Untangling geochronological complexity in organic spring deposits using multiple dating methods
Introduction
Perennial or ephemeral springs are important palaeoenvironmental archives, particularly in otherwise arid or semi-arid environments (Boyd and Luly, 2005). The use of springs for palaeoenvironmental research requires the preservation of a record within the sediments or organic deposits that accumulate around the spring vent. Organic deposits are particularly valuable in this context as they preserve a range of palaeoecological proxies, including palynomorphs (pollen and non-pollen), diatoms, molluscs, charcoal, plant macrofossils and humic acids (e.g. Scott, 1982a, Van de Geer et al., 1986, Dodson and Wright, 1989, Boyd, 1990b, Macphail et al., 1999, Owen et al., 2004, Dobrowloski et al., 2005, Dobrowolski et al., 2012, McGowan et al., 2012, Fluin et al., 2013, Backwell et al., 2014, Dobrowolski et al., 2016, Field et al., 2017). Typically, organic sediments such as peat are formed in saturated and low oxygen conditions since this minimises the rate of decomposition. However, in many springs artesian pressure sustains a water-table “dome” higher than the surrounding terrain, allowing organic material to accumulate above ground at spring vents when evaporation rates are lower than spring discharge (Ponder, 1986, Boyd, 1990a, McCarthy et al., 2010). Whilst under artesian pressure, capillary creep and slow outward diffusion of groundwater throughout these deposits is sufficient to maintain moisture levels minimising oxidation and desiccation (McCarthy et al., 2010). In addition to palaeoecological indicators organic spring deposits can preserve a range of other palaeoenvironmental information such as stable isotope ratios, geochemical tracers and rates of dust deposition through time (e.g. Dobrowloski et al., 2005, Dobrowolski et al., 2012, Dobrowolski et al., 2016, McGowan et al., 2012, Mazurek et al., 2014).
In arid and semi-arid environments establishing continuous records of Quaternary environmental change is challenging due to the lack of perennial water sources (e.g. lakes) that preserve robust sedimentary records. Records from those water bodies that do exist, such as ephemeral lakes, commonly contain chronological hiatuses, for example during dry periods when deposition may cease, or unconformities due to aeolian deflation and scouring during heavy rainfall and floods (Head and Fullager, 1992, Magee et al., 2004). In addition, these environments have limited fossil preservation due to oxidation during dry periods. As such, important palaeoenvironmental and climatic information has been derived from organic springs in the arid and semi-arid environments of Australia (e.g. Dodson and Wright, 1989, Boyd, 1990b, McGowan et al., 2012, Fluin et al., 2013, Field et al., 2017), Kenya (e.g. Owen et al., 2004), Jordan (Rambeau, 2010) and South Africa (e.g. Scott, 1982a, Scott, 1988, Scott, 1982b, Scott and Vogel, 1983, Van Zinderen Bakker, 1989, Scott and Cooremans, 1990, Scott and Nyakale, 2002, Truc et al., 2013), with some records spanning up to ∼37,000 years (e.g. Fluin et al., 2013). For instance, pollen analysis at Wonderkrater spring mound has yielded a ∼35,000 year record of climatic evolution (Scott, 1982b), and more recently quantitative climatic estimates from pollen-transfer functions (Truc et al., 2013), whilst diatom-transfer functions from a spring at the Campground Wetlands in Kenya have been used to provide pH, conductivity and temperature estimates for part of the late Holocene (Owen et al., 2004). Spring deposits have also been shown to be viable traps for aeolian dust. For example, a ∼6500 year record of dust accumulation was constructed for Black Springs, one of the study sites investigated in this paper (McGowan et al., 2012). In that study, patterns in dust flux through time were used to investigate changing moisture regimes which were linked to changing patterns in the Australian monsoon. Outside of arid and semi-arid regions, springs within larger mire complexes have also been used to reconstruct valuable palaeoenvironmental records, for example in Poland (e.g. Dobrowloski et al., 2005, Dobrowolski et al., 2012, Dobrowolski et al., 2016, Mazurek et al., 2014) and Australia (e.g. Van de Geer et al., 1986, Macphail et al., 1999), with some records encompassing at least ∼52,000 years (e.g. Van de Geer et al., 1986). Since springs are a focus of human and faunal activity in water scarce regions they are also an important source of palaeontological and archaeological information. Consequently, occupational and faunal remains have been found surrounding springs (e.g. Hughes and Lambert, 1985, Veth, 1989) and within their sediments (e.g. Potezny, 1978, Kuman and Clarke, 1986, Backwell et al., 2014) up to ∼138,000 years before present (BP) (e.g. Backwell et al., 2014). Springs have particular importance in arid Australia and feature prominently in the songlines and myths of Indigenous Australians (Harris, 2002). They were also of strategic importance to early European explorers in Australia, acting as vital “stepping stones” across the arid interior of the continent (Harris, 1981).
Despite the potential of organic spring deposits as excellent archives of palaeoenvironmental and climatic information, there are a number of complications in the application of standard radiocarbon (14C) techniques. This, in many cases, has led to confusing chronologies in springs (e.g. Scott and Vogel, 1983, Kuman and Clarke, 1986, Scott, 1987, Boyd, 1990a, Macphail et al., 1999, Scott et al., 2003, Dobrowolski et al., 2012, Mazurek et al., 2014, Field et al., 2017) (see Section 1.2; Fig. 2, Fig. 3) which may compromise the integrity of environmental, climatic, palaeontological and archaeological reconstructions.
The aim of this study was to establish a protocol for building reliable chronologies for organic spring deposits through the analysis of dates generated using multiple methodologies, utilising springs from the Kimberley region of northwestern Australia as an example. The Kimberley is also of importance as there remains a limited understanding of palaeoenvironmental and climatic change in Australia's tropical savannah (Reeves et al., 2013). It is also an area within which multiple organic springs have been identified (e.g. DEC, 2012, Field et al., 2017) and is an environment where there are few other opportunities to construct detailed chronological records of the recent past. Ages from three springs in the Kimberley are obtained via accelerator mass spectrometry (AMS) and standard radiometric 14C dating with acid-alkali-acid (AAA) and hydrogen pyrolysis (HyPy) pre-treatments on a number of carbon fractions (pollen concentrate, macro-charcoal, bulk organics, stable polycyclic aromatic carbon (SPAC) and roots). For the uppermost sediments, chronological control is attempted using lead-210 (210Pb) and plutonium-239 + 240 (239+240Pu) analysis. Supporting chronological information is provided by analysis of the polonium-210 (210Po) and radium-226 (226Ra) activities of groundwater, macro-charcoal investigation, and the application of luminescence techniques at high spatial resolution (natural sensitivity-corrected luminescence (Ln/Tn)). The ability to build reliable chronologies for organic spring sediments is important to enable researchers to obtain accurate, high resolution palaeoclimate and environmental records. This is of particular concern in arid and semi-arid regions which are typically challenging for Quaternary research. Additionally, reliable chronologies from organic spring deposits will facilitate improved interpretation of the age of cultural and faunal artefacts found in their sediments.
Within organic springs there are a number of processes that have the potential to complicate the geochronological integrity of sedimentary deposits (Fig. 1). These processes include aeolian and alluvial input, root growth, upwelling groundwater, weathering, bioturbation, and mass movement within the spring.
Allochthonous transport of biological material to the spring by aeolian or alluvial processes has the potential to introduce “old” carbon, affecting 14C ages. This can include “old” charcoal which may remain in the surrounding landscape due to its presumed long-term environmental persistence, despite some evidence suggesting that it may not be as stable as originally thought (see Bird, 2007 and references within). The protuberant morphology of the spring mound above the surrounding landscape reduces the likelihood of “old” carbon washing into the spring sediments. However, alluvial input may occur during the spring's early stages before development of a mound (Ponder, 1986, Field et al., 2017), or during large floods when the mound is small but prone to inundation. Over time charcoal in the landscape may be broken down into finer fractions (Walker, 2005, Bird, 2007) which can then be blown into the spring during any stage of its development. Transport of “old” carbon into springs by these mechanisms may compromise their age structure.
Perennial springs, or those where oxygenated ground water supports enhanced plant growth, can contain deep rooted vegetation (Scott and Vogel, 1983) introducing young carbon into the spring deposits (Boyd, 1990b). Depending on the fraction dated, this young carbon cannot easily be removed from other biological fractions since separation methods (e.g. density separation or acid digestion) will also affect the organics targeted for 14C dating. Removing rootlets by hand is extremely time consuming and not possible in all cases. Rootlets can also penetrate large charcoal fragments selected for dating by growing along the original wood fibres (Harkness et al., 1994), thereby compromising the age of the charcoal fragment. In addition, the decay of roots can facilitate transport of young biological material or sediment via root channels affecting geochronology. These root channels may also allow the movement of other radiological sediment markers such as 210Pb and 239+240Pu to shift downwards in the profile.
In springs upwelling groundwater is related to meteoric recharge (McCarthy et al., 2010). During periods of decreased rainfall a lower water table can facilitate desiccation of the organic sediments (Scott, 2016). If springs then become dry, aeolian deflation may remove part of the deposit, while onset of rainfall could result in downward transport of sediment and biological substances through the peat matrix. In addition, springs may also experience cracking and self-mulching leading to mixing and overturning of mound stratigraphy. Alternatively, during periods of high spring discharge there is potential for upward movement of sediment and biological material, whilst sediments may also be removed via scouring during the spring's early stages before a mound develops. These processes may cause 14C, 210Pb, fallout radionuclide and luminescence results to be unreliable.
Water table fluctuations can cause in-situ alterations to charcoal preserved in the sediments (e.g. Bird et al., 2002) affecting returned 14C ages. Because of the large surface area and porous nature of charcoal it lends itself to adsorption of various organic and inorganic compounds (Bird, 2007). This makes it an ideal substrate for microbial colonisation which allows continued carbon cycling between fossilised charcoal and the modern environment (Zackrisson et al., 1996), and facilitates replacement of the original carbon by various oxides and oxihydroxides (Bird et al., 2002). Charcoal may also undergo oxidative degradation (Cohen-Ofri et al., 2006), particularly in coarser matrixes, as a result of bacterial activity and weathering, which may occur during periods when the water table is lower. It has been demonstrated that removal of charcoal contamination with 14C pre-treatments such as AAA or even the more rigorous acid-base-oxidation with stepped combustion (ABOX-SC) methodology is not always effective (e.g. Gillespie et al., 1992, Harkness et al., 1994, Bird et al., 2002, Higham et al., 2009a, Higham et al., 2009b), particularly in high weathering environments, where contaminants can irreversibly react with, and alter, the charcoal surface (Bird, 2007). Charcoal in organic spring deposits may therefore be contaminated with young carbon, particularly in cases where there have been large water table fluctuations.
Upwelling groundwater through organic spring deposits can also render them open systems in regards to uranium, which may be leached from host rocks/sediments and transported in solution within groundwater (Sirocko et al., 2007). This can affect U-series dating, including 210Pb. Organic decomposition products such as fulvic and humic acids are capable of adsorbing large quantities of uranium, which bind strongly to clay minerals in organic sediments (Szalay, 1969, Van der Wijk et al., 1986). Therefore the decomposition of organic matter and low redox potential of peat deposits make them ideal environments for the geochemical trapping of uranium (Shotyk, 1988). This can result in uranium-enriched organic deposits (e.g. Heijnis, 1992, Heijnis and van der Plicht, 1992, Zayre et al., 2006) even where relatively low concentrations of uranium are observed in the groundwater (e.g. Armands, 1967). Uranium and associated daughter products such as 210Po and 226Ra can therefore be high in organic spring sediments which may complicate the use of 210Pb as a geochronological tool by masking unsupported 210Pb activities.
Internal physical processes within springs can also influence their age stratigraphy. This includes slumping of the spring mound, if they develop too steeply, or if the vent position changes. As a focal point for animals, springs can also be affected by bioturbation (Scott, 2016). In Australia, this may be more pronounced after the arrival of Europeans and the introduction of domestic livestock, which occurred from the 1800s in the Kimberley.
Existing palaeoenvironmental records from organic springs overwhelmingly rely on 14C dating, however the majority of published 14C chronologies are complicated by age-depth reversals and/or erroneously young dates. This is demonstrated in Fig. 2, Fig. 3, which depict age-depth profiles characteristic of organic springs presented in earlier published works, both from isolated dryland springs (Fig. 2) and from springs within larger mire complexes (Fig. 3) (see figure captions for publication details). The plotted profiles come from a variety of global locations (Fig. 4), in which the carbon fraction dated consists predominantly of either bulk organics or pollen concentrate, with the exception of Dalhousie Springs, Scot, Meriba Springs and Rietvlei (Fig. 2A, G, 2I and 2J) where the carbon fraction dated is unknown, and Zawadówka, Mowbray Swamp and Pulbeena Swamp (Fig. 3M, T and 3U) where a number of other fractions were analysed including macrofossils, root stumps and humic extracts.
The chronologies of the plotted records are predominantly characterised by either sudden increases in age over narrow depth intervals (Fig. 2, Fig. 3N), unreasonably high sedimentation rates (Fig. 2, Fig. 3O), or more commonly age reversals (Fig. 2, Fig. 3 R). The latter implies complexity of carbon sources within the springs, while the former could represent contamination of the preceding layer with younger carbon, or perhaps indicate unconformities or growth hiatuses in the springs as a result of aeolian deflation or scouring. Age reversals are also seen in the Mowbray Swamp and Pulbeena Swamp chronologies (Fig. 3T and U) and whilst the authors note the presence of trace contamination at these sites, these reversals may be due in part to dates being near the 14C barrier (Van de Geer et al., 1986). There are also a number of examples where ages are contradictory between corresponding cores or pits surveyed at the same site and similar depth (Fig. 2B, C, 2G and 2L).
The extent of complications in 14C chronologies within organic springs is further illustrated by the fact that, to the best of our knowledge, there are just four published studies where the spring chronology is free from complications (Fig. 2, Fig. 3S). In one of these studies at Florisbad Spring, South Africa, an earlier study had previously found erratic 14C ages which were attributed to contamination of 14C by roots (Kuman and Clarke, 1986). In the later study of Scott and Nyakale (2002), a concerted effort to remove root fragments from bulk organic 14C was undertaken, resulting in a coherent age-depth profile for the site, i.e. ages increased with depth (Fig. 2K). Therefore, although existing studies have highlighted the value of organic spring deposits as palaeoenvironmental archives, to the best of our knowledge, in most cases the information contained is convoluted by the problematic geochronology.
Section snippets
Physical setting
Sediment cores from three springs in the northwest Kimberley are analysed in this study; Black Springs, Fern Pool and Gap Springs. The springs are located along a broadly north-south transect approximately 100 km in length (Fig. 5), with Black Springs furthest north at 15.633°S, 126.389°E, Fern Pool at 15.937°S, 126.284°E, and Gap Springs furthest south at 16.404°S, 126.134°E. The climate is monsoonal with a mean annual rainfall of ∼1000 mm/yr. When combined with a mean annual potential
Core collection and description
The three springs examined in this study (Black Springs, Fern Pool and Gap Springs) were cored during the Kimberley's dry season in June, 2015 using a 50 cm long Russian D-section corer. A core was also collected from Black Springs in 2005 (BSP00) and was previously described in McGowan et al. (2012) and Field et al. (2017). Ages from that core are presented here for comparison. Cores BSP02, FRN02 and ELZ01 were photographed on the ITRAX micro X-ray Fluorescence (μXRF) core scanner at the
Core lithology
Sedimentary sequences for BSP02, FRN02 and ELZ01 are shown in Fig. 6. The BSP02 core (242 cm) collected from Black Springs consists of approximately 80 cm of black, well decomposed peat (70% organic content by weight). Between 80 and 175 cm the organic content decreases down-core whilst the clay content increases. A marked change to organic silty clay occurs at 175 cm (organic content <20%), with the silty clay content increasing below 200 cm (organic content <10%).
The FRN02 core (204 cm)
The behaviour of 210Pb and 239+240Pu in organic spring deposits and their utility for developing chronologies
Used in combination, 210Pb ages and 239+240Pu activities have the potential to provide detailed chronologies for the past 100 years of spring development. However, in BSP02 use of 239+240Pu as a chronostratigraphic maker is complex because no clear 239+240Pu activity peak conforming to peak fallout at 1963 CE could be identified. Furthermore, the onset of atmospheric nuclear weapon testing and the corresponding appearance of a 239+240Pu signal found in southern hemisphere environmental archives
Conclusions
This study demonstrates that in complex hydro-geological settings such as springs, there is a need to understand how different dating components are behaving. It also shows value in utilising multiple radionuclides for understanding spring behaviour and developing accurate chronologies for their sediments. This is important since, despite the complexity of organic spring deposits as highlighted by this study, they are critical palaeoenvironmental archives in many settings. Therefore, developing
Acknowledgements
This research was supported by the Kimberley Foundation Australia with 2015 fieldwork and the majority of the laboratory research completed whilst the first author was the recipient of an Australian Postgraduate Award, an ANSTO Postgraduate Research Award (ANSTO PGRA No. 2158493), and a Kimberley Foundation Top-Up Scholarship. We acknowledge the financial support from the Australian Government for the Centre for Accelerator Science at ANSTO through the National Collaborative Research
References (94)
- et al.
The calculation of 210Pb dates assuming a constant rate of supply of unsupported 210Pb to the sediment
Catena
(1978) - et al.
Hydropyrolysis as a new tool for radiocarbon pre-treatment and the quantification of black carbon
Quat. Geochronol.
(2009) - et al.
Multiproxy record of late Quaternary climate change and Middle Stone Age human occupation at Wonderkrater, South Africa
Quat. Sci. Rev.
(2014) Charcoal
- et al.
Radiocarbon analysis of the early archaeological site of Nauwalabila I, Arnhem Land, Australia: implications for sample suitability and stratigraphic integrity
Quat. Sci. Rev.
(2002) - et al.
The efficiency of charcoal decontamination for radiocarbon dating by three pre-treatments – ABOX, ABA and HyPy
Quat. Geochronol.
(2014) Quaternary pollen analysis in the arid zone of Australia: Dalhousie Springs, Central Australia
Rev. Palaeobot. Palynology
(1990)- et al.
Nuclear data Sheets for A = 236
Nucl. Data Sheets
(2006) - et al.
Nuclear data Sheets for A = 239
Nucl. Data Sheets
(2014) - et al.
Modern and fossil charcoal: aspects of structure and diagenesis
J. Archaeol. Sci.
(2006)
Humid to arid to subhumid vegetation shift on Pilliga Sandstone, Ulungra springs, New south Wales
Quat. Res.
A late Quaternary record of monsoon variability in the northwest Kimberley, Australia
Quat. Int.
The ANTARES AMS facility at ANSTO
Nucl. Instrum. Methods Phys. Res. Sect. B
The uptake of uranium by organic substances in a peat bog environment on a granite bedrock
Chem. Geol.
Uranium/thorium dating of Late Pleistocene peat deposits in NW Europe, uranium/thorium isotope systematics and open-system behaviour of peat layers
Chem. Geol.
Problems with radiocarbon dating the middle to upper palaeolithic transition in Italy
Quat. Sci. Rev.
Depositional history of artificial radionuclides in the Ross Ice Shelf, Antarctica
Earth Planet. Sci. Lett.
The 240Pu/239Pu ratio, a potential geochronometer
Earth Planet. Sci. Lett.
The character, origin and palaeoenvironmental significance of the Wonderkrater spring mound, South Africa
J. Afr. Earth Sci.
Assessment of hydropyrolysis as a method for the quantification of black carbon using standard reference materials
Geochimica Cosmochimica Acta
Palynology and its application to geomorphology
Rapid equivalent dose estimation for aeolian dune sands using a portable OSL reader and polymineralic standardised luminescence growth curves: expedited sample screening for OSL dating
Quat. Geochronol.
Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol
Radiat. Meas.
Palaeoenvironmental change in tropical Australasia over the last 30,000 years - a synthesis by the OZ-INTIMATE group
Quat. Sci. Rev.
Determination of recent sedimentation rates in Lake Michigan using Pb-210 and Cs-137
Geochimica Cosmochimica Acta
Using simple portable OSL measurements and laboratory characterisation to help understand complex and heterogeneous sediment sequences for luminescence dating
Quat. Geochronol.
A late quaternary pollen record from the Transvaal Bushveld, South Africa
Quat. Res.
Late quaternary forest history in Venda, southern Africa
Rev. Palaeobot. Palynology
Fluctuations of vegetation and climate over the last 75000 years in the Savanna Biome, South Africa: Tswaing Crater and Wonderkrater pollen sequences reviewed
Quat. Sci. Rev.
Review of the inorganic geochemistry of peats and peatland waters
Earth Sci. Rev.
Quantification of climate change for the last 20,000 years from Wonderkrater, South Africa: implications for the long-term dynamics of the Intertropical Convergence Zone
Palaeogeogr. Palaeoclimatol. Palaeoecol.
Dating peat with U/Th disequilibrium: some geochemical considerations
Chem. Geol.
From carbon to actinides: a new universal 1MV accelerator mass spectrometer at ANSTO
Nucl. Instrum. Methods Phys. Res. B
A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols
Radiat. Meas.
Geochemical prospecting of a uraniferous bog deposit at Masugnsbyn, northern Sweden
Hydropyrolysis: implications for radiocarbon pre-treatment and characterization of black carbon
Radiocarbon
Stability of elemental carbon in a savanna soil
Glob. Biogeochem. Cycles
Mound springs
Inland mound springs
OxCal Program, V.4.2.2. Radiocarbon Accelerator Unit
Nuclear data Sheets for A = 137
Nucl. Data Sheets
Climate Data Online
Archaeogeology and Quaternary environment in the interior of southern Africa
Late quaternary environments in South Africa
High sensitivity analysis of Plutonium isotopes in environmental samples using Accelerator Mass Spectrometry (AMS)
J. Anal. Atomic Spectrom.
Understanding wetlands
Chronostratigraphy of calcareous mire sediments at Zawadówka (eastern Poland) and their use in paleogeographical reconstruction
Geochronometria
Cited by (12)
A radiocarbon chronology for Sanamere Lagoon, Cape York Peninsula, using multiple organic fractions
2022, Quaternary GeochronologyA late quaternary palaeoenvironmental record from Ntsikeni Wetland, KwaZulu-Natal Maloti-Drakensberg, South Africa
2022, Quaternary InternationalCitation Excerpt :The raw, uncalibrated AMS dates for the core range from 5750 ± 30 BP (6560-6410 cal BP) at a mean depth of 111.75 cm, to 21,190 ± 70 BP (25,645–25,345 cal BP) at a mean depth of 292.5 cm (Table 1). The date of 6390 BP (7320-7250 cal BP) for the sample representing 200–202.5 cm was identified as an outlier by BACON (Fig. 3), possibly due to contamination in-situ or during extrusion, or bioturbation (Field et al., 2018). The BACON model approximates a mean accumulation rate of 0.01 cm a−1 for the Ntsikeni sequence (Fig. 3).
Monsoon driven ecosystem and landscape change in the 'Top End' of Australia during the past 35 kyr
2021, Palaeogeography, Palaeoclimatology, PalaeoecologyExamining the response of an eastern Australian mangrove forest to changes in hydro-period over the last century
2020, Estuarine, Coastal and Shelf ScienceHolocene environmental changes in northern Poland recorded in alkaline spring-fed fen deposits – A multi-proxy approach
2019, Quaternary Science ReviewsCitation Excerpt :Results of the radiocarbon dating are presented in Table 1 and the final age-depth models are shown in Fig. 3. The bulk sediment dating, although sometimes complicated by the presence of inconsistencies like age reversals (Field et al., 2018), can lead to acceptable age-depth relationships when the dates are scrutinised in the light of lithology, geochemistry and palaeoenvironmental datasets (Dobrowolski et al., 2012, 2016). Several age reversals appeared in cores BOB-8 and OGA-7 and were subjected to critical evaluation before age-depth modelling (Fig. 3).
Palynology of Middle Stone Age spring deposits in grassland at the Florisbad hominin site, South Africa
2019, Review of Palaeobotany and PalynologyCitation Excerpt :com) (Scott and Rossouw, 2005; Rossouw, 2009). Constraints will, however, always be the poor pollen concentration in some layers and the accompanying problem of modern pollen and root contamination (Field et al., 2018). It is essential to couple such research with micromorhological studies of cracks by testing horizontal sections to determine how far the surface should be cut back in renewed sampling of pollen from existing exposures.