Investigating the potential of oxygen-isotope records from anthropogenic lakes as tracers of 20th century climate change

Abstract

Historical climate change in southern England was investigated using ostracod oxygen-isotope (δ¹⁸O) records from two anthropogenic lakes in Hampshire, southern England. A strong relationship is observed between δ¹⁸O_ostracod, δ¹⁸O_{precipitation} and δ¹⁸O_{lake_water} in the contemporary environment and therefore δ¹⁸O_ostracod from the sedimentary record of these systems has the potential to reflect past climate variability. The possibility of these sites to act as archives of climate change through δ¹⁸O_ostracod analysis is explored through the study of lake sediment cores that cover the period from the early 20th century onwards. Both lakes showed similar directionality of shifts in δ¹⁸O_ostracod over this period, suggesting common driving mechanisms. Comparing δ¹⁸O_ostracod timeseries to meteorological data is challenging in part because of the complexity with which climate parameters are recorded in the δ¹⁸O_{lake_water} and consequently within lacustrine carbonates. Our findings highlight the potential of sediments from anthropogenic lakes to act as archives of past climate and indicate they may be an important resource for generating climatic reconstructions across the medieval to instrumental period, which the sediments of many anthropogenic lakes cover. Such climate reconstructions would greatly improve our spatial and temporal understanding of climate variability where instrumental data are unavailable and other natural archives are scarce.

Keywords

anthropogenic lakes climate change historical Late-Holocene ostracods oxygen isotopes

Introduction

Investigating climate variability over the past millennium is critical for understanding the mechanisms and impacts of climate changes under scenarios of changing forcing factors (i.e. natural vs anthropogenic). However, obtaining paleoclimatic records for this period is challenging, not least because of the limited availability of archives that are able to produce high-resolution, chronologically well-constrained climate reconstructions that extend beyond the limit of instrumental datasets. The Medieval Climate Anomaly (MCA: ca. 950–1400 CE) and Little Ice Age (LIA: ca. 1200–1850 CE) are known natural climatic events during this time-frame that were driven by a range of decadal-scale variables. These include solar variability, volcanism and atmospheric controls and feedbacks, such as the North Atlantic Oscillation, when many controls on the Earth system (i.e. ice volume, orbital configuration, sea level, etc.) were much the same as the present day, including the influence of anthropogenic activity (Diaz et al., 2011; Goosse et al., 2012; Nesje and Dahl, 2003; Owens et al., 2017; Ruddiman, 2003; Trouet et al., 2009, 2012; Wanner et al., 2011). Despite the importance of these oscillations for understanding the scale and impact of such climate variability, there are limited high-resolution climate reconstructions that cover this time period, even in well studied areas such as northwest Europe. Study of the post-medieval to instrumental period is hindered by a lack of natural environmental archives and a greater reliance on documentary evidence, which is inconsistent in terms of its detail and continuity despite its contextual importance (e.g. Brönnimann et al., 2019; Burgdorf, 2022). Thus, there is a need for additional environmental archives across the last millennium that should ideally meet the following conditions: (1) have the potential for high-resolution (ideally decadal or better) proxy reconstruction in order to capture short-timescale climate variability; (2) contain a range of proxies, enabling human impacts and catchment processes to be distinguished from regional climate signals; and (3) a robust chronology.

Lake sediments have the potential to meet these requirements, but natural lakes covering the post-medieval onwards are quite rare in northwest Europe for a host of reasons. For example, many natural lakes that formed during the last glacial period have infilled and are no longer extant by the Late-Holocene. Equally it is likely that sedimentation that covers this period is strongly influenced by anthropogenic landscape management as much as by climate variability and untangling these impacts is challenging. Tree rings and speleothems have the potential to fill this temporal gap, at high resolution (potentially annually), but they too have limitations. For instance, to achieve interdecadal climate variability from tree rings requires tree-ring width data to be statistically treated and their chronologies sometimes produce unclear climate data (Dobrovolný et al., 2018; Freund et al., 2023). Furthermore, speleothems are restricted to regions of carbonate bedrock. Consequently, there are large spatial gaps in climate records for the past millennium, limiting our understanding of spatial variation in climate change and its impact over this period. The British Isles has an abundance of ornamental lakes often associated with period estates. These are anthropogenic lakes that became popular in garden design from the early 18th century in England with 450 on National Trust property in the UK as well as 1048 legacy Medieval fishing ponds listed by Historic England (2023). Similar anthropogenic lakes also exist across Europe (Bishop, 2021; Freshwater Habitats Trust, 2022). Although the latter may not all have extant water, the numbers show the huge opportunity to explore these features. Such lakes may be able to provide records of past climate across the historical period, helping to overcome the gaps in spatial coverage. This, alongside the variety of proxies applicable in lake settings and the likelihood of historical documentary evidence on such lakes recording periods of human disturbance, suggests anthropogenic lakes may be an important archive. However, thus far no studies have investigated the potential of the use of such sites as the basis for paleoclimate archives. The aim of this paper is to evaluate the potential of these anthropogenic lakes as archives of past climate through a study of two sites in Hampshire, southern England, both of which are located on chalk geology, namely Old Alresford Pond and The Grange Lake.

Oxygen isotopes in lake carbonates as a paleoclimate proxy

Oxygen isotopes in lake-sediment carbonates are a potential climate proxy in anthropogenic lakes. Their use as a climate proxy has been widely demonstrated in natural lakes in humid-temperate mid-latitude regions (Leng and Marshall, 2004). In such settings, the oxygen-isotope composition of endogenic or biogenic carbonate is a function of the oxygen-isotope composition of lake water (δ¹⁸O_{lake_water}) from which the carbonate precipitates and the temperature at which the carbonate mineralises (Leng and Marshall, 2004). The δ¹⁸O_{lake_water}, in turn, is controlled by the isotopic composition of inflows (i.e. rainfall, surface inflow and groundwater) together with any in-basin processes that lead to fractionation, particularly evaporation. In lakes with a high rate of recharge and a short water residence time, evaporitic modification may be minimal, meaning that the lake water isotope composition is close to that of precipitation (δ¹⁸O_{precipitation}). The isotope composition of precipitation is controlled by a range of factors including air temperature, and the source, transport pathway and rain-out history of rain-bearing air masses (Dansgaard, 1964; Rozanski et al., 1992, 1993). Providing the controls on the oxygen isotopic composition of lake sediment carbonates can be adequately constrained, such records therefore have the potential to provide us with information about climate. The shells of ostracods (small aquatic crustaceans that secrete shells of calcite) are good sources of carbonate for stable-isotope analysis. Some species have well-defined seasonality, so that the oxygen isotope signal can be tied to a specific time of year. Ostracods do not form shells that are in isotopic equilibrium with water, but the offsets are well understood and precisely quantified for some taxa (Holmes and Chivas, 2002; von Grafenstein et al., 1999). Finally, by using ostracod shells for isotope analyses (δ¹⁸O_ostracod), we can be sure that the calcite was produced within the lake and does not include any catchment-derived (or exogenic) material.

Here we investigate how well δ¹⁸O_ostracod reflects prevailing climate, including seasonal variables in the present day as a baseline for our historical reconstruction by monitoring lake water temperature and δ¹⁸O_{lake_water} composition, supported by meteorological and rainfall-isotope data. Isotope analyses of modern ostracods from each lake were undertaken to understand how δ¹⁸O_{lake_water} and water temperature is encoded in the ostracod shell calcite. Isotope analyses of sub-fossil ostracods sampled from lake sediment cores from each lake were then undertaken to produce a historical climate record from this part of southern England. Although our primary focus is on oxygen-isotopes, we use the carbon-isotope (δ¹³C) records that are also produced as supporting information. The δ¹³C of ostracod shells records the carbon-isotope composition of dissolved inorganic carbon (DIC), which in turn is controlled by the δ¹³C of groundwater, together with in-lake modification. Stuart and Smedley (2009) report δ¹³C value of −17.4 to −9.36 ‰ in groundwater of the Hampshire Chalk, with the more ¹³C-deplete values relating to dominance of solid-derived carbon and the least deplete to carbon from bedrock source. Enhanced uptake of ¹²C during aquatic photosynthesis as well as contributions of atmospheric CO₂, will also raise the δ¹³C value of lake DIC (Kelts and Talbot, 1990). Clearly, the δ¹³C of ostracods within the lakes can be influenced by both regional factors and lake-specific controls.

Our investigation focusses on approximately the last 100 years since this period overlaps in time with instrumental records. Our study also acts as a proof of concept for the use of such constructed lakes to record a climate signal over a longer period, potentially covering much of the past millennium at some sites.

Study sites

Old Alresford Pond (51°5′37.43″N, 1°9′30.92″W) and The Grange Lake (51°7′12.32″N, 1°11′46.66″W) are two constructed, historical, neighbouring lakes lying about 4 km apart in Hampshire, UK (Figure 1). The area is underlain by Cretaceous chalk bedrock and both lakes are fed by chalk streams that have been artificially dammed to generate lakes. The chalk streams are groundwater fed, with their flow greatest in winter and spring (Environment Agency, 2013). Both sites are shallow, with a maximum depth of <1 m. Old Alresford Pond is mostly infilled and extremely shallow (<0.5 m) for much of its area.

Figure 1.

Location maps. (a) Old Alresford Pond in relation to The Grange Lake with inset showing these in the context of Hampshire. (b) The Grange Lake with core location and flow direction (top) and sampling locations (bottom). (c) Old Alresford Pond sampling and coring locations. This study uses cores ALRE19 and GRA16. Imagery from Google Earth.

Old Alresford Pond

Constructed in ca. 1189 CE, Old Alresford Pond is a former medieval fishing pond that began use ca. 1208–1209 CE (Rollason, 2017; Turpin, 2008). The lake was created by constructing a 6 m high, 76 m long earth dam to produce a reservoir for mills along the River Itchen (Binnie, 1974; Environment Agency, 2013). Today the lake is 0.08 km² in area, smaller than its former extent of ca. 0.24 km² in the 13th century (Environment Agency, 2020; Roberts, 1986). The lake’s water residence time is estimated at less than one day (Wang, 2013). Water sources include surface runoff from the catchment via the stream network, direct precipitation into the lake and groundwater. Boreholes to supply groundwater for the watercress beds upstream of Old Alresford Pond were introduced in the 20th century and contribute to water within the pond. Documentary evidence records dredging of the lake sometime between 1252 and 1254 CE, clearing built up sediment and then deepening one side of the lake in 1254 CE for easier eel fishing (Roberts, 1986). However, there is no record of more recent dredging. Peat from the nearby watercress farms was flushed into the lake at times during the 20th century (Tony Chambers and Rosemary Chambers, personal communication).

The Grange Lake

The Grange Lake is an ornamental lake on The Grange Estate, an addition typical of 18th century grounds design, and was built by damming the Candover Brook (also known as the Candover Stream) (Currie, 2001; Deveson, 2005). It is believed the lake was first ordered sometime between 1764 and 1772 CE by Robert Henley, Earl of Northington (Geddes, 1983). It is first seen on Milne’s map of Hampshire in 1791 CE as a long, narrow lake and the southern part was later enlarged by Baron Ashburton, Alexander Baring, as shown on the ca. 1817 CE Ordnance Survey and the 1826 CE Greenwood’s maps of the area (Norgate and Norgate, 2002a, 2002b, 2002c; Parks and Gardens, 2021). Today, the lake is divided by a waterfall and analyses refer to the ‘Upper Lake’, from where the core was recovered and some modern water and ostracod samples collected, and the ‘Lower Lake’ from where only modern water and ostracod samples were collected (Figure 1).

Methods

Modern limnology

Measurements of modern water oxygen and hydrogen isotope composition and temperature at each lake were undertaken to understand the relationship between the oxygen-isotope composition of rainfall and that of lake water and, consequently, how the lake water isotope signal is encoded into ostracod shells from the lake (Figure 1). Measurements and collections were made from May 2019 to November 2020. Samples at Old Alresford Pond were taken in May 2019, August 2019, November 2019, February and November 2020. Samples at The Grange Lake were taken in August 2019, November 2019, February 2020 and November 2020. Sampling had to be paused between February 2020 and November 2020 owing to the COVID-19 pandemic.

Hourly measurements of water temperature were made from August 2019 to November 2020, using TinyTag water temperature sensors installed in both lakes (Figure 1). There is an additional, short, temperature record from May to August 2019 at Old Alresford Pond that was discontinued because of device dysfunction.

Water samples were taken from various locations (Figure 1) at both lakes approximately every 3 months from May 2019 (Old Alresford Pond) or August 2019 (The Grange Lake) to November 2020 to help identify seasonal variations in water isotope composition. On each visit, 50 mL samples were collected in polyethylene bottles, which were sealed with electrical tape, and refrigerated below 5°C until analysis. In addition, an automated water sampler collected water samples approximately weekly from both lakes. These samples were ‘capped’ by a thin layer of paraffin oil that was placed in each bottle before sampling to minimise any evaporative effect on the water sample following IAEA/GNIP (2014) procedures for precipitation isotope sample collection. During each site visit these samples were transferred to 500 mL bottles, which were capped, sealed with electrical tape, and refrigerated below 5°C until analysis. The paraffin oil was removed before analysis by standing the water in a separation funnel for at least 2 h and allowing the oil to float. The separation funnel was covered to minimise evaporation. Once the oil had been displaced, a 50 mL water aliquot of water was collected in a centrifuge tube, sealed with electrical tape, and refrigerated below 5°C until analysis. Any samples that contained detrital material were additionally filtered through grade 1 filter paper and a grade 1 syringe filter. All water samples were analysed for oxygen and hydrogen isotopes at the British Geological Survey, Keyworth, using a CO₂ equilibrium method with an Isoprime 100 mass spectrometer plus Aquaprep device for oxygen isotope composition. Hydrogen isotopes were analysed on an online Cr reduction method with a EuroPyrOH-311-system coupled to a Micromass Isoprime mass spectrometer. Measurements were made against internal standards that are calibrated against the international standard VSMOW2 and presented in standard delta units. The typical uncertainties are ±0.05 ‰ δ¹⁸O and ±1.0 ‰ for δD.

Ostracod samples were collected from the surface sediments and from a combination of submerged, floating and emergent macrophytes, depending on availability, at both lakes, on each visit using a 250 µm sieve and/or net. Samples were stored in sealed bottles prior to wet sieving through a 250 µm mesh with tap water in the laboratory and then oven drying at ca. 40°C–60°C in a drying cabinet. Specimens of Candona neglecta Sars, 1887 and Candona candida (Müller, 1776) were picked for oxygen and carbon isotope analysis. These species complete their lifecycles within a year and calcify their adult shells in the winter months, which reduces the influence of any summer evaporation of lake water on the recorded oxygen isotope signal (Decrouy, 2009; Dole-Olivier et al., 2000). The collected valves are likely to be from specimens that died at varying times before their collection and so the isotope signal will represent an average of the last few years. Carapaces in these samples were typically small and not consistently abundant across the samples and so valves were preferentially picked. The species were chosen because they are common in both modern lakes and their sediment records and because the offsets from oxygen-isotope equilibrium are well known (+2.2 ± 0.15 ‰ for members of the family Candondidae: Holmes and Chivas, 2002; von Grafenstein et al., 1999). Moreover, both species have similar, broad ecological tolerances and so can be used interchangeably for isotope analyses (Dole-Olivier et al., 2000; Külköylüoĝlu and Vinyard, 2000). The selected lake sites are so shallow that they are entirely littoral in habitat and therefore habitat variability is not a concern. For oxygen and carbon isotope analysis, samples consisted of multiple shells, representing a time average of the sample core slice, to a mass of between 34 and 100 μg of carbonate. Analyses were performed using an IsoPrime dual inlet mass spectrometer plus Multiprep device. Isotope values (¹³C, ¹⁸O) are reported as per mille (‰) deviations of the isotopic ratios (¹³C/¹²C, ¹⁸O/¹⁶O) calculated to the VPDB scale using a within-run laboratory standard (KCM) calibrated against international standard NBS-19. The calcite-acid fractionation factor applied to the gas values is 1.00813. Due to the long run time a drift correction is applied across the run, calculated using the standards that bracket the samples. The Craig correction (Craig, 1957) is also applied to account for δ¹⁷O. The standard calcite values for KCM are for +2.00 ‰ for δ¹³C, and −1.73 ‰ δ¹⁸O with an average analytical reproducibility of 0.05 ‰.

Air temperature data from 2019 to 2020 CE were obtained from the MetOffice Southampton National Oceanography Centre, Odiham and Otterbourne weather stations (Figure 1), the closest weather stations to the two lakes, to compare with water temperature measurements. Precipitation isotope data (δ¹⁸O_{precipitation}) for comparison with the lake water values (δ¹⁸O_{lake_water}) were taken from Wallingford, Oxfordshire (Figure 1) the nearest GNIP (Global Network of Isotopes in Precipitation) station, which lies 69 km north of Alresford. A local meteoric water line (LMWL) was calculated for oxygen and hydrogen isotopes using the precipitation amount weighted least squares regression and given by δD = 7.34 and δ¹⁸O_{precipitation} + 4.13 (IAEA/WMO, 2023). There may be an offset in the isotope composition of precipitation at Wallingford compared to our study sites given the distance between them, but we would expect this to be smaller than the error. We also compare measured δ¹⁸O_{lake_water} with modelled δ¹⁸O_{precipitation} at Old Alresford Pond using the Piso_AI software (Nelson et al., 2021). This allows a comparative estimate of δ¹⁸O_precipitaton at the core location that also extends past the earliest GNIP measurement at Wallingford. The modelled output at both Old Alresford Pond and The Grange Lake were very similar and so we use the Old Alresford Pond output in our presented results.

Core recovery

Cores were collected from Old Alresford Pond and The Grange Lake (Figure 1). The lakes were cored with a UWITEC gravity corer for the uppermost sediments (0–20 cm), while the deeper sediments were collected using a 52 mm diameter Russian corer in 50 cm segments. Old Alresford Pond was originally cored in 2008 (ARL1) and then both lakes were cored in 2013 (ARL13 and GRA13), with an additional core collected from The Grange Lake in 2016 (GRA16: 51°7′12.32″N, 1°11′46.66″W) and Old Alresford Pond in 2017 (ARL17). These cores were collected as part of other studies focused on the lake’s palaeoecology and the sediments were predominantly composed of marl (35–54 % of the inorganic fraction being calcium carbonate). In this study we use chronological markers and magnetic susceptibility tie points from the ARL13 and GRA13 cores (Supplemental Materials). A new sediment core was taken from Old Alresford Pond in 2019 (ALRE19: 51°5′37.43″N, 1°9′30.92″W) in three sections, using an adapted Livingstone corer to account for the shallow water depth (<1 m) and to collect the sediment-water interface, retrieving 260 cm of new sediment. The core was extruded into 1 cm slices and stored at <5°C. Here, we present data from core sediments between 0 and 93 cm, equivalent to ca. 1904–2019 CE.

Samples from ALRE19 and GRA16 were taken from both cores at 1 cm intervals, sieved at 250 µm with tap water and the coarse residue dried at a low temperature. Individual shells of Candona neglecta and Candona candida were picked from the dried >250 µm fraction and detrital material removed with a fine paint brush, under a low power biological microscope. Samples for analysis were composed of multiple (typically 5) shells to provide enough material for analysis and ensure that the results were representative of average lake conditions over the time-interval covered by a 1 cm increment of sediment. Analytical methods followed those for the modern ostracods.

Chronology

Chronologies for the cores were produced using a combination of short-lived radioisotopes (ALRE19) and spheroidal carbonaceous particle (SCP) counts (ARL13 and GRA13). Chronological determinations undertaken on the two cores used for isotope analysis in this study (ALRE19 and GRA16) were supplemented by tie points with parallel cores ARL13 and GRA13 using magnetic susceptibility data (Supplemental Figure 1). Tie points between ARL13 and ALRE19 occur at 44.5 cm (1958.1 ± 10 CE), 66.5 cm (1943.4 ± 13 CE) and 82.5 cm (1927 ± 15.5 CE) on the ARL13 age-depth scale reported using the intcal20_t age (years CE) and half the total years in the 95.4 % probability age range produced by the model. The age model for ALRE19 uses both independent ages from short-lived radioisotopes and SCP ages from ARL13. GRA16 relies on magnetic susceptibility tie points with GRA13 as there was not enough core material available to produce independent ages. There are 26 tie points between GRA13 and GRA16 detailed in the Supplemental Material.

Short-lived radioisotope analysis was applied to the ALRE19 core only as there was not enough sediment available from GRA16. Approximately 3 g of dried sediment were taken from ALRE19, crushed into a powder using a pestle and mortar and then analysed for ²¹⁰Pb, ²²⁶Ra, ¹³⁷Cs and ²⁴¹Am on an ORTEC HPGe GWL series well-type coaxial low background intrinsic germanium detector by direct gamma assay at the Environmental Radiometric Facility, University College London. ²¹⁰Pb ages were then calculated using the constant rate supply (CRS) model (Appleby, 2001).

SCP ages were produced from the neighbouring ARLE13 and GRA13 cores. Samples were taken every 8 cm from Old Alresford Pond (ARLE13) and every 5 cm from The Grange Lake (GRA13) for SCP analyses (following Rose, 2008), to identify broad trends compared with the South and Central England calibration curve, following cumulative accumulation methods (Rose and Appleby, 2005). Subsequent samples were taken to clarify the depths of specific SCPs concentration peaks, which can be used to track industrial pollution over time, and hence these are used as a chronological marker for the last ca. 150+ years. SCP counts were expressed as cumulative percentage frequencies, which could then be related to ages determined using the South and Central England SCP age dataset, produced by Rose and Appleby (2005) that is based on ²¹⁰Pb chronologies from multiple lakes in the region. These data were used to produce age models for ARL13 and GRA13 using a Bayesian model in OxCal v4.4 (Bronk-Ramsey, 2009, 2021). All chronological data are in the Supplemental Material.

Magnetic susceptibility data for ALRE19 were produced by taking ca. 6 g of freeze-dried sediment and 1 cm intervals and placing in individual pots topped with cling film where sediment did not fill the pot. Low frequency magnetic susceptibility by sample mass (g) was recorded for each sample. ALRE13 was measured using mass specific magnetic susceptibility in HF mode. GRA16 data used volumetric susceptibility and GRA16 mass specific magnetic susceptibility in HF mode. These data were used for core correlation to produce single master cores for each site and date.

Age models for GRA13, GRA16, ARL13 and ALRE19 were generated using a Bayesian model in OxCal v4.4 (Bronk-Ramsey, 2009, 2021). Figures 2 and 3 in the Supplemental Figure document show the final age model for GRA16 and ALRE19 that are used in this study, and all data are available in the Supplemental Data File.

Figure 2.

Water temperature at Old Alresford Pond lake centre (a) and under bridge (b) at outflow (Figure 1c). Water temperature at The Grange Lake Upper (c) and Lower Lake (d). a, c and d = August 2019 to November 2020; b = May 2019 to August 2019 (logger broke). Red = daily maximum temperature, blue = daily minimum temperature, black = daily mean temperature. Monthly mean temperatures (yellow) average the daily mean temperature over the period of one calendar month where values are available. (e) Monthly mean water temperature from Old Alresford Pond and The Grange Lake (‘Upper’ and ‘Lower’) compared to monthly mean air temperature from three MetOffice weather stations in the region (Otterbourne, Odiham, Southampton National Oceanography Centre (NOC)). Air temperature data provided by the National Meteorological Library and Archive – MetOffice UK. Site locations are shown in Figure 1.

Figure 3.

(a) δ¹⁸O_{lake_water} samples taken across Old Alresford Pond (circles) and The Grange Lake (squares – includes upper and lower lake measurements) across the monitoring period May 2019 to November 2020, with months depicted by colour and outliers removed. (b) δ¹⁸O_ostracod and δ¹³C_ostracod data from ostracod valves collected at Old Alresford Pond (circles) and The Grange Lake (squares) across the monitoring period. UL = Upper Lake, LL = Lower Lake, S = Sediment sample, M = Macrophyte sample. The linear trend line and R² use all data points. Outliers removed. (c) Compares Piso_AI modelled δ¹⁸O_{precipitation} with measured δ¹⁸O_{lake_water} values from the Old Alresford Pond across the monitoring period May 2019 to November 2020. (d) compares Piso_AI modelled δ¹⁸O_{precipitation} with measured δ¹⁸O_{lake_water} values from The Grange Lake across the monitoring period May 2019 to November 2020. The error on modelled values is ±1.7 ‰ δ¹⁸O and ±13 ‰ δD_{precipitation} as shown by error bars (Nelson et al., 2021). (e) Compares Piso_AI modelled δ¹⁸O_{precipitation} from the Wallingford GNIP site across the monitoring period May 2019 to November 2020 with measured values from May to December 2019. (f) Compares Piso_AI modelled δ¹⁸O_{precipitation} and δD_{precipitation} values from Wallingford, Old Alresford Pond and The Grange Lake across the monitoring period May 2019 to November 2020.

Results

Monthly recorded water temperature variations at both sites show expected seasonality and are similar at both sites. For example, maximum monthly temperatures were 18.68°C (2019) and 17.85°C (2020) at Old Alresford Pond, with similar values being recorded at The Grange Lake ‘Upper’ (18.79°C, 16.30°C in 2019–2020 CE respectively; Figure 2a –d). The monthly mean air temperatures recorded at Odiham, Otterbourne and Southampton NOC show the same seasonal trends as the recorded water temperatures (Figure 2e).

When comparing the δ¹⁸O_{lake_water} and δD_{lake_water} from both Old Alresford Pond and The Grange Lake across 2019–2020 CE, the values show a tight clustering of data points (Figure 3a). All the lake water isotope data lie close to, but slightly above, the LMWL for Wallingford (Figure 3a). For the 2019 data that overlap the monitoring period (Table 1), the δ¹⁸O_{precipitation} values at Wallingford are within error of the δ¹⁸O_{lake_water} values at both Old Alresford Pond and The Grange Lake, despite the offset of the LMWL. The Piso_AI modelled data for both study sites and Wallingford demonstrate that the measured δ¹⁸O_{lake_water} and δ¹⁸O_{precipitation} lie within those expected by the model for those locations (Figure 3c –e).

Table 1.

Summary of δ¹⁸O_{lake_water} and δD_{lake_water} data from Old Alresford Pond and The Grange Lake May to Dec 2019 and of δ¹⁸O_{precipitation} and δD_{precipitation} values from Wallingford for the same period (IAEA/WMO, 2023).

Location	δ¹⁸O_{lake_water}		δD_{lake_water}		δ¹⁸O_{precipitation}		δD_{precipitation}		Piso_AI δ¹⁸O_{precipitation}		Piso_AI δD_{precipitation}		δ¹⁸O_groundwater
	Mean	1 SD	Mean	1 SD	Mean	1 SD	Mean	1 SD	Mean	1 SD	Mean	1 SD	Mean
Old Alresford Pond	−6.39	0.9	−39.98	6.4					−5.9	1.4	−39.95	9.9
The Grange Lake	−6.33	0.3	−40.13	1.9					−5.89	1.4	−40.21	9.8
Wallingford GNIP					−6.63	2.8	−44.75	21	−6.1	1.5	−41.05	10.9
Wessex Basin Chalk groundwaters													−6.6

SD: standard deviation.

Source: Piso_AI model data from Nelson et al. (2021) and Wessex Chalk groundwater data from Allen (2017).

The modern δ¹⁸O_ostracod and δ¹³C_ostracod values from Old Alresford Pond and The Grange Lake are consistent with each other (Figure 3b). δ¹⁸O_ostracod values at The Grange Lake span −1.17 ‰ (GRA-NOV19-LL-S) to −4.90 ‰ (GRA-FEB20-UL-M). These δ¹⁸O_ostracod values vary from −3.89 ‰ (ALRE-NOV20-S) to −5.13 ‰ (ALRE-FEB20-S). There is no relationship between the month of collection and the δ¹⁸O_ostracod value.

The ALRE19 core has a maximum δ¹⁸O_ostracod value of −2.18 ‰ and a minimum of −4.71 ‰ compared to GRA16 that has a maximum δ¹⁸O_ostracod value of −1.96 ‰ and a minimum of −4.29 ‰. The range of δ¹³C_ostracod values in ALRE19 is from −9.94 ‰ to −5.90 ‰, and for GRA16 is −10.59 ‰ to −7.24 ‰. δ¹⁸O_ostracod values are largely in the inverse direction of δ¹³C_ostracod in both the ALRE19 and GRA16 records (Figure 4). A gradual negative shift in δ¹⁸O_ostracod values from ca. 91 to 38 cm in ALRE19 followed by a positive shift from ca. 38 to ca. 33 cm, and then a decrease in values is recorded from ca. 38 cm to the top of the core (Figure 4a). Similar trends are observed in the GRA16 δ¹⁸O_ostracod record with a negative trend observed from ca. 105 to 55 cm, followed by a positive shift from ca. 55 to 40 cm, which is followed by a decrease in values identified by just a few data points (Figure 4b). Overall, the trends across ALRE19 and GRA16 δ¹⁸O_ostracod are similar.

Figure 4.

Downcore oxygen and carbon isotope data from (a) ALRE19 with (c) cross-plot of ALRE19 isotope data and (b) GRA16 core with (d) cross-plot of GRA16 isotope data. Data correspond to ca. 1904–2019 CE. Grey circles = individual data points with outliers removed. Red line = five-point moving average.

Interpretation

Do modern lake waters at Old Alresford Pond and The Grange Lake record prevailing climatic conditions?

Oxygen-isotope values in ostracod-shell calcite are controlled by the temperature and isotopic composition of lake water at the time of ostracod shell secretion. To assess how modern δ¹⁸O_ostracod values in the Alresford and Grange sequences record present-day climate conditions in the Hampshire lakes, we first assess how well lake temperature and δ¹⁸O_{lake_water} values record air temperature and δ¹⁸O_{precipitation}, respectively. For water temperature, Old Alresford Pond and The Grange Lake water temperature records show seasonal variability, indicating that the temperature of lake water responds quickly to air temperature (Figure 2a –d). The water temperatures show good agreement with air-temperature data from three local Met Office weather stations over the monitoring period (Figure 2e), indicating that we can use modern water temperature values as a good approximations of modern air temperature.

Secondly, we consider the relationship between δ¹⁸O_{lake_water} and δ¹⁸O_{precipitation}. The δ¹⁸O_{lake_water} values from the two lakes are similar and show minimal seasonal variation. The values lie parallel to the LMWL for Wallingford ca. 80 km away, but are slightly offset, suggesting that the data form a LMWL for each site (Figure 3a). To investigate this further, we generated a δ¹⁸O_{precipitation} time series using the Piso_AI model (Nelson et al., 2021) (Figure 3c –e). Measured δ¹⁸O_{precipitation} values at Wallingford lie within error of the modelled δ¹⁸O_{precipitation} values (Figure 3e), indicating the Piso_AI model does produce a reasonable approximation of true δ¹⁸O_{precipitation}. We therefore use the model to estimate δ¹⁸O_{precipitation} values at Old Alresford Pond and The Grange Lake and compare these to the measured δ¹⁸O_{lake_water} values. Then we test if these measured δ¹⁸O_{lake_water} values can be used as a proxy of δ¹⁸O_{precipitation} values (3c–d). At both study sites, measured δ¹⁸O_{lake_water} values lie within error of the modelled δ¹⁸O_{precipitation} values at each location, and we therefore conclude that δ¹⁸O_{lake_water} can be used as a proxy for δ¹⁸O_{precipitation} (3c–d). The fact that the δ¹⁸O_{lake_water} values from both systems closely approximates the mean modelled δ¹⁸O_{precipitation} values supports the suggestion that lake water values reflect the weighted mean annual δ¹⁸O_{precipitation} value.

What controls the δ¹⁸O_ostracod of modern ostracod valves?

Having demonstrated that water temperature is a good approximation of air temperature (Figure 2) and that δ¹⁸O_{lake_water} is a good reflection of δ¹⁸O_{precipitation} (Figure 3), we use the Kim and O’Neil (1997) equation to predict the oxygen-isotope value of carbonate precipitated in these lakes, referred to here as δ¹⁸O_synthetic, adding 2.2 ‰ for the offset from oxygen-isotope equilibrium for Candonidae (equation (1)). This equation is often used in ostracod studies and has good reproducibility in experimental conditions (Keatings et al., 2002). We use the mean annual δ¹⁸O_{lake_water} values for each site as a representation of mean annual δ¹⁸O_{precipitation} and mean measured winter temperature (DJF) values since both C. candida and C. neglecta calcify in the cooler, winter months (Decrouy, 2009; Dole-Olivier et al., 2000). We then compare the δ¹⁸O_synthetic value to measured δ¹⁸O_ostracod values over the monitoring period (Figure 3b and Table 2), with the caveat that the latter may represent the last few years because the ostracods were not living at the time of collection.

\begin{array}{l} δ^{18} O_{c a r b o n a t e (K i m & O^{'} N e i l (1997))} = δ^{18} O_{l a k e_w a t e r} + (28.625 - (S Q R T \\ ((16.5604 + (0.32 * D J F \\ m e a n w a t e r t e m p e r a t u r e))) / 0.16)) \\ δ^{18} O_{s y n t h e t i c} = δ^{18} O_{c a r b o n a t e (K i m & O^{'} N e i l (1997))} + 2.2 \end{array}

At Old Alresford Pond mean δ¹⁸O_ostracod was −4.07 ± 0.29 ‰, which falls outside the range of the δ¹⁸O_synthetic value using DJF temperatures at −2.73 ‰. The difference between δ¹⁸O_synthetic and δ¹⁸O_ostracod is closer when using SON temperatures (−3.77 ‰). Whilst we know that the two Candona species that we analysed typically calcify in the winter, the individuals collected could represent several years, as they were not necessarily living at the time of collection, whereas our measured lake temperatures are from just 1 year. It is also possible that some individuals may have calcified slightly earlier in the year than others. Moreover, there is additional uncertainty around how representative the measured modern δ¹⁸O_ostracod values at Old Alresford Pond are, because only three samples were analysed from this site. Furthermore, the measured lake temperatures across SON and DJF have quite large standard deviations (±3.94°C and ±1.41°C, respectively), so that average temperatures from 2019 to 2020 CE are not a perfect representation of the conditions in which specimens analysed had calcified. These reasons could explain the larger offset between δ¹⁸O_synthetic and δ¹⁸O_ostracod in DJF at Old Alresford Pond as this difference is reduced when considering SON temperatures, also representing the cooler part of the year. The ‘Upper Lake’ at The Grange Lake has a δ¹⁸O_synthetic value that lies within the range of δ¹⁸O_ostracod value, indicating that δ¹⁸O_ostracod are predominately a reflection of temperature and δ¹⁸O_{precipitation}. The proximity of the lakes to one another means we expect similar δ¹⁸O_ostracod trends if these values are being driven by climate variables. The range of δ¹⁸O_ostracod values from The Grange Lake lie largely within the range of those from Old Alresford Pond (Figure 3b). These similarities between the two lakes further support our conclusion that in the modern day δ¹⁸O_ostracod values at these lakes are primarily a reflection of the temperature of the coldest part of the year, albeit with greater uncertainty from the Old Alresford Pond data presented, and mean δ¹⁸O_{precipitation}.

Table 2.

Comparing δ¹⁸O_synthetic to δ¹⁸O_ostracod at Old Alresford Pond and The Grange Lake in the 2019–2020 monitoring period.

Site	Season	Mean annual δ¹⁸O_{lake_water}	δ¹⁸O_{lake_water} SD	Number of mean annual δ¹⁸O_{lake_water} samples	Mean water temperature (°C)	Water temperature SD	Number of water temperature samples	Mean δ¹⁸O_synthetic	Mean δ¹⁸O_ostracod	Number of δ¹⁸O_ostracod samples	δ¹⁸O_ostracod SD	δ¹⁸O_ostracod range	Difference between δ¹⁸O_synthetic and δ¹⁸O_ostracod
Old Alresford Pond	DJF	−6.39	0.89	28	7.29	1.41	90	−2.73	−4.07	3	0.29	−3.78 to −4.36	1.34
Old Alresford Pond	SON	−6.39	0.89	28	11.89	3.94	90	−3.77	−4.07	3	0.29	−3.78 to −4.36	0.3
The Grange Lake (Upper)	DJF	−6.28	0.31	19	8.44	0.95	90	−2.88	−3.14	6	0.83	−2.31 to −3.97	0.26

All measured values from The Grange Lake come from the ‘Upper’, lake to align with core location. Outliers have been removed. All data available in Supplemental Material.

DJF: December, January, February; SON: September, October, November.

Does sub-fossil δ¹⁸O_ostracod from Old Alresford Pond and The Grange Lake reflect 20th century climate variability?

If the modern temperature and isotope chemistry of lake waters in the Old Alresford Pond and The Grange Lake systems reflect prevailing climatic conditions, and both factors control δ¹⁸O_ostracod, then downcore changes in δ¹⁸O_ostracod should record past climate variability. To test this, we compare the measured δ¹⁸O_ostracod values derived from downcore analysis of these core sequences with a ‘synthetic’ δ¹⁸O_ostracod record calculated using the fractionation equation of Kim and O’Neil (1997) to combine the annual DJF temperature record from the Southampton MetOffice station and annual δ¹⁸O_{precipitation} values simulated using Piso_AI. A standard value of 2.2 ‰ is added to these values to allow for the vital offsets from oxygen-isotope equilibrium in Candonidae. We refer to this downcore synthetic record here as δ¹⁸O_synthetic. We present the δ¹⁸O_synthetic record for Old Alresford Pond (Figure 6c) as this site has the most δ¹⁸O_ostracod datapoints to compare to, and we have already demonstrated the similarity between the δ¹⁸O_ostracod at the two lakes (Figure 5). We assume that δ¹⁸O_{precipitation} was a close approximation of δ¹⁸O_{lake_water} for the 20th century, as it is at present, and that neither of the lakes has been significantly impacted by evaporative enrichment, for example as a result of local hydrological changes, in the past. Several lines of evidence suggest that this assumption is reasonable. First, the δ¹⁸O_ostracod records for the two lakes broadly agree for the period of temporal overlap. Had local hydrological effects, such as anthropogenic catchment modifications, led to significant evaporative enrichment, this would be most likely to have affected one or the other of the two catchments and not both. Climatic influences, in particular a reduction of effective moisture, could have led to significant evaporative enrichment, but there is no evidence for this in the meteorological data. Secondly, there is no covariance between δ¹⁸O_ostracod and δ¹³C_ostracod in the down-core data (Figure 4c and d). Although not an unequivocal indicator of increased residence time and evaporative enrichment in mid-latitude lakes (Drummond et al., 1995; Talbot, 1990), the lack of covariance does lend some support to our assumption that such enrichment has not been a significant component of the lakes’ hydrological balance during the 20th century.

Figure 5.

(a) δ¹⁸O_ostracod data from ALRE19 (grey triangles) and GRA16 (grey circles) with a five-point moving average (ALRE19 = red and GRA16 = blue); (b) boxplot summarising the δ¹⁸O_ostracod data from ALRE19 (red) and GRA16 (blue) from 1904 CE. Outliers have been removed (all data are presented in Supplemental Material).

Comparison of the structure of the δ¹⁸O_ostracod and δ¹⁸O_synthetic records for the 20th century shows reasonable agreement given the chronological uncertainties of δ¹⁸O_ostracod (Figure 6c–d). The key characteristic of the δ¹⁸O_ostracod record is the positive shift from the 1960s, peaking ca. 1971 CE (Figure 6d). This is followed by a fluctuation of values around a stable mean, which is lower than the 1960s–70s peak. We also observe a peak in the δ¹⁸O_synthetic record ca. 1960s, but the stabilising of values post-δ¹⁸O_synthetic are not necessarily more negative. The long-term negative shift in δ¹⁸O_ostracod values from 1904 to 1960 CE is interrupted by a positive shift in the 1940s CE, which is also evident in the δ¹⁸O_synthetic record. Whilst the absolute values of the two records are offset on average by ca. 0.6 ‰, the range between maximum and minimum of the moving average values is ca. 1 ‰. The overall long term negative trend in the δ¹⁸O_ostracod is not as clearly captured in the δ¹⁸O_synthetic record (Figure 6c and d). This is unlikely to be a result of basin-specific effect because this difference is observed in both the Old Alresford Pond and The Grange Lake records, as well as both δ¹⁸O_ostracod and δ¹³C_ostracod datasets (Figure 4).

Figure 6.

(a) Modelled δ¹⁸O_{precipitation} at Old Alresford Pond from Piso_AI model (Nelson et al., 2021) 1904 to 2020 CE where grey circles = all data points and red line = five-point moving average. (b) Mean annual temperature in Southampton (grey triangles) with five-point moving average (red line) and mean winter (December, January, February) temperature in Southampton (grey circles) with five-point moving average 1904–2019 CE from MetOffice. (c) δ¹⁸O_synthetic at Old Alresford Pond (grey circles) with red line = five-point moving average. (d) Measured δ¹⁸O_ostracod values from ALRE19 (grey triangles) with five-point moving average (red line) and GRA16 (grey circles) with five-point moving average (blue line). (e) Measured δ¹⁸O_{precipitation} at Wallingford GNIP site 1979–2019 CE and five-point moving average (IAEA/WMO, 2023). (f) Summer (June, July, August) rainfall amount (grey triangles) with five-point moving average (red line) and winter (December, January, February) rainfall amount (grey circles) with five-point moving average (blue line) from Southampton MetOffice, 1904–2000 CE. All data in Supplemental Material. MetOffice data made available from the National Meteorological Library and Archive.

The consistency between these two records is important for the ‘ground-truthing’ of the data. For example, modern monitoring indicates that δ¹⁸O_ostracod at Old Alresford Pond is controlled by δ¹⁸O_{lake_water} (which in this instance is effectively δ¹⁸O_{precipitation}) and autumn/winter temperatures. The reasonable agreement between the δ¹⁸O_ostracod and δ¹⁸O_synthetic, which is generated using winter temperatures and δ¹⁸O_{precipitation}, indicates that this relationship holds true for the 20th century at least. Equally, the good agreement between the modelled δ¹⁸O_{precipitation} data, which is incorporated in the δ¹⁸O_synthetic values, and the empirical data provides added confidence that the modelled outputs are robust. For instance, the positive shift in δ¹⁸O_{precipitation} in the 1960s–1970s corresponds with positive shifts observed in both δ¹⁸O_ostracod and δ¹⁸O_synthetic. The uncertainty surrounding the ALRE19 age model (with age uncertainties as much as ±23.5 years at some levels) make the statistical testing of this similarity challenging. However, to gain some insight into the similarity, we split the record based on the sampling density in the ALRE19 record; from ca. 1904 to 1940 CE the Pearson correlation coefficient (r) value between δ¹⁸O_ostracod and δ¹⁸O_synthetic is 0.04, ca. 1941 to 1970 CE equal to 0.23 and from ca. 1971 to 2019 CE at 0.27. This demonstrates some statistical similarity throughout the record and that there is some similarity in the direction of correlation between the δ¹⁸O_ostracod and δ¹⁸O_synthetic prior to 1980 CE. The difference post-1980 CE may be a factor of the distribution of δ¹⁸O_ostracod data points, particularly how few δ¹⁸O_ostracod data points are in the ca. 1990s, as the moving averages in Figure 6c and 6d indicate that a slight negative trend is present in both δ¹⁸O_ostracod and δ¹⁸O_synthetic.

Whilst a reasonable agreement exists between the δ¹⁸O_ostracod and δ¹⁸O_synthetic records there is less of a clear relationship between δ¹⁸O_ostracod and any single climatic parameter. For example, both mean annual and winter temperatures have increased across the 20th century in the Southampton dataset (Figure 6b), but no comparable, simple trend exists within the isotopic dataset. It is likely that this discrepancy reflects the fact that both the δ¹⁸O_ostracod and δ¹⁸O_synthetic records reflect an amalgamation of multiple climate factors not just autumn/winter temperatures but also the different climate factors that control δ¹⁸O_{precipitation} (i.e. air temperature, the amount/seasonality rainfall, air mass trajectory, etc.; see Rozanski et al., 1992, 1993). Only the trend in summer rainfall amount shows any comparability with the trend seen in δ¹⁸O_ostracod with a value of 0.15 between ca. 1904 and 2000 CE (Figure 6f). Summer rainfall variability has the potential to produce the negative/positive trends seen in the δ¹⁸O_ostracod record, that is, as summer rainfall typically has the most positive δ¹⁸O_{precipitation} values of any rainfall occurring throughout the year, its increase/decrease will potentially result in more positive/negative δ¹⁸O_ostracod values respectively by its contribution to average δ¹⁸O_{lake_water} values. The magnitude of changes seen in the δ¹⁸O_ostracod record are, however, unlikely to be explained by the changes in summer rainfall based on the Southampton weather station data.

Discussion

Can we identify individual climate drivers from the δ¹⁸O_ostracod record?

In many Pleistocene studies in Northwestern Europe the variability in the δ¹⁸O of lake carbonates is frequently interpreted as reflecting temperature changes (Blockley et al., 2018; Candy et al., 2016). In the present study, it is unlikely that temperature is the only driving factor, however, because the magnitude of the temperature changes that are observed in some of these Late Pleistocene transitions, such as the Younger Dryas, is so large that it is often interpreted as the primary control and is assumed to overwhelm the influence of other environmental factors. The present study has demonstrated, however, that during the Late-Holocene, and in particular the 20th century when the magnitude of temperature change is much smaller (Figure 6b), several climatic variables interact to control the δ¹⁸O of lake carbonates, in this case ostracods. Furthermore, these interactions may not be straightforward, as measured temperature and rainfall amount do not necessarily follow the same trends (Figure 6b and f). In this example there is a weak correlation between winter temperature (the season when the candonids used for this study calcify) and summer rainfall (thought to have a greater influence on δ¹⁸O_{precipitation} and therefore δ¹⁸O_{lake_water}). The Pearson correlation coefficient (r) from 1904 to 2000 CE for these variables is −0.02. Whilst it is possible to use anthropogenic lakes, such as Old Alresford Pond, as an archive for isotopic changes across the last 1000 years, interpreting this proxy record based on a single climate variable is problematic. This study has shown that such lakes do have a key role, however, in validating the output of isotope enabled climate models. These archives allow high-resolution isotopic records to be generated that can be directly compared with modelled δ¹⁸O_{precipitation} values. The coupling of model and empirical isotope data, through studies such as these, therefore, has the potential to validate and improve modelling studies in the future.

Can anthropogenic lakes be used as climate archives?

The British Isles hosts many anthropogenic lakes including ornamental lakes and medieval fishing or stew ponds. These are typically overlooked as archives of past climate change, largely because of the influence of human activity on the record and the challenges faced in creating robust chronologies over recent centuries. However, such sites are likely to have written histories that may help overcome some of these challenges, as well as extant water bodies that allow us to disentangle climate and catchment variability captured in our chosen proxy, as in the present study. This study has shown that there is great potential in using anthropogenic lakes, in this instance those that precipitate carbonate, as archives of past climatic and environmental change over the historical period. Consequently, using anthropogenic lakes in this way has the potential to make a significant contribution to our knowledge and understanding of the regional impact of known climate changes in this period (e.g. MCA and LIA). Considering the known variability in the expression of these events across NW Europe, an ability to produce proxy records that may be able to quantify climate change variability at a more localised level, will be of significant interest to the scientific community. From the perspective of the British Isles, decades of Holocene paleoclimatic research have demonstrated that there is often regional variability and asynchronicity in response to climate change (e.g. Charman, 2010). As we continue to understand what modern climate change may look like in the British Isles it becomes increasingly important to develop our understanding of the magnitude and variability of past climate change, especially in the Late-Holocene where external climate driving mechanisms have changed little. Therefore, improving the spatial and temporal resolution of such records will provide a useful reference point. Future work should investigate additional anthropogenic lakes across the British Isles to assess their potential for paleoclimatic reconstruction. Oxygen isotopes in particular have the advantage of yielding quantified climate reconstructions and are therefore a valuable proxy to employ in such investigations. A network of sites across the region would provide the opportunity to compare sites over a wider spatial scale and identify possible trends, synchronicities and complexities in the expression of climate variability over this time. This may provide important insights into the expression and impact of known events, such as the LIA, on the region. Furthermore, based on the success of this study in comparing to modelled values, it may be possible to further explore the use of a proxy-model comparison, employing isotope-enabled climate models to help identify possible driving mechanisms behind observed climate trends recorded in proxy oxygen-isotope datasets.

Conclusions

This work has brought together modern limnological monitoring alongside palaeoclimatological methods to investigate the potential of constructed lakes as archives of 20th century climate. We demonstrated for our Hampshire sites that δ¹⁸O_{lake_water} is representative of local δ¹⁸O_{precipitation}, and this is recorded in δ¹⁸O_ostracod. Furthermore, we have shown that water temperature is in equilibrium with air temperature in the present day, and therefore assume that this is true in the past. We observe similarities in the direction of δ¹⁸O_ostracod trends in both Old Alresford Pond and The Grange Lake over the 20th century, indicating they are most likely driven by a common climate driver rather than basin-specific effects. We explore this further by comparing δ¹⁸O_ostracod trends to recorded changes in temperature and rainfall amount, two climatic factors that may drive changes in δ¹⁸O_ostracod. These comparisons demonstrated that there is no simple relationship between measurable climate variables (temperature and rainfall amount) and δ¹⁸O_ostracod and so we argue that a complex interaction of climatic variables drive the observed δ¹⁸O_ostracod shifts in the 20th century. We demonstrate a good relationship between measured and modelled isotopic values in this study and discuss the potential role anthropogenic lakes containing paleoclimatic proxy records may have in ‘ground-truthing’ modelled oxygen-isotope outputs. Future work that employs a more extensive data-model comparison approach using isotope-enabled climate models in lacustrine environments may allow for a greater variety of climate driving mechanisms on δ¹⁸O_ostracod to be assessed, given that there is no simple driver in our 20th century record. Overall, we have shown the sediments of anthropogenic lakes are valuable archives of climatic variability over the last millennium.

Supplemental Material

sj-docx-1-hol-10.1177_09596836241266435 – Supplemental material for Investigating the potential of oxygen-isotope records from anthropogenic lakes as tracers of 20th century climate change

Supplemental material, sj-docx-1-hol-10.1177_09596836241266435 for Investigating the potential of oxygen-isotope records from anthropogenic lakes as tracers of 20th century climate change by JM Tindall, J Holmes, I Candy, MJ Leng, P Langdon, T Fonville and DA Sear in The Holocene

Research Data

sj-xlsx-2-hol-10.1177_09596836241266435 – Research Data for Investigating the potential of oxygen-isotope records from anthropogenic lakes as tracers of 20th century climate change

Research Data, sj-xlsx-2-hol-10.1177_09596836241266435 for Investigating the potential of oxygen-isotope records from anthropogenic lakes as tracers of 20th century climate change by JM Tindall, J Holmes, I Candy, MJ Leng, P Langdon, T Fonville and DA Sear in The Holocene

This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).

Footnotes

Acknowledgements

We thank Ian Patmore, James Shilland, Lucy Roberts, Branaavan Sivarajah for help with the field work. The late Tony Chambers and Rosemary Chambers for access to Old Alresford Pond and the late Lord Baring and Zam Baring for access to The Grange. Hilary Sloane, Kotryna Savickaite and Carol Arrowsmith for the isotope analyses. Cath D’Alton for producing the maps in Figure 1 and Jim Davy and Miles Irving for SEM images (Supplemental Materials). The meteorological datasets from Odiham, Otterbourne and Southampton NOC were kindly provided by the National Meteorological Library and Archive – MetOffice UK. The Wallingford GNIP dataset was collected by the British Geological Survey and accessed through the database.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Natural Environment Research Council [grant number NE/L002485/1] studentship through the London NERC DTP and the British Geological Survey NERC Isotope Facility application IP-1851-1118. Initial coring was supported by funding from the Environment Agency for which Tim Sykes (EA Biodiversity lead) is thanked.

ORCID iDs

JM Tindall

DA Sear

Supplemental material

Supplemental material for this article is available online.

References

Allen

(ed.) (2017) The Chalk aquifer of the Wessex Basin. British Geological Survey Research Report No. RR/11/02. p. 118.

Appleby

(2001) Chronostratigraphic techniques in recent sediments. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments Volume 1: Basin Analysis, Coring and Chronological Techniques. Dordrecht, Netherlands: Kluwer Academic Publishers, pp.171–203.

Binnie

(1974) The evolution of British dams. Transactions of the Newcomen Society 47(1): 207–224.

Bishop

(2021) Ornamental Lakes: Their Origins and Evolution in English Landscapes. Abingdon: Routledge.

Blockley

Candy

Matthews

, et al. (2018) The resilience of postglacial hunter-gatherers to abrupt climate change. Nature Ecology and Evolution 2: 810–818.

Bronk-Ramsey

(2009) Bayesian analysis of radiocarbon dates. Radiocarbon 51(1): 337–360.

Bronk-Ramsey

(2021) OxCal 4.4. Available at: https://c14.arch.ox.ac.uk/oxcal/OxCal.html (accessed 21 April 2024).

Brönnimann

Allan

Ashcroft

, et al. (2019) Unlocking pre-1850 instrumental meteorological records: A Global Inventory. Bulletin of the American Meteorological Society 100(12): ES390–ES413.

Burgdorf

A-M

(2022) A global inventory of historical documentary evidence related to climate since the 15th century. Climate of the Past 18: 1407–1428.

10.

Candy

Abrook

Elliot

, et al. (2016) Oxygen isotopic evidence for high-magnitude, abrupt climatic events during the Lateglacial Interstadial in north-west Europe: Analysis of a lacustrine sequence from the site of Tirinie, Scottish Highlands. Journal of Quaternary Science 31(6): 607–621.

11.

Charman

(2010) Centennial scale variability in the British Isles during the mid-late-Holocene. Quaternary Science Reviews 29: 1534–1554.

12.

Craig

(1957) Isotopic standards for carbon and oxygen & correction factors for mass spectrometric analysis. Geochemica et Cocmochemica Acta 12: 133–149.

13.

Currie

(2001) Archaeological recording at The Grange, Northington, Hampshire. Report, CKC Archaeology. https://doi.org/10.5284/1006608

14.

Dansgaard

(1964) Stable isotopes in precipitation. Tellus 16(4): 436–468.

15.

Decrouy

(2009) Environmental and biological controls on the geochemistry (δ¹⁸O, δ¹³C, Mg/Ca, and Sr/Ca) of living ostracods from Lake Geneva. PhD Thesis, Université de Lausanne, Lausanne, Switzerland.

16.

Deveson

(2005) The early history of The Grange, Northington. Proceedings of Hampshire Field Club Archaeology Society 60: 198–208.

17.

Diaz

Trigo

Hughes

, et al. (2011) Spatial and temporal characteristics of climate in medieval times revisited. American Meteorological Society 92(11): 1487–1500.

18.

Dobrovolný

Rybníček

Kolář

, et al. (2018) May-July precipitation reconstruction from oak tree-rings for Bohemia (Czech Republic) since AD 1040. International Journal of Climatology 38: 1910–1924.

19.

Dole-Olivier

Galassi

DMP

Marmonier

, et al. (2000) The biology and ecology of lotic microcrustaceans. Freshwater Biology 44(1): 63–91.

20.

Drummond

Patterson

Walker

JCG

(1995) Climatic forcing of carbon-oxygen isotopic covariance in temperate-region marl lakes. Geology 23(11): 1031–1034.

21.

Environment Agency (2013) Test & Itchen river restoration strategy technical report. https://www.therrc.co.uk/sites/default/files/files/Designated_Rivers/Test_Itchen/technical_report_issue_5_final.pdf

22.

Environment Agency (2020) Catchment data explorer: Old Alresford Pond. Available at: https://environment.data.gov.uk/catchment-planning/WaterBody/GB30744431 (accessed 17 November 2020).

23.

Freshwater Habitats Trust (2022) Ornamental lakes with the National Trust. Available at: https://content.freshwaterhabitats.org.uk/2019/11/Ornamental-Lakes-National-Trust.pdf? (accessed 21 July 2022).

24.

Freund

Helle

Balting

, et al. (2023) European tree-ring isotopes indicate unusual recent hydroclimate. Communications Earth and Environment 4: 26.

25.

Geddes

(1983) The Grange, Northington. Architectural History 26: 35.

26.

Goosse

Guiot

Mann

, et al. (2012) The medieval climate anomaly in Europe: Comparison of the summer and annual mean signals in two reconstructions and in simulations with data assimilation. Global and Planetary Change 84–85: 35–47.

27.

Historic England (2023) 1,048 search results for ‘medieval fishpond’. Available at: https://historicengland.org.uk/listing/the-list/results/?searchType=NHLE+Simple&search=medieval+fishpond (accessed 11 August 2023).

28.

Holmes

Chivas

(2002) Ostracod shell chemistry – overview. In: Holmes

Chivas

(eds) The Ostracoda: Applications in Quaternary Research. Geophysical Monograph Series. Washington, DC: American Geophysical Union, pp.118–204.

29.

IAEA/GNIP (2014) IAEA/GNIP precipitation sampling guide Introduction: The Global Network of Isotopes in Precipitation (GNIP). Available at: https://nucleus.iaea.org/sites/ihn/Pages/GNIP.aspx

30.

IAEA/WMO (2023) Global network of isotopes in precipitation. The GNIP Database. Available at: https://nucleus.iaea.org/wiser

31.

Keatings

Heaton

THE

Holmes

(2002) Carbon and oxygen isotope fractionation in non-marine ostracods: Results from a ‘natural culture’ environment. Geochimica et Cosmochimica Acta 66(10): 1701–1711.

32.

Kelts

Talbot

(1990) Lacustrine carbonates as geochemical archives of environmental change and biotic/abiotic interactions. In: Tilzer

Serruya

(eds) Large Lakes: Ecology, Structure and Function. Netherlands, Dordrecht: Springer, pp.288–315.

33.

Kim

S-T

O’Neil

(1997) Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochimica et Cosmochimica Acta 61(16): 3461–3475.

34.

Külköylüoĝlu

Vinyard

(2000) Distribution and ecology of freshwater Ostracoda (Crustacea) collected from Springs of Nevada, Idaho, and Oregon: A preliminary study. Western North American Naturalist 60(3): 291–303.

35.

Leng

Marshall

(2004) Paleoclimate interpretation of stable isotope data from lake sediment archives. Quaternary Science Reviews 23(7–8): 811–831.

36.

Nelson

Basler

Kahmen

(2021) Precipitation isotope time series predictions from machine learning applied in Europe. Proceedings of the National Academy of Sciences 118(26): e2024107118.

37.

Nesje

Dahl

(2003) The ‘Little Ice Age’ – only temperature? The Holocene 13(1): 139–145.

38.

Norgate Mand Norgate

(2002a) Milne’s Hampshire 1791. Old Hampshire Mapped. Available at: https://www.oldhampshiremapped.org.uk/hantsmap/milne1/MLN55.htm (accessed 21 June 2024).

39.

Norgate

(2002b) Ordnance Survey ca. 1817 CE map. Old Hampshire Mapped. Available at: https://www.oldhampshiremapped.org.uk/hantsmap/ordnce6/OSS27.htm (accessed 27 July 2024).

40.

Norgate

(2006c) Greenwood’s Hampshire 1826. Old Hampshire Mapped. Available at: https://www.oldhampshiremapped.org.uk/hantsmap/grnwood2/GRW65F.htm (accessed 21 July 2024).

41.

Owens

Lockwood

Hawkins

, et al. (2017) The Maunder minimum and the Little Ice Age: An update from recent reconstructions and climate simulations. Journal of Space Weather and Space Climate 7: 1–10.

42.

Parks and Gardens (2021) The Grange, Northington. Available at: https://www.parksandgardens.org/places/grange-the-northington#:~:text=SirRobertHenley%2C first Lord Northington%2C is credited, out either side of a long%2C narrow lake (accessed 29 June 2021).

43.

Roberts

(1986) The Bishop of Winchester’s fishponds in Hampshire, 1150–1400: Their development, function and management. Proceedings of Hampshire Field Club Archaeological Society 42: 125–138.

44.

Rollason

(2017) Princes of the Church: Bishops and Their Palaces, 1st edn. Abington: Taylor & Francis.

45.

Rose

(2008) ‘Quality control in the analysis of lake sediments for spheroidal carbonaceous particles’. Limnology and Oceanography: Methods 6: 172–179. https://doi.org/10.4319/lom.2008.6.1

46.

Rose

Appleby

(2005) Regional applications of lake sediment dating by spheroidal carbonaceous particle analysis I: United Kingdom. Journal of Paleolimnology 34: 349–361.

47.

Rozanski

Araguás-Araguás

Gonfiantini

(1992) Relation between long-term trends of oxygen-18 isotope composition of precipitation and climate. Science 258(5084): 981–985.

48.

Rozanski

Araguás-Araguás

Gonfiantini

(1993) Isotopic patterns in modern global precipitation. In: Swart

Lohmann

McKenzie

, et al. (eds) Climate Change in Continental Isotopic Records. Washington, DC: American Geophysical Union, pp.1–36.

49.

Ruddiman

(2003) The anthropogenic greenhouse era began thousands of years ago. Climatic Change 61: 261–293.

50.

Stuart

Smedley

(2009) Baseline groundwater chemistry: The Chalk aquifer of Hampshire.Open Report OR/09/052. Nottingham, UK: British Geological Survey, p.61.

51.

Talbot

(1990) A review of the palaeohydrological interpretation of carbon and oxygen isotopic ratios in primary lacustrine carbonates. Chemical Geology: Isotope Geoscience Section 80(4): 261–279.

52.

Trouet

Esper

Graham

, et al. (2009) Persistent positive North Atlantic Oscillation mode dominated the medieval climate anomaly. Science 324(5923): 78–80.

53.

Trouet

Scourse

Raible

(2012) North Atlantic storminess and Atlantic Meridional Overturning Circulation during the last Millennium: Reconciling contradictory proxy records of NAO variability. Global and Planetary Change 84–85: 48–55.

54.

Turpin

(2008) Dam. London: Reaktion Books.

55.

von Grafenstein

Erlernkeuser

Trimborn

(1999) Oxygen and carbon isotopes in modern fresh-water ostracod valves: Assessing vital offsets and autecological effects of interest for paleoclimate studies. Palaeogeography, Palaeoclimatology, Palaeoecology 148(1–3): 133–152.

56.

Wang

(2013) Simulating complex hydro-geomorphic changes in lake-catchment systems. PhD Thesis, University of Southampton, Southampton.

57.

Wanner

Solomina

Grosjean

, et al. (2011) Structure and origin of Holocene cold events. Quaternary Science Reviews 30(21–22): 3109–3123.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.11 MB

0.18 MB