04, 67.4% in WCS01 and 73% in WS01.

Figure 6.

MARR04’s age–depth profile of organic content (LOI), Ca, Mn, Ti and Fe. Ratios of Mo incoherent/coherent, Ca/Ti, Mn/Ti, Sr/Ti and Fe/Ti and zonations identified in the lithostratigraphic figure are included.

Figure 7.

WS01’s age–depth profile of organic content (LOI), Ca, Mn, Ti and Fe. Ratios of Mo incoherent/coherent, Ca/Ti, Mn/Ti, Sr/Ti and Fe/Ti and zonations identified in the lithostratigraphic figure are included.

Figure 8.

WCS01’s age–depth profile of organic content (LOI), Ca, Mn, Ti and Fe. Ratios of Mo incoherent/coherent, Ca/Ti, Mn/Ti, Sr/Ti and Fe/Ti and zonations identified in the lithostratigraphic figure are included.

Results

Stratigraphy and chronology

Radiocarbon and lead-210 ages provide a chronology for each of the sediment cores (see Table 1 and Figures 2 –4). Each core is divided into zones based on optical images, x-radiographs and lithogenic changes. Zone 1 in MARR04 (1300–750 cal. a BP) indicates a high energy environment with bimodal, very poorly sorted and coarse silt particles (Figure 2). In zone 2 of MARR04 (750–450 cal. a BP), the median grain size decreases and sediments become mostly unimodal and poorly sorted as the clay and sand component is replaced by silts. Organic material fluctuates in zone 3 of MARR04 (500 cal. a BP – AD 1960) with high organic content present in zone 4 (AD 1960 to present).

Zone 1 of WS01 (450–300 cal. a BP) is characterised by a decreasing median grain size with a coarse, very poorly sorted silt transitioning to a medium, poorly sorted silt (Figure 3). In zone 2 (300–50 cal. a BP), sediment briefly becomes a very coarse and poorly sorted silt accompanied by an abrupt change in the median grain size around 300 cal. a BP. WCS01 zone 1 (550–350 cal. a BP) and zone 2 (350–40 cal. a BP) contain oxidised clays with low organic content (Figure 4). Sediments remain medium, poorly sorted silt throughout the WCS01 record with median grain size increasing in zone 3 (40 cal. a BP to present). Zones 3 of WS01 (50 cal. a BP to present) and WCS01 (40 cal. a BP to present) are characterised by high organic content. The results are discussed using zonations identified in the lithostratigraphic units.

Geochemical analysis

Variations downcore in the elemental data produced by ITRAX µXRF analysis are useful indicators of palaeoenvironmental change. The correlation matrix identifies the closely associated elements (r value ≥ 0.7 or ≤ −0.7; supplementary material, available online). Ti, K and Fe are strongly correlated across all sites, associated with clay minerals and detrital input (Kylander et al., 2011). Sr and Ca are also strongly correlated at all sites, indicating silicate and carbonate weathering in the catchment and/or authigenic precipitation (Cohen, 2003).

PCA explores the trends in elemental composition within cores. The first two axes presented in Figure 5 explain 70.5% of the variance in MARR04, 67.4% in WCS01 and 73% in WS01. The bi-plot displays every 10 mm of data with elemental composition driving the distribution of samples. Detrital elements Ti, Fe, Zn, Rb, K and Zr define the base of all cores. The importance of lithogenic elements decreases at all sites and transitions to Ca, Sr and Mn exerting the most influence on samples. This change suggests the development of ephemeral, groundwater-fed wetlands as Ca and Sr minerals are produced authigenically during increased evaporative concentration and carbonate precipitation, and Mn deposits under oxic conditions (Kylander et al., 2011; Moreno et al., 2007). Br indicates increased organic content in lakes (Carrevedo et al., 2015; Gilfedder et al., 2011) and characterises the last zone of all cores suggesting increased productivity and possibly a wetter period. The PCA results from each core follow a similar pattern suggesting comparable stages of wetland development across the island.

Variations in elements through time are shown at MARR04 (Figure 6), WS01 (Figure 7) and WCS01 (Figure 8). Normalised element age profiles of Mn, Ca, Ti and Fe show similar trends across sites. In zone 1 of MARR04, Fe and Ti are initially high with low or no presence of Mn and Ca, despite the presence of shell hash. Ca and Mn increase significantly after 800 cal. a BP replacing Fe and Ti. Zone 1 of WS01 and WCS01 has high Fe and Ti compared to variations seen in MARR04, which declines towards the top of both cores. Zones 1 and 3 at WS01 have minimum values of Ca and Mn which increases in zone 2 (300–50 cal. a BP). In zone 2 of WCS01 (350–40 cal. a BP), high values of Ca and Mn decrease from 150 cal. a BP to minimum values in zone 3.

Ti is often used to normalise data as it indicates allochthonous inputs from the catchment (Cohen, 2003), immobile in sediments and resists post-depositional influences (Montero-Serrano et al., 2010). Ratios of Ca, Sr, Mn and Fe over Ti can identify within-site processes such as biogenic production or evaporative concentrations rather than detrital deposition (Davies et al., 2015). Mn/Ti and Ca/Ti profiles of all cores show a general increasing trend from the base. Significant peaks and troughs are seen in the ratios at WCS in the past 100 cal. a BP. Fe/Ti is highest in zone 1 of WCS01 and MARR04, while WS01 sees little change. LOI and Mo incoherent/coherent reflect the organic content of sediments, identifying areas where elemental results may be diluted. Organic content peaks at the top of all cores (between 5% and 20%) and may affect elemental trends (Calvert, 1983; Rollinson, 2014).

Discussion

Wetland development in the Holocene

Elemental data are used to reconstruct the palaeoenvironmental history of wetlands across Bentinck Island by recording relative changes in detrital inputs, sediment characteristics, redox conditions and hydrological indicators. Results are discussed for each of the three sites (MARR04, WS01 and WCS01) based on the zonations defined by the lithogenic chronology (Figures 3 –5). First, sediment characteristics and detrital elements are discussed together as they are closely linked, followed by the variations in redox sensitive elements which identify periods of aerobic and anaerobic environments in the wetlands history and indicate a mangrove phase in MARR04. Finally, elements representing site productivity and hydrology are discussed as important indicators of local site development and regional environmental change during the Holocene.

Sediment characteristics and detrital inputs: Fe, K, Rb, Si and Ti

The Wellesley Islands are built upon a lateritic formation composed of ferric oxides, aluminium and titanium bearing rocks (Gardner, 1957; Grimes, 1979). The eroding laterite is distributed by runoff in the wet season and strong southeasterly winds during the dry season. Fe, K, Ti and, at MARR04, Rb represent the terrigenous input into the catchment. Elements occurring in terrigenous silicates and oxides characterise the detrital component and are often associated with grain characteristics: Fe, K, Ti and Rb with clays and fine grained materials (Cuven et al., 2010; Kylander et al., 2011) and Si with coarse silt and sand. Detrital indicators are a proxy for changes in runoff within a catchment (Corella et al., 2012; Metcalfe et al., 2010), detrital input (Balascio et al., 2011) or aeolian deposition (Bakke et al., 2009).

The Marralda wetland on the south east coast of Bentinck Island is protected from strong southeasterly winds by a low beach ridge. Zone 1 of MARR04 (1250–750 cal. a BP) is characterised by >50% sand content, pisolith inclusion and marine shell hash (Figure 2). The correlation matrix (supplementary material, available online) and PCA of elemental profiles (Figure 5) at MARR04 find K, Ti, Fe, Zr and Rb correlate with silicon (Si) reflecting the silicate component of the basal unit which is absent from other more protected sites. The zone 1 sediment is very poorly sorted (5–6 σ_G) with a large median grain size (>20 µm) indicating rapid deposition of sediment with minimal sorting. The grain size distribution is very finely skewed (−0.5 Sk_G) and very platykurtic suggesting selective sifting of particles. Pisoliths are likely transported to the site via runoff or fluvial deposition. The presence of marine shells, grain size and sorting suggests zone 1 of MARR04 was a supratidal environment with wave-washed sediment deposition. Maximum Ti and Fe content in zone 1 indicates a greater contribution of detrital material and fine grained sediments, possibly from the river now further south discharging sediment onto tidal mudflats which later formed the Marralda wetlands (Figure 2). Alternatively, storm surges may have washed over a low frontal dune depositing sand, shell hash and detrital material in the developing swale. This interpretation is supported by the basal sediment found across the Marralda wetlands (see Moss et al., 2015).

Zone 2 (750–425 cal. a BP) of MARR04 is defined by significantly decreasing grain size and increasing silt (>50%) which suggests a low energy environment. In zone 2, Ti and Fe decrease with the clay content. Reduced counts of detrital elements and changes in the grain size support the hypothesis of river diversion, and seaward ridge formation caused a reduction in fluvial and marine input to the system (Woodroffe, 1992). Poorly sorted, symmetrical mesokurtic and leptokurtic grain distributions indicate aeolian sorted sediment deposition increased in zone 2. At MARR04, the declining importance of detrital components and increased aeolian input marks the boundary between zones 1 and 2.

WS01 (450 cal. a BP) and WCS01 (550 cal. a BP) are located behind a series of parallel dunes and swales within 1 km of the coast. Elemental and grain size data suggest comparatively stable systems without the significant changes in catchment morphology seen in the MARR04 record. Sediment characteristics of zone 1 WS01 (450–300 cal. a BP) is platykurtic with a very coarsely skewed (>0.5 Sk_G) profile indicating detrital material is the primary component and possible soil formation (Figure 3). WCS01 zone 1 (550–350 cal. a BP) remains poorly sorted, medium silt with platykurtic distribution, while kurtosis increases slightly from zone 1–3 suggesting shallowing of the system (Figure 4).

Ti and Fe peaks in zone 1 of WS01 and WCS01 are associated with slightly higher clay content and low median grain size. At WS01, sampling with a sand auger below the reach of the D-section revealed an unconsolidated, sandy sediment with shell hash and pisoliths compared to zone 1 of MARR04. Increased Ti is often cited as an indicator of increased rainfall or runoff (see Davies et al., 2015). However, high levels of Ti in WS01 and WCS01 between 500 and 300 cal. a BP indicates detrital input to sites from erosion and weathering of exposed lateritic bedrock. At WS01, a distinct peak in sand (>40%), median grain size and a shift to finely skewed grain distribution around 300 cal. a BP suggests a storm event on the north-west coastline. Interestingly, this is not recorded in the elemental data highlighting the importance of comparing sediment analysis to geochemical records. An increase in the grain size in zone 3 at WCS01 associated with sand content may indicate increasing erosion and transportation of sediment on the south coast in the last 100 years.

Redox conditions and mangrove environments: Fe/Ti and Mn/Ti

Post-depositional mobilisation of elements can affect the distribution of Fe and Mn within a sediment profile and overwhelm an environmental signal. In this study, Fe correlates with Ti across all sites and is primarily sourced from weathering of the lateritic bedrock. Sharp peaks in the Fe/Ti ratio identifies reducing conditions and diagenetic iron as Fe is redox sensitive and Ti is inert (Rothwell and Croudace, 2015). If Mn/Ti inversely correlates with Fe/Ti, the reduced iron species (Fe²⁺) likely donated an electron in the reduction of Mn. This reaction produces Mn²⁺ which is mobile and Fe³⁺ which forms immobile complexes in the sediment (Kylander et al., 2013). The dissimilatory reduction in Fe and Mn can occur in anaerobic environments such as those created by decomposing organic material consuming oxygen in sediments (Davison, 1993). Waterlogged sediments are also depleted in oxygen allowing redox transformation of susceptible elements to occur at the oxic/anoxic boundary.

In zone 1 of MARR04, Fe and Ti peaks with the high clay content (15–25%) and iron pisoliths are present. Fe/Ti peaks significantly, while Mn and Mn/Ti are absent in zone 1 indicating an anoxic event between 1250 and 800 cal. a BP. Mangrove forests are waterlogged anaerobic environments, with sulphate ions and reduced Fe³⁺ accumulating as iron sulphides in sediments (e.g. Berner, 1970; Clark et al., 1998; Wada and Seisuwan, 1986). Mangroves form on wave-dominated coasts where rivers discharge fluvial sediment onto the shoreline (Woodroffe, 1992). The PCA analysis of MARR04 data shows sulphur defines the sediment profile between 1250 and 800 cal. a BP, with a peak in the organic material (LOI), Mo ratio and clay content during the anoxic zone, suggesting a mangrove phase. In the Gulf of Carpentaria, the late-Holocene sea-level highstand (+2 m above pmsl) receded in the last c. 2000 years (Lewis et al., 2013; Nakada and Lambeck, 1989; Reeves et al., 2008). The basal sediment in the Marralda wetlands (both MARR04 and MARR01 published by Moss et al., 2015) initially records a supratidal environment, with palynological results finding a diverse mangrove community occurring around 500 cal. a BP. The mangrove community at MARR01 is replaced by a freshwater swamp dominated by Typha in the last ~70 years, suggesting variable timing in mangrove and swamp development across the catchment depending on proximity to the coast (Moss et al., 2015).

At WCS01 and WS01, the Fe/Ti and Mn/Ti records vary independently. At WCS01, between 550 and 450 cal. a BP, Fe/Ti is at its maximum and Mn/Ti at a minimum, as found in zone 1 of MARR04. However, there is no accompanying peak in organic material to suggest anaerobic conditions or a mangrove community.

Freshwater availability and wetland productivity: Ca, Mn and Sr

Freshwater is a scarce and valuable resource in the semi-arid South Wellesley Islands. Ca and Sr indicate regional carbonate weathering within the catchment or in situ precipitation of CaCO and co-precipitation of SrCO₃ during evaporative concentration and saturation. Wetlands on Bentinck Island are fed by groundwater high in Ca and Sr, which collects on the buried lateritic bedrock during the monsoon season and percolates through the porous soils, feeding freshwater sites during the dry season. Normalising Ca and Sr by Ti removes the silicate mineral component introduced with detrital material, with Ca/Ti and Sr/Ti indicating authigenic carbonates or changes in the catchments’ sediment source and size (Kylander et al., 2013). Small freshwater molluscs are occasionally present in the Marralda Wetland core and may contribute biogenic carbonates. However, Ca/Ti and Sr/Ti values are unaffected across sites by marine shells in the basal unit, and the ratios are thought to predominantly record authigenic carbonates. Variations in Mn can identify a climatic signal if unaffected by post-depositional processes. Except for zone 1 of MARR04, Mn/Ti varies independently from Fe/Ti in all cores suggesting the elemental signal is intact. Manganese oxides precipitate in aerobic conditions, migrating up to the oxic transition zone (Davison, 1993; Kylander et al., 2011; Lowemark et al., 2008). In a wetland environment, increased oxygenation may occur due to mixing by winds, reduced water levels or by excessive photosynthetic activity in the surface waters (Davison, 1993).

At MARR04, Ca and Sr significantly increase at 800 cal. a BP as the grain size and sediment source change, with an absence of shell hash. After 800 cal. a BP, Ca/Ti and Sr/Ti signals remain relatively high suggesting the site was fed by groundwater, with authigenic deposition of elements occurring during periods of low water levels or drying out of the wetlands. At the beginning of zone 2 (750–450 cal. a BP), wetter conditions and a more permanent wetland environment are suggested by relatively low Mn and Mn/Ti counts. Ca/Ti increases indicating a decrease in detrital input as the wetlands are supplied with groundwater rather than rainfall/runoff, and the catchments’ sediment source comes primarily from biotic processes within the site and aeolian deposition.

In zone 3 of MARR04 (450 cal. a BP – AD 1960), Ca/Ti, Sr/Ti and Mn/Fe ratios increase suggesting a drying phase encompassed the last 500 years. Mn/Ti peaks in zone 3 correlate with the Mo ratio and LOI results, indicating short periods of increased organic material possibly associated with a shallow vegetated wetland. At WS01, Ca/Ti and Sr/Ti increased from 300 cal. a BP. Fe and Ti increases between 200 and 150 cal. a BP with declining Sr/Ti, Ca/Ti and Mn/Fe, suggesting a period of increased water depth which is absent from other sites. Wetland indicators (Ca/Ti, Sr/Ti and Mn/Ti) at WCS01 begin to increase from 400 cal. a BP, significantly increasing in the last c. 100 years. The drying trend beginning 500 cal. a BP at MARR04 occurs at 150 and 100 cal. a BP at WS01 and WCS01, respectively, indicating shallower water resources across the island. From AD 1950 to present, cores record higher organic content across the island suggesting increased local biomass and productivity and a recent wet phase within wetlands.

Palaeoenvironmental context

Micro-XRF analysis of palaeoenvironmental records from the South Wellesley Islands records wetland development, local hydrological changes and regional climate signals during the late-Holocene. The three sites examined in this study began to accumulate sediment 1200 years ago with wetlands developing 800 and 400 cal. a BP. Initially, MARR04 and WS01 record a basal unit containing siliciclastic sediment with abundant marine shell fragments, suggesting a high-energy, coastal environment predates wetland development. The different timing of site development at MARR04 (1250 cal. a BP) and WS01 (475 cal. a BP) likely reflects local rates of coastal progradation, with the south east coast developing earlier due to the availability of sediment. The recent development of sites can be explained by a higher than present sea level in the mid- to late-Holocene, restricting coastal and wetland development. Sea levels higher than present dropped to modern levels only in the last 2000 years around northern and north-eastern Australia (Lewis et al., 2013; Sloss et al., 2007; Woodroffe, 2002). Hydro-isostatic flexing of northern Australia’s wide continental shelf is thought to explain the variable timing and magnitude of the mid-Holocene sea-level highstand and its subsequent fall to present levels (Chappell et al., 1982; Lambeck, 2002; Lambeck and Nakada, 1990; Lewis et al., 2013; Woodroffe, 1993). In the north west Joseph Bonaparte Gulf, sea levels reached modern heights around 6000 cal. a BP (Clarke and Ringis, 2000; Lewis et al., 2013; Yokoyama et al., 2000), followed by a 1–2 m highstand (Jennings, 1975; Lees, 1992; Lessa and Masselink, 2006). Beach ridges on the Sir Edward Pellew Group record emergence of around 1.6 m in the last 5100 years (Chappell et al., 1982). Close to the South Wellesley Islands, the supratidal chenier ridges at Karumba in the southern Gulf of Carpentaria suggest sea levels were +2.5 m above pmsl at 6400 cal. a BP before smoothly falling to modern levels after 1000 cal. a BP (Chappell et al., 1982; Lewis et al., 2013; Rhodes, 1982; Rhodes et al., 1980). The accuracy of chenier deposits to record past sea levels is questioned (Lewis et al., 2013); however, emergence in the southern and eastern parts of the Gulf of Carpentaria is supported by deltaic studies of the McArthur and Gilbert Rivers (Jones, 2003; Nott, 1996; Woodroffe and Chappell, 1993).

Hydro-isostatic adjustment can explain emergent shorelines in the Gulf of Carpentaria (Chappell, 1983; Chappell et al., 1982) and suggests that the Wellesley Islands experienced a sea-level highstand between 1 and 2 m which gradually fell to modern levels in the last 1000 years. As sea levels fell to modern levels, prograding coastlines developed and wetlands formed behind dunes, swales and cheniers (Chappell, 1988; Woodroffe, 1988). Geochemical records from Bentinck Island fit within this regional model, suggesting sea level reached modern datum in the last 1000 years in the southern Gulf of Carpentaria.

Palaeoenvironmental records from tropical northern Australia are often further removed from direct sea-level influence, either in distance or elevation, with sites recording both local geomorphic processes and climate change during the Holocene. Site development and expansion in the late-Holocene occurred during a period of intensified ENSO and increased variability in effective precipitation (Shulmeister, 1992; Shulmeister and Lees, 1995). On the Cape York Peninsula, swamp vegetation expanded after 2700 cal. a BP, suggesting an increase in freshwater availability (Butler, 1998; Stephens and Head, 1995). A recent study from Big Willum Swamp on the Cape York Peninsula found a permanent body of water developed in the last 2200 cal. a BP with the greatest depth and extent occurring in the last 600–400 years (Stevenson et al., 2015). Further north in the Torres Strait Islands, freshwater swamps developed by 2600 cal. a BP, with palynological records dominated by Melaleuca and the herbaceous swamp taxa Leptocarpus and Cyperaceae (Rowe, 2007a, 2015). Records from the Torres Strait Islands find dark, organic mud replaced a sandy profile within the last 981 ± 47 yr BP at Boigu Gawat 1 and approximately 671 ± 19 yr BP at Boigu Gawat 2, suggesting wetlands became more productive, organic material increased and coastal influence was reduced (Rowe, 2015). In the south west Gulf of Carpentaria on Vanderlin Island, a lake developed by 4500 cal. a BP with organic sediments increasing in the last 2000 years (Prebble et al., 2005). Further west in the Gulf of Carpentaria on Groote Eylandt, a seasonal swamp formed 5000 cal. a BP, with swamp indicators including Melaleuca, Restionaceae and Cyperaceae indicating lake expansion in the last 1000 years as effective precipitation increased (Shulmeister, 1992). Black Spring in the Kimberley’s fluvial sediments and pollen from the aquatic Triglochin increased from 1300 cal. a BP to present suggesting the modern summer Australian monsoon climate developed, increasing summer rainfall to the region (McGowan et al., 2012). Across the Magela floodplains, Clark and Guppy (1988) found mangrove forests transitioned to freshwater swamps by 1300 yr BP. The recent development of wetlands on Bentinck Island fits within the wider context of site development and increased effective precipitation in the last 1000 years across northern Australia. Geochemical analysis suggests freshwater wetlands developed in the last 800 years at MARR04 and across the island in the last 400 years. The regional development and expansion of freshwater sites found on Bentinck Island and across northern Australia is driven by sea-level fluctuation, coastal progradation and climate during the late-Holocene.

Conclusion

This paper presents high-resolution µXRF records from coastal wetlands in tropical northern Australia. Geochemical proxies and physical sediment characteristics from wetlands across the South Wellesley Islands reconstruct variations in the detrital input from local weathering and erosion (Fe, K, Rb, Ti and Zr), wetland productivity (Br, Ca, Mn, Mo ratio, Ca/Ti and Mn/Ti) and redox environments (Mn/Ti and Fe/Ti) during the late-Holocene. Results suggest sea-level fluctuation and coastal progradation determined the timing of wetland development on Bentinck Island, with sites transitioning from coastal environments to wetlands between 800 and 400 cal. a BP. The formation of wetlands reflects the regional trend of increasing freshwater availability across northern Australia during the late-Holocene.

Footnotes

Acknowledgements

The authors thank AINSE Ltd for providing financial assistance (award: John Ferris Memorial Scholar PGRA-10903) and Jack Goralewski and Atun Zawadzki for their assistance. The authors acknowledge the R team for the use of RStudio and thank Patricia Smith and Claire Kain for writing and refining the code. The authors acknowledge Kaiadilt traditional owners of the South Wellesley Islands as research partners. The Kaiadilt Aboriginal Corporation collaborated in establishing the research framework for this project. The authors also thank Sean Ulm, Craig Sloss, Daniel Rosendahl, Lincoln Steinberger, Duncan Kelly, Rene Simpson, Carl and Eunice Oberdorf, John and Melinda Barton, and Tex and Lyn Battle for support and advice.

Funding

This research was supported by the Australian Research Council’s Discovery Projects funding scheme (DP120103179).

References

Abram

McGregor

Gagan

. (2009) Oscillations in the southern extent of the Indo-Pacific Warm Pool during the mid-Holocene. Quaternary Science Reviews 28: 2794–2803.

Appleby

Oldfield

(1978) The calculation of lead-210 dates assuming a constant rate of supply of unsupported ²¹⁰Pb to the sediment. CATENA 5: 1–8.

Bakke

Lie

Heegaard

. (2009) Rapid oceanic and atmospheric changes during the Younger Dryas cold period. Nature Geoscience 2: 202–205.

Balascio

Zhang

Bradley

. (2011) A multi-proxy approach to assessing isolation basin stratigraphy from the Lofoten Islands, Norway. Quaternary Research 75: 288–300.

Berner

(1970) Sedimentary pyrite formation. American Journal of Science 268: 1–23.

Blaauw

(2010) Methods and code for ‘classical’ age-modelling of radiocarbon sequences. Quaternary Geochronology 5: 512–518.

Blott

Pye

(2001) GRADISTAT: A grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surface Processes and Landforms 26: 1237–1248.

Boyle

(2002) Inorganic geochemical methods in palaeolimnology. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments. Dordrecht: Springer, pp. 83–141.

Bureau of Meteorology (2014a) Historical Impacts in the Gulf of Carpentaria. Melbourne: Bureau of Meteorology. Available at: http://www.bom.gov.au/cyclone/history/gulf.shtml

10.

Bureau of Meteorology (2014b) Monthly Climatic Statistics for Mornington Island, Queensland. Melbourne: Bureau of Meteorology. Available at: http://www.bom.gov.au/climate/averages/tables/cw_029039.shtml

11.

Burn

Palmer

(2014) Solar forcing of Caribbean drought events during the last millennium. Journal of Quaternary Science 29: 827–836.

12.

Butler

(1998) Environmental change in the Quaternary In: David

(ed.) Ngarrabullgan: Geographical Investigations in Djungan Country, Cape York Peninsula. Melbourne: Monash University, Monash Publications in Geography and Environmental Science, pp. 78–97.

13.

Calvert

(1983) Geochemistry of Pleistocene sapropels and associated sediments from the Eastern Mediterranean. Oceanologica Acta 6: 255–267.

14.

Carrevedo

Frugone

Latorre

. (2015) A 700-year record of climate and environmental change from a high Andean lake: Laguna del Maule, central Chile (36°S). The Holocene 25: 956–972.

15.

Chappell

(1983) Evidence for smoothly falling sea-level relative to north Queensland, Australia, during the past 6000 yr. Nature 302: 406–408.

16.

Chappell

(1988) Geomorphological dynamics and evolution of tidal river and floodplain systems in northern Australia. Northern Australia: Progress and Prospects 2: 34–57.

17.

Chappell

Rhodes

Thom

. (1982) Hydro-isostasy and the sea-level isobase of 5500 B.P. in north Queensland, Australia. Marine Geology 49: 81–90.

18.

Chazdon

Chao

Colwell

. (2011) A novel statistical method for classifying habitat generalists and specialists. Ecology 92: 1332–1343.

19.

Clark

McConchie

Lewis

. (1998) Redox stratification and heavy metal partitioning in Avicennia-dominated mangrove sediments: A geochemical model. Chemical Geology 149: 147–171.

20.

Clark

Guppy

(1988) A transition from mangrove forest to fresh-water wetland in the monsoon tropics of Australia. Journal of Biogeography 15: 665–684.

21.

Clarke

JDA

Ringis

(2000) Late Quaternary stratigraphy and sedimentology of the inner part of southwest Joseph Bonaparte Gulf. Australian Journal of Earth Sciences 47: 715–732.

22.

Cohen

(2003) Paleolimnology: The History and Evolution of Lake Systems. New York: Oxford University Press.

23.

Corella

Brauer

Mangili

. (2012) The 1.5-ka varved record of Lake Montcortès (southern Pyrenees, NE Spain). Quaternary Research 78: 323–332.

24.

Croudace

Rindby

Rothwell

(2006) ITRAX: Description and evaluation of a new multi-function X-ray core scanner. Geological Society of London, Special Publications 267: 51–63.

25.

Crowley

(1996) Late quaternary mangrove distribution in northern Australia. Australian Systematic Botany 9: 219–225.

26.

Crowley

Gagan

(1995) Holocene evolution of coastal wetlands in wet-tropical northeastern Australia. The Holocene 5: 385–399.

27.

Crowley

Anderson

Kershaw

. (1990) Palynology of a Holocene marine transgressive sequence, lower Mulgrave River Valley, north-east Queensland. Australian Journal of Ecology 15: 231–240.

28.

Cuven

Francus

Lamoureux

(2010) Estimation of grain size variability with micro X-ray fluorescence in laminated lacustrine sediments, Cape Bounty, Canadian High Arctic. Journal of Paleolimnology 44: 803–817.

29.

Davies

Lamb

Roberts

(2015) Micro-XRF core scanning in palaeolimnology: Recent developments. In: Croudace

Guy Rothwell

(eds) Micro-XRF Studies of Sediment Cores. Dordrecht: Springer, pp. 189–226.

30.

Davison

(1993) Iron and manganese in lakes. Earth-Science Reviews 34: 119–163.

31.

Folk

Ward

(1957) Brazos River bar: A study in the significance of grain size parameters. Journal of Sedimentary Research 27: 3–26.

32.

Francus

Lamb

Nakagawa

. (2009) The potential of high-resolution X-ray fluorescence core scanning: Applications in paleolimnology. PAGES (Past Global Changes) News 17: 93–95.

33.

Gagan

Hendy

Haberle

. (2004) Post-glacial evolution of the Indo-Pacific Warm Pool and El Niño-Southern oscillation. Quaternary International 118–119: 127–143.

34.

Gardner

(1957) Laterite in Australia. Unpublished record no. 67, Australian Bureau of Mineral Resources. Available at: https://d28rz98at9flks.cloudfront.net/10266/Rec1957_067.pdf.

35.

Gilfedder

Petri

Wessels

. (2011) Bromine species fluxes from Lake Constance’s catchment, and a preliminary lake mass balance. Geochimica et Cosmochimica Acta 75: 3385–3401.

36.

Goldberg

Gamble

Griffin

. (1977) Pollution history of Narragansett Bay as recorded in its sediments. Estuarine and Coastal Marine Science 5: 549–561.

37.

Grimes

(1979) Mornington-Cape Van Dieman Queensland: 1:250,000 Geological Series Explanatory Notes. Canberra: Australian Government Publishing Service.

38.

Grimes

Sweet

(1979) Westmoreland Queensland: 1:250,000 Geological Series Explanatory Notes. Canberra: Australian Government Publishing Service.

39.

Grindrod

Rhodes

(1984) Holocene sea-level history of a tropical estuary: Missionary Bay, north Queensland. In: Thom

(ed.) Coastal Geomorphology in Australia. London: Academic Press, pp. 151–178.

40.

Grindrod

Moss

Van der Kaars

(1999) Late Quaternary cycles of mangrove development and decline on the north Australian continental shelf. Journal of Quaternary Science 14: 465–470.

41.

Harrison

Heijnis

Caprarelli

(2003) Historical pollution variability from abandoned mine sites, Greater Blue Mountains World Heritage Area, New South Wales, Australia. Environmental Geology 43: 680–687.

42.

Haug

Hughen

Sigman

. (2001) Southward migration of the intertropical convergence zone through the Holocene. Science 293: 1304–1308.

43.

Hogg

Hua

Blackwell

. (2013) SHCal13 Southern Hemisphere calibration, 0–50,000 years cal BP. Radiocarbon 55: 1–15.

44.

Jennings

(1975) Desert dunes and estuarine fill in the Fitzroy estuary (North-Western Australia). CATENA 2: 215–262.

45.

Jones

Woodroffe

Martin

(2003) Deltas in the Gulf of Carpentaria, Australia: Forms, processes, and products. In: Hasan Sidi

Nummedal

Imbert

. (eds) Tropical Deltas of Southeast Asia; Sedimentology, Stratigraphy and Petroleum Geology. SEPM Special Publication, 76. Tulsa, OK: Society for Sedimentary Geology, pp. 21–43.

46.

Kylander

Ampel

Wohlfarth

. (2011) High-resolution X-ray fluorescence core scanning analysis of Les Echets (France) sedimentary sequence: New insights from chemical proxies. Journal of Quaternary Science 26: 109–117.

47.

Kylander

Klaminder

Wohlfarth

. (2013) Geochemical responses to paleoclimatic changes in southern Sweden since the late glacial: The Hasseldala Port lake sediment record. Journal of Paleolimnology 50: 57–70.

48.

Lambeck

(2002) Sea level change from Mid-Holocene to recent time: An Australian example with global implications. Ice Sheets, Sea Level and the Dynamic Earth 29: 33–50.

49.

Lambeck

Nakada

(1990) Late Pleistocene and Holocene sea-level change along the Australian coast. Global and Planetary Change 89: 143–176.

50.

Lees

(1992) The development of a chenier sequence on the Victoria Delta, Joseph Bonaparte Gulf, northern Australia. Marine Geology 103: 215–224.

51.

Lessa

Masselink

(2006) Evidence of a mid-Holocene sea level highstand from the sedimentary record of a macrotidal barrier and paleoestuary system in northwestern Australia. Journal of Coastal Research 22: 100–112.

52.

Lewis

Sloss

Murray-Wallace

. (2013) Post-glacial sea-level changes around the Australian margin: A review. Quaternary Science Reviews 74: 115–138.

53.

Wheeler

Meinke

. (2007) Probabilistic forecasts of the onset of the north Australian wet season. Monthly Weather Review 135: 3506–3520.

54.

Lowemark

Chen

Yang

. (2011) Normalizing XRF-scanner data: A cautionary note on the interpretation of high-resolution records from organic-rich lakes. Journal of Asian Earth Sciences 40: 1250–1256.

55.

Lowemark

Jakobsson

Morth

. (2008) Arctic Ocean manganese contents and sediment colour cycles. Polar Research 27: 105–113.

56.

Luly

Grindrod

Penny

(2006) Holocene palaeoenvironments and change at Three-Quarter Mile Lake, Silver Plains Station, Cape York Peninsula, Australia. The Holocene 16: 1085–1094.

57.

McBride

Nicholls

(1983) Seasonal relationships between Australian rainfall and the Southern Oscillation. Monthly Weather Review 111: 1998–2004.

58.

McGowan

Marx

Moss

. (2012) Evidence of ENSO mega-drought triggered collapse of prehistory Aboriginal society in northwest Australia. Geophysical Research Letters 39: L22702.

59.

Markgraf

Díaz

(2000) The past ENSO record: A synthesis. In: Diaz

Markgraf

(eds) El Nino and the Southern Oscillation: Multiscale Variability and Global and Regional Impacts. Cambridge: Cambridge University Press, pp. 465–488.

60.

Memmott

Round

Rosendahl

. (2016) Fission, fusion and syncretism: Linguistic and environmental changes amongst the Tangkic people of the southern Gulf of Carpentaria, northern Australia. In: Verstraete

J-C

Hafner

(eds) Land and Language in Cape York Peninsula and the Gulf Country (Culture and Language Use, Vol. 18). Amsterdam: John Benjamins Publishing Company, pp. 105–136.

61.

Metcalfe

Jones

Davies

. (2010) Climate variability over the last two millennia in the North American Monsoon region, recorded in laminated lake sediments from Laguna de Juanacatlan, Mexico. The Holocene 20: 1195–1206.

62.

Montero-Serrano

Palarea-Albaladejo

Martín-Fernández

. (2010) Sedimentary chemofacies characterization by means of multivariate analysis. Sedimentary Geology 228: 218–228.

63.

Moreno

Giralt

Valero-Garcés

. (2007) A 14 kyr record of the tropical Andes: The Lago Chungará sequence (18°S, northern Chilean Altiplano). Quaternary International 161: 4–21.

64.

Moss

Petherick

Neil

(2011) Environmental change at Myora Springs, North Stradbroke island over the last millennium. Proceedings of the Royal Society of Queensland 117: 113–140.

65.

Moss

Mackenzie

Ulm

. (2015) Environmental context for Late-Holocene human occupation of the South Wellesley Archipelago, Gulf of Carpentaria, northern Australia. Quaternary International 385: 136–144.

66.

Nakada

Lambeck

(1989) Late Pleistocene and Holocene sea-level change in the Australian region and mantle rheology. Geophysical Journal International 96: 497–517.

67.

Nicholls

(1992) Historical El Nino/Southern oscillation variability in the Australasian region. In: Diaz

Markgraf

(eds) El Niño: Historical and Paleoclimatic Aspects of the Southern Oscillation. Cambridge: Cambridge University Press, pp. 151–173.

68.

Nott

(1996) Late Pleistocene and Holocene sea-level highstands in northern Australia. Journal of Coastal Research 12: 907–910.

69.

Prebble

Sim

Finn

. (2005) A Holocene pollen and diatom record from Vanderlin Island, Gulf of Carpentaria, lowland tropical Australia. Quaternary Research 64: 357–371.

70.

Proske

Heslop

Haberle

(2014) A Holocene record of coastal landscape dynamics in the eastern Kimberley region, Australia. Journal of Quaternary Science 29: 163–174.

71.

Reeves

Bostock

Ayliffe

. (2013) Palaeoenvironmental change in tropical Australasia over the last 30,000 years – A synthesis by the OZ-INTIMATE group. Quaternary Science Reviews 74: 97–114.

72.

Reeves

Chivas

Garcia

. (2008) The sedimentary record of palaeoenvironments and sea-level change in the Gulf of Carpentaria, Australia, through the last glacial cycle. Quaternary International 183: 3–22.

73.

Rhodes

(1982) Depositional model for a chenier plain, Gulf of Carpentaria, Australia. Sedimentology 29: 201–221.

74.

Rhodes

Polach

Thom

. (1980) Age structure of Holocene coastal sediments-Gulf of Carpentaria, Australia. Radiocarbon 22: 718–727.

75.

Robbins

Edgington

(1975) Determination of recent sedimentation rates in Lake Michigan using Pb-210 and Cs-137. Geochimica et Cosmochimica Acta 39: 285–304.

76.

Rollinson

(2014) Using Geochemical Data: Evaluation, Presentation, Interpretation. London: Routledge.

77.

Rothwell

Croudace

(2015) Twenty years of XRF core scanning marine sediments: What do geochemical proxies tell us? In: Croudace

Rothwell

(eds) Micro-XRF Studies of Sediment Cores: Applications of a Non-Destructive Tool for the Environmental Sciences. Dordrecht: Springer, pp. 25–102.

78.

Rowe

(2007a) A palynological investigation of Holocene vegetation change in Torres Strait, seasonal tropics of northern Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 251: 83–103.

79.

Rowe

(2007b) Vegetation change following mid-Holocene marine transgression of the Torres Strait shelf: A record from the island of Mua, northern Australia. The Holocene 17:927–937.

80.

Rowe

(2015) Late-Holocene swamp transition in the Torres Strait, northern tropical Australia. Quaternary International 385: 56–68.

81.

RStudio Team (2015) RStudio: Integrated Development for R. Boston, MA: RStudio, Inc. Available at: http://www.rstudio.com/.

82.

Ryzak

Bieganowski

(2011) Methodological aspects of determining soil particle-size distribution using the laser diffraction method. Journal of Plant Nutrition and Soil Science 174: 624–633.

83.

Shulmeister

(1992) A Holocene Pollen record from Lowland Tropical Australia. The Holocene 2: 107–116.

84.

Shulmeister

(1999) Australasian evidence for mid-holocene climate change implies precessional control of Walker Circulation in the Pacific. Quaternary International 57–58: 81–91.

85.

Shulmeister

Lees

(1992) Morphology and chronostratigraphy of a coastal dunefield; Groote Eylandt, northern Australia. Geomorphology 5: 521–534.

86.

Shulmeister

Lees

(1995) Pollen evidence from tropical Australia for the onset of an ENSO-dominated climate at C 4000 BP. The Holocene 5: 10–18.

87.

Sloss

Murray-Wallace

Jones

(2007) Holocene sea-level change on the southeast coast of Australia: A review. The Holocene 17: 999–1014.

88.

Stephens

Head

(1995) Palaeoecology of archaeological and swamp sites in S.E. Cape York Peninsula. Tempus 3: 18–32.

89.

Stevenson

Brockwell

Rowe

. (2015) The palaeo-environmental history of Big Willum Swamp, Weipa: An environmental context for the archaeological record. Australian Archaeology 80: 17–31.

90.

Stott

Cannariato

Thunell

. (2004) Decline of surface temperature and salinity in the western tropical Pacific Ocean in the Holocene epoch. Nature 431: 56–59.

91.

Sturman

Tapper

(1996) The Weather and Climate of Australia and New Zealand. Melbourne: Oxford University Press.

92.

Turney

CSM

Hobbs

(2006) ENSO influence on Holocene aboriginal populations in Queensland, Australia. Journal of Archaeological Science 33: 1744–1748.

93.

Ulm

Evans

Rosendahl

. (2010) Radiocarbon and linguistic dates for occupation of the South Wellesley Islands, Northern Australia. Archaeology in Oceania 45: 39–43.

94.

Van der Weijden

(2002) Pitfalls of normalization of marine geochemical data using a common divisor. Marine Geology 184: 167–187.

95.

Wada

Seisuwan

(1986) The process of pyrite formation in mangrove soils. In: Dorst

(ed.) Selected Papers of the Dakar Symposium on Acid Sulphate Soils. Wagening: IRLI Publications, pp. 24–37.

96.

Williams

Santoro

Smith

. (2008) The impact of ENSO in the Atacama Desert and Australian Arid Zone: Exploratory time-series analysis of Archaeological records. Chungara-Revista De Antropologia Chilena 40: 245–259.

97.

Williams

Ulm

Goodwin

. (2010) Hunter-gatherer response to Late-Holocene climatic variability in northern and central Australia. Journal of Quaternary Science 25: 831–838.

98.

Woodroffe

(1988) Changing mangrove and wetland habitats over the last 8000 years, Northern Australia and Southeast Asia. Northern Australia: Progress and Prospects 2: 1–33.

99.

Woodroffe

(1992) Mangrove sediments and geomorphology. In: Robertson

Alongi

(eds) Tropical Mangrove Ecosystems. Washington, DC: American Geophysical Union, pp. 7–41.

100.

Woodroffe

(1993) Late Quaternary evolution of coastal and lowland riverine plains of southeast-Asia and northern Australia – An overview. Sedimentary Geology 83: 163–175.

101.

Woodroffe

(2002) Coasts: Form, Process and Evolution. Cambridge: Cambridge University Press.

102.

Woodroffe

Chappell

(1993) Holocene emergence and evolution of the McArthur River delta, southwestern Gulf of Carpentaria, Australia. Sedimentary Geology 83: 303–317.

103.

Woodroffe

Thom

Chappell

(1985) Development of widespread mangrove swamps in mid-Holocene times in Northern Australia. Nature 317: 711–713.

104.

Woodroffe

Chappell

Thom

. (1989) Depositional model of a macrotidal estuary and floodplain, South Alligator River, Northern Australia. Sedimentology 36: 737–756.

105.

Woodroffe

Thom

Chappell

. (1987) Relative sea-level in the South Alligator River region, north-Australia, during the Holocene. Search 18: 198–200.

106.

Yokoyama

Lambeck

De Deckker

. (2000) Timing of the Last Glacial Maximum from observed sea-level minima. Nature 406: 713–716.