JOURNAL OF QUATERNARY SCIENCE (2004) 19(3) 251–262 Copyright ß 2004 John Wiley & Sons, Ltd. Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jqs.842

Climatic change recorded by stable isotopes and trace elements in a British Holocene tufa E. R. GARNETT,1 J. E. ANDREWS,1* R. C. PREECE2 and P. F. DENNIS1 1 School of Environmental Sciences, University of East Anglia, Norwich, NR4 7TJ, England 2 Department of Zoology, University of Cambridge, Downing Street, Cambridge, CB2 3EJ, England Garnett, E. R., Andrews, J. E., Preece, R. C. and Dennis, P. F. 2004. Climatic change recorded by stable isotopes and trace elements in a British Holocene tufa. J. Quaternary Sci., Vol. 19 pp. 251–262. ISSN 0267–8179. Received 29 October 2003; Revised 26 January 2004; Accepted 29 January 2004

ABSTRACT: Combined stable isotope (
18O and
13C) and trace element (Mg, Sr) geochemistry from bulk tufa calcite and ostracod shell calcite from an early Holocene British tufa reveal clear records of Holocene palaeoclimatic change. Variation in
18O is caused principally by change in the isotopic composition of Holocene rainfall (recharge), itself caused mainly by change in air temperature. The
13C variability through much of the deposit reflects increasing influence of soil-zone CO2, owing to progressive woodland soil development. Bulk tufa Mg/Ca and Sr/Ca are controlled by their concentrations in the spring water. Importantly, Mg/Ca ratios are not related to
18O values and thus show no temperature dependence. First-order sympathetic relationships between
13C values and Mg/Ca and Sr/Ca are controlled by aquifer processes (residence times, CO2 degassing and calcite dissolution/reprecipitation) and probably record intensity of palaeorainfall (recharge) effects. Stable isotope records from ostracod shells show evidence of vital effects relative to bulk tufa data. The ostracod isotopic records are markedly ‘spiky’ because the ostracods record ‘snapshots’ of relatively short duration (years), whereas the bulk tufa samples record averages of longer time periods, probably decades. The
18O record appears to show early Holocene warming, a thermal maximum at ca. 8900 cal. yr BP and the global 8200 yr BP cold event. Combined
13C, Mg/Ca and Sr/Ca data suggest that early Holocene warming was accompanied by decreasing rainfall intensity. The Mg/Ca data suggest that the 8200 yr BP cold event was also dry. Warmer and wetter conditions were re-established after the 8200 yr BP cold event until the top of the preserved tufa sequence at ca. 7100 cal. yr BP. Copyright ß 2004 John Wiley & Sons, Ltd. KEYWORDS: tufa; ostracods; Holocene; stable isotopes; trace elements.

Introduction There is much current research interest in attempting to extract palaeoclimatic and palaeoenvironmental information from terrestrial sediments. When compared with both marine (typically sediment cores) and polar (ice core) environments, information from terrestrial records is typically of short temporal duration and geographically dispersed. It is therefore important to establish both the quality and availability of as many terrestrial climatic and environmental proxies as possible. Research on terrestrial carbonate sediments, typically speleothems and lake sediments (primary inorganic carbonate precipitates, calcareous shells (ostracods, molluscs) or charophytes), has demonstrated that long and detailed palaeoclimate records can be gained using stable isotopes (see e.g. Hammarlund et al., 1999; Schwalb et al., 1999; von Grafenstein et al., 1999a; McDermott et al., 2001; Ayalon et al., 2002) and trace elements (e.g. Ayalon et al., 1999; Fairchild et al., 2000). * Correspondence to: J. Andrews, School of Environmental Sciences, University of East Anglia, Norwich, NR4 7TJ, England. E-mail: [email protected]

In addition to speleothems and lake carbonates, it is very likely that tufas may also contain both palaeoclimatic and palaeoenvironmental information in their geochemical records. Moreover, tufa deposits are widespread on limestone bedrock, and their surface occurrence makes them, on the whole, much easier to study and recover than either speleothems or lake sediments. Indeed, early research undertaken by Pazdur et al. (1988) concluded that the oxygen isotope composition of fossil Polish tufa appeared to be sensitive to temperature and humidity, in reasonable agreement with other temperature proxies for southern Poland. Subsequent research by Andrews et al. (1993, 1997, 2000), Janssen (2000), Matsuoka et al. (2001) and Ihlenfeld et al. (2003) concluded that the isotope geochemistry of recent tufa contains an interpretable palaeoclimatic and palaeoenvironmental signal; a signal that may also be preserved in fossil tufa (Andrews et al., 1994). The aims of this study were first, to establish if stable isotope geochemistry of well-characterised Holocene tufa (i.e. of known palaeoenvironment and age) can be used unequivocally to decipher palaeoclimatic change. We set out to study this using both bulk tufa and coexisting ostracod shell calcite

252

JOURNAL OF QUATERNARY SCIENCE

data, allowing detailed intercomparison for the first time. In addition, we set out to test whether trace element data from Holocene tufas could provide valuable additional climatic/ environmental information, particularly given the encouraging results of Ihlenfeld et al. (2003).

Study materials and site Tufa deposits with simple layer-cake stratigraphy, and hence clear historical records, are typically found in pool deposits in barrage sequences, or spring-fed paludal environments (Andrews et al., 2000). These types of deposits are typically micritic and of low porosity, so avoiding the problems of pervasive secondary cementation by sparitic pore cements prevalent in waterfall and barrage tufas. Pool deposits are, however, potentially prone to contamination with both detrital clay/silt and reworked tufa carbonate, a potentially serious problem for geochemical and geochronological studies (Garnett et al., 2004). We chose to study a paludal Holocene tufa site at Wateringbury in southern England (Fig. 1). Although not well exposed at the present-day, this site has simple stratigraphy, and most important, an unusually good existing Holocene chronology

Figure 1 Wateringbury tufa deposit displaying radiocarbon dates (six on carbonate, one on wood), and calibrated radiocarbon dates. The calibrated ages are the age ranges that contain 95.4% of the area under the probability curve. All calibrated ages were determined with CALIB 4.3 (Stuiver and Reimer, 2000) using the bidecadal atmospheric curve (Stuiver and Reimer, 1993, and references therein). ‘T’ ¼ soft white tufa, ‘O’ ¼ oncoids, vertical lines ¼ grey tufa—soil horizon? horizontal lines ¼ iron pan, diagonal lines ¼ Atherfield Clay Copyright ß 2004 John Wiley & Sons, Ltd.

Figure 2 Depth versus age profile for radiocarbon ages from Fig. 1. Diamonds show linear relationship for carbonate 14C ages, suggesting relatively constant deposition rate. Carbon-14 age on wood (square) and dashed line show probable age relationship for carbonate dates corrected for ‘hard water’ effects

based on 14C age dates and supported by molluscan biostratigraphy (Kerney et al., 1980). The Wateringbury tufa is situated in the small village of Wateringbury (NGR TQ 6876 5344), 6 km southwest of Maidstone in southeast southeast England (Fig. 1). The tufa deposit is elliptical in shape, covering an area of approximately 100 by 73 m, and is about 2.5 m thick (Kerney et al., 1980). The Holocene deposit commences with an oncoidal facies that passes upward into soft white to grey tufa (Fig. 1). The sequence is punctuated by an iron-stained horizon (Fig. 1), probably marking the position of the former water table, and by grey horizons reflecting pauses in carbonate formation and possibly incipient soil development. The environment of deposition was a spring-fed vegetated calcareous swamp with small pools and shallow films of water trickling across wet ground (Kerney et al., 1980). The springs that precipitated the Holocene tufa are today artificially channelled to a pond. Active tufa formation, although restricted to small surface crusts and moss tufas at the outlet of this pond does, however, allow comparative geochemical study. The Holocene chronology is based on 14C dating of both wood and tufa carbonate (Kerney et al., 1980; Preece, 1991: Fig. 1). Radiocarbon dating of terrestrial carbonate is, of course, particularly prone to well-known ‘hard water/dead carbon’ effects (e.g. Genty et al., 1999) that can result in an older than expected age. This may be the case with some of the Wateringbury 14C ages, e.g. where calcite sample Gd-5538 at 1.9 m depth gives the same calibrated age (9490
187 cal. years BP) as wood sample Q-1425 some 0.7 m deeper in the section (Fig. 1). An attempt to ‘correct’ for the ‘hard water/dead carbon’ effect can be seen in Fig. 2. Overall, however, the dates are stratigraphically consistent, suggest a relatively constant deposition rate (Fig. 2) and are close to expected ages based on molluscan biostratigraphical data (Kerney et al., 1980). Attempts to corroborate and expand the existing 14C chronology with U-series dating (Garnett et al., 2004) were unsuccessful owing to low initial U contents, short time for ingrowth of radiogenic 230Th and detrital contamination, although this study did confirm the early Holocene age. Approximate ages given in this paper are interpolations based on the apparent linear tufa accumulation rate (Fig. 2) and are shown in italics. Errors on the interpolated ages are clearly large (>200 yr minimum), at least as large as those on the bracketing 14C ages. The geochemical work was based mainly on archived tufa collected by Kerney et al. (1980). The rather coarse sampling regime of Kerney et al. (1980) is not ideal for high-resolution climatic or environmental investigation. For example, individual isotope analyses of bulk Holocene tufa were based on J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

CLIMATIC CHANGE RECORDED FROM BRITISH HOLOCENE TUFA

homogenised samples of centimetric thickness, which probably translate into decadal resolution. However, this drawback is greatly outweighed by the excellent chronology and biostratigraphy available. Ostracods ‘picked’ for the original study (Kerney et al., 1980) were unfortunately no longer available; the tufa samples were thus ‘re-picked’ for ostracods, except the cored section (below 3 m; Fig. 1) where the preserved oncoidal material was devoid of ostracods. To remedy this, the basal metre was re-cored in July 2001, approximately half a metre from the original site. Six samples were acquired, correlated with the original samples and ‘picked’ for ostracods.

Screening for diagenetic alteration Before geochemical study the tufa samples were examined for evidence of diagenetic alteration. The tufas are soft and friable with dense or clotted microbial micritic fabrics in impregnated thin-section. There is no evidence of pervasive diagenetic recrystallisation or sparry calcite cementation of void space. Scanning electron microscopy showed groovecast and millstone textures indicative of carbonate precipitation around root hairs and ridged diamond textures reminiscent of calcified root epidermis in calcretes (Klappa 1980). Above the iron-stained horizon at 2.2 m depth (Fig. 1) micron-scale crusts of fungalrelated needle-fibre calcite (Verrecchia and Verrecchia, 1994) occur around the edges of micropores, but do not constitute significant volumes of calcite. The micrites are nonluminescent (under cathodoluminescence, CL); molluscan shell aragonite is still pristine (checked by X-ray diffraction) and ostracod shell calcite is very well preserved; all of which suggest that these tufas have suffered minimal diagenetic alteration. Cathodoluminescence revealed little or no evidence of detrital carbonate clasts in the micrite.

Analytical methods Stable isotopes For bulk tufa, approximately 0.1 g of sample was dried overnight at 110  C and then ground to a fine powder. Volatile organic matter was removed by low temperature (<80  C) oxygen plasma ashing for 3 h at 300 W forward power in a Bio-Rad PT 7300 plasma barrel etcher. With the exception of the Wateringbury re-cored samples, stable isotope analyses were performed on CO2 derived from 3 to 5 mg tufa sample reacted with anhydrous H3PO4 at 25  C overnight. Isotope ratios were measured on a VG Sira Series II mass spectrometer at the University of East Anglia. Replicate analyses of the laboratory standard (n ¼ 24) gave a 2 precision of 0.11% for oxygen and 0.06% for carbon. For the re-cored samples, 0.3
0.03 mg of dried plasma ashed tufa was reacted with anhydrous H3PO4 at 70  C until the entire sample had reacted. Isotope ratios were measured on a Europa Sigma Hybrid, with an in-house automated sampler, at the University of East Anglia. Replicate analyses of the laboratory standard (n ¼ 11) gave a 2 precision of
0.11% for oxygen and
0.19% for carbon. Extraction of ostracods from the bulk tufa was done exactly as suggested by Griffiths and Holmes (2000). Before geochemical analysis ostracod shells were cleaned using a fine paintbrush and high purity Milli-QTM water (Holmes, 1992; Griffiths and Holmes, 2000). Stable isotope analyses were performed on CO2 derived from one to six adult valves of Copyright ß 2004 John Wiley & Sons, Ltd.

253

Psychrodromus olivaceus, which were reacted with anhydrous H3PO4 at 80  C for 20 minutes. Isotope ratios were measured on a PDZ-Europa Geo 20/20 mass spectrometer with CAPS automatic carbonate system at the School of Ocean Sciences, University of Wales, Bangor. Replicate analyses of the laboratory standard (n ¼ 14) gave a 2 precision of 0.08% for oxygen and 0.2% for carbon. The
18O of modern water samples were analysed following equilibration with CO2 of known isotopic composition (Epstein and Mayeda, 1953) and
13C DIC (dissolved inorganic carbon) was collected and fixed using the method of Bishop et al. (1990). Isotopic measurements were made on the UEA Europa Sigma Hybrid with 2 precision of
0.11% (
18O water) and
0.19% (
13C DIC). Isotopic data in this paper are reported relative to the V-PDB scale unless stated otherwise.

Trace elements For bulk tufa, 500 mg of dried homogenised sample was dissolved in 25 ml of 10% acetic acid in ultrapure Milli-QTM water. Trials showed that dilute acetic acid caused minimal leaching of elements from acid insoluble residues (IR). Solutions were then filtered with an in-line syringe GFC filter and made up to 100 ml. Trace elements (only Mg and Sr discussed here) were determined using a Varian Vista pro CCD simultaneous ICP–AES and Ca was determined by EGTA titration (Jeffery et al., 1989). The ICP data were drift and blank corrected; mean 1 standard deviations of 12.3 ppm (Mg) and 3.3 ppm (Sr) were propagated from weighing errors and errors associated with the ICP concentration–intensity regression.

Results Limited stable isotope data for the water and active modern tufa deposits are given in Tables 1 and 2. The bulk Holocene tufa
18O values range between 4.8 and 6.3% (mean of

Table 1 Modern Wateringbury spring water
18O and
13C DIC. Errors are 1 sigma Date

09/10/01 13/01/02 03/04/02 11/07/02 Mean


18O % V-SMOW

DIC
13C % V-PDB

Water temperature (  C)

7.01
0.04 7.04
0.04 7.07
0.04 7.18
0.04 7.25
0.04 7.32
0.04 7.35
0.04 7.17
0.02

13.0
0.24 12.3
0.1

12.8 8.0

11.0
0.1

12.8

12.2
0.1

17.0

12.1
0.07

12.6

Table 2
18O and
13C (V-PDB) for modern Wateringbury tufa precipitates. Both samples contained a living ‘green’ biofilm layer assumed to contain the most recent tufa precipitates. Errors are to the one sigma level 13/1/02-1

13/1/02-1 (biofilm)

13/01/02-2 13/01/02-2 Mean (biofilm)


18O% 6.0
0.1 5.6
0.13 5.3
0.3 4.9
0.06 5.4
13C% 9.5
0.03 9.2
0.02 12.1
0.18 12.1
0.18 10.7 J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

254

JOURNAL OF QUATERNARY SCIENCE

Table 3 Wateringbury stable isotope and trace element data summaries. Element data are for bulk fufa

Range Mean


18O % Tufa


13C% Tufa

4.8 to 6.6 5.5

8.1 to 11.4 9.4


18O% Ostracods 3.6 to 5.7 4.4


13C% Ostracods

Mg ppm

Sr ppm

7.6 to 9.6 333–732 108–186 8.8 588 158

Mg/Ca (molar)

Sr/Ca (molar)

0.0019 to 0.0031 1.56  104 to 2.28  104 0.0026 1.94  104

Figure 3 Oxygen isotope values for Wateringbury bulk tufa and ostracod Psychrodromus olivaceus, plotted against the tufa profile (see Fig. 1 for key)

5.5%; Table 3), whereas the
13C values range between 8.1 and 11.4% (mean of 9.4%; Table 3): these data are shown in stratigraphical order in Figs 3 and 4. The
18O profile in Fig. 3 shows an overall trend (albeit with up to about 0.5% variability) of increasing values upward from the base of the deposit (around 6%) to 1.85 m (values around 5%). In the upper part of the profile there is a low in the
18O values, between 1.85 and 0.7 m, where values decline from 4.8% to 5.9% at 1.15 m, and then rise to 5.2%. The profile of
13C values (Fig. 4) shows more variability (up to 2.5%) than the
18O values, with an overall trend of
13C values increasing from < 10% at the base until 2.9 m, where values > 9% first occur. After this a general trend to more negative values with time is recorded upwards but with significantly less negative excursions (Fig. 4). The Holocene ostracod calcite
18O values range between 3.6 and 5.7% (mean of 4.4%; Table 3) whereas the
13C values range between 7.6 to 9.6% (mean of 8.8%; Table 3); these data are shown in stratigraphical order in Figs 3 and 4. There is little clear structure in the profile of the ostracod
18O data, variability being generally around 0.5%, with the exception of a very negative value of 5.7% at 1.15 m (Fig. 3). Valves were reanalysed from this horizon, giving a repeat negative value of 5.4%. These values at 1.15 m are clearly distinctly negative relative to the rest of the data. Copyright ß 2004 John Wiley & Sons, Ltd.

Overall, ostracod
13C values decrease up-profile from near 8% to 9.5% (Fig. 4). Superimposed on this trend there is slight increase in
13C (from 9.2 to 8.2%) between 1.65 and 1.15 m, with two clear high values (7.6 and 8.1%) at 1.3 and 1.15 m respectively. Although a suite of trace elements were measured in this study (Mg, Sr, Ba, Fe, Mn, K) only Mg and Sr are discussed here. Trace elements were also measured in ostracod shell calcites, but the results do not contribute any additional insights and are not presented here (see Garnett (2003) for the full data set). The other elements were measured mainly to assess diagenetic effects, and the effects of non-carbonate impurities in the tufas. Most important in this context, a positive relationship between K and IR (r ¼ 0.65), but weak negative relationships between K–IR and Mg–Sr (data not presented), suggests that leaching of impurities during acid digestion was insignificant. Concentrations of Mg and Sr range between 330 and 730 ppm and 100 and 200 ppm, respectively (Table 3). The data are represented as profiles of Mg/Ca and Sr/Ca molar ratios in Fig. 5. Both Mg/Ca and Sr/Ca increase upward in the basal metre of the deposit (Fig. 5) to values >0.003(Mg/Ca) and >0.0002 (Sr/Ca) respectively. Above 1.5 m both ratios decrease again toward the top of the deposit. The overall shape of these profiles is similar to those of both the bulk tufa
13C and >2 mm sediment fraction (Fig. 5). J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

CLIMATIC CHANGE RECORDED FROM BRITISH HOLOCENE TUFA

255

Figure 4 Carbon isotope values for Wateringbury bulk tufa and ostracod Psychrodromus olivaceus, plotted against the tufa profile (see Fig. 1 for key)

Figure 5 Wateringbury Mg/Ca, Sr/Ca,
13C and sediment >2 mm plotted against the tufa profile (see Fig. 1 for key). 1 sigma errors are smaller than data points Copyright ß 2004 John Wiley & Sons, Ltd.

J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

256

JOURNAL OF QUATERNARY SCIENCE

Interpretation Stable isotopes Bulk tufa
18O The
18O data values of both the Holocene and active bulk tufas are consistent with those from modern riverine northwest European tufas and freshwater microbial carbonates as a whole (Andrews et al., 1993, 1997). There is no significant relationship between the
18O and
13C in the bulk tufa data (Table 4); isotopic covariance is not expected in temperate climate riverine or paludal tufas where there is sufficient throughflow for evaporation and residence time effects to be negligible (Andrews et al. 1993, 1997). The mean
18O of the modern spring water (7.2% VSMOW; Table 1) is essentially a record of the isotopic composition of the regional rainfall. Contour maps of
18O in modern British groundwater (Darling et al., 1996) show the 7% VSMOW contour passing through Kent. Darling et al. (1996) also demonstrated that the long-term average rainfall
18O records for stations in Oxfordshire and Nottinghamshire are comparable to the groundwater signal. The measured calcite
18O for modern Wateringbury tufa crusts associated with living biofilms collected in January 2002 is between 4.9 and 5.6% (Table 2). At the time of sampling the water temperature was 8  C and the water
18O value was 7.0% V-SMOW (Table 1). Meteoric calcite precipitated in isotopic equilibrium with this water should be around 5.2% (using the Hays and Grossman (1991) equation), indicating that the tufa
18O is close to isotopic equilibrium. We expected small deviation from the equilibrium value because it is not possible to prove that the tufa and water samples are truly cogenetic, and it is unlikely that a measured
18O value in natural tufa will relate exactly to the water conditions at the time of sampling. Assuming similar isotopic equilibrium between Holocene tufa calcite formation and spring water, long-term variations in
18O are likely to have been caused mainly by: (i) changes in the temperature at the time of calcite precipitation; (ii) variation in the isotopic composition of the tufa precipitating water; or (iii) a combination of both. A 6  C change in the spring water temperature would have been required to explain the total measured variation in Holocene bulk tufa
18O (1.5%; Fig. 3), as a 1  C change in the temperature of calcite precipitation results in a 0.24% alteration in
18O (Craig, 1965). This mechanism is rejected as the main cause of Holocene bulk tufa
18O variability, as groundwater temperature variation of this magnitude did not occur during the Holocene. Temperature variation during the Holocene, including the climatic optimum (9 to 4 ka), was probably only around 1–2  C (Bell and Walker, 1992; Andrews et al., 1994). Temperature variation of around 6  C is more in line with the rapid climate amelioration between the Loch Lomond stadial (Younger Dryas) and the Holocene (Atkinson et al., 1987; Hammarlund et al., 1999). Moreover, if the variation in Holocene bulk tufa
18O were solely the result of changes in the temperature of calcite precipitation, early Holocene climate amelioration (Atkinson et al., 1987; Dansgaard et al., 1993) should be reflected in more negative tufa
18O values. The clear overall trend to more positive tufa
18O in the basal 2 m does not record this. Change in the isotopic composition of meteoric water feeding groundwater and the Wateringbury spring is the other obvious mechanism for explaining the Holocene bulk tufa
18O variability. Moreover, it has been established that mean rainfall
18O in Europe is strongly related to mean microbial (tufa) calcite
18O (Andrews et al., 1997). Variability in the Copyright ß 2004 John Wiley & Sons, Ltd.

isotopic composition of meteoric water in temperate regions at a given latitude and altitude are caused principally by air temperature change and the ‘amount effect’ (Lawrence and White, 1991; Andrews et al., 1994). For example, at Groningen in The Netherlands, temperature effects mainly control rainfall
18O during the cold months, whereas the amount-isotope effect seems to be significant in the summer months (Lawrence and White, 1991). This pattern is corroborated to some extent in the Malham area of Britain, where both large volume and cold winter rainfall tends to have an isotopically light composition (Lawler, 1987). In the British Isles, groundwater has been shown to represent the mean
18O in rainfall (Lawler, 1987; Darling et al., 1996) so anomalous and seasonal rainfall events tend to be smoothed out by groundwater mixing as a result of long aquifer residence times. Therefore the
18O signature of groundwater feeding a spring, such as that at Wateringbury, will be strongly influenced by the dominant controlling factor over
18O in rainfall, which appears to be air temperature (Rozanski et al., 1993), in addition to some influence from rainout and source effects (Andrews et al., 1994). Although some influence from the air mass source direction can be detected in the
18O of modern rainfall (Lawler, 1987), the dominant source of British Holocene air masses is not known with certainty, but is likely to have been similar to present (Andrews et al., 1994). If change in the Holocene tufa
18O were controlled predominantly by the air temperature (
source/amount effect), a trend toward more positive
18O would correspond to increasing temperatures (Andrews et al., 1994). Interpreted on this basis the bulk tufa
18O data in Fig. 3 record increasing temperature from the base of the deposit (12 221
538 cal. yr. BP; Fig. 1) to 1.85 m (ca. 8900 cal. yr BP). The ‘trough’ in
18O values between 1.85 and 0.7 m is thus likely to indicate a decline in temperature, with the minimum
18O centred on 1.15 m (ca. 8000 cal. yr BP), followed by a rise in temperature. When surface air temperatures range between 1 and 10  C (early Holocene temperatures are likely to fall within the upper part of this range (Atkinson et al., 1987)), a 1  C increase in temperature results in a 0.58% increase in
18O (Rozanski et al., 1993). The 1.5% variation in tufa
18O seen from near the base of the deposit to 1.85 m would thus correspond to a temperature variation around 2.6  C and the 1% decline in
18O in the ‘trough’ between 1.85 and 0.7 m would correspond to a 1.7  C decrease in temperature. Note also that the temperature effects on
18O related to calcite precipitation (see above) work in isotopic opposition to changes caused by meteoric water
18O variation. Any isotopic effect caused by the temperature of calcite precipitation would thus dampen (reduce) the ‘climatic temperature signal’. The magnitude of temperature changes recorded by the bulk tufa
18O data are similar to those derived from other British Holocene climate proxies (Atkinson et al., 1987; Guiot et al., 1993; Andrews et al., 1994; Coope et al., 1998; Rousseau et al., 1998) and to those from across Europe (Barnett et al., 2001; Barnekow and Sandgren, 2001; Rose´n et al., 2001). The suggestion that climatic temperature is the main cause of bulk tufa
18O variation is supported when the Wateringbury
18O profile is compared with the GISP2 (Greenland Ice Sheet Project)
18O profile (Fig. 6). The profiles were correlated using both the calibrated 14C and ‘dead carbon corrected’ Wateringbury chronologies (Figs 1 and 2), which differ by some 400 yr, but do not change the correlation significantly. Using the calibrated 14C chronology, tufa deposition occurred between 12 221
538 and 7092
281 cal. yr BP (Fig. 1), with the negative
18O value at 1.15 m corresponding to an age of ca. 8000 cal. yr BP. The ‘dead carbon corrected’ chronology places tufa deposition between ca. 11 360 and ca. 6520 cal. J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

CLIMATIC CHANGE RECORDED FROM BRITISH HOLOCENE TUFA

257

Figure 6 Correlation of Greenland ice-core (Greenland Ice Sheet Project 2—GISP2)
18O (Steig et al., 1994; Stuiver et al., 1995, 1997) and Wateringbury tufa
18O. The calendar ages from GISP2 are from annual ice-layer counting (Messe et al., 1994). This comparison shows similar overall structure in
18O profiles including early Holocene warming, and possibly the 8200 yr BP cold event and Preboreal oscillation

yr BP, with the negative
18O value at 1.15 m, corresponding to ca. 7740 cal. yr BP. Clearly there is similarity between the first-order shapes of the profiles in Fig. 6. The 1% decrease in tufa
18O centred on 1.15 m (ca. 8000 cal. yr BP) equates to about a 1.7  C decrease in temperature. This event might be related to what is now known to have been a globally significant cold event at 8200 cal. yr BP. This event was originally detected in the GISP2 (Fig. 6) and GRIP Greenland ice-cores (Dansgaard et al., 1993; O’Brien et al., 1995; Alley et al., 1997). More recently, however, it has been identified in speleothems (Baldini et al., 2002), terrestrial mollusc sequences (Rousseau et al., 1998) and lacustrine sediments via loss-on-ignition (Nesje and Dahl, 2001), diatoms (Korhola et al., 2000), pollen (Magny et al., 2001), ostracods (von Grafenstein et al., 1998) and chironomids (Rose´n et al., 2001). The cause of the 8200 yr BP cold event is not yet known with certainty: it may relate to periodicity within the climate system (O’Brien et al., 1995; Bond et al., 1997; Campbell et al., 1998) or pulses of freshwater into the North Atlantic (Alley et al., 1997; Klitgaard-Kristensen et al., 1998; von Grafenstein et al., 1998; Barber et al., 1999). The 1% decline (1.7  C cooling) in
18O defined in the Wateringbury tufa is consistent with other northwest European estimates of cooling at this time (e.g. von Grafenstein et al., 1998; Rousseau et al., 1998; Magny et al., 2001). More problematic is the apparent duration of the Wateringbury event, which is recorded by 1.15 m of tufa deposit representing approximately 1500 calendar years assuming constant sedimentation rates. This time estimate is very poorly constrained by the present chronology, but is much longer than other European proxies that suggest up to 500 yr duration for the Copyright ß 2004 John Wiley & Sons, Ltd.

event (e.g. Rousseau et al., 1998; von Grafenstein et al., 1998). It is possible that the gradual decline in
18O between 9000 and 8330 cal. yr in GISP2 (Fig. 6) is recorded by the 0.6% decline at Wateringbury (between 1.85 and 1.2 m). The rapid decrease in
18O to the 8200 yr BP cold event seen in the GISP2 core might then correspond to the 0.4% decline between 1.2 and 1.15 m at Wateringbury. This rapid decline would equate to about 0.7  C of cooling and is consistent with similar estimates of cooling for southeast England between 8000 and 8500 cal. yr BP (1  C for both summer and winter temperatures) based on terrestrial mollusc assemblages (Rousseau et al., 1998). Moreover, the groundwater-fed Wateringbury spring tufa record would tend to be dampened by the total variation in temperature owing to residence time effects and mixing in the aquifer. The 1% decrease in bulk tufa
18O between 3.9 m and 3.75 m (ca. 11 400–11 250 cal. yr BP; Fig. 6) could be a record of the Preboreal Oscillation (ca. 11 300 cal. yr BP) (Dansgaard et al., 1993), which has been interpreted as a temperature decrease of 1–3  C in southern Sweden (Hammarlund et al., 1999). Although first-order variations in bulk tufa
18O appear to be related to variations in temperature, there are distinct ‘spikes’ that deviate from the dominant trends. For example, two distinct negative spikes in
18O occur at 3.15 and 2.05 m (Fig. 3) within an overall trend towards more positive
18O. These ‘spikes’ are clearly a record of isotopically lighter
18O being introduced to the system, possibly from changes in dominant airmass source and/or amount effects, or perhaps more likely, from regional changes in climate. For example, establishment of larger volumes of cold isotopically light winter rainfall (Lawler, 1987), that temporarily altered the isotopic composition of J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

258

JOURNAL OF QUATERNARY SCIENCE

Table 4 Correlation coefficient (r) and P values for Wateringbury isotope and trace element data


18O tufa
13C tufa
18O ostracod
13C ostracod Mg/Ca


13C tufa


18O ostracod

0.09 —

0.33* n/a —


13C ostracod n/a 0.40** 0.29 —

Mg/Ca

Sr/Ca

0.02 0.58*** n/a n/a —

0.09 0.53*** n/a n/a 0.54***

*P  0.05, **P  0.01, ***P  0.001 (P  0.05 is statistically significant at the 95% confidence level, etc.).

the groundwater and spring. We do note, however, that these
18O ‘spikes’ do not appear to be represented by other geochemical proxies associated with palaeohydrology, for example, trace element molar ratios (see below).

Ostracod shell calcite
18O Compared with the
18O record in the bulk tufa, the shell calcite
18O values of the ostracod Psychrodromus olivaceus are displaced by, on average, 1% (range 0.4–1.8%) toward more positive values (Fig. 3). This is consistent with earlier work showing that ostracods produce biological vital effects that shift their shell
18O from equilibrium values (Xia et al., 1997, von Grafenstein et al., 1999b, Keatings et al., 2002). Vital effects have been found to be constant within species and possibly genus too (von Grafenstein et al., 1999b), but there is no published data for P. olivaceus, the ostracod used in this study. The correlation between bulk tufa and ostracod
18O is low (r ¼ 0.33; Table 4), although the data below 1.1 m has a stronger relationship (r ¼ 0.56). It is not, however, surprising that there is no strong correlation between the bulk tufa and ostracod
18O. The tufa samples were of centimetric size and represent time-averaged records (decade(s)) whereas the ostracod records are specific to the duration of shell calcification. Psychrodromus olivaceus, for example, probably calcifies throughout the year (Meisch, 2000) with a life expectancy of about 1 yr. The one to six valves analysed per sample thus represent a mean value for a maximum of 6 yr. The ostracod records are thus specific ‘snapshots’ of relatively short duration. Visual evaluation of the data in Fig. 3 suggests that, although more ‘spiky’, the ostracods record shows some similarity to the bulk tufa. For example, early Holocene climatic warming is shown and the trend to more negative values at the postulated 8200 yr BP cold event is clear, including the very negative value at 1.15 m. Above this horizon the ostracod
18O profile is noticeably ‘spiky’ probably because ostracods were less abundant in this part of the deposit such that number of valves per sample analysed fell to about two to three.

mean measured tufa value was 10.7% (Table 2) demonstrating close correspondence to empirical values. Changes in tufa
13C on Holocene time-scales are likely to reflect the relative proportion of isotopically light CO2 derived from soil organic matter, mixed with isotopically heavier carbon derived from the dissolution of the aquifer limestone, and equilibration of the aquifer and spring water with atmospheric CO2 (Andrews et al., 1993, 1997). The basal part of the Wateringbury tufa (below 2.3 m) is oncoid-rich (Fig. 1), which suggests flowing spring-fed water, probably with flow rates decreasing with time. The ostracod assemblages (Garnett, 2003) and particle size data (coarsest at base; Fig. 5) support this interpretation, and the mollusc assemblage indicates an open grassy environment with thin soils below 3.65 m (Kerney et al., 1980). Bulk tufa
13C values in this part of the deposit increase from around 10 to 8.5%. This may indicate increasing influence of aquifer limestone dissolution as aquifer-water residence times increased and spring flow rates decreased. Alternatively the
13C increase could indicate more complete equilibration between spring-water and atmospheric CO2 as flow rates decreased. The very negative basal
13C (Fig. 4) may be anomalous owing to the presence of macroscopic plant detritus found toward the base of the deposit (Kerney et al., 1980). Above 3 m the
13C profile (ca. 10 300 cal. yr BP) begins an overall upward trend to decreasing values, from 8.5 to 10.5% at the top of the deposit. This suggests a steadily increasing input of isotopically light soil-zone CO2, which agrees with changes in the molluscan fauna, indicating woodland development at the expense of grassland (Kerney et al., 1980). Establishment of woodland would provide leaf litter and humus development, increasing root respiration and microbial organic decay, which in turn increased the generation of isotopically light soil carbon. The bulk tufa
13C profile is noticeably more ‘spiky’ above 2 m (ca. 8900 cal. yr BP), perhaps related to variable photosynthetic activity from aquatic vegetation on the margins of the tufa.

Ostracod shell calcite
13C Compared with the bulk tufa
13C, the shell calcite
13C values of the ostracod Psychrodromus olivaceus are displaced on average by about 0.6% (range of 1.9% more positive to 0.76% more negative) to more positive values, a mean offset probably resulting from vital effects (von Grafenstein et al., 1999b). The overall shape of the two
13C profiles are similar (Fig. 4), particularly the trend with time towards more isotopically light
13C (as discussed and interpreted above). Some of the distinct spikes in the bulk tufa
13C record are repeated in the ostracod profile (e.g. at 0.25 m and 1.15 m; Fig. 4), although others clearly are not. This is probably a function of the more time-specific ostracod record, compared with the rather timeaveraged tufa record (as discussed above for
18O), which also explains the weak statistical relationship between the two
13C records (Table 4).

Bulk tufa
13C The mean modern spring water
13C DIC value of 12.1% (Table 1) reflects principally the input of isotopically light CO2 derived from the decomposition of soil organic matter (Andrews et al., 1993, 1997). Equilibrium precipitation of CaCO3 from this DIC should result in tufa calcite
13C about 1.0% higher than the DIC, an enrichment factor independent of kinetic effects and temperature (Romanek et al., 1992). The Copyright ß 2004 John Wiley & Sons, Ltd.

Bulk tufa trace elements: possible palaeorainfall and residence time indicators In this study, a statistically significant (r ¼ 0.54; Table 4) sympathetic relationship was found between Mg/Ca and Sr/Ca molar ratios in the bulk tufa calcites, resulting in similar profiles (Fig. 5). In addition a similar statistically significant relationship J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

CLIMATIC CHANGE RECORDED FROM BRITISH HOLOCENE TUFA

was found between these element ratios and the bulk tufa
13C (Table 4 and Fig. 5). Fractionation of
13C during calcite precipitation from dissolved bicarbonate is insensitive to precipitation rate and temperature (Romanek et al., 1992). The positive relationships between bulk tufa Mg/Ca and Sr/Ca molar ratios and
13C, and the lack of correlation between Mg/Ca and
18O (Table 4) suggest strongly that Mg incorporation into these tufa calcites was not controlled by temperature (discussed in further detail by Garnett (2003)). Covariation between Mg/Ca and Sr/Ca are also seen in speleothem calcites (Fairchild et al., 2000; Verheyden et al., 2000, Huang and Fairchild, 2001), highlighting the importance of palaeohydrological effects (Fairchild et al., 1996, 2000; Roberts et al., 1998; Hellstrom and McCulloch, 2000; Verheyden et al., 2000). Ratios of Mg/Ca in groundwater have long been known to increase with aquifer residence times (Langmuir, 1971; Plummer, 1977), an effect ascribed to incongruent dissolution of dolomite in the presence of calcite (Langmuir, 1971), where short reaction times resulted in preferential calcite dissolution over dolomite (Fairchild et al., 2000). Once saturation with respect to calcite is achieved, continued dissolution of dolomite increases the Mg concentration in the water, whereas Ca concentrations are held constant by precipitation of calcite (Langmuir, 1971; Roberts et al., 1998). Consequently Mg/Ca ratios increase until dolomite saturation is reached. Low Mg/ Cawater ratios are thus associated with winter recharge and short residence times during wet periods. High Mg/Cawater ratios are achieved during summer dryness and longer residence times during dry episodes. The Cretaceous limestone aquifer rock at Wateringbury is not dolomitic (Worssam, 1963), so preferential dissolution of dolomite during longer groundwater residence time is prohibited. However, Fairchild et al. (2000) demonstrated that selective leaching of Mg and Sr over Ca can occur in dolomite-free limestone. Their experimental work showed that strong enrichments in both Mg/Ca and Sr/Ca occur, representing up to five-fold selective leaching of Mg and Sr compared with the composition of the source limestone. They suggested this selective leaching process is due to either non-congruent calcite dissolution, or the accumulation of soluble Mg and Sr related to cycles of dissolution and calcite precipitation along the water flowpath. Both effects should be enhanced during dry conditions with lower water availability in the aquifer and a sympathetic relationship between Mg/Ca and Sr/Ca occurs (Fairchild et al., 2000). The positive correlation between
13C and Mg/Ca and Sr/Ca is unsurprising if variability in both trace elements (as above) and
13C is controlled by aquifer processes. For example, decreased recharge (dry) will increase aquifer water residence time, allowing longer contact and dissolution of aquifer limestone moving
13C to more positive values. Lower recharge also increases air space in the aquifer, encouraging CO2 degassing and calcite precipitation, a process that also increases
13C of the remaining water (Ihlenfled et al., 2003). Dry periods thus increase Mg/Ca, Sr/Ca and
13C together. Values of Mg/Ca, Sr/Ca and
13C show an increasing trend from the base of the deposit until 3.15 m (ca. 10 400 cal. yr BP) suggesting increased aquifer-water residence time owing to drier conditions. This is corroborated by the sedimentology in this part of the tufa profile, where progressive decrease in size and loss of oncoids up-section indicates declining spring flow rates. The small amount of data for >2 mm sediment content (Fig. 5) in this part of the section also suggests an abundance of coarse tufa particles, perhaps concentrated as a lag during declining flow. Above 3 m (ca. 10 300 cal. yr BP), a short episode of apparently dry stable conditions (ca. 10 300– 9050 cal. yr BP) occurred, perhaps supported by evidence of Copyright ß 2004 John Wiley & Sons, Ltd.

259

dry conditions in Cumbrian (northwest Britain) peat bogs around 10 000 yr BP (Hughes et al., 2000). This is followed by decreasing Mg/Ca and Sr/Ca ratios and increasingly negative
13C (Fig. 5) in the upper 2 m (ca. 9050–6700 cal. yr BP) of the Wateringbury deposit. These changes indicate increased wetness, strengthening spring flow and increasing development of woodland vegetation respectively (see above). Polish landslide records suggest increased fluviatile process at the Atlantic–Boreal transition between 8400 and 7700 14C yr BP (Alexandrowicz and Alexandrowicz, 1999). The possible 8200 yr BP cold event, seen in the
18O records at 1.15–1.2 m (see above), corresponds to distinct peaks in Mg/ Ca and
13C (Fig. 5) suggesting much lower spring flow rates and increased aquifer-water residence time. A relatively dry 8200 yr BP cold event also has been identified from speleothem data from southwest Ireland (Baldini et al., 2002). Increased aridity at low latitudes has been further documented for the 8200 yr BP cold event from the GRIP ice-core methane record (Blunier et al., 1995). There is, however, no corresponding peak in Wateringbury Sr/Ca at this horizon. Neither is there a clear geochemical signal associated with the three grey tufa horizons, interpreted as pauses in sedimentation (Figs 1 and 5).

Wider discussion The combined isotopic and trace element data presented here show beyond reasonable doubt that Holocene tufa deposits contain clear palaeoclimatic records. The
18O values in the bulk tufa contain the clearest record of temperature change, which is smoothed and dampened somewhat by the sample size. The results of this study are supported by earlier work on Holocene tufa from Alport in Derbyshire (Andrews et al., 1994), which suggested that early Holocene warming was well underway by ca. 9500 cal. yr BP (ca. 8600 14C yr) as recorded by
18O in bulk tufa calcites. We also know now that the chronology of the Alport tufa above the ca. 8600 14C yr biostratigraphical marker (Andrews et al., 1994) is highly suspect (see Garnett, 2003) and this precludes further detailed comparison of the
18O profiles. The interpretation offered here, that changes in Holocene bulk tufa
18O are caused by climatic-induced variability in the isotopic composition of rainfall, contrasts with recent studies on modern laminated tufa deposits. These studies have concluded that the main driver of seasonal variability in modern tufa
18O is change in stream-water temperature (Chafetz et al., 1991; Matsuoka et al., 2001; Ihlenfeld et al., 2003). These differences are due to two factors. Firstly, the time-scale of sampling is totally different between the two types of study: work on modern tufas is able to resolve seasonal variability (Matsuoka et al., 2001), whereas in studies of Holocene tufas to date, the resolution is much coarser, probably decadal at best (see above). It is thus important that fossil deposits with long records of seasonal laminae are located (and dated) to test whether seasonal records of
18O (and other geochemical proxies) are preserved. Secondly, the strong seasonal records typically come from riverine tufas collected hundreds of metres (Matsuoka et al., 2001) to kilometres (Ihlenfeld et al., 2003), downstream of springs. The longer sample distances are from springs the more chance the system has to be affected by solar insolation (and evaporation in arid climates), such that one might expect a stronger signal from stream-water temperature variability. Sites closer to springs (such as Wateringbury) should be less affected by solar insolation and hence seasonality. It is, however, important to establish that the spring site J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

260

JOURNAL OF QUATERNARY SCIENCE

systems were in isotopic equilibrium, as tufas precipitating very close to springs are well known to suffer disequilibrium isotopic effects. The combined Mg/Ca, Sr/Ca and
13C data from Wateringbury appear to help resolve the palaeorecord of rainfall intensity (recharge) and water residence time effects in the aquifer and support, in part, the findings of Ihlenfeld et al. (2003). These authors studied recent laminated riverine tufa in northern Australia and concluded that sympathetic variations in
13C, Sr and Ba were related to changes in water chemistry, amounts of calcite precipitation in the vadose zone and also to past rainfall variability, consistent with the Wateringbury data. There are, however, some significant differences between these studies. Ihlenfeld et al. (2003) also noted anticorrelation between
18O and Mg, where both records were thought to be recording primarily the temperature of calcite precipitation, which was found to oscillate by ca. 10  C. Their finding, that Mg incorporation into tufa is controlled by temperature, is clearly not appropriate for the long-term temperate-zone Holocene tufa data described here, and may not be appropriate for long-term records generally. The interesting concept of using Mg as a palaeothermometer in fossil tufas is likely to be problematic and we feel that long-term Mg concentration data in most temperate-zone fossil tufas will reflect mainly the Mg/Ca of the water.

Conclusions 1 Combined stable isotope (
18O and
13C) and trace element (Mg, Sr) data from a well-characterised Holocene tufa show, for the first time, a clear record of palaeoclimatic changes. Variation in bulk tufa
18O was caused principally by change in the isotopic composition of Holocene rainfall (recharge), itself caused mainly by change in air temperature. Bulk tufa Mg/Ca and Sr/Ca are controlled by their concentrations in water: Mg/Ca ratios are not correlated with
18O and thus show no temperature dependence. Firstorder relationships between
13C and Mg/Ca and Sr/Ca suggest control by aquifer processes (residence time, CO2 degassing, calcite dissolution/reprecipitation): dry periods (low recharge) increasing Mg/Ca and Sr/Ca and
13C. These parameters therefore may be a record of palaeorainfall intensity. 2 Ostracod shell calcites sampled from the tufa record similar isotopic variability to the bulk tufas. Ostracod
18O values are offset (to more positive values) by about 1% relative to bulk tufa and
13C values are offset (to more positive values) by about 0.6% relative to bulk tufa. These offsets probably result from vital effects. The ostracod isotopic records are more ‘spiky’ because the ostracods record ‘snapshots’ of relatively short duration (years), whereas the bulk tufa samples record averages of longer time periods, probably decades. The ostracod stable isotope data do not contribute additional palaeoenvironmental information and are probably not necessary in future geochemical studies of tufa. 3 Values of
18O in the Wateringbury tufa record temperature trends that are consistent with other Holocene palaeoclimatic records. Early Holocene warming of about 2.6  C occurred until about ca. 8900 cal. yr BP, when temperatures reached their Holocene maximum. Combined
13C, Mg/Ca and Sr/Ca data suggest this warming was accompanied by decreasing rainfall intensity and thus decreasing spring flow. The 8200 yr BP cold event may also be recorded by
18O, and the Mg/Ca data suggest that this event was dry. The
13C record through much of the deposit Copyright ß 2004 John Wiley & Sons, Ltd.

reflects primarily the increasing influence of soil-zone CO2, owing to progressive woodland soil development. The combined data also suggest that warmer and wetter conditions established after the 8200 yr BP cold event until ca. 7100 cal. yr BP where the tufa sequence terminates.

Acknowledgements We thank Paul Kennedy (University of Wales, Bangor) and Sarah Dennis (UEA) who helped us with aspects of the stable isotope analysis. The late Huw Griffiths (University of Hull) confirmed ostracod identifications and Jonathan Holmes (UCL) advised on other aspects of ostracod geochemistry. Charles Turner (OU) helped with the re-coring. This work was part of a NERC (GT 04/99/ES/65) studentship to E. Garnett.

References Alexandrowicz SW, Alexandrowicz Z. 1999. Recurrent Holocene landslides: a case study of the Krynica landslide in the Polish Carpathians. Holocene 9: 91–99. Alley RB, Mayweski PA, Sowers T, Stuiver M, Taylor KC, Clark PU. 1997. Holocene climatic instability: a prominent, widespread event 8200 yr ago. Geology 25: 483–486. Andrews JE, Riding R, Dennis PF. 1993. Stable isotopic compositions of Recent freshwater cyanobacterial carbonates from the British Isles: local and regional environmental controls. Sedimentology 40: 303–314. Andrews JE, Pedley HM, Dennis PF. 1994. Stable isotope record of palaeoclimatic change in a British Holocene tufa. Holocene 4: 349–355. Andrews JE, Riding R, Dennis PF. 1997. The stable isotope record of environmental and climatic signals in modern terrestrial microbial carbonates from Europe. Palaeogeography, Palaeoclimatology and Palaeoecology 129: 171–189. Andrews JE, Pedley MP, Dennis PF. 2000. Palaeoenvironmental records in Holocene Spanish tufas: a stable isotope approach in search of reliable climatic archives. Sedimentology 47: 961–978. Atkinson TC, Briffa KR, Coope GR. 1987. Seasonal temperatures in Britain during the past 22 000 years, reconstructed using beetle remains. Nature 325: 587–592. Ayalon A, Bar-Matthews M, Kaufman A. 1999. Petrography, strontium, barium and uranium concentrations, and strontium and uranium isotopes ratios in speleothems as palaeoclimatic proxies: Soreq Cave, Israel. Holocene 9: 715–722. Ayalon A, Bar-Matthews M, Kaufman A. 2002. Climatic conditions during marine oxygen isotope stage 6 in the eastern Mediterranean region from the isotopic composition of speleothems of Soreq Cave, Israel. Geology 30: 303–306. Baldini JUL, McDermott F, Fairchild IJ. 2002. Structure of the 8200year cold event revealed by a speleothem trace element record. Science 296: 2203–2206. Barber DC, Dyke A, Hillaire-Marcel C, Jennings AE, Andrews JT, Kerwin MW, Bilodeau G, McNeely R, Southon J, Morehead MD, Gagnon JM. 1999. Forcing of the cold event of 8200 years ago by the catastrophic drainage of Laurentide lakes. Nature 400: 344–348. Barnekow L, Sandgren P. 2001. Palaeoclimate and tree-line changes during the Holocene based on pollen and plant macrofossil records from six lakes at different altitudes in northern Sweden. Reviews of Palaeobotany and Palynology 117: 109–118. Barnett C, Dumanyne-Peaty L, Matthews JA. 2001. Holocene climatic change and tree-line response in Leirdalen, central Jotunheimen, south central Norway. Reviews of Palaeobotany and Palynology 117: 119–137. Bell M, Walker MJC. 1992. Late Quaternary Environmental Change: Physical and Human Perspectives. Longman, Harlow. Bishop PK. 1990. Precipitation of dissolved carbon species from natural waters for
13C analysis—a critical appraisal. Chemical Geology (Isotope Geoscience Section) 80: 251–259. J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

CLIMATIC CHANGE RECORDED FROM BRITISH HOLOCENE TUFA Blunier T, Chappallaz J, Schwander J, Stauffer B, Raynaud D. 1995. Variations in atmospheric methane concetration during the Holocene epoch. Nature 374: 46–49. Bond G, Showers W, Chesby M, Lotti R, Almasi P, deMenocal P, Priore P, Cullen H, Hajdas I, Bonani G. 1997. A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278: 1257–1266. Campbell ID, Campbell C, Apps MJ, Rutter NW, Bush ABG. 1998. Late Holocene-1500yr climatic periodicities and their implications. Geology 26: 471–473. Chafetz HS, Utech NM, Fitzmaurice SP. 1991. Differences in the
18O and
13C signatures of seasonal laminae comprising travertine stromatolites. Journal of Sedimentary Petrology 61: 1015–1028. Coope GR, Lemdahl G, Lowe JJ, Walkling A. 1998. Temperature gradients in northern Europe during the last glacial–Holocene transition (14–9 14C kyr BP) interpreted from coleopteran assemblages. Journal of Quaternary Science 13: 419–433. Craig H. 1965. The measurement of oxygen isotope paleotemperatures. In Stable Isotopes in Oceanographic Studies and Paleotemperatures, Tongiorgi E (ed.). Consiglio Nazionale delle Richerche, Laboratorio de Geologia Nucleare: Pisa; 161–182. Dansgaard W, Johnsen SJ, Clausen HB, Dahl-Jensen D, Gundestrup NS, Hammer CU, Hvidberg CS, Stefferson JP, Sveinbjo¨rnsdottir AE, Jouzel J, Bond G. 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364: 218–220. Darling WG, Talbot JC, Brownless MA. 1996. Stable isotopic content of rainfall and groundwaters in the British Isles. In Isotopes in Water Resources Management: Proceedings, International Symposium, Vienna, March, 1995. International Atomic Energy Agency: Vienna. Epstein S, Mayeda T. 1953. Variations of the 18O content of waters from natural sources. Geochimica Cosmochimica Acta 4: 213–224. Fairchild IJ, Tooth AF, Huang Y, Frisia S, McDermott F. 1996. Spatial and temporal variations in water and stalactite chemistry in currently active caves: a precursor to interpretations of past climate. In Proceedings of the Fourth International Symposium on the Geochemistry of the Earth’s Surface, Bottrell S (ed.). University of Leeds: Ilkley, Yorkshire; pp. 229–233. Fairchild IJ, Borsato A, Tooth AF, Frisia S, Hawkesworth CJ, Huang Y, McDermott F, Spiro B. 2000. Controls on trace element (Sr–Mg) compositions of carbonate cave waters: implications for speleothem climatic records. Chemical Geology 166: 255–269. Garnett ER. 2003. Geochemical Records of Climatic and Environmental Change in British Holocene Tufas. Unpublished PhD thesis, University of East Anglia, Norwich, 276 pp. Garnett ER, Gilmour MA, Rowe PJ, Andrews JE, Preece RC. 2004. 230 Th/234U dating of Holocene tufas: possibilities and problems. Quaternary Science Reviews 23: 947–958. Genty D, Massault M, Gilmour M, Baker A, Verheyden S, Kepens E. 1999. Calculation of past dead carbon proportion and variability by the comparison of AMS 14C and TIMS U/Th ages on two Holocene stalagmites. Radiocarbon 41: 251–270. Griffiths HI, Holmes JA. 2000. Non-marine Ostracods and Quaternary Palaeoenvironments. Technical Guide 8, Quaternary Research Association: London. Guiot J, Harrison SP, Prentice IC. 1993. Reconstruction of Holocene precipitation patterns in Europe using pollen and lake-level data. Quaternary Research 40: 139–149. Hammarlund D, Edwards TWD, Bjorck S, Buchardt B, Wohlfarth B. 1999. Climate and environment during the Younger Dryas (GS-1) as reflected by composite stable isotope records of lacustrine carbonates at Torreberga, southern Sweden. Journal of Quaternary Science 14: 17–28. Hays PD, Grossman EL. 1991. Oxygen isotopes in meteoric calcite cements as indicators of continental paleoclimate. Geology 19: 441–444. Hellstrom JC, McCulloch MT. 2000. Multi-proxy constraints on the climatic significance of trace element records from a New Zealand speleothem. Earth and Planetary Science Letters 179: 287–297. Holmes JA. 1992. Trace-element chemistry of non-marine ostracod shells: a preliminary evaluation of cleaning methods. Journal of Micropalaeontology 11: 36. Huang Y, Fairchild IJ. 2001. Partitioning of Sr2þ and Mg2þ into calcite under karst-analogue experimental conditions. Geochimica Cosmochimica Acta 65: 47–62. Copyright ß 2004 John Wiley & Sons, Ltd.

261

Hughes PDM, Mauquoy D, Barber KE, Langdon PG. 2000. Miredevelopment pathways and palaeoclimate records from a full Holocene peat archive at Walton Moss, Cumbria, England. Holocene 10: 465–479. Ihlenfeld C, Norman MD, Gagan MK, Drysdale RN, Maas R, Webb J. 2003. Climatic significance of seasonal trace element and stable isotope variations in a modern freshwater tufa. Geochimica Cosmochimica Acta 67: 2341–2357. Janssen A. 2000. Petrography and geochemistry of active and fossil tufa deposits from Belgium. Unpublished PhD thesis, Katholieke Universiteit Leuven, Leuven, Belgium; 392. Jeffery GH, Bassett J, Mendham J, Denny RC. 1989. Vogel’s Text Book of Quantitative Chemical Analysis. ELBS, Longman: Harlow. Katz A. 1973. The interaction of magnesium with calcite during crystal growth at 25–90 deg C and one atmosphere. Geochimica Cosmochimica Acta 37: 1563–1586. Keatings KW, Heaton THE, Holmes JA. 2002. Carbon and oxygen isotope fractionation in non-marine ostracods: results from a ‘natural culture’ environment. Geochimica Cosmochimica Acta 66: 1701– 1711. Kerney MP, Preece RC, Turner C. 1980. Molluscan and plant biostratigraphy of some late Devensian and Flandrian deposits in Kent. Transactions of the Royal Society Series B 291: 1–43. Klappa CF. 1980. Rhizoliths in terrestrial carbonates: classification, recognition, genesis and significance. Sedimentology 27: 613–629. Klitgaard-Kristensen D, Sejrup HP, Haflidason H, Johnsen S, Spurk M. 1998. A regional 8200 cal. yr BP cooling event in northwest Europe, induced by final stages of the Laurentide ice-sheet deglaciation? Journal Quaternary Science 13: 165–169. Korhola A, Weckstrom J, Holmstrom L, Erasto P. 2000. A quantitative Holocene climatic record from diatoms in northern Fennoscandia. Quaternary Research 54: 284–294. Langmuir D. 1971. The geochemistry of some carbonate ground waters in central Pennsylvania. Geochimica Cosmochimica Acta 35: 1023– 1045. Lawler HA. 1987. Sampling for isotopic responses in surface waters. Earth Surface Processes and Landforms 12: 551–559. Lawrence JR, White JWC. 1991. The elusive climate signal in the isotopic composition of precipitation. In Stable Isotope Geochemistry: A Tribute to Samuel Epstein, Taylor HP, O’Neil JR, Kaplan IR (eds). Geochemical Society Special Publication 3: San Antonio, Texas; 169–185. Magny M, Guiot J, Schoellammer P. 2001. Quantitative reconstruction of Younger Dryas to Mid-Holocene paleoclimates at Le Locle, Swiss Jura, using pollen and lake-level data. Quaternary Research 56: 170–180. Matsuoka J, Kano A, Oba T, Watanabe T, Sakai S, Seto K. 2001. Seasonal variation of stable isotopic compositions recorded in a laminated tufa, SW Japan. Earth and Planetary Science Letters 192: 31–44. McDermott F, Mattey DP, Hawkesworth C. 2001. Centennial-scale Holocene climate variability revealed by a high-resolution speleothem
18O record from SW Ireland. Science 294: 1328–1331. Meisch C, 2000. Freshwater ostracoda of Western and Central Europe. In Suesswasserfauna von Mitteleuropa 8/3, Schwoerbel J, Zwick P (eds). Spektrum Akademischer Verlag: Heidelberg; 522 pp. Messe DA, Alley RB, Fiacco RJ, Germani MS, Gow AJ, Grootes PM, Illing M, Mayewki PA, Morrison MC, Ram M, Taylor KC, Yang Q, Zielinski GA. 1994. Preliminary Depth–Age Scale of the GISP2 Ice Core. Special Report 94-1 CRREL: Hanover, New Hampshire. Nesje A, Dahl SO. 2001. The Greenland 8200 cal. yr BP event detected in loss-on-ignition profiles in Norwegian lacustrine sediment sequences. Journal of Quaternary Science 16: 155–166. O’Brien SR, Mayewski PA, Meeker LD, Meese DA, Twickler MS, Whitlow SI. 1995. Complexity of Holocene climate as reconstructed from a Greenland ice core. Science 270: 1962–1964. Pazdur A, Pazdur MF, Starkel L, Szulc J. 1988. Stable isotopes of Holocene calcareous tufa in southern Poland as paleoclimatic indicators. Quaternary Research 30: 177–189. Plummer LN. 1977. Defining reactions and mass transfer in part of the Floridan aquifer. Water Resources Research 13: 801–812. Preece RC. 1991. Accelerator and radiometric radiocarbon dates on a range of materials from colluvial deposits at Holywell Coombe, J. Quaternary Sci., Vol. 19(3) 251–262 (2004)

262

JOURNAL OF QUATERNARY SCIENCE

Folkestone. In Radiocarbon Dating: Recent Applications and Future Potential, Lowe JJ (ed.). Wiley: Chichester; 45–53. Roberts MS, Smart PL, Baker A. 1998. Annual trace element variations in a Holocene speleothem. Earth and Planetary Science Letters 154: 237–246. Romanek CS, Grossman EL, Morse JW. 1992. Carbon isotope fractionation in synthetic aragonite and calcite: effects of temperature and precipitation rate. Geochimica Cosmochimica Acta 56: 419–430. Rose´n P, Segerstro¨m U, Eriksson L, Renberg I, Birks HJB. 2001. Holocene climatic change reconstructed from diatoms, chironomids, pollen and near-infrared spectroscopy at an alpine lake (Sjuodjijaure) in northern Sweden. Holocene 11: 551–562. Rousseau D-D, Preece R, Limondin-Lozouet N. 1998. British late glacial and Holocene climatic history reconstructed from land snails assemblages. Geology 26: 651–654. Rozanski K, Araguas-Araguas L, Gonfiantini R, 1993. Isotopic patterns in modern global precipitation. In Climate Change in Continental Isotopic Records, Swart PK, Lohmann KC, Mckenzie J, Savin S (eds). American Geophysical Union: Washington, DC; 1–36. Schwalb A, Burns SJ, Kelts K. 1999. Holocene environments from stable isotope stratigraphy of ostracods and authigenic carbonate in Chilean Altiplano Lakes. Palaeogeography, Palaeoclimatology and Palaeoecology 148: 153–168. Steig EJ, Grootes PM, Stuiver M. 1994. Seasonal precipitation timing and ice core records. Science 266: 1885–1886. Stuiver M, Reimer PJ. 1993. Extended 14C database and revised CALIB 3.0 14C calibration. Radiocarbon 35: 215–230. Stuiver M, Reimer PJ, 2000. CALIB v. 4.3. University of Washington: Washington, DC. Stuiver M, Grootes PM, Braziunas TF. 1995. The GISP2 18O climate record of the past 16 500 years and the role of the sun, ocean and volcanoes. Quaternary Research 44: 341–354.

Copyright ß 2004 John Wiley & Sons, Ltd.

Stuiver M, Braziunas TF, Grootes PM, Zielinski GA. 1997. Is there evidence for solar forcing of climate in the GISP2 oxygen isotope record. Quaternary Research 48: 341–354. Verheyden S, Keppens E, Fairchild IJ, McDermott F, Weis D. 2000. Mg, Sr and Sr isotope geochemistry of a Belgian Holocene speleothem: implications for paleoclimate reconstructions. Chemical Geology 169: 131–144. Verrecchia EP, Verrecchia KE. 1994. Needle-fiber calcite: a critical review and a proposed classification. Journal of Sedimentary Research A64: 650–664. Von Grafenstein U, Erlenkeuser H, Muller J, Jouzel J, Johnsen S. 1998. The cold event 8200 years ago documented in oxygen isotope records of precipitation in Europe and Greenland. Climate Dynamics 14: 73–81. Von Grafenstein U, Erlenkeuser H, Brauer A, Jouzel J, Johnsen SJ. 1999a. A mid-European decadal isotope-climate record from 15 500 to 5000 years BP. Science 284: 1654–1657. Von Grafenstein U, Erlernkeuser H, Trimborn P. 1999b. Oxygen and carbon isotopes in modern fresh-water ostracod valves: assessing vital offsets and autecological effects of interest for palaeoclimate studies. Palaeogeography Palaeoclimatology Palaeoecology 148: 133–152. Wansard G, Roca JR, Mezquita F. 1999. Experimental determination of strontium and magnesium partitioning in calcite of the freshwater ostracod Herpetocypris intermedia. Archiv fur Hydrobiologie 145: 237–253. Worssam BC. 1963. Geology of the Country Around Maidstone: Memoir for 1:50 000 Geological Sheet 288. Her Majesty’s Stationary Office: London. Xia J, Ito E, Engstrom DR. 1997. Geochemistry of ostracode calcite: Part 1—An experimental determination of oxygen isotope fractionation. Geochimica Cosmochimica Acta 61: 377–382.

J. Quaternary Sci., Vol. 19(3) 251–262 (2004)