- Email: [email protected]
Fluvial sediments, correlations and palaeoenvironmental reconstruction: The development of robust radiocarbon chronologies
Journal of Archaeological Science 36 (2009) 2680–2688
Contents lists available at ScienceDirect
Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas
Fluvial sediments, correlations and palaeoenvironmental reconstruction: The development of robust radiocarbon chronologies Andy J. Howard a, *, Ben R. Gearey b, Tom Hill c, William Fletcher d, Peter Marshall e a
Institute of Archaeology & Antiquity, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK Birmingham Archaeo-Environmental, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK c School of Science and the Environment, Bath Spa University, Newton Park Campus, Bath, BA2 9BN, UK d Suffolk County Council Archaeological Service, Shire Hall, Bury St Edmunds, Suffolk, IP33 2AR, UK e Chronologies, 25 Onslow Road, Sheffield, S11 7AF, UK b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 September 2008 Received in revised form 21 July 2009 Accepted 31 August 2009
Multiple sequences of radiocarbon dates extracted from organic materials are increasingly being used to provide robust chronologies for landscape development, particularly the timing and correlation of significant climatic and ‘process’ events. Whilst the validity of using such frameworks in sedimentary environments such as lake basins and raised bogs has been the focus of much attention, such debates have not extended to fluvial systems. Using examples from three lowland, vertically accreting river valleys in East Anglia, UK, this paper assesses the robustness of their associated radiocarbon chronologies by assessing the contrasting age estimates that are obtained by dating different parts of the same organic sample (humic, humin and plant macrofossils) at a variety of stratigraphic levels. Overall, the humin and humic acid fraction results were statistically consistent, whilst the plant macrofossil remains were found to be of a slightly younger age. In these examples, it is argued that the younger ages appear to be the result of Phragmites (common reed) roots pushing plant macrossils through the sedimentary sequence or opening up voids for material to fall through, although studies in lacustrine and mire environments suggest alternative explanations may also be possible. Whichever explanation is preferred, this study demonstrates clearly that the complexity of valley floor stratigraphy and processes is such, that using single radiocarbon dates, whether AMS or bulk samples to reconstruct chronologies of ‘geomorphic system response’ may need to be refined and subjected to the same level of assessment that has been applied in other sedimentary systems. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Fluvial Geoarchaeology Holocene Radiocarbon
1. Introduction Within the geoarchaeological community, advances in the understanding of radiocarbon techniques (Bronk Ramsey, 2008a), as well as the application of advanced statistics such as Bayesian modelling to improve resolution (Blockley et al., 2004; Bronk Ramsey, 2008b) has resulted in the increasing use of these methods to provide chronologies for landscape development, particularly the timing and correlation of significant climatic and ‘process’ events. However, assessment of the robustness of primary radiocarbon data has varied significantly with respect to different depositional environments. A number of workers who have constructed chronologies from organic materials extracted from lacustrine environments (Turney et al., 2000; Wohlfarth et al.,
* Corresponding author. Tel.: þ44(0)121 414 8564; fax: þ44(0)121 414 3595. E-mail address: [email protected] (A.J. Howard). 0305-4403/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2009.08.006
1998), other small basins (Walker et al., 2001), more extensive lowland raised bogs (Lowe et al., 2004) and upland raised mosses/ soligenous mires (Shore et al., 1995) have all demonstrated the variability that can result from analyzing different parts of the carbon fraction (humic and humin) and different types of remains (plant macrofossils). However, such debates have not extended to the fluvial record where radiocarbon chronologies have become a key research tool in elucidating the timing of climate change and the role that human activity and changing natural cycles (flood frequency and magnitude) have played in Holocene river development not only in the UK (Chiverrell et al., 2007; Howard et al., 1999; Johnstone et al., 2006; Macklin et al., 2005; Tipping, 1995), but across other parts of Europe (Anderson et al., 2004; Macklin et al., 2006; Starkel et al., 2006; Thorndycraft and Benito, 2006) and the rest of the globe (Daniels and Knox, 2005; Knox, 2006). In terms of the British record, early syntheses of ‘Holocene system response’ were based almost entirely on databases
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
constructed from relatively few radiocarbon dated samples amassed from both published and unpublished literature pertaining to individual river catchments (Macklin and Lewin, 1993; Macklin, 1999). Usually, this material comprised ‘bulk samples’; or where individual plant fragments were analysed, little attention was paid to precisely what part of the plant was being dated. The growth in geoarchaeological studies of alluvial systems has provided the opportunity for more recent reviews to incorporate many more dates (Johnstone et al., 2006; Macklin et al., 2005; Macklin and Lewin, 2003) and apply a range of quantitative analyses, though current studies suggest that more robust methodologies are required to fully understand the potential of the data (Chiverrell et al., 2009). These latter age estimates are usually derived from identified fragments of material analysed by AMS. Furthermore, the validity of these age estimates is more rigorously examined in the light of advances in our understanding of the depositional context of dated materials (Lewin and Macklin, 2003; Lewin et al., 2005). However, despite such refinements to radiocarbon chronologies, few if any of these individual database samples from fluvial environments have been the subject of multiple dating programmes, which have analysed different parts of the carbon fraction (e.g. humic, humin, individual macrofossils) in an attempt to establish, which fraction may provide the most ‘robust’chronology for any given sequence. It is fair to point out that for the majority of geoarchaeological studies undertaken in alluvial environments over the past thirty years the aspiration to conduct large-scale dating programmes to produce robust radiocarbon chronologies has been unachievable through lack of funding. However, since 2003 the Aggregates Levy Sustainability Fund (ALSF), administered in part by English Heritage, has commissioned a significant number of geoarchaeological projects in aggregate-rich, river valley floors, which have included significant programmes of radiocarbon dating (Bayliss et al., 2007, 2008). Although the primary aims of these ASLF sponsored radiocarbon analyses has not been methodological development, the routine availability of multiple dates for individual fluvial sequences has provided an opportunity for palaeoenvironmentalists to become more critical of primary data in these demonstrably complex environments and to highlight anomalies in the record. In 2006, the analysis of thirty-six AMS radiocarbon samples across three river reaches as part of the Suffolk River Valleys Project produced a significant number of dating anomalies within seemingly secure biostratigraphic contexts (Hill et al., 2008a). A subsequent re-evaluation of the sequences using duplicate multifraction dating analyses aimed to explain these inconsistencies in the original record (Hill et al., 2008b) and this paper reports the results of this subsequent study. Whilst this project was not designed or conceived to test the validity of radiocarbon methodologies, the results have important implications for the use of such dating techniques in fluvial environments and highlights a number of issues that many geoarchaeologists and environmental archaeologists should consider, when using single dates for environmental reconstruction, or assembling multiple dates to develop chronologies of landscape development. 2. Methods The aim of the ALSF sponsored Suffolk River Valleys Project was to enhance the palaeoenvironmental record and hence curatorial knowledge of river valleys under threat from aggregate extraction and to dovetail this information with the cultural record. Work focused on reaches of the River Waveney at Beccles, the Blackbourn at Ixworth and the River Lark at Hengrave (Fig. 1). All of these rivers are typical of lowland river systems with cohesive channel banks
2681
with thick sequences of vertically accreted organic-rich sediments deposited by channels, which have been relatively stable within their respective valley floors throughout the Holocene (Howard and Macklin, 1999). Within each reach, single cores were extracted from organic sediments using a standard pattern Russian Corer. Each core was described using Troels-Smith (1955) system and assessed for a range of environmental proxies including pollen and insects. In order to provide a chronological framework, single AMS radiocarbon measurement was undertaken on plant macrofossil remains selected from key sedimentary horizons in each core. Dating of the three sequences produced anomalous results including inversions and modern ages deep within some of the sequences (Table 1). In the light of these results, a critical review by the project team and chronology specialists from English Heritage concluded that neither the field sampling methods nor laboratory protocols such as pre-treatment methods (see Cook et al., 1998) of well established radiocarbon facilities could adequately explain the majority of the anomalies observed within the dataset, especially since the pollen biostratigraphy at all of the sites appeared secure. Therefore, in 2007, English Heritage commissioned a second phase of the project with the specific aim of obtaining duplicate sample cores from the three sites and re-dating the sequences; fieldwork was undertaken during 2008 (Table 2). At each radiocarbon sample depth, individual plant macrofossil samples were collected in addition to bulk samples. At Beccles, new cores were taken at a distance of around 50 m (Beccles 2008 Core 1) and 100 m (Beccles 2008 Core 2) along a transect trending east of Beccles 2007 Core 1. At each core site, the upper 1.2 m of sediment was exposed through machine trenching and monolith tins were used to sample the sequence below the plough zone. Below this depth to the base of the gravels, samples were extracted using a standard pattern Russian Corer. At Hengrave, the new core (Hengrave 2008) was taken 2 m east of the initial core. A hand-dug trench was excavated through 0.38 m of ‘made ground’ to a depth of 0.65 m. Monolith tins were used to sample the underlying 0.27 m of well-humified peat with a further 3 m of peat collected to the top of the basal gravels using the standard pattern Russian corer. At Ixworth, a slot was opened to a depth of 0.5 m immediately adjacent to the 2006 intervention; the (2008) organic sequence was collected using a Russian corer to a depth of 2.4 m. In the laboratory, all the sampled sequences were described using Troels-Smith (1955) and each new core was deemed to stratigraphically correlate with those taken during Phase 1. It is recognized that replication of core sequences can be problematic since floodplains have added levels of complexity than either lacustrine or mire systems owing to the levels of energy and associated processes of erosion, reworking and redeposition of sediments. However, such complexity is likely to be significantly less in these lowland, essential stable, vertically accreting river systems than within their upland and piedmont equivalents (Brown et al., 1994; Howard and Macklin, 1999). As part of the second phase analyses, detailed notes were made recording the variation in abundance and diversity of plant macrofossil remains as well as the degree of humification. Prior to radiocarbon sample submission, duplicate samples were taken from each horizon and assessed for their pH, carbonate content and overall organic content, the latter determined by loss on ignition (Table 3). These tests were used to assess whether carbon content was unusually high or low indicating possible contamination through the influence of hard water and/or in wash of allochthonous material. Table 3 indicates that the carbonate content of the sediments was low and ph was generally between weak acid and neutral.
2682
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Fig. 1. Location of the study areas.
In total, fourteen macrofossil samples were submitted to the Scottish Universities Environmental Research Centre (SUERC), East Kilbride (Table 2). These included the dating of a number of duplicate macrofossil fragments from the same sample depth for radiocarbon comparison (notably in Beccles 2008 Cores 1 and 2). Samples were pre-treated by the acid-base-acid protocol (Stenhouse and Baxter, 1983) and CO2 obtained by combustion in precleaned sealed quartz tubes (Vandeputte et al., 1996). The purified CO2 was converted to graphite (Slota et al., 1987) for subsequent AMS analysis. The sample 14C/13C ratios were measured on the SUERC AMS, as described by Xu et al. (2004). In addition, twelve bulk peat samples (weighing between 72 g and 103 g) were submitted to the Centre for Isotope Studies, University of Groningen, The Netherlands (GrN). These latter samples were pre-treated using the acid/alkali/acid method (Mook and Waterbolk, 1985) and measured using gas proportional counting (Mook and Streurman, 1983). In all cases the acid insoluble/alkali soluble (humic acid) and alkali/acid insoluble (humin) fractions of the samples were separated after pre-treatment, combusted and measured. Each separation was carried out in a quantitative manner thus the total budget of carbon in the peat sample was conserved within the component fractions recovered for gas proportional counting. Both laboratories maintain continual programmes of quality assurance procedures, in addition to participation in international inter-comparisons (Scott, 2003). 3. Results The results of these new analyses are provided in Table 2 and Figs. 2–5 and are quoted in accordance with the international standard known as the Trondheim Convention (Stuiver and Kra, 1986). They are conventional radiocarbon ages (Stuiver and Polach, 1977) and have been calibrated using the curves of Reimer et al. (2004) and the computer program OxCal (4.0.5; Bronk Ramsey, 1995, 1998, 2001, 2009). The calibrated date ranges cited in the text and tables are those for 95% confidence. They are quoted in the
form recommended by Mook (1986), with the end points rounded outwards to 10 years. The ranges in Table 2 have been calculated according to the maximum intercept method (Stuiver and Reimer, 1986). A summary of the radiocarbon dating results for Beccles (2008) Core 1 is provided in Fig. 2 and Table 2. All submitted plant macrofossil samples were alder (Alnus) remains (twig/wood fragments). Duplicate radiocarbon dating of separate macrofossil samples was undertaken at both 0.84 m and 3.30 m depths. The four results from the 0.84 m level are statistically consistent and could be of the same actual age (T0 ¼ 6.9; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson,1978). At 3.30 m, the four measurements are not statistically consistent (T0 ¼ 347.155; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson, 1978) and thus represent material of different ages. However, both the humin/humic acid fractions (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8) and two Alnus fragments (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8) are statistically consistent. At 4.60 m, the three measurements are not statistically consistent (T0 ¼ 2576.412; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978) indicating that the sample comprises material of different ages. In all three samples, the humin acid contains most of the carbon and therefore has the greatest influence on a combined age. Dating results for Beccles (2008) Core 2 are provided in Fig. 3 and Table 2. Duplicate radiocarbon dating of separate macrofossil samples was undertaken at 1.37 m depth. At the 1.37 m level, all four measurements are not statistically consistent (T0 ¼ 84.1; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson, 1978). Separated by sample type, this indicates that the peat fractions are not statistically consistent (T0 ¼ 63.1; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978), whilst the two Alnus twigs are: (T0 ¼ 1.5; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). With the humic acid faction removed, the three remaining measurements from 1.37 m are statistically consistent (T0 ¼ 1.6; n ¼ 2; T0 (5%) ¼ 6; Ward and Wilson, 1978). Again at 3.59 m, the three measurements are not statistically consistent (T0 ¼ 47.663; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978), although the humin and humic acid fractions of the peat sample do show consistency (T0 ¼ 0.0; n ¼ 1; T0 (5%) 3.8: Ward and Wilson, 1978). At 4.30 m, the
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
2683
Table 1 Phase 1 Radiocarbon dates from the Suffolk River Valleys Project. Sample ID
Lab code
Material
d13C (%)
Radiocarbon Age (BP)
Calibrated date (95% confidence)
Beccles 1–0.84 m Beccles 1–0.99 m Beccles 1–1.15 m Beccles 1–1.18 m A Beccles 1–1.18 m B Beccles 1–1.99 m Beccles 1–2.02 m Beccles 1–3.50 m Beccles 1–5.34 m Beccles 2–1.34 m Beccles 2–2.22 m Beccles 2–2.50 m Beccles 2–2.54 m Beccles 2–2.83 m Hengrave-0.26 m Hengrave-0.59 m Hengrave-0.99 m Hengrave-1.50 m Hengrave-1.63 m
GrA-33471 GrA-33472 SUERC-12035 SUERC-12036 GrA-33473 GrA-33475 SUERC-12037 GrA-33476 GrA-33477 SUERC-12038 GrA-35050 SUERC-12039 GrA-33479 GrA-35067 SUERC-12027 GrA-35051 SUERC-12028 GrA-33481 SUERC-12029
27.3 27.3 24.2 28.1 28.3 27.6 27.5 28.2 27.1 27.9 27.7 24.2 29.7 26.4 27.8 28 25.9 27.4 27.4
2080 50 2090 70 1595 35 1975 35 2215 40 2695 40 2835 35 2785 40 9960 130 915 35 225 40 1770 35 105.45 1.09%Mod. 1445 40 125 35 1025 45 955 35 1965 40 1750 35
350 cal BC–cal AD 30 360 cal BC–cal AD 60 cal AD 390–550 50 cal BC–cal AD 120 390–170 cal BC 920–790 cal BC 1120–900 cal BC 1030–830 cal BC 10040–9220 cal BC cal AD 1020–1220 cal AD 1530–1950 cal AD 130–380 cal AD 1956–1957 cal AD 540–660 cal AD 1660–1955 cal AD 890–1150 cal AD 1010–1170 50 cal BC–cal AD 130 cal AD 210–390
Hengrave-1.99 m Hengrave-2.32 m Hengrave-2.56 m Hengrave-2.99 m Ixworth-0.86 m Ixworth-1.37 m Ixworth-1.40 m Ixworth-1.49 m Ixworth-2.49 m Ixworth-2.63 m Ixworth-3.44 m
GrA-35054 SUERC-12030 GrA-33482 SUERC-12031 GrA-35055 SUERC-12021 GrA-35056 SUERC-12025 GrA-33485 SUERC-12026 GrA-33483
Alnus glutinosa, small wood fragment Alnus glutinosa, small wood fragment Poaceae fragments Unidentified wood Poaceae fragments and internode Alnus glutinosa, stem Bark fragment, inidentified Alnus glutinosa, stem unidentified plant fragments Poaceae fragment unidentified plant fragments herbaceous stems Monocotyledon stem unidentified plant fragments Poaceae stems Poaceae stems Poaceae fragments Poaceae stems and internode Unidentified plant remains cf. seed/ flower head Poaceae stems Poaceae fragments Poaceae fragments unidentified plant stems unidentified seed Poaceae fragments Poaceae fragments unidentified wood fragments Alnus glutinosa, stem Alnus glutinosa, wood fragment Alnus glutinosa, stem
plant macrofossil sample did not produce enough carbon for dating purposes. However, the humin and humic acid fractions did and are statistically consistent (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). In all three samples of Beccles (2008) Core 2, the humin contains most of the carbon and therefore has the greatest influence on a combined age. At Hengrave (Fig. 4; Table 2), all three measurements from the 0.47 m level are statistically inconsistent (T0 ¼ 12.198; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978), although the plant macrofossil and humin fraction are statistically consistent (T0 ¼ 1.4; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). At 1.61 m, the humin and humic acid fractions are statistically consistent (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978), whilst the plant macrofossil sample yielded too little carbon for analysis. At 2.76 m, the three measurements are not statistically consistent (T0 ¼ 21.697; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson,1978), although the humin and humic acid fractions are statistically consistent (T0 ¼ 0.2; n ¼ 1; (5% 3.8)). Two of the samples show the humin fraction contains most of the carbon and therefore has the greatest influence on a combined age. At Ixworth (Fig. 5; Table 2), the high level of humification encountered within the organic deposits prevented the separation and dating of macroscopic plant remains. A single fragment of Alnus was initially identified from within the basal sample (2.39 m depth), but yielded too little carbon for analysis. However at 0.71 m, 1.24 m and 2.39 m, the humin and humic acid fractions are statistically consistent (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8; T0 ¼ 1.5; n ¼ 1; T0 (5%) ¼ 3.8; T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8 respectively; Ward and Wilson, 1978). Two of the samples show the humin fraction contains most of the carbon and therefore has the greatest influence on a combined age. The very low organic content of the sample from 1.24 m (Table 3/LOI) however does not seem to have had an influence of the % carbon content.
27 26.1 27.3 27.3 26.8 27 25.4 27.4 29 26.6 27.7
1740 45 1000 35 1620 35 1720 35 1935 40 1560 35 1465 45 2905 35 6265 45 5980 40 9900 60
cal AD 130–420 cal AD 980–1160 cal AD 340–540 cal AD 230–420 40 cal BC–cal AD 140 cal AD 410–590 cal AD 530–660 1260–990 cal BC 5330–5070 cal BC 4990–4770 cal BC 9660–9250 cal BC
4. Discussion The second phase results of radiocarbon dating of samples from the top, middle and bottom of each core proved to be more successful than those obtained during the first phase, primarily because of an improved understanding of sediment stratigraphy, including the composition of organic horizons. Overall, the results for all three sites are more robust with no age inversions or modern age estimates being encountered within the sequences. In general, it appears that the humin and humic acid fraction dating results are statistically consistent, whilst the plant macrofossil remains were found to be of a slightly younger age. Within the literature, whilst a number of studies have assessed the importance of different components of organic material (humic, humin, plant macrofossils) to the accuracy and precision of radiocarbon dating, there is a level of debate regarding which is the most significant. For example, whilst this study suggests that humin is particularly important in shaping the chronology of alluvial valley floors, Shore et al. (1995), found that the humic acid component contained most of the carbon from upland peat samples recovered from Lanshaw Moss and White Moss (West Yorkshire). The differences observed between these lowland river valleys in Suffolk and upland sites in northern England might be explained by the contrasting environmental settings, geomorphic processes and erosion and sedimentation histories of the sites (i.e. soligenuous mire and raised mire complexes being compared directly with floodplain environments). However, unfortunately, whilst a number of studies have identified similar issues and are not unique to this paper (Bartley and Chambers, 1992; Johnson et al., 1990; Lowe et al., 2004; Walker et al., 2001; Wohlfarth et al., 1998), with the exception of this paper, all other studies discussing this conundrum have focused on mires, blanket peats and lacustrine basins of varying size where
2684
Table 2 Phase 2 Radiocarbon dates from the Suffolk River Valleys Project. Sample ID
Lab code
Material
Organic Content
pH
d13C (%)
Radiocarbon age (BP)
Weighted mean
Calibrated date (95% confidence)
Beccles 1–0.84 m Beccles 1–0.84 m Beccles 1–0.84 m A
GrN-31116 GrN-31151 SUERC15973 SUERC15974 GrN-31117 GrN-31152 SUERC15975 SUERC15976 GrN-31118 GrN-31153 SUERC15981
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus twig,
79%
3.90
28.9 28.9 29.5
2130 40 2160 50 2065 35
2142 32 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
360–50 cal BC
28.1
2015 40
28 28.7 30.8
4590 30 4590 50 3885 35
30.7
3970 35
Peat (Humin) Peat (Humic acid) Plant macrofossil: cf. Alnus twig
72%
28.4 28 28.4
8460 50 8340 80 5660 35
GrN-31119 GrN-31154 SUERC15982 SUERC15983 GrN-31120 GrN-31155 SUERC15984 GrN-31121 GrN-31156 GU-6796
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus 1 growth ring Plant macrofossil: Alnus 1 growth ring Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus roundwood, c. 8 growth Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus 1 growth ring
76% twig,
28.7 28.6 28.7
2230 30 1830 40 2275 35
390–200 cal BC cal AD 70–320 400–210 cal BC
twig,
28.4
2215 35
390–180 cal BC
28.6 28 29
5060 30 5060 40 4765 35
5060 24 BP (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8)
7735 35 BP (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8)
GrN-31113 GrN-31148 SUERC16385 GrN-31114 GrN-31149 GU-6786 GrN-31115 GrN-31150 SUERC15972
Peat (Humin) Peat (Humic acid) Plant macrofossil: stem fragment
GrN-31110 GrN-31145 GrN-31111 GrN-31146 GU-6798 GrN-31112 GrN-31147
Peat (Humin) Peat (Humic acid) Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus wood Peat (Humin) Peat (Humic acid)
Beccles 1–0.84 m B Beccles 1–3.30 m Beccles 1–3.30 m Beccles 1–3.30 m A
Beccles 1–4.60 m Beccels 1–4.60 m Beccels 1–4.60 m Beccels 2–1.37 m Beccles 2–1.37 m Beccles 2–1.37 m A Beccles 2–1.37 m B Beccles 2–3.59 m Beccles 2–3.59 m Beccles 2–3.59 m Beccles 2–4.30 m Beccles 2–4.30 m Beccles 2–4.30 m Hengrave-0.47 m Hengrave-0.47 m Hengrave-0.47 m Hengrave-1.61 m Hengrave-1.61 m Hengrave-1.61 m Hengrave-2.76 m Hengrave-2.76 m Hengrave-2.76 m Ixworth-0.71 m Ixworth-0.71 m Ixworth-1.24 m Ixworth-1.24 m Ixworth-1.24 m Ixworth-2.39 m Ixworth-2.39 m
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus wood
78%
6.20
Plant macrofossil: Alnus wood
70%
4.80
5.60
6.00
160–70 cal BC 4590 26 BP (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8)
3500–3340 cal BC 2480–2210 cal BC 2580–2350 cal BC
8427 43 BP (T0 ¼ 1.6; n ¼ 1; T0 (5%) ¼ 3.8)
7580–7370 cal BC 4560–4400 cal BC
3960–3785 cal BC 3650–3380 cal BC
rings 31%
4.30
27.6 27.7
7740 40 7720 70 Sample failed
44%
6.00
60%
6.40
29.1 29.7 25.1
715 30 540 40 660 35
28.9 28.5
1430 35 1450 30 Sample failed 2310 40 2340 60 2095 35
1442 23 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
cal AD 570–655
2319 34 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
410–360 cal BC
1740 35 1830 40 2670 40 2730 40 Sampled failed 7530 50 7510 50
1779 27 BP (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8) 2700 29 BP (T0 ¼ 1.1; n ¼ 1; T0 (5%) ¼ 3.8)
cal AD 130–340
7520 36 BP (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8)
6460–6260 cal BC
twig, c.
Peat (Humin) Peat (Humic acid) Plant macrofossil: herbaceous stem Peat (Humin) Peat (Humic acid) Plant macrofossil: monocot culm
47%
5.70 29.8 30.5 27.5
55%
5.60
14%
7.00
53%
5.90
29.6 29.2 29.3 29.3 28.9 28.3
6640–6480 cal BC
cal AD 1260–1380 cal AD 1300–1450 cal AD 1270–1400
210–1 cal BC
910–800 cal BC
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Beccles 1–3.30 m B
Plant macrofossil: Alnus twig
190–10 cal BC
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688 Table 3 Summary of Loss on Ignition (LOI), carbonate content and pH measurements of the 12 sedimentary sampled for radiocarbon dating during Phase 2. For LOI, samples were oven dried at 105 degrees for 12hrs, organic content was then fired at 550 degrees for 4 hours; carbonate content was fired at 950 degrees for 2 hours. Measurements for pH were achieved following the procedure of Catt (1990). Sample No.
Sample Name
Organic content %
Carbonate content %
pH
1 2 3 4 5 6 7
Hengrave 0.47 m Hengrave 1.61 m Hengrave 2.76 m Ixworth 0.71 m Ixworth 1.24 m Ixworth 2.39 m Beccles 2008 Core 1: 0.84 m Beccles 2008 Core 1: 3.30 m Beccles 2008 Core 1: 4.60 m Beccles 2008 Core 2: 1.37 m Beccles 2008 Core 2: 3.59 m Beccles 2008 Core 2: 4.30 m
43.98 59.22 46.60 54.88 14.75 52.79 79.12
2.76 2.68 2.91 2.56 1.34 3.35 2.98
6.04 6.43 5.65 5.64 6.95 5.82 3.85
78.25
2.73
6.15
71.77
3.61
4.83
76.71
3.45
5.61
70.25
1.74
5.98
31.35
1.36
4.29
8 9 10 11 12
geomorphological and depositional processes are relatively simple. Until further methodological research is undertaken with respect to radiocarbon chronologies in a range of complex sedimentary systems, comparison of such datasets between environments must be approached with caution. Within Beccles (2008) Core 1, it was notable that in two cases (3.30 m and 4.60 m) Alnus fragment(s) were younger than the bulk sediment measurements from the same level. The stratigraphic consistency of both sets of data when analysed independently thus raises the possibility that either could be accurate. However,
2685
detailed sedimentary analysis revealed that Phragmites remains were present in the units immediately overlying both of the horizons in which the dating discrepancies were evident. This therefore raises the possibility that the Alnus fragments were intrusive and that the bulk samples provide the reliable age estimates. The invasive nature of Phragmites roots suggests that they have the strength to push larger organic material (e.g. small twigs, seeds) through the sediment or provide (root) cavities for material to fall down during dry periods. Deep penetration of Alnus roots into floodplain soils may be in response to down-draw of the water table (see Brayshay and Dinnin, 1999). As a cautionary note, it is possible that the Alnus fragments provide the true age and that the bulk sediment measurements are inaccurate; however, the consistency of the humic and humin measurements suggest that this is not the case since if the Alnus ages were indeed correct, the humin fraction ages would be closer in age to the macrofossil dates since the humins are composed of organic detritus. At Hengrave, the Alnus fragment from 2.76 m was also younger than the bulk sediment sample from the same horizon although the offset between the bulk and macrofossil dating is noticeably much smaller. This can perhaps be explained by the much lower incidence of wood remains from this core, although there is a much greater incidence of Phragmites. The overall consistency between humin and humic acid fractions is amply demonstrated by Beccles 2008 (Core 2) and suggests that these data provide the more reliable age estimates for each dated sample depth. At Hengrave, there is also statistical consistency between the humin and humic acid fractions analysed at 1.61 m and 2.76 m. The consistency of humic and humin fractions almost certainly reflects the stability of floodplain water-tables preventing movement of such fractions in this part of the vadose zone in these studies. In contrast however, explanations for the age difference between the humic acid and humin/Alnus fragment at 0.47 m depth may include the upward movement of humic acid or
Fig. 2. Probability distributions of dates from Beccles (2008) Core 1. Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
2686
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Fig. 3. Probability distributions of dates from Beccles (2008) Core 2. Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
the intrusion of younger rootlets from above. However, the caveats explored above with respect to the variability of humic and humin fractions must still apply. This study demonstrates very clearly that the complexity of valley floor stratigraphy and processes is such, that using single radiocarbon dates, whether AMS or bulk samples to reconstruct chronologies of ‘geomorphic system response’ may be unreliable. Whilst more recent studies have demonstrated the need for caution
when considering precisely what event/sedimentary context is being dated by the submission of a particular sample (Lewin and Macklin, 2003; Lewin et al., 2005), little consideration has been given to anomalies that might be introduced from the actual fractions of material being dated. Whilst such methodological considerations of radiocarbon dating have been undertaken in environments that might be classified as geomorphologically ‘less complex’ (Lowe et al., 2004; Walker et al., 2001), these studies have
Fig. 4. Probability distributions of dates from Hengrave (2008). Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
2687
Fig. 5. Probability distributions of dates from Ixworth (2008). Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
demonstrated that the issue requires more research, especially comparison between differing sedimentary environments. However, equally, this paper adds to the wider radiocarbon debate by suggesting that the dating of multiple fractions within the same sample can yield consistent and reliable results, which can be used to reconstruct valley floor chronologies. Furthermore, advances in the application of Bayesian approaches, especially using the framework advocated recently by Chiverrell et al. (2009) provide an opportunity to improve the quality and application of such datasets significantly.
5. Conclusions In geoarchaeological and palaeoenvironmental research, radiocarbon dates are used all too frequently to reconstruct landscape chronologies without full consideration being given to the validity of the data used. Whilst researchers working in selected environments including lake basins and mire complexes have sought to examine the robustness of radiocarbon chronologies and integrity of the data being collected, those working within fluvial systems have been less rigorous in considering the differences in dating results that can be obtained by analyzing different fractions of carbon (humic, humin and plant macrofossils). This paper has described the chronological problems encountered during one particular dating programme and demonstrated the value of taking a multi-fraction approach to radiocarbon analysis. In this particular study, it is suggested that plant macrofossils of Alnus produced consistently younger dates than either the humic and humin fractions. Whilst a number of other studies have demonstrated the variable importance of humic and humin fractions, associated sedimentary evidence in this instance suggest that the intrusion of plant macrofossils into older sediments may be a result of Phragmites root penetration. On this evidence, it is suggested that where there is evidence of Phragmites within sedimentary sequences, macrofossils should only be dated if there is evidence that they have grown in-situ and the submission of unidentified plant remains, monocot, Poaceae fragments and other similar material should be avoided. In this example, the statistical consistency
present between the dating results from the humin and humic acid fractions suggests that the dating of bulk radiometric sediment samples may provide accurate age estimates. Bulk AMS size samples are very unreliable since a small amount of contamination has a disproportionate effect compared with a radiometric sized sample. With the exception of this study, the corpus of literature addressing the issue of organic fractions used for radiocarbon dating has all been undertaken in environments that might be classified as geomorphologically ‘less complex’ (Lowe et al., 2004; Walker et al., 2001). If this issue is to be addressed more fully, our study clearly demonstrates the need for more research in a range of sedimentary environments of varying complexity, especially where groundwater chemistry and flow-paths may vary. Finally, this research has important implications for environmental reconstruction and indicates that single dates should never be used as the basis for the development of robust chronologies. Acknowledgements This research was funded by the Aggregates Levy Sustainability Fund administered by English Heritage as part of the Suffolk River Valleys Projects (PN 4772; PN 4772/ANL). Particular thanks go to the following EH staff for supporting our work: Kath Buxton, Dr Jen Heathcote, Dr Jane Sidell, Tom Cromwell and John Ette. Thanks are also due to Derek Hamilton (formerly of the EH Dating Team). Site access was granted by Suffolk Wildlife Trust, Beccles Town Council and Mr Philip Aitken. The comments of three anonymous referees greatly improved the clarity of this manuscript. References Anderson, E., Harrison, S., Passmore, D.G., Mighall, T., Wathan, S., 2004. Late Quaternary river terrace development in the Macgillycuddy0 s Reeks, southwest Ireland. Quaternary Science Reviews 23, 1785–1801. Bartley, D.D., Chambers, C., 1992. A pollen diagram, radiocarbon ages and evidence of agriculture on Extwistle Moor, Lancashire. New Phytologist 121, 311–320. Bayliss, A., Cook, G., Bronk Ramsey, C., van der Plicht, J., McCormac, G., 2008. Radiocarbon Dates from Samples Funded by English Heritage Under the Aggregates Levy Sustainability Fund 2004-7. English Heritage, Swindon.
2688
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Bayliss, A., Bronk Ramsey, C., Cook, G., van der Plicht, J., 2007. Radiocarbon Dates from Samples Funded by English Heritage Under the Aggregates Levy Sustainability Fund 2002-4. English Heritage, Swindon. Blockley, S.P.E., Lowe, J.J., Walker, J.J., Asioli, A., Trincardi, F., Coope, G.R., Donahue, R.E., Pollard, A.M., 2004. Bayesian analysis of radiocarbon chronologies: examples from the European Lateglacial. Journal of Quaternary Science 19, 159–175. Brayshay, B.A., Dinnin, M.H., 1999. Integrated palaeoecological evidence for biodiversity at the floodplain-forest margin. Journal of Biogeography 26, 115–131. Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51, 337–360. Bronk Ramsey, C., 2008a. Radiocarbon dating: revolutions in understanding. Archaeometry 50, 249–275. Bronk Ramsey, C., 2008b. Deposition models for chronological records. Quaternary Science Reviews 27, 42–60. Bronk Ramsey, C., 2001. Development of the radiocarbon calibration program. Radiocarbon 43, 355–363. Bronk Ramsey, C., 1998. Probability and dating. Radiocarbon 40, 461–474. Bronk Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy. Radiocarbon 36, 425–430. Brown, A.G., Keough, M., Rice, R.J., 1994. Floodplain evolution in the East Midlands, United Kingdom: the Lateglacial and Flandrian alluvial record from the Soar and Nene valleys. Philosophical Transactions of the Royal Society of London A348, 261–293. Catt, J.A., 1990. Palaeopedology manual. Quaternary International 6, 1–95. Chiverrell, R.C., Foster, G., Thomas, G.S.P., Marshall, P., Hamilton, D., 2009. Robust chronologies for landform development. Earth Surface Processes and Landforms 34, 319–328. Chiverrell, R.C., Harvey, A.M., Foster, G.C., 2007. Hillslope gullying in Northwest England and Southwest Scotland: response to human impact and/or climatic downturn? Geomorphology 84, 317–343. Cook, G.T., Dugmore, A.J., Shore, J.S., 1998. The influence of pretreatment on humic acid yield and 14C age of Carex peat. Radiocarbon 40, 21–27. Daniels, J.M., Knox, J.C., 2005. Alluvial stratigraphic evidence for channel incision during the Medieval Warm Period, central Great Plains, USA. The Holocene 15, 736–747. Hill, T.C.B., Fletcher, W., Gearey, B.R., Howard, A.J., 2008a. The Suffolk River Valleys Project: an assessment of the potential and character of the palaeoenvironmental and geoarchaeological resource of Suffolk river valleys affected by aggregate extraction. Unpublished report to English Heritage, Birmingham Archaeo-Environmental, University of Birmingham. Hill, T.C.B, Fletcher, W., Gearey, B.R., Howard, A.J., 2008b. The Suffolk River Valleys Project Phase 2: An assessment of the potential and character of the palaeoenvironmental and geoarchaeological resource of Suffolk river valleys affected by aggregate extraction. Unpublished report to English Heritage, Birmingham Archaeo-Environmental, University of Birmingham. Howard, A.J., Macklin, M.G., 1999. A generic geomorphological approach to archaeological interpretation and prospection in British river valleys: a guide for archaeologists investigation Holocene landscapes. Antiquity 73, 527–541. Howard, A.J., Macklin, M.G., Black, S., Hudson-Edwards, K.A., 1999. Holocene river development and environmental change in Upper Wharfedale, Yorkshire Dales, England. Journal of Quaternary Science 15, 239–252. Johnson, R.H., Tallis, J.H., Wilson, P., 1990. The Seal Edge Coombes, north Derbyshire – a study of their erosional and depositional history. Journal of Quaternary Science 5, 83–94. Johnstone, E., Macklin, M.G., Lewin, J., 2006. The development and application of a database of radiocarbon-dated Holocene fluvial deposits in Great Britain. Catena 66, 14–23. Knox, J.C., 2006. Floodplain sedimentation in the Upper Mississippi Valley: natural versus human accelerated. Geomorphology 79, 286–310. Lewin, J., Macklin, M.G., Johnstone, E., 2005. Interpreting alluvial archives: sedimentological factors in the British Holocene fluvial record. Quaternary Science Reviews 24, 1873–1889. Lewin, J., Macklin, M.G., 2003. Preservation potential for Late Quaternary alluvium. Journal of Quaternary Science 18, 107–120. Lowe, J.J., Walker, M.J.C., Scott, E.M., Harkness, D.D., Bryant, C., Davies, S.M., 2004. A coherent high-precision radiocarbon chronology for the Late-glacial sequence at Sluggan Bog, Co. Antrim, Northern Ireland. Journal of Quaternary Science 19, 147–158. Macklin, M.G., 1999. Holocene river environments in prehistoric Britain: human interaction and impact. In: Edwards, K.J., Sadler, J.P. (Eds.), Holocene Environments of Prehistoric Britain. Quaternary Proceedings, vol. 7, pp. 521–530.
Macklin, M.G., Benito, G., Gregory, K.J., Johnstone, E., Lewin, J., Michczyn´ska, D.J., Soja, R., Starkel, L., Thorndycraft, V.R., 2006. Past hydrological events reflected in the Holocene hydrological record of Europe. Catena 66, 145–154. Macklin, M.G., Johnstone, E., Lewin, J., 2005. Pervasive and long-term forcing of Holocene river instability and flooding by centennial-scale climate change. The Holocene 15, 937–943. Macklin, M.G., Lewin, J., 2003. River sediments, great floods and centennial scale Holocene climate change. Journal of Quaternary Science 18, 101–105. Macklin, M.G., Lewin, J., 1993. Holocene river alluviation in Britain. In: Douglas, I., Hagedorn, J. (Eds.), Geomorphology and Geoecology, Fluvial Geomorphology. Zeitscrift fur Geomorphologie, vol. 85 (Suppl.), pp. 109–122. Mook, W.G., 1986. Business meeting: recommendations/resolutions adopted by the twelfth international radiocarbon conference. Radiocarbon 28, 799. Mook, W.G., Waterbolk, H.T., 1985. Radiocarbon Dating. Handbook for Archaeologists, 3. European Science Foundation, Strasbourg. Mook, W.G., Streurman, H.J., 1983. Physical and chemical aspects of radiocarbon dating. In: Mook, W.G., Waterbolk, H.T. (Eds.), Proceedings of the First International Symposium 14C and Archaeology, vol. 8. PACT, pp. 31–55. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04-terrestrial radiocarbon age calibration, 0– 26 cal kyr BP. Radiocarbon 46 (3), 1029–1058. Scott, E.M. (Ed.), 2003. The Third International Radiocarbon Intercomparison (TIRI) and the Fourth International Radiocarbon Intercomparison (FIRI) 1990–2002: results, analysis, and conclusions. Radiocarbon, vol. 45, pp. 135–408. Shore, J.S., Bartley, D.D., Harkness, D.D., 1995. Problems encountered with the 14C dating of peat. Quaternary Science Reviews 14, 373–383. Slota Jr., P.J., Jull, A.J.T., Linick, T.W., Toolin, L.J., 1987. Preparation of small samples for 14C accelerator targets by catalytic reduction of CO. Radiocarbon 29, 303–306. Starkel, L., Soja, R., Michczynska, D.J., 2006. Past hydrological events reflected in Holocene history of Polish rivers. Catena 66, 24–33. Stenhouse, M.J., Baxter, M.S., 1983. 14C Dating Reproducibility: Evidence from Routine Dating of Archaeological Samples. PACT. 8, 147–161. Stuiver, M., Reimer, P.J., 1993. Extended 14C data base and revised CALIB 3.0 14C age calibration program. Radiocarbon 35, 215–230. Stuiver, M., Kra, R.S., 1986. Editorial comment. Radiocarbon 28 (2B), ii Stuiver, M., Reimer, P.J., 1986. A computer program for radiocarbon age calculation. Radiocarbon 28, 1022–1030. Stuiver, M., Polach, H.A., 1977. Reporting of 14C data. Radiocarbon 19, 355–363. Thorndycraft, V.R., Benito, G., 2006. The Holocene fluvial chronology of Spain: evidence from a newly compiled radiocarbon database. Quaternary Science Reviews 25, 223–234. Tipping, R., 1995. Holocene evolution of a lowland Scottish landscape: Kirkpatrick Fleming. III. Fluvial history. The Holocene 5, 184–195. Troels-Smith, J., 1955. Karakterisering af lose jordater. (characterisation of unconsolidated sediments). Denmarks Geologiske Undersogelse Series IV 3 (10), 73. Turney, C.S.M., Coope, G.R., Harkness, D.D., Lowe, J.J., Walker, M.J.C., 2000. Implications for the dating of Wisconsinan (Weichselian) Late-Glacial events of systematic radiocarbon age differences between terrestrial plant macrofossils from a site in SW Ireland. Quaternary Research 53, 114–121. Vandeputte, K., Moens, L., Dams, R., 1996. Improved sealed-tube combustion of organic samples to CO2 for stable isotopic analysis, radiocarbon dating and percent carbon determinations. Analytical Letters 29, 2761–2774. Walker, M.J.C., Bryant, C., Coope, G.R., Harkness, D.D., Lowe, J.J., Scott, E.M., 2001. Towards a radiocarbon chronology of the Late-Glacial: sample selection strategies. Radiocarbon 43, 1007–1019. Ward, G.K., Wilson, S.R., 1978. Procedures for comparing and combining radiocarbon age determinations: a critique. Archaeometry 20, 19–31. Wohlfarth, B., Skog, G., Possnert, G., Holmquist, B., 1998. Pitfalls in the AMS radiocarbon-dating of terrestrial macrofossils. Journal of Quaternary Science 13, 137–145. Xu, S., Anderson, R., Bryant, C., Cook, G.T., Dougans, A., Freeman, S., Naysmith, P., Schnabel, C., Scott, E.M., 2004. Capabilities of the new SUERC 5MV AMS facility for 14C dating. Radiocarbon 46, 59–64.
Contents lists available at ScienceDirect
Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas
Fluvial sediments, correlations and palaeoenvironmental reconstruction: The development of robust radiocarbon chronologies Andy J. Howard a, *, Ben R. Gearey b, Tom Hill c, William Fletcher d, Peter Marshall e a
Institute of Archaeology & Antiquity, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK Birmingham Archaeo-Environmental, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK c School of Science and the Environment, Bath Spa University, Newton Park Campus, Bath, BA2 9BN, UK d Suffolk County Council Archaeological Service, Shire Hall, Bury St Edmunds, Suffolk, IP33 2AR, UK e Chronologies, 25 Onslow Road, Sheffield, S11 7AF, UK b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 September 2008 Received in revised form 21 July 2009 Accepted 31 August 2009
Multiple sequences of radiocarbon dates extracted from organic materials are increasingly being used to provide robust chronologies for landscape development, particularly the timing and correlation of significant climatic and ‘process’ events. Whilst the validity of using such frameworks in sedimentary environments such as lake basins and raised bogs has been the focus of much attention, such debates have not extended to fluvial systems. Using examples from three lowland, vertically accreting river valleys in East Anglia, UK, this paper assesses the robustness of their associated radiocarbon chronologies by assessing the contrasting age estimates that are obtained by dating different parts of the same organic sample (humic, humin and plant macrofossils) at a variety of stratigraphic levels. Overall, the humin and humic acid fraction results were statistically consistent, whilst the plant macrofossil remains were found to be of a slightly younger age. In these examples, it is argued that the younger ages appear to be the result of Phragmites (common reed) roots pushing plant macrossils through the sedimentary sequence or opening up voids for material to fall through, although studies in lacustrine and mire environments suggest alternative explanations may also be possible. Whichever explanation is preferred, this study demonstrates clearly that the complexity of valley floor stratigraphy and processes is such, that using single radiocarbon dates, whether AMS or bulk samples to reconstruct chronologies of ‘geomorphic system response’ may need to be refined and subjected to the same level of assessment that has been applied in other sedimentary systems. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Fluvial Geoarchaeology Holocene Radiocarbon
1. Introduction Within the geoarchaeological community, advances in the understanding of radiocarbon techniques (Bronk Ramsey, 2008a), as well as the application of advanced statistics such as Bayesian modelling to improve resolution (Blockley et al., 2004; Bronk Ramsey, 2008b) has resulted in the increasing use of these methods to provide chronologies for landscape development, particularly the timing and correlation of significant climatic and ‘process’ events. However, assessment of the robustness of primary radiocarbon data has varied significantly with respect to different depositional environments. A number of workers who have constructed chronologies from organic materials extracted from lacustrine environments (Turney et al., 2000; Wohlfarth et al.,
* Corresponding author. Tel.: þ44(0)121 414 8564; fax: þ44(0)121 414 3595. E-mail address: [email protected] (A.J. Howard). 0305-4403/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2009.08.006
1998), other small basins (Walker et al., 2001), more extensive lowland raised bogs (Lowe et al., 2004) and upland raised mosses/ soligenous mires (Shore et al., 1995) have all demonstrated the variability that can result from analyzing different parts of the carbon fraction (humic and humin) and different types of remains (plant macrofossils). However, such debates have not extended to the fluvial record where radiocarbon chronologies have become a key research tool in elucidating the timing of climate change and the role that human activity and changing natural cycles (flood frequency and magnitude) have played in Holocene river development not only in the UK (Chiverrell et al., 2007; Howard et al., 1999; Johnstone et al., 2006; Macklin et al., 2005; Tipping, 1995), but across other parts of Europe (Anderson et al., 2004; Macklin et al., 2006; Starkel et al., 2006; Thorndycraft and Benito, 2006) and the rest of the globe (Daniels and Knox, 2005; Knox, 2006). In terms of the British record, early syntheses of ‘Holocene system response’ were based almost entirely on databases
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
constructed from relatively few radiocarbon dated samples amassed from both published and unpublished literature pertaining to individual river catchments (Macklin and Lewin, 1993; Macklin, 1999). Usually, this material comprised ‘bulk samples’; or where individual plant fragments were analysed, little attention was paid to precisely what part of the plant was being dated. The growth in geoarchaeological studies of alluvial systems has provided the opportunity for more recent reviews to incorporate many more dates (Johnstone et al., 2006; Macklin et al., 2005; Macklin and Lewin, 2003) and apply a range of quantitative analyses, though current studies suggest that more robust methodologies are required to fully understand the potential of the data (Chiverrell et al., 2009). These latter age estimates are usually derived from identified fragments of material analysed by AMS. Furthermore, the validity of these age estimates is more rigorously examined in the light of advances in our understanding of the depositional context of dated materials (Lewin and Macklin, 2003; Lewin et al., 2005). However, despite such refinements to radiocarbon chronologies, few if any of these individual database samples from fluvial environments have been the subject of multiple dating programmes, which have analysed different parts of the carbon fraction (e.g. humic, humin, individual macrofossils) in an attempt to establish, which fraction may provide the most ‘robust’chronology for any given sequence. It is fair to point out that for the majority of geoarchaeological studies undertaken in alluvial environments over the past thirty years the aspiration to conduct large-scale dating programmes to produce robust radiocarbon chronologies has been unachievable through lack of funding. However, since 2003 the Aggregates Levy Sustainability Fund (ALSF), administered in part by English Heritage, has commissioned a significant number of geoarchaeological projects in aggregate-rich, river valley floors, which have included significant programmes of radiocarbon dating (Bayliss et al., 2007, 2008). Although the primary aims of these ASLF sponsored radiocarbon analyses has not been methodological development, the routine availability of multiple dates for individual fluvial sequences has provided an opportunity for palaeoenvironmentalists to become more critical of primary data in these demonstrably complex environments and to highlight anomalies in the record. In 2006, the analysis of thirty-six AMS radiocarbon samples across three river reaches as part of the Suffolk River Valleys Project produced a significant number of dating anomalies within seemingly secure biostratigraphic contexts (Hill et al., 2008a). A subsequent re-evaluation of the sequences using duplicate multifraction dating analyses aimed to explain these inconsistencies in the original record (Hill et al., 2008b) and this paper reports the results of this subsequent study. Whilst this project was not designed or conceived to test the validity of radiocarbon methodologies, the results have important implications for the use of such dating techniques in fluvial environments and highlights a number of issues that many geoarchaeologists and environmental archaeologists should consider, when using single dates for environmental reconstruction, or assembling multiple dates to develop chronologies of landscape development. 2. Methods The aim of the ALSF sponsored Suffolk River Valleys Project was to enhance the palaeoenvironmental record and hence curatorial knowledge of river valleys under threat from aggregate extraction and to dovetail this information with the cultural record. Work focused on reaches of the River Waveney at Beccles, the Blackbourn at Ixworth and the River Lark at Hengrave (Fig. 1). All of these rivers are typical of lowland river systems with cohesive channel banks
2681
with thick sequences of vertically accreted organic-rich sediments deposited by channels, which have been relatively stable within their respective valley floors throughout the Holocene (Howard and Macklin, 1999). Within each reach, single cores were extracted from organic sediments using a standard pattern Russian Corer. Each core was described using Troels-Smith (1955) system and assessed for a range of environmental proxies including pollen and insects. In order to provide a chronological framework, single AMS radiocarbon measurement was undertaken on plant macrofossil remains selected from key sedimentary horizons in each core. Dating of the three sequences produced anomalous results including inversions and modern ages deep within some of the sequences (Table 1). In the light of these results, a critical review by the project team and chronology specialists from English Heritage concluded that neither the field sampling methods nor laboratory protocols such as pre-treatment methods (see Cook et al., 1998) of well established radiocarbon facilities could adequately explain the majority of the anomalies observed within the dataset, especially since the pollen biostratigraphy at all of the sites appeared secure. Therefore, in 2007, English Heritage commissioned a second phase of the project with the specific aim of obtaining duplicate sample cores from the three sites and re-dating the sequences; fieldwork was undertaken during 2008 (Table 2). At each radiocarbon sample depth, individual plant macrofossil samples were collected in addition to bulk samples. At Beccles, new cores were taken at a distance of around 50 m (Beccles 2008 Core 1) and 100 m (Beccles 2008 Core 2) along a transect trending east of Beccles 2007 Core 1. At each core site, the upper 1.2 m of sediment was exposed through machine trenching and monolith tins were used to sample the sequence below the plough zone. Below this depth to the base of the gravels, samples were extracted using a standard pattern Russian Corer. At Hengrave, the new core (Hengrave 2008) was taken 2 m east of the initial core. A hand-dug trench was excavated through 0.38 m of ‘made ground’ to a depth of 0.65 m. Monolith tins were used to sample the underlying 0.27 m of well-humified peat with a further 3 m of peat collected to the top of the basal gravels using the standard pattern Russian corer. At Ixworth, a slot was opened to a depth of 0.5 m immediately adjacent to the 2006 intervention; the (2008) organic sequence was collected using a Russian corer to a depth of 2.4 m. In the laboratory, all the sampled sequences were described using Troels-Smith (1955) and each new core was deemed to stratigraphically correlate with those taken during Phase 1. It is recognized that replication of core sequences can be problematic since floodplains have added levels of complexity than either lacustrine or mire systems owing to the levels of energy and associated processes of erosion, reworking and redeposition of sediments. However, such complexity is likely to be significantly less in these lowland, essential stable, vertically accreting river systems than within their upland and piedmont equivalents (Brown et al., 1994; Howard and Macklin, 1999). As part of the second phase analyses, detailed notes were made recording the variation in abundance and diversity of plant macrofossil remains as well as the degree of humification. Prior to radiocarbon sample submission, duplicate samples were taken from each horizon and assessed for their pH, carbonate content and overall organic content, the latter determined by loss on ignition (Table 3). These tests were used to assess whether carbon content was unusually high or low indicating possible contamination through the influence of hard water and/or in wash of allochthonous material. Table 3 indicates that the carbonate content of the sediments was low and ph was generally between weak acid and neutral.
2682
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Fig. 1. Location of the study areas.
In total, fourteen macrofossil samples were submitted to the Scottish Universities Environmental Research Centre (SUERC), East Kilbride (Table 2). These included the dating of a number of duplicate macrofossil fragments from the same sample depth for radiocarbon comparison (notably in Beccles 2008 Cores 1 and 2). Samples were pre-treated by the acid-base-acid protocol (Stenhouse and Baxter, 1983) and CO2 obtained by combustion in precleaned sealed quartz tubes (Vandeputte et al., 1996). The purified CO2 was converted to graphite (Slota et al., 1987) for subsequent AMS analysis. The sample 14C/13C ratios were measured on the SUERC AMS, as described by Xu et al. (2004). In addition, twelve bulk peat samples (weighing between 72 g and 103 g) were submitted to the Centre for Isotope Studies, University of Groningen, The Netherlands (GrN). These latter samples were pre-treated using the acid/alkali/acid method (Mook and Waterbolk, 1985) and measured using gas proportional counting (Mook and Streurman, 1983). In all cases the acid insoluble/alkali soluble (humic acid) and alkali/acid insoluble (humin) fractions of the samples were separated after pre-treatment, combusted and measured. Each separation was carried out in a quantitative manner thus the total budget of carbon in the peat sample was conserved within the component fractions recovered for gas proportional counting. Both laboratories maintain continual programmes of quality assurance procedures, in addition to participation in international inter-comparisons (Scott, 2003). 3. Results The results of these new analyses are provided in Table 2 and Figs. 2–5 and are quoted in accordance with the international standard known as the Trondheim Convention (Stuiver and Kra, 1986). They are conventional radiocarbon ages (Stuiver and Polach, 1977) and have been calibrated using the curves of Reimer et al. (2004) and the computer program OxCal (4.0.5; Bronk Ramsey, 1995, 1998, 2001, 2009). The calibrated date ranges cited in the text and tables are those for 95% confidence. They are quoted in the
form recommended by Mook (1986), with the end points rounded outwards to 10 years. The ranges in Table 2 have been calculated according to the maximum intercept method (Stuiver and Reimer, 1986). A summary of the radiocarbon dating results for Beccles (2008) Core 1 is provided in Fig. 2 and Table 2. All submitted plant macrofossil samples were alder (Alnus) remains (twig/wood fragments). Duplicate radiocarbon dating of separate macrofossil samples was undertaken at both 0.84 m and 3.30 m depths. The four results from the 0.84 m level are statistically consistent and could be of the same actual age (T0 ¼ 6.9; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson,1978). At 3.30 m, the four measurements are not statistically consistent (T0 ¼ 347.155; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson, 1978) and thus represent material of different ages. However, both the humin/humic acid fractions (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8) and two Alnus fragments (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8) are statistically consistent. At 4.60 m, the three measurements are not statistically consistent (T0 ¼ 2576.412; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978) indicating that the sample comprises material of different ages. In all three samples, the humin acid contains most of the carbon and therefore has the greatest influence on a combined age. Dating results for Beccles (2008) Core 2 are provided in Fig. 3 and Table 2. Duplicate radiocarbon dating of separate macrofossil samples was undertaken at 1.37 m depth. At the 1.37 m level, all four measurements are not statistically consistent (T0 ¼ 84.1; n ¼ 3; T0 (5%) ¼ 7.8; Ward and Wilson, 1978). Separated by sample type, this indicates that the peat fractions are not statistically consistent (T0 ¼ 63.1; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978), whilst the two Alnus twigs are: (T0 ¼ 1.5; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). With the humic acid faction removed, the three remaining measurements from 1.37 m are statistically consistent (T0 ¼ 1.6; n ¼ 2; T0 (5%) ¼ 6; Ward and Wilson, 1978). Again at 3.59 m, the three measurements are not statistically consistent (T0 ¼ 47.663; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978), although the humin and humic acid fractions of the peat sample do show consistency (T0 ¼ 0.0; n ¼ 1; T0 (5%) 3.8: Ward and Wilson, 1978). At 4.30 m, the
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
2683
Table 1 Phase 1 Radiocarbon dates from the Suffolk River Valleys Project. Sample ID
Lab code
Material
d13C (%)
Radiocarbon Age (BP)
Calibrated date (95% confidence)
Beccles 1–0.84 m Beccles 1–0.99 m Beccles 1–1.15 m Beccles 1–1.18 m A Beccles 1–1.18 m B Beccles 1–1.99 m Beccles 1–2.02 m Beccles 1–3.50 m Beccles 1–5.34 m Beccles 2–1.34 m Beccles 2–2.22 m Beccles 2–2.50 m Beccles 2–2.54 m Beccles 2–2.83 m Hengrave-0.26 m Hengrave-0.59 m Hengrave-0.99 m Hengrave-1.50 m Hengrave-1.63 m
GrA-33471 GrA-33472 SUERC-12035 SUERC-12036 GrA-33473 GrA-33475 SUERC-12037 GrA-33476 GrA-33477 SUERC-12038 GrA-35050 SUERC-12039 GrA-33479 GrA-35067 SUERC-12027 GrA-35051 SUERC-12028 GrA-33481 SUERC-12029
27.3 27.3 24.2 28.1 28.3 27.6 27.5 28.2 27.1 27.9 27.7 24.2 29.7 26.4 27.8 28 25.9 27.4 27.4
2080 50 2090 70 1595 35 1975 35 2215 40 2695 40 2835 35 2785 40 9960 130 915 35 225 40 1770 35 105.45 1.09%Mod. 1445 40 125 35 1025 45 955 35 1965 40 1750 35
350 cal BC–cal AD 30 360 cal BC–cal AD 60 cal AD 390–550 50 cal BC–cal AD 120 390–170 cal BC 920–790 cal BC 1120–900 cal BC 1030–830 cal BC 10040–9220 cal BC cal AD 1020–1220 cal AD 1530–1950 cal AD 130–380 cal AD 1956–1957 cal AD 540–660 cal AD 1660–1955 cal AD 890–1150 cal AD 1010–1170 50 cal BC–cal AD 130 cal AD 210–390
Hengrave-1.99 m Hengrave-2.32 m Hengrave-2.56 m Hengrave-2.99 m Ixworth-0.86 m Ixworth-1.37 m Ixworth-1.40 m Ixworth-1.49 m Ixworth-2.49 m Ixworth-2.63 m Ixworth-3.44 m
GrA-35054 SUERC-12030 GrA-33482 SUERC-12031 GrA-35055 SUERC-12021 GrA-35056 SUERC-12025 GrA-33485 SUERC-12026 GrA-33483
Alnus glutinosa, small wood fragment Alnus glutinosa, small wood fragment Poaceae fragments Unidentified wood Poaceae fragments and internode Alnus glutinosa, stem Bark fragment, inidentified Alnus glutinosa, stem unidentified plant fragments Poaceae fragment unidentified plant fragments herbaceous stems Monocotyledon stem unidentified plant fragments Poaceae stems Poaceae stems Poaceae fragments Poaceae stems and internode Unidentified plant remains cf. seed/ flower head Poaceae stems Poaceae fragments Poaceae fragments unidentified plant stems unidentified seed Poaceae fragments Poaceae fragments unidentified wood fragments Alnus glutinosa, stem Alnus glutinosa, wood fragment Alnus glutinosa, stem
plant macrofossil sample did not produce enough carbon for dating purposes. However, the humin and humic acid fractions did and are statistically consistent (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). In all three samples of Beccles (2008) Core 2, the humin contains most of the carbon and therefore has the greatest influence on a combined age. At Hengrave (Fig. 4; Table 2), all three measurements from the 0.47 m level are statistically inconsistent (T0 ¼ 12.198; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson, 1978), although the plant macrofossil and humin fraction are statistically consistent (T0 ¼ 1.4; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978). At 1.61 m, the humin and humic acid fractions are statistically consistent (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8; Ward and Wilson, 1978), whilst the plant macrofossil sample yielded too little carbon for analysis. At 2.76 m, the three measurements are not statistically consistent (T0 ¼ 21.697; n ¼ 2; T0 (5%) ¼ 6.0; Ward and Wilson,1978), although the humin and humic acid fractions are statistically consistent (T0 ¼ 0.2; n ¼ 1; (5% 3.8)). Two of the samples show the humin fraction contains most of the carbon and therefore has the greatest influence on a combined age. At Ixworth (Fig. 5; Table 2), the high level of humification encountered within the organic deposits prevented the separation and dating of macroscopic plant remains. A single fragment of Alnus was initially identified from within the basal sample (2.39 m depth), but yielded too little carbon for analysis. However at 0.71 m, 1.24 m and 2.39 m, the humin and humic acid fractions are statistically consistent (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8; T0 ¼ 1.5; n ¼ 1; T0 (5%) ¼ 3.8; T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8 respectively; Ward and Wilson, 1978). Two of the samples show the humin fraction contains most of the carbon and therefore has the greatest influence on a combined age. The very low organic content of the sample from 1.24 m (Table 3/LOI) however does not seem to have had an influence of the % carbon content.
27 26.1 27.3 27.3 26.8 27 25.4 27.4 29 26.6 27.7
1740 45 1000 35 1620 35 1720 35 1935 40 1560 35 1465 45 2905 35 6265 45 5980 40 9900 60
cal AD 130–420 cal AD 980–1160 cal AD 340–540 cal AD 230–420 40 cal BC–cal AD 140 cal AD 410–590 cal AD 530–660 1260–990 cal BC 5330–5070 cal BC 4990–4770 cal BC 9660–9250 cal BC
4. Discussion The second phase results of radiocarbon dating of samples from the top, middle and bottom of each core proved to be more successful than those obtained during the first phase, primarily because of an improved understanding of sediment stratigraphy, including the composition of organic horizons. Overall, the results for all three sites are more robust with no age inversions or modern age estimates being encountered within the sequences. In general, it appears that the humin and humic acid fraction dating results are statistically consistent, whilst the plant macrofossil remains were found to be of a slightly younger age. Within the literature, whilst a number of studies have assessed the importance of different components of organic material (humic, humin, plant macrofossils) to the accuracy and precision of radiocarbon dating, there is a level of debate regarding which is the most significant. For example, whilst this study suggests that humin is particularly important in shaping the chronology of alluvial valley floors, Shore et al. (1995), found that the humic acid component contained most of the carbon from upland peat samples recovered from Lanshaw Moss and White Moss (West Yorkshire). The differences observed between these lowland river valleys in Suffolk and upland sites in northern England might be explained by the contrasting environmental settings, geomorphic processes and erosion and sedimentation histories of the sites (i.e. soligenuous mire and raised mire complexes being compared directly with floodplain environments). However, unfortunately, whilst a number of studies have identified similar issues and are not unique to this paper (Bartley and Chambers, 1992; Johnson et al., 1990; Lowe et al., 2004; Walker et al., 2001; Wohlfarth et al., 1998), with the exception of this paper, all other studies discussing this conundrum have focused on mires, blanket peats and lacustrine basins of varying size where
2684
Table 2 Phase 2 Radiocarbon dates from the Suffolk River Valleys Project. Sample ID
Lab code
Material
Organic Content
pH
d13C (%)
Radiocarbon age (BP)
Weighted mean
Calibrated date (95% confidence)
Beccles 1–0.84 m Beccles 1–0.84 m Beccles 1–0.84 m A
GrN-31116 GrN-31151 SUERC15973 SUERC15974 GrN-31117 GrN-31152 SUERC15975 SUERC15976 GrN-31118 GrN-31153 SUERC15981
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus twig,
79%
3.90
28.9 28.9 29.5
2130 40 2160 50 2065 35
2142 32 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
360–50 cal BC
28.1
2015 40
28 28.7 30.8
4590 30 4590 50 3885 35
30.7
3970 35
Peat (Humin) Peat (Humic acid) Plant macrofossil: cf. Alnus twig
72%
28.4 28 28.4
8460 50 8340 80 5660 35
GrN-31119 GrN-31154 SUERC15982 SUERC15983 GrN-31120 GrN-31155 SUERC15984 GrN-31121 GrN-31156 GU-6796
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus 1 growth ring Plant macrofossil: Alnus 1 growth ring Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus roundwood, c. 8 growth Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus 1 growth ring
76% twig,
28.7 28.6 28.7
2230 30 1830 40 2275 35
390–200 cal BC cal AD 70–320 400–210 cal BC
twig,
28.4
2215 35
390–180 cal BC
28.6 28 29
5060 30 5060 40 4765 35
5060 24 BP (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8)
7735 35 BP (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8)
GrN-31113 GrN-31148 SUERC16385 GrN-31114 GrN-31149 GU-6786 GrN-31115 GrN-31150 SUERC15972
Peat (Humin) Peat (Humic acid) Plant macrofossil: stem fragment
GrN-31110 GrN-31145 GrN-31111 GrN-31146 GU-6798 GrN-31112 GrN-31147
Peat (Humin) Peat (Humic acid) Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus wood Peat (Humin) Peat (Humic acid)
Beccles 1–0.84 m B Beccles 1–3.30 m Beccles 1–3.30 m Beccles 1–3.30 m A
Beccles 1–4.60 m Beccels 1–4.60 m Beccels 1–4.60 m Beccels 2–1.37 m Beccles 2–1.37 m Beccles 2–1.37 m A Beccles 2–1.37 m B Beccles 2–3.59 m Beccles 2–3.59 m Beccles 2–3.59 m Beccles 2–4.30 m Beccles 2–4.30 m Beccles 2–4.30 m Hengrave-0.47 m Hengrave-0.47 m Hengrave-0.47 m Hengrave-1.61 m Hengrave-1.61 m Hengrave-1.61 m Hengrave-2.76 m Hengrave-2.76 m Hengrave-2.76 m Ixworth-0.71 m Ixworth-0.71 m Ixworth-1.24 m Ixworth-1.24 m Ixworth-1.24 m Ixworth-2.39 m Ixworth-2.39 m
Peat (Humin) Peat (Humic acid) Plant macrofossil: Alnus wood
78%
6.20
Plant macrofossil: Alnus wood
70%
4.80
5.60
6.00
160–70 cal BC 4590 26 BP (T0 ¼ 0.0; n ¼ 1; T0 (5%) ¼ 3.8)
3500–3340 cal BC 2480–2210 cal BC 2580–2350 cal BC
8427 43 BP (T0 ¼ 1.6; n ¼ 1; T0 (5%) ¼ 3.8)
7580–7370 cal BC 4560–4400 cal BC
3960–3785 cal BC 3650–3380 cal BC
rings 31%
4.30
27.6 27.7
7740 40 7720 70 Sample failed
44%
6.00
60%
6.40
29.1 29.7 25.1
715 30 540 40 660 35
28.9 28.5
1430 35 1450 30 Sample failed 2310 40 2340 60 2095 35
1442 23 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
cal AD 570–655
2319 34 BP (T0 ¼ 0.2; n ¼ 1; T0 (5%) ¼ 3.8)
410–360 cal BC
1740 35 1830 40 2670 40 2730 40 Sampled failed 7530 50 7510 50
1779 27 BP (T0 ¼ 2.9; n ¼ 1; T0 (5%) ¼ 3.8) 2700 29 BP (T0 ¼ 1.1; n ¼ 1; T0 (5%) ¼ 3.8)
cal AD 130–340
7520 36 BP (T0 ¼ 0.1; n ¼ 1; T0 (5%) ¼ 3.8)
6460–6260 cal BC
twig, c.
Peat (Humin) Peat (Humic acid) Plant macrofossil: herbaceous stem Peat (Humin) Peat (Humic acid) Plant macrofossil: monocot culm
47%
5.70 29.8 30.5 27.5
55%
5.60
14%
7.00
53%
5.90
29.6 29.2 29.3 29.3 28.9 28.3
6640–6480 cal BC
cal AD 1260–1380 cal AD 1300–1450 cal AD 1270–1400
210–1 cal BC
910–800 cal BC
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Beccles 1–3.30 m B
Plant macrofossil: Alnus twig
190–10 cal BC
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688 Table 3 Summary of Loss on Ignition (LOI), carbonate content and pH measurements of the 12 sedimentary sampled for radiocarbon dating during Phase 2. For LOI, samples were oven dried at 105 degrees for 12hrs, organic content was then fired at 550 degrees for 4 hours; carbonate content was fired at 950 degrees for 2 hours. Measurements for pH were achieved following the procedure of Catt (1990). Sample No.
Sample Name
Organic content %
Carbonate content %
pH
1 2 3 4 5 6 7
Hengrave 0.47 m Hengrave 1.61 m Hengrave 2.76 m Ixworth 0.71 m Ixworth 1.24 m Ixworth 2.39 m Beccles 2008 Core 1: 0.84 m Beccles 2008 Core 1: 3.30 m Beccles 2008 Core 1: 4.60 m Beccles 2008 Core 2: 1.37 m Beccles 2008 Core 2: 3.59 m Beccles 2008 Core 2: 4.30 m
43.98 59.22 46.60 54.88 14.75 52.79 79.12
2.76 2.68 2.91 2.56 1.34 3.35 2.98
6.04 6.43 5.65 5.64 6.95 5.82 3.85
78.25
2.73
6.15
71.77
3.61
4.83
76.71
3.45
5.61
70.25
1.74
5.98
31.35
1.36
4.29
8 9 10 11 12
geomorphological and depositional processes are relatively simple. Until further methodological research is undertaken with respect to radiocarbon chronologies in a range of complex sedimentary systems, comparison of such datasets between environments must be approached with caution. Within Beccles (2008) Core 1, it was notable that in two cases (3.30 m and 4.60 m) Alnus fragment(s) were younger than the bulk sediment measurements from the same level. The stratigraphic consistency of both sets of data when analysed independently thus raises the possibility that either could be accurate. However,
2685
detailed sedimentary analysis revealed that Phragmites remains were present in the units immediately overlying both of the horizons in which the dating discrepancies were evident. This therefore raises the possibility that the Alnus fragments were intrusive and that the bulk samples provide the reliable age estimates. The invasive nature of Phragmites roots suggests that they have the strength to push larger organic material (e.g. small twigs, seeds) through the sediment or provide (root) cavities for material to fall down during dry periods. Deep penetration of Alnus roots into floodplain soils may be in response to down-draw of the water table (see Brayshay and Dinnin, 1999). As a cautionary note, it is possible that the Alnus fragments provide the true age and that the bulk sediment measurements are inaccurate; however, the consistency of the humic and humin measurements suggest that this is not the case since if the Alnus ages were indeed correct, the humin fraction ages would be closer in age to the macrofossil dates since the humins are composed of organic detritus. At Hengrave, the Alnus fragment from 2.76 m was also younger than the bulk sediment sample from the same horizon although the offset between the bulk and macrofossil dating is noticeably much smaller. This can perhaps be explained by the much lower incidence of wood remains from this core, although there is a much greater incidence of Phragmites. The overall consistency between humin and humic acid fractions is amply demonstrated by Beccles 2008 (Core 2) and suggests that these data provide the more reliable age estimates for each dated sample depth. At Hengrave, there is also statistical consistency between the humin and humic acid fractions analysed at 1.61 m and 2.76 m. The consistency of humic and humin fractions almost certainly reflects the stability of floodplain water-tables preventing movement of such fractions in this part of the vadose zone in these studies. In contrast however, explanations for the age difference between the humic acid and humin/Alnus fragment at 0.47 m depth may include the upward movement of humic acid or
Fig. 2. Probability distributions of dates from Beccles (2008) Core 1. Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
2686
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Fig. 3. Probability distributions of dates from Beccles (2008) Core 2. Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
the intrusion of younger rootlets from above. However, the caveats explored above with respect to the variability of humic and humin fractions must still apply. This study demonstrates very clearly that the complexity of valley floor stratigraphy and processes is such, that using single radiocarbon dates, whether AMS or bulk samples to reconstruct chronologies of ‘geomorphic system response’ may be unreliable. Whilst more recent studies have demonstrated the need for caution
when considering precisely what event/sedimentary context is being dated by the submission of a particular sample (Lewin and Macklin, 2003; Lewin et al., 2005), little consideration has been given to anomalies that might be introduced from the actual fractions of material being dated. Whilst such methodological considerations of radiocarbon dating have been undertaken in environments that might be classified as geomorphologically ‘less complex’ (Lowe et al., 2004; Walker et al., 2001), these studies have
Fig. 4. Probability distributions of dates from Hengrave (2008). Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
2687
Fig. 5. Probability distributions of dates from Ixworth (2008). Each distribution represents the relative probability that an event occurred at a particular time. These distributions are the result of simple radiocarbon calibration (Stuiver and Reimer, 1993).
demonstrated that the issue requires more research, especially comparison between differing sedimentary environments. However, equally, this paper adds to the wider radiocarbon debate by suggesting that the dating of multiple fractions within the same sample can yield consistent and reliable results, which can be used to reconstruct valley floor chronologies. Furthermore, advances in the application of Bayesian approaches, especially using the framework advocated recently by Chiverrell et al. (2009) provide an opportunity to improve the quality and application of such datasets significantly.
5. Conclusions In geoarchaeological and palaeoenvironmental research, radiocarbon dates are used all too frequently to reconstruct landscape chronologies without full consideration being given to the validity of the data used. Whilst researchers working in selected environments including lake basins and mire complexes have sought to examine the robustness of radiocarbon chronologies and integrity of the data being collected, those working within fluvial systems have been less rigorous in considering the differences in dating results that can be obtained by analyzing different fractions of carbon (humic, humin and plant macrofossils). This paper has described the chronological problems encountered during one particular dating programme and demonstrated the value of taking a multi-fraction approach to radiocarbon analysis. In this particular study, it is suggested that plant macrofossils of Alnus produced consistently younger dates than either the humic and humin fractions. Whilst a number of other studies have demonstrated the variable importance of humic and humin fractions, associated sedimentary evidence in this instance suggest that the intrusion of plant macrofossils into older sediments may be a result of Phragmites root penetration. On this evidence, it is suggested that where there is evidence of Phragmites within sedimentary sequences, macrofossils should only be dated if there is evidence that they have grown in-situ and the submission of unidentified plant remains, monocot, Poaceae fragments and other similar material should be avoided. In this example, the statistical consistency
present between the dating results from the humin and humic acid fractions suggests that the dating of bulk radiometric sediment samples may provide accurate age estimates. Bulk AMS size samples are very unreliable since a small amount of contamination has a disproportionate effect compared with a radiometric sized sample. With the exception of this study, the corpus of literature addressing the issue of organic fractions used for radiocarbon dating has all been undertaken in environments that might be classified as geomorphologically ‘less complex’ (Lowe et al., 2004; Walker et al., 2001). If this issue is to be addressed more fully, our study clearly demonstrates the need for more research in a range of sedimentary environments of varying complexity, especially where groundwater chemistry and flow-paths may vary. Finally, this research has important implications for environmental reconstruction and indicates that single dates should never be used as the basis for the development of robust chronologies. Acknowledgements This research was funded by the Aggregates Levy Sustainability Fund administered by English Heritage as part of the Suffolk River Valleys Projects (PN 4772; PN 4772/ANL). Particular thanks go to the following EH staff for supporting our work: Kath Buxton, Dr Jen Heathcote, Dr Jane Sidell, Tom Cromwell and John Ette. Thanks are also due to Derek Hamilton (formerly of the EH Dating Team). Site access was granted by Suffolk Wildlife Trust, Beccles Town Council and Mr Philip Aitken. The comments of three anonymous referees greatly improved the clarity of this manuscript. References Anderson, E., Harrison, S., Passmore, D.G., Mighall, T., Wathan, S., 2004. Late Quaternary river terrace development in the Macgillycuddy0 s Reeks, southwest Ireland. Quaternary Science Reviews 23, 1785–1801. Bartley, D.D., Chambers, C., 1992. A pollen diagram, radiocarbon ages and evidence of agriculture on Extwistle Moor, Lancashire. New Phytologist 121, 311–320. Bayliss, A., Cook, G., Bronk Ramsey, C., van der Plicht, J., McCormac, G., 2008. Radiocarbon Dates from Samples Funded by English Heritage Under the Aggregates Levy Sustainability Fund 2004-7. English Heritage, Swindon.
2688
A.J. Howard et al. / Journal of Archaeological Science 36 (2009) 2680–2688
Bayliss, A., Bronk Ramsey, C., Cook, G., van der Plicht, J., 2007. Radiocarbon Dates from Samples Funded by English Heritage Under the Aggregates Levy Sustainability Fund 2002-4. English Heritage, Swindon. Blockley, S.P.E., Lowe, J.J., Walker, J.J., Asioli, A., Trincardi, F., Coope, G.R., Donahue, R.E., Pollard, A.M., 2004. Bayesian analysis of radiocarbon chronologies: examples from the European Lateglacial. Journal of Quaternary Science 19, 159–175. Brayshay, B.A., Dinnin, M.H., 1999. Integrated palaeoecological evidence for biodiversity at the floodplain-forest margin. Journal of Biogeography 26, 115–131. Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51, 337–360. Bronk Ramsey, C., 2008a. Radiocarbon dating: revolutions in understanding. Archaeometry 50, 249–275. Bronk Ramsey, C., 2008b. Deposition models for chronological records. Quaternary Science Reviews 27, 42–60. Bronk Ramsey, C., 2001. Development of the radiocarbon calibration program. Radiocarbon 43, 355–363. Bronk Ramsey, C., 1998. Probability and dating. Radiocarbon 40, 461–474. Bronk Ramsey, C., 1995. Radiocarbon calibration and analysis of stratigraphy. Radiocarbon 36, 425–430. Brown, A.G., Keough, M., Rice, R.J., 1994. Floodplain evolution in the East Midlands, United Kingdom: the Lateglacial and Flandrian alluvial record from the Soar and Nene valleys. Philosophical Transactions of the Royal Society of London A348, 261–293. Catt, J.A., 1990. Palaeopedology manual. Quaternary International 6, 1–95. Chiverrell, R.C., Foster, G., Thomas, G.S.P., Marshall, P., Hamilton, D., 2009. Robust chronologies for landform development. Earth Surface Processes and Landforms 34, 319–328. Chiverrell, R.C., Harvey, A.M., Foster, G.C., 2007. Hillslope gullying in Northwest England and Southwest Scotland: response to human impact and/or climatic downturn? Geomorphology 84, 317–343. Cook, G.T., Dugmore, A.J., Shore, J.S., 1998. The influence of pretreatment on humic acid yield and 14C age of Carex peat. Radiocarbon 40, 21–27. Daniels, J.M., Knox, J.C., 2005. Alluvial stratigraphic evidence for channel incision during the Medieval Warm Period, central Great Plains, USA. The Holocene 15, 736–747. Hill, T.C.B., Fletcher, W., Gearey, B.R., Howard, A.J., 2008a. The Suffolk River Valleys Project: an assessment of the potential and character of the palaeoenvironmental and geoarchaeological resource of Suffolk river valleys affected by aggregate extraction. Unpublished report to English Heritage, Birmingham Archaeo-Environmental, University of Birmingham. Hill, T.C.B, Fletcher, W., Gearey, B.R., Howard, A.J., 2008b. The Suffolk River Valleys Project Phase 2: An assessment of the potential and character of the palaeoenvironmental and geoarchaeological resource of Suffolk river valleys affected by aggregate extraction. Unpublished report to English Heritage, Birmingham Archaeo-Environmental, University of Birmingham. Howard, A.J., Macklin, M.G., 1999. A generic geomorphological approach to archaeological interpretation and prospection in British river valleys: a guide for archaeologists investigation Holocene landscapes. Antiquity 73, 527–541. Howard, A.J., Macklin, M.G., Black, S., Hudson-Edwards, K.A., 1999. Holocene river development and environmental change in Upper Wharfedale, Yorkshire Dales, England. Journal of Quaternary Science 15, 239–252. Johnson, R.H., Tallis, J.H., Wilson, P., 1990. The Seal Edge Coombes, north Derbyshire – a study of their erosional and depositional history. Journal of Quaternary Science 5, 83–94. Johnstone, E., Macklin, M.G., Lewin, J., 2006. The development and application of a database of radiocarbon-dated Holocene fluvial deposits in Great Britain. Catena 66, 14–23. Knox, J.C., 2006. Floodplain sedimentation in the Upper Mississippi Valley: natural versus human accelerated. Geomorphology 79, 286–310. Lewin, J., Macklin, M.G., Johnstone, E., 2005. Interpreting alluvial archives: sedimentological factors in the British Holocene fluvial record. Quaternary Science Reviews 24, 1873–1889. Lewin, J., Macklin, M.G., 2003. Preservation potential for Late Quaternary alluvium. Journal of Quaternary Science 18, 107–120. Lowe, J.J., Walker, M.J.C., Scott, E.M., Harkness, D.D., Bryant, C., Davies, S.M., 2004. A coherent high-precision radiocarbon chronology for the Late-glacial sequence at Sluggan Bog, Co. Antrim, Northern Ireland. Journal of Quaternary Science 19, 147–158. Macklin, M.G., 1999. Holocene river environments in prehistoric Britain: human interaction and impact. In: Edwards, K.J., Sadler, J.P. (Eds.), Holocene Environments of Prehistoric Britain. Quaternary Proceedings, vol. 7, pp. 521–530.
Macklin, M.G., Benito, G., Gregory, K.J., Johnstone, E., Lewin, J., Michczyn´ska, D.J., Soja, R., Starkel, L., Thorndycraft, V.R., 2006. Past hydrological events reflected in the Holocene hydrological record of Europe. Catena 66, 145–154. Macklin, M.G., Johnstone, E., Lewin, J., 2005. Pervasive and long-term forcing of Holocene river instability and flooding by centennial-scale climate change. The Holocene 15, 937–943. Macklin, M.G., Lewin, J., 2003. River sediments, great floods and centennial scale Holocene climate change. Journal of Quaternary Science 18, 101–105. Macklin, M.G., Lewin, J., 1993. Holocene river alluviation in Britain. In: Douglas, I., Hagedorn, J. (Eds.), Geomorphology and Geoecology, Fluvial Geomorphology. Zeitscrift fur Geomorphologie, vol. 85 (Suppl.), pp. 109–122. Mook, W.G., 1986. Business meeting: recommendations/resolutions adopted by the twelfth international radiocarbon conference. Radiocarbon 28, 799. Mook, W.G., Waterbolk, H.T., 1985. Radiocarbon Dating. Handbook for Archaeologists, 3. European Science Foundation, Strasbourg. Mook, W.G., Streurman, H.J., 1983. Physical and chemical aspects of radiocarbon dating. In: Mook, W.G., Waterbolk, H.T. (Eds.), Proceedings of the First International Symposium 14C and Archaeology, vol. 8. PACT, pp. 31–55. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04-terrestrial radiocarbon age calibration, 0– 26 cal kyr BP. Radiocarbon 46 (3), 1029–1058. Scott, E.M. (Ed.), 2003. The Third International Radiocarbon Intercomparison (TIRI) and the Fourth International Radiocarbon Intercomparison (FIRI) 1990–2002: results, analysis, and conclusions. Radiocarbon, vol. 45, pp. 135–408. Shore, J.S., Bartley, D.D., Harkness, D.D., 1995. Problems encountered with the 14C dating of peat. Quaternary Science Reviews 14, 373–383. Slota Jr., P.J., Jull, A.J.T., Linick, T.W., Toolin, L.J., 1987. Preparation of small samples for 14C accelerator targets by catalytic reduction of CO. Radiocarbon 29, 303–306. Starkel, L., Soja, R., Michczynska, D.J., 2006. Past hydrological events reflected in Holocene history of Polish rivers. Catena 66, 24–33. Stenhouse, M.J., Baxter, M.S., 1983. 14C Dating Reproducibility: Evidence from Routine Dating of Archaeological Samples. PACT. 8, 147–161. Stuiver, M., Reimer, P.J., 1993. Extended 14C data base and revised CALIB 3.0 14C age calibration program. Radiocarbon 35, 215–230. Stuiver, M., Kra, R.S., 1986. Editorial comment. Radiocarbon 28 (2B), ii Stuiver, M., Reimer, P.J., 1986. A computer program for radiocarbon age calculation. Radiocarbon 28, 1022–1030. Stuiver, M., Polach, H.A., 1977. Reporting of 14C data. Radiocarbon 19, 355–363. Thorndycraft, V.R., Benito, G., 2006. The Holocene fluvial chronology of Spain: evidence from a newly compiled radiocarbon database. Quaternary Science Reviews 25, 223–234. Tipping, R., 1995. Holocene evolution of a lowland Scottish landscape: Kirkpatrick Fleming. III. Fluvial history. The Holocene 5, 184–195. Troels-Smith, J., 1955. Karakterisering af lose jordater. (characterisation of unconsolidated sediments). Denmarks Geologiske Undersogelse Series IV 3 (10), 73. Turney, C.S.M., Coope, G.R., Harkness, D.D., Lowe, J.J., Walker, M.J.C., 2000. Implications for the dating of Wisconsinan (Weichselian) Late-Glacial events of systematic radiocarbon age differences between terrestrial plant macrofossils from a site in SW Ireland. Quaternary Research 53, 114–121. Vandeputte, K., Moens, L., Dams, R., 1996. Improved sealed-tube combustion of organic samples to CO2 for stable isotopic analysis, radiocarbon dating and percent carbon determinations. Analytical Letters 29, 2761–2774. Walker, M.J.C., Bryant, C., Coope, G.R., Harkness, D.D., Lowe, J.J., Scott, E.M., 2001. Towards a radiocarbon chronology of the Late-Glacial: sample selection strategies. Radiocarbon 43, 1007–1019. Ward, G.K., Wilson, S.R., 1978. Procedures for comparing and combining radiocarbon age determinations: a critique. Archaeometry 20, 19–31. Wohlfarth, B., Skog, G., Possnert, G., Holmquist, B., 1998. Pitfalls in the AMS radiocarbon-dating of terrestrial macrofossils. Journal of Quaternary Science 13, 137–145. Xu, S., Anderson, R., Bryant, C., Cook, G.T., Dougans, A., Freeman, S., Naysmith, P., Schnabel, C., Scott, E.M., 2004. Capabilities of the new SUERC 5MV AMS facility for 14C dating. Radiocarbon 46, 59–64.