The magnetic record of inorganic fly ash deposition in lake sediments and ombrotrophic peats

Abstract

Interest in identifying a geological marker signifying the starting point for the Anthropocene has prompted an exploration of the stratigraphic record of inorganic particulates generated by industrial activities. Magnetic measurements of recent lake sediments and ombrotrophic peats are here used to reconstruct the history of deposition of inorganic fly-ash spheres resulting mainly from solid fuel combustion and metal smelting. The chronologies used have been based on moss-increment counting, radioisotope dating and pollen analysis. The sites come from the United Kingdom, Scandinavia and North America. In several cases where detailed chronologies of both fly-ash deposition and local industry can be compared, the sequence of concentration-linked magnetic measurements appears to capture accurately the record of industrial development despite incontrovertible evidence from other peat-based records for some selective dissolution of magnetic minerals. The dates at which magnetic concentration increases begin range from the 16th century in the peat profiles around the head of Morecambe Bay, South Cumbria, in North-West England where early iron manufacture using charcoal-fuelled bloomery hearths is well documented, to the mid-20th century at the remotest sites in Arctic Scandinavia. The lake sediment profiles used here come mainly from the United Kingdom and, in most cases, they date increases to the late 19th century or the first decades of the 20th century. Any attempt to use the magnetic record of fly-ash deposition in lake sediments and/or peats to mark the date chosen as the onset of the Anthropocene would require careful choice of site location and archive, bearing in mind the issue of selective magnetic mineral dissolution.

Keywords

Anthropocene fly-ash industrial pollution lake sediments magnetic measurements peat

Introduction

The proposal that the Anthropocene should be recognised as a new geological epoch raises the question of what might constitute a suitable marker for its inception. In the original definition of Crutzen and Stoermer (2000), the onset is placed during the early stages of the Industrial Revolution which were, prior to the emergence of oil as a dominant fossil fuel-based energy source and the development of the internal combustion engine, literally fuelled by coal. Coal combustion and many associated heavy industries, such as iron and steel manufacture and metal smelting, generate residual, un-combusted mineral particles many of which leave the site of combustion as particulate pollutants in the atmosphere. These particulates are often referred to under the generic term ‘fly-ash’. Rose (1994) divides fly-ash into organic and inorganic fractions, the former comprising spherical carbonaceous particles (SCPs) and the latter inorganic ash spheres (IASs). It is with these latter (Figure 1) that the present article is concerned. Rose (1990) details a chemical extraction technique for isolating IAS, but here, we focus on non-destructive, magnetic methods of detection and measurement.

Figure 1.

Pictures of inorganic fly-ash spheres from Ely Lake Bog (Figure 10). (a and b) Show fly-ash spheres from the 2–3 cm depth and (c and d) are from 14–16 cm depth and show signs of surface degradation.

Previous publications (Anderson et al., 1986; Battarbee et al., 1988; Fritz et al., 1989; Oldfield, 1991; Oldfield and Richardson, 1990; Oldfield et al., 1978, 1981; Patrick et al., 1989; Renberg et al., 1990; Richardson, 1986; Snowball et al., 2013; Thompson and Oldfield, 1986; Thompson et al., 1980; Tolonen and Oldfield, 1986) confirm that magnetic records from both lake sediments and peats can shed light on the history of atmospheric deposition from industrial activities. Moreover, these records also show that (1) the spatial distribution of the inventories of atmospheric deposition captured by magnetic measurements (Thompson and Oldfield, 1986) reflects the spatial distribution of heavy industry, solid fuel-based power generation and coal combustion in Europe, (2) where direct comparisons can be made between dated records of magnetic deposition and local industry there can be a compelling match between the two (Oldfield et al., 1981; Tolonen and Oldfield, 1986) and (3) dates for the onset of the magnetic record of atmospheric deposition sometimes lie close to Crutzen and Stoermer’s preferred date (ad 1800) for the start of the Anthropocene. At the same time, Williams (1992) has shown and subsequent work has confirmed that despite the apparent credibility of some magnetic records of atmospheric deposition in peat, there is incontrovertible evidence for partial dissolution of the record at some sites.

Particles of IAS arising from coal combustion are rich in both haematite and magnetite (Bhattacharjee et al., 2013; Sharonova et al., 2003; Tishmack and Burns, 2004; Winburn et al., 2000), as well as heavy metals such as lead and zinc. The former two can be recognised and their presence partially quantified by means of magnetic measurements (Liu et al., 2012; Oldfield, 1991, 2012). Particulates rich in these components have been widely dispersed in the atmosphere from the beginning of the Industrial Revolution. They are readily incorporated in accumulating lake sediments and peats, both of which are widespread in many industrialised areas and both of which often accumulate sufficiently rapidly to provide a detailed stratigraphic record. Whereas lake sediments usually also contain magnetic minerals derived from the lake catchment and/or from biological productivity within the lake, the so-called ombrotrophic (rain-fed) peat bogs (Clymo, 1984) which are, by definition, raised above the ground water-table, are nourished entirely by atmospheric deposition. They therefore occupy a prime position in archives of particulate pollution history, although one that is not without problems and challenges, as noted below.

The text that follows presents a summary of lake sediment and peat analyses carried out mainly during the 1980s and 1990s. The examples chosen are representative of 36 lake sediment records and over 200 peat cores, together comprising a total of over 70 sites, mainly in the United Kingdom, Europe and, to a lesser extent, in North America. Many are from unpublished PhD dissertations (Gedye, 1998; Jones, 1985, 1986) by graduate students from the then Department of Geography, University of Liverpool. Wherever possible, the records used in the present account include chronological information giving, at the very least, some indication of the depth at which the record of industrial deposition begins. The dating methods used are based on a range of techniques.

Annual moss-increment counting

This relies on the strong seasonality that controls the growth of moss (notably Sphagnum and Polytrichum) species in continental climates (Pakarinen and Tolonen, 1977). It can give accurate and precise chronologies down to the depth at which the annual growth increments are no longer recognisable owing to decomposition. Below this depth, an accumulation model detailed in Pakarinen and Tolonen (1977) extends the chronology back in time, but with less precision and accuracy.

Short-lived radioisotope dating. The main radioisotope used is lead 210 (²¹⁰Pb) which, under ideal circumstances, can be used to date the last 100–150 years. Its application to lake sediments is outlined in Appleby (2001, 2008) and to peats in Appleby et al. (1997) and Oldfield et al. (1995). A single dated profile on its own and without independent constraints is often problematical and ²¹⁰Pb is best used in conjunction with the isotopes caesium 137 (¹³⁷Cs) and americium 241 (²⁴¹Am). Both have been around since post-1945 nuclear weapons testing in the 1950s and both had peak deposition in 1963 in the northern hemisphere and a year later in the southern hemisphere. Caesium, but not americium, was also widely distributed in the northern hemisphere as a result of the Chernobyl accident in 1986 (Van der Post et al., 1997). Both often provide reliable constraints on ²¹⁰Pb dates in lake sediments, but ¹³⁷Cs especially can become mobile in acid peat, which greatly reduces its value as a chronological marker. Even more valuable as constraints on ²¹⁰Pb based chronologies are independent markers as detailed below and illustrated in Figure 2.

Pollen analysis. Some of the changes in agricultural practices that can be securely dated from documentary evidence can also be recognised in pollen diagrams spanning the last few centuries. They can therefore be used as important constraints on radioisotope dates. In NW England especially, one of the most useful in the present context is the demise of hemp (Cannabis sativa) cultivation during the second half of the 18th century as a consequence of the increasing importance of jute from abroad (Houseman, 1800; Oldfield and Statham, 1963; Richardson, 1986). In N America, the spread of weeds in the wake of European settlement has also left an important marker in the pollen record – the appearance and increasing representation of the weed Ambrosia artemsiifolia (Richardson, 1986).

Radiocarbon dating. Individual radiocarbon measurements are of little use in dating sediments and peats spanning the last two to three centuries. Chronologies can, however, be established by mapping a sequence of closely spaced ‘dates’ onto the calibration curve derived from comparing radiocarbon activity variations with tree rings (Figure 2).

Figure 2.

Chronology of recent peat accumulation based on wiggle-matched ¹⁴C, ²¹⁰Pb, ²⁴¹Am, pollen correlations and historical evidence: Ellergower Moss, Galloway, SW Scotland (Oldfield et al., 1995). (a) Successive, contiguous ¹⁴C ‘dates’ for Core EM3, mapped onto the tree-ring based calibration curve. Pollen data from the same core show the increase in Pinus pollen frequency between inflexions B and C, with an interpolated date of c. ad 1835. (b) The wiggle-matched ¹⁴C dated inflexions and the Pinus pollen increase in Core EM3 plotted against depth and Year ad. (c) The results of ²¹⁴Am determinations for Core EM4. (d) Plots of ²¹⁰Pb derived age–depth curve for Core EM4 with, superimposed on it, the Pinus increase date based on pollen correlation between EM3 and EM4, the ²⁴¹Am dates and the age of the final increase in Pinus pollen frequency reflecting recent, commercial afforestation. Ages are from ad 1990.

Magnetic measurements

The advantages of the magnetic approach include non-destructive measurement with minimal pre-treatment, detection of magnetic minerals at concentrations orders of magnitude below those achieved by conventional techniques and some partially quantitative discrimination between different magnetic mineral contributions. The disadvantages include the fact that the relationship between the measurements and magnetic mineral concentrations is not precisely quantifiable. During the period when the reported results were obtained, the equipment available for magnetic measurements changed, and the methodology evolved in response both to these changes and to increasing experience of what, at the time, was still a new approach to environmental research. The following magnetic properties are used and interpreted as follows.

Saturation isothermal remanent magnetisation

Saturation isothermal remanent magnetisation (SIRM) measured in a DC field of 1 T or, in the case of some early analyses, 850 mT. This is used here to provide the best available indication of changes in total magnetic mineral concentrations. In the present measurements, SIRM values mainly reflect changes in the concentration of magnetite since it is some two orders of magnitude more magnetic than the other main component, haematite. In Figures 5 –8 and 10, SIRM values are shown as proxies for concentration. In Figures 3 and 9, the values have been recalculated as an accumulation rate per unit area.

Hard magnetic remanence

Hard magnetic remanence (HIRM) expresses that part of the original SIRM that remains unreversed in a high reverse field (here usually −100 mT or 300 mT). In Figure 4a, it is used as a proxy for changing haematite concentrations; in Figure 3, the HIRM values have been recalculated as an accumulation rate. In Figures 5 –8 and 10, they are shown alongside the SIRM profiles and expressed as percentages of SIRM unreversed, indicating the changing contribution of haematite relative to the total magnetic mineral record.

Figure 3.

SIRM accumulation rates in three lake sediment cores from Scotland and Wales. (a) Lochnagar in the Cairngorms, NE Scotland; (b) Loch nah Achlaise, SW Highlands, Scotland; and (c) Llyn Dulyn, Snowdonia, N Wales.

Methods of measurement and calculation as well as the basis for interpreting these measurements are detailed in Oldfield (1999).

Depositional histories

Lake sediment records

Examples referred to are from sites in Great Britain (Battarbee et al., 1988; Oldfield and Richardson, 1990), Sweden (Renberg et al., 1990) and the United States (Oldfield, 1990). Each of the cores analysed, with the exception of that from Grey Heugh Slack (Battarbee et al., 2015) was taken in the 1980s either as part of the Surface Water Acidification Project (SWAP) in Europe (Battarbee et al., 1990), a UK Department of Environment Programme, or the Paleocological Investigations of Recent Lake Acidification (PIRLA) project in North America. All were carried out with a view to establishing the history of industrial atmospheric deposition in relation to the onset and progress of lake acidification. In all cases where dates are given, they are based on ²¹⁰Pb analyses.

British sites. A total of 39 suites of magnetic measurements from 32 sites were completed as part of the SWAP and related programmes (Battarbee et al., 1988). Three well-dated cores, from Scotland (Lochnagar; Loch nah Achlaise) and Wales (Llyn Dulyn), are used in Figure 3 as examples.

At Lochnagar, as in all the dated profiles from the most remote Scottish sites, there is no evidence for any significant increase before ad 1900. Slight increases occur between 1900 and 1950, with the most dramatic increases after 1950. Llyn Dulyn in Snowdonia, NW Wales, Loch nah Achlaise 60 km north of the industrialised Clyde Estuary and also Scoat Tarn in the English Lake District (Battarbee et al., 1990) are less remote from areas of heavy 19th-century industrialisation in S Wales, Scotland and N England, respectively. In these cases, there are clear indications of increasing magnetic concentrations prior to 1900, steady increases around 1910–1925 and, in the case of Llyn Dulyn, a further sharp increase in the mid-20th century. At Grey Heugh Slack on the North York Moors in NE England, magnetic deposition begins to increase in the early 19th century alongside other geochemical markers of industrial activity (Battarbee et al., 2015).

Lilla Oresjon, SW Sweden was one of the key sites in the SWAP programme. As Figure 4a shows, there is some indication of a low, sustained increase in haematite concentrations from ad 1900 onwards followed by a sharp increase after 1960. There are similar indications from the record of SCPs, poly-aromatic hydrocarbons (PAHs) and lead concentrations in the same core (Renberg et al., 1990).

Big Moose Lake, New York State, USA, lies in the Adirondack Mountains. It was part of the N American PIRLA project (Charles and Whitehead, 1986). The detailed ²¹⁰Pb chronology allows the results to be expressed as an accumulation rate. Comparison between two analysed cores led Oldfield (1990) to conclude that the ‘hard’ remanence broadly representative of haematite deposition was the least ambiguous indicator of atmospheric deposition at the site. Figure 4b shows the changing accumulation rates of this component in two cores from the site. There is no sign of any significant increase prior to ad 1890, after which a steady increase is recorded in both cores, peaking at the end of the record in the late 1970s.

Figure 4.

Lake sediment records from S Sweden and the Adirondacks, New York State, USA. (a) Selected data from Lille Oresjon (Renberg et al., 1990). (b) The record of ‘Hard’ IRM accumulation in two cores from Big Moose Lake (Oldfield, 1990). In both plots, the ‘hard’ IRM (remanence remaining at −300 MT) has been isolated to avoid the effects of magnetite derived from the catchment. Whereas plot (a) shows ‘hard’ IRM as a proxy for changing concentrations, the detailed chronology allows the values in plot (b) to be recalculated as a changing accumulation rate.

Ombrotrophic peatland records

North-West England – Morecambe Bay peatlands, South Cumbria. Even before the beginning of the 18th century small scale iron production took place in the area, using local iron ore and charcoal-fuelled ‘bloomeries’. During the second decade of the 18th century, the local iron industry began to expand rapidly with coke-powered blast furnaces common by ad 1750 (Awty and Philips, 1979–1980). Heavy industry based on nearby iron and coal continued until the early decades of the 20th century.

Richardson (1986) demonstrated that a strong increase in magnetic mineral concentrations at Heathwaite Moss, one of several ombrotrophic bogs in South Cumbria, round the head of Morecambe Bay began shortly after the end of hemp (Cannabis) cultivation around ad 1800. Using a more comprehensive approach to chronology involving ²¹⁰Pb, ¹³⁷Cs, ²⁴¹Am, wiggle-matched ¹⁴C and a range of pollen markers, Gedye (1998) produced 12 dated profiles of magnetic measurements from three peat bogs in the region. In all but one of these, the first stage of SIRM increase occurs before ad 1800 and, in five cases, before ad 1550 (Figure 5).

Figure 5.

Helsington Moss, South Cumbria; a dried out ‘rampart’ of peat. Core HM1. (a) Mapping of the ¹⁴C ‘dates’ onto the dendrocalibration curve and also notes the depth at which the steep decline in Cannabis pollen indicates a date close to ad 1800. (b) The changing SIRM values against depth, with dates added. The decline in SIRM values above 15 cm reflects dilution by increasingly undecomposed peat rather than changes in atmospheric deposition. The unreversed IRM% values at −100 mT and −300 mT show no change through the period of increasing SIRM values.

Figure 6 shows the results from two cores taken more recently from the region. The cores were selected to represent a contrast in moisture status, with the peat at the marginal site dried out. The SIRM values are over an order of magnitude higher in the latter. There is also a contrast in the reverse field percentage values with the marginal site showing no change in unreversed hard IRM% and the wettest site showing a strong peak. This increase in the relative importance of the likely haematite contribution to the SIRM is a common feature of many of the magnetic profiles from peat (Figures 7, 8 and 10) and its significance is considered in the ‘Discussion’.

Figure 6.

Fishhouse Moss, South Cumbria, UK: profiles of SIRM and unreversed IRM% values (at −100 mT). Note the contrast in peak SIRM values and in the unreversed IRM% curves.

Figure 7.

Ellergower Moss, Galloway, SW Scotland. Core EM4. SIRM values are shown against depth, also the unreversed IRM% values at −100 mT. The dates added are transferred from the chronology shown in Figure 2. Note the major increase in unreversed IRM% above 25 cm.

Figure 8.

Ringinglow Bog, S Yorkshire, N England (Jones, 1985). HIRM plotted against depth. The slight increase at 36 cm coincides with a major increase in unreversed IRM%.

Galloway, SW Scotland. Of the 12 cores from Ellergower Moss included in the SWAP programme and summarised in Clymo et al. (1990), the results from well-dated core EM 4 (see Figure 2) are used to illustrate the record (Figure 7). Although the main increase in SIRM lies at 20 cm, above the ad 1835 level, there are indications of slight increase at 27 cm, just above the increase in the percentage representation of HIRM at 28 cm. The ²¹⁰Pb profile indicates a date close to ad 1800.

Bloak Moss, eastern lowland Scotland. Three cores from this site (Richardson, 1986) allow direct comparison between the magnetic record and dated pollen horizons. In each case, magnetic concentrations begin to increase just before early increases in Pinus and Ulmus pollen representation reflecting late 18th- and early 19th-century plantation.

Ringinglow Bog, N England, lies on the southern edge of Sheffield which, along with industrial centres in N Derbyshire, was a major centre of iron and steel based industry during the Industrial Revolution. Coke fired iron smelting in the region dates back to the mid-18th century and the industry increased to the point where, by 1856, there were 135 steel-making firms in Sheffield. Figure 8 illustrates the record from Ringinglow Bog. The first increase in SIRM is preceded by an increase in the percentage contribution of hard remanence (haematite) to the total SIRM, typical of all five of the cores analysed (Jones, 1985). Using a combination of ²¹⁰Pb, pollen and ¹⁴C analyses, Jones dates the base of the HIRM feature to 1830–1850 and the main increase in SIRM to 1900–1920.

Peat profiles from Finland. Of the four sites studied by Oldfield et al. (1981), the most detailed record comes from Karpansuo bog. Four moss-increment dated profiles present a generally consistent picture with indications of very slight increases in deposition around ad 1860, followed by steady increases after 1900. The record from Kunonniemensuo, a site further east in Finnish Karelia provides a less detailed but parallel record. Shorter profiles from other sites show varied records depending on proximity to industrial sources. The Harpar Lilltrask site lies only 7.8 km north-west of the steel works at Koverhar which began iron production in 1961 and steel production in 1971. The magnetic record appears to reflect this industrial development very clearly.

Additional profiles from Finland, all dated by moss-increment counting are reported in Jones (1986). At Kaurastensuo Bog in C Finland, in the four cores with good chronology and in which both an initial increase and a subsequent more rapid increase to peak values can be recognised, dates for the earlier feature range from 1840 to 1900 and, for the later feature, from 1900 to 1940. Jones (1986) also records the results of analyses of six dated peat profiles from sites close to and north of the Arctic Circle. In those cores for which a clear change in magnetic accumulation rates can be detected, the dates range from 1920 to 1960. In six of the eight cores analysed, there is a marked increase from c. 1950 onwards. In Norway, Jones (1986) also analysed a dated profile from 36 km south of Mo-I-Rana, the site of a major iron ore complex. Here, the SIRM values begin to increase around 1860 and rise steeply from 1940 onwards. In Denmark, Jones (1986) also obtained data from a ²¹⁰Pb dated core from Draved Moss. Here, there are slight increases in SIRM between c. 1810 and 1890, but the first clear rise in values is around 1920 with a much steeper increase from 1950 onwards.

Regent Street Bog, Fredricton, New Brunswick, Canada, yielded two cores which were analysed both for magnetic mineral deposition and a range of heavy metals (Tolonen and Oldfield, 1986). Figure 9 shows some of the data expressed as accumulation rates per year, alongside the dates of the main stages in industrial development in the area. The record shows magnetic concentrations increasing just before ad 1800, after the establishment of lumber mills. The second, larger increase is around the time of establishment of a foundry and brickworks. Closing of the foundry around 1930 led to a steep decline in magnetic deposition.

Figure 9.

Regent Street Bog, Fredricton, New Brunswick. Canada. SIRM values plotted as accumulation rates against a timescale derived from moss-increment counting. Dated events in the industrial history of the local area are added (Tolonen and Oldfield, 1986).

An additional N American site, Ely Lake Bog within the Mesabi Iron Range, Minnesota, was studied by Richardson (1986). The characteristic increase in HIRM% coincides with the increase of Ambrosia pollen representation, dated to the same time as the start of iron mining in ad 1888. SIRM values increase from this time onwards (Figure 10). Figure 1 includes SEM images of ash spheres from the surface and from the depth of maximum HIRM% values. The latter images show clear signs of dissolution.

Figure 10.

Ely lake Bog, Minnesota, USA. Pollen and IRM data from two cores. In both, the increase in IRM_850 m coincides with the first increase in Ambrosia pollen representation, dated to ad 1890 (Richardson, 1986). It is also marked by a strong relative increase in hard remanence unreversed IRM%. The SEM images in Figure 1 come from the upper, hummock profile.

Adjacent lake sediment and peat records

The record from Llyn Hir, a small lake in the uplands of Central Wales may be compared directly with that from nearby peat profiles taken from the lake catchment (Fritz et al., 1989). Both the lake sediment profiles and the record from the adjacent peat bog show an early stage in inferred atmospheric deposition marked by higher unreversed IRM% values. In the case of the peat record, the SIRM values are much higher in the more aerated ‘hummock’ core, relative to the more persistently water-logged ‘pool’ core.

Discussion

Background values

Since, by definition, ombrotrophic peats lack any input of magnetic minerals other than those resulting from atmospheric deposition, the question of background values does not arise, although in some peat profiles, there are tiny traces of a ferrimagnetic sulphide mineral, greigite, formed in situ by the transformation of non-magnetic iron compounds (Oldfield, 2012). The concentrations are so low that they do not significantly affect the record of atmospheric deposition.

In the case of lake sediment records, where the signs of atmospheric deposition have to be distinguished from magnetic minerals arising from other sources, problems arise both from the presence of eroded material from the surrounding land surfaces and from magnetic minerals formed in the sediments themselves. These latter include magnetite ‘chains’ formed by bacteria, as well as greigite. Only where these various types of magnetic mineral sources are negligible or at the very least low and constant is a clear record of atmospheric deposition decipherable.

Problems of depths and accumulation rates

The ideal record is one where the sequence of dates in an analysed profile is sufficiently detailed to permit the calculation of mineral deposition rates between reliable dates (Figures 3, 4b and 9). The next best alternative is where the magnetic measurements, as indicators simply of changing concentrations, can be plotted against a timescale or isolated dates (Figures 4a, 5 –8 and 10). In this case, two major variables control the shape of the plot – the rate of deposition/preservation of the magnetic minerals and the rate of accumulation of the sediment or peat. In the case of peat especially, this can lead to changes strongly controlled by a steep decomposition gradient across the transition from undecomposed peat in the upper, acrotelm and much more decomposed peat in the catotelm, below within which the rate of accumulation is often around 10% of that for the fresh peat nearer the surface. This has the effect of diluting concentrations near the surface relative to those at greater depths. Thus, although peat profiles may record the start of deposition, the changes above this may not represent stages in deposition history unless they can be converted into deposition rates.

Links to local industrial history

Figure 9 shows the only example used here where the deposition history has been reconstructed and dated in sufficient detail at a site close to a location for which the industrial history is well known. In this and in examples from Finland (Oldfield et al., 1981), the records come from ombrotrophic bogs dated by annual moss-increment counting in the upper layers. In each case, they appear to reflect the local, industrial history in a remarkably accurate manner. Most of the other records are dated with less precision and accuracy, lack close links to nearby industry or are the result of industrial changes for which detailed records are not available. For many of these, the key question concerns the evidence for survival or progressive degradation of the magnetic record.

Questions of survival

The first essential assumption needed in interpreting the results is that the magnetic measurements, used as surrogates for concentrations of the actual minerals, are reliable for the present purposes of establishing the beginning and main trends in magnetic mineral, hence IAS deposition. Williams (1992) has shown that dissolution of magnetic minerals occurs alongside the decomposition of organic matter as fresh bog vegetation is converted to peat, especially across the acrotelm/catotelm transition noted above.

One of the most frequent characteristics of the peat records is an increase in unreversed IRM% roughly coinciding with the earliest stages of SIRM increase. This feature is illustrated in Figures 6 –8 and 10. It reflects an increase in the importance of haematite relative to magnetite during the earliest stages of industrial deposition. As Figure 6 shows, there is a gradient linking the strength of the feature to the wettest site. The profile from Helsington Moss shown in Figure 5 comes from a dried out, residual rampart of peat surrounded by lower, drained, agricultural land from which peat has been removed for fuel. As in the driest site shown in Figure 6, there is no inflexion in the unreversed IRM% curve. The SEM images shown in Figure 1 show strong evidence for dissolution of the fly-ash spherules at the base of the ‘kink’ in unreversed IRM% values at Ely lake Bog (Figure 10). Taken together, all these lines of evidence point to a likely link between the dissolution process identified by Williams (1992) and the greater relative importance of haematite in the earliest stages of fly-ash deposition. Rothwell and Lindsay’s (2007) evidence from Pennine peats also points to a close link between survival of the magnetic record and the degree to which peat has dried out along the edge of drainage gullys. Hubbard et al. (1984) note that whereas magnetite forms during the coal combustion process and is characteristic of the outer rim of the spheres produced, haematite is more likely to be a residual survivor from the original mineral content of the coal. This may contribute to selective survival. Despite the effects of selective dissolution, the magnetic mineral record of atmospheric deposition in peat, although incompletely preserved in many cases, may, for several centuries in the case of the peat records from South Cumbria, retain a clear indication of the earliest stages in the deposition of industrially generated IASs. Longer term survival of the magnetic evidence cannot be confirmed. The parallels between the adjacent lake sediment and peat profiles at LLyn Hir (Fritz et al., 1989) suggest that selective dissolution may also affect acid, organic lake sediment sequences.

Dating the beginning of ‘fly-ash’ accumulation

From the wealth of data summarised above and in Table 1, it is clear that no single synchronous horizon common to all sites can be defined using magnetic records of fly-ash deposition. Where sufficiently detailed comparisons can be made between the chronology of magnetic deposition and local industrial history, the correspondence is quite close and convincing, but industrial histories vary between sites and regions. The earliest clearly documented and reliably dated increases in deposition of inorganic fly-ash come from NW England and just predate the mid-16th century. This roughly corresponds with the time that Fischer-Kowalski, Krausmann and Pallua (2014) favour as the socio-metabolic onset of the Anthropocene. Limited chronological data from, for example, Ringinglow Bog in the industrial heartland of northern England suggest that at such sites, the onset of magnetic accumulation corresponds quite closely to the early industrial date favoured by Crutzen and Stoermer (2000), but questions regarding the long-term survival and detection of the onset of magnetic deposition in peat remain to be resolved. In sites more remote from early industrial regions, the record of detectable deposition begins any time between the beginning and middle of the 20th century, and many sites show a major increase in magnetic deposition roughly coinciding with the mid-20th century ‘Great Acceleration’ in energy use documented by Steffen et al. (2007). The pattern revealed here is therefore more complex than that proposed by Snowball et al. (2013). The results as whole confirm the value of magnetic measurements as providing a temporal and spatial record of a dominant energy source and stage in industrialisation. Their record is complementary to that obtainable from a range of other industrial markers, for example, heavy metals, SCPs, PAHs and pesticide derivatives.

Table 1.

Summary of the dates when magnetic records suggest the first deposition of inorganic fly-ash in each area and, where detectable, any subsequent steeper increase.

	First detectable increase	Subsequent steeper increase^a
Lake sediments
Scottish Highlands	Between 1900 and 1950	~1950
English Lake District	Pre-1900	~1910
Upland Wales	Pre-1900	~1910
SW Sweden	~1900	1960
Adirondack Mts, NE USA	Steady increase from ~1890 onwards
Ombrotrophic peat
Morecambe Bay, NW England^b	Initial increases mostly pre-1800 and often pre-1550
Galloway, SW Scotland^b	~1800 with some acceleration around 1835
Eastern lowlands Scotland^b	~1800 or slightly earlier
Ringinglow Bog, N England^b	~1835 with steep increase around 1900–1920
S Finland	~1860	~1860
SE Finland (near Koverhar steel)		1961–1971
Central Finland	1840–1900	post-1900
N Finland	post-1920 with steep increases ~1950 in most cases
N Norway (near Mo-I-Rana iron)	1860	1940
Fredricton, New Brunswick, Canada	Just pre-1800	1850–1880
Ely Lake Bog, Minnesota, USA	~1888 followed by steady increase

Wherever possible the start of rapid acceleration in deposition is dated.

In these cases, undated changes in peat accumulation/preservation make it difficult to date any second change in magnetic accumulation.

The signatures derived from peats and lake sediments are essentially regional, save perhaps in the case of the results from the most northerly Scandinavian sites (Table 1). Early attempts to compile a magnetic record from ice cores (Sahota et al., 1996) failed to detect any signal above the noise level of the instrumentation then available, save in the case of Himalayan samples. More recent research in both Greenland and Antarctica (Lanci and Delmonte, 2013; Lanci et al., 2001, 2008, 2012) confirm that current instrumentation can detect and characterise effectively the magnetic properties of ice both in Greenland and Antarctica, but the thrust of the work so far has been towards documenting dust sources characteristic of glacial and interglacial periods. These results together with those summarised in the present account and the records of heavy metal deposition retained in Greenland ice (Candelone et al., 1995) suggest that Greenland ice cores spanning the industrial period will contain, at the very least, a northern hemispheric magnetic record of 20th century fly-ash deposition.

Conclusion

Depending on the site and its proximity to sources of heavy industrial activity, the date of onset of fly-ash deposition can vary from the 16th to the mid-20th century. Thus, seeking a so-called ‘golden spike’ with which to define the onset of the Anthropocene by using the deposition record of IASs (fly-ash) in peat and lake sediments poses a complex mixture of problems and opportunities. Selection of an appropriate site to provide a record capable of defining the onset of the Anthropocene first requires a choice between the early stages of the Industrial Revolution around ad 1800 and the start of the Great Acceleration around ad 1950. The record preserved in lake sediments and peats appears to be capable of defining either, depending on the site and type of environmental archive chosen. That said, questions remain in the case of peat-based records as to the long-term survival of the magnetic minerals used here as proxies for fly-ash deposition. This calls for further geochemical research coupled with magnetic measurements, scanning electron and more detailed chronologies of deposition history in relation to the documented history of industrial activity in the region selected. Future priorities could also include extension of this approach more widely, in Asia, for example, as well as analysis of recent ice cores.

Footnotes

Acknowledgements

FO’s contribution to this paper is presented to John Birks with thanks for over 50 years of friendship and support. We thank Suzanne Yee for producing the diagrams.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

Anderson

Battarbee

Appleby

. (1986) Palaeolimnological Evidence for the Acidification of Loch Fleet (PRU research papers, issue 17). London: Palaeoecology Research Unit, University College London.

Appleby

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

Birks

Last

(eds) Tracking Environmental Change Using Lake Sediments, Volume 1: Basin Analysis, Coring and Chronological Techniques. Berlin: Springer, pp. 171–203. DOI: 10.1016/S0341-8162(78) 80002–2.

Appleby

(2008) Three decades of dating recent sediments by fallout radionuclides: A review. The Holocene 18: 83–93.

Appleby

Shotyk

Frankhauser

(1997) 210Pb age dating of three peat cores in the Jura Mountains, Switzerland. Water, Air and Soil Pollution 100: 223–231. DOI: 10.1023/A:1018380922280.

Awty

Philips

(1979–1980) The Cumbrian bloomery forge in the 17^th century. Transactions of the Newcomen Society 51: 25–40.

Battarbee

Anderson

Appleby

. (1988) Lake Acidification in the United Kingdom. London: ENSIS Ltd.

Battarbee

Lamb

Bennett

. (2015) John Birks: Pioneer in quantitative palaeoecology. The Holocene 25(1). DOI: 10.1177/0959683614556390.

Battarbee

Mason

Renberg

. (1990) Palaeolimnology and Lake Acidification (Philosophical Transactions of the Royal Society of London Series B – Biological Sciences 327). London: Royal Society, 445 pp.

Bhattacharjee

Mandal

Roy

. (2013) Physical characteristics of fly ashes from three thermal power plants in West Bengal, India: A comparative study. International Journal of ChemTech Research 5: 836–843.

10.

Candelone

Hong

Pellone

. (1995) Post-industrial revolution changes in large-scale atmospheric pollution of the northern Hemisphere by heavy metals as documented in central Greenland by snow and ice. Journal of Geophysical Research: Atmospheres 100: 16605–16616.

11.

Charles

Whitehead

(1986) The PIRLA project: Paleocological investigations of recent lake acidification. Hydrobiologia 143: 13–20.

12.

Clymo

(1984) Sphagnum-dominated peat-bog: A naturally acid ecosystem. Proceedings of the Royal Society of London Series A 305: 487–499.

13.

Clymo

Oldfield

Appleby

Pearson

Ratnesar

Richardson

(1990) The record of atmospheric deposition on a rainwater-dependent peatland. Philosophical Transactionsof the Royal Society of London Series B – Biological Sciences 327: 331–338.

14.

Crutzen

Stoermer

(2000) The ‘Anthropocene’. IGBP Newsletter no. 41, pp. 17–18. Available at: http://www.igbp.net/download/18.316f18321323470177580001401/1376383088452/NL41.pdf

15.

Fischer-Kowalski

Krausmann and Pallua

(2014) A socio-metabolic reading of the Anthropocene: Modes of subsistence, population size and human impact on Earth. The Anthropocene Review 1: 8–33.

16.

Fritz

Stevenson

Patrick

. (1989) Paleolimnological evidence for the recent acidification of Llyn Hir, Dyfed, Wales. Journal of Paleolimnology 2: 245–262.

17.

Gedye

(1998) Mass balance in recent peats. Unpublished PhD Dissertation, University of Liverpool.

18.

Houseman

(1800) A Topographical Description of Cumberland and Westmorland, Lancashire and a Part of the West Riding of Yorkshire. Carlisle: F. Jollie.

19.

Hubbard

McGill

Dhir

. (1984) Clay and pyrite transformations during ignition of pulverized coal. Mineralogical Magazine 48: 251–256.

20.

Jones

(1985) Magnetic minerals and heavy metals in ombrotrophic peat. Unpublished PhD Dissertation, University of Liverpool.

21.

Jones

MDH

(1986) The magnetic deposition record in some Scandinavian peat profiles. Unpublished PhD Dissertation, University of Liverpool.

22.

Lanci

Delmonte

(2013) Magnetic properties of aerosol dust in peripheral and inner Antarctic ice cores as a proxy for dust provenance. Global and Planetary Change. Available at: https://dx-doi-org.web.bisu.edu.cn/10.1016/j.gloplacha.2013.05.003

23.

Lanci

Delmonte

Kent

. (2012) Magnetization of polar ice: A measurement of terrestrial dust and extra-terrestrial fallout. Quaternary Science Reviews 33: 20–31.

24.

Lanci

Delmonte

Maggi

. (2008) Ice magnetization in the EPICA-Dome C ice core: Implication for dust sources during glacial and interglacial periods. Journal of Geophysical Research. Epub ahead of print 19 July. DOI: 10.1029/2007JD009678.

25.

Lanci

Kent

Biscaye

. (2001) Isothermal remanent magnetization of Greenland ice: Preliminary results. Geophysical Research Letters 28: 1639–1642.

26.

Liu

Roberts

Larrasoana

. (2012) Environmental magnetism: Principles and applications. Reviews of Geophysics. Epub ahead of print 6 November. DOI: 10.1029/2012RG000393.

27.

Oldfield

(1990) Magnetic measurements of recent sediments from Big Moose Lake, Adirondack Mountains, N.Y., USA. Journal of Paleolimnology 4: 93–102.

28.

Oldfield

(1991) Environmental magnetism – A personal perspective. Quaternary Science Reviews 10: 73–85.

29.

Oldfield

(1999) The rock magnetic identification of magnetic mineral and magnetic grain size assemblages. In: Walden

Oldfield

Smith

(eds) Environmental Magnetism: A Practical Guide (Technical Guide). London: Quaternary Research Association, pp. 98–112.

30.

Oldfield

(2012) Mud and magnetism: Records of late Pleistocene and Holocene environmental change recorded by magnetic measurements. Journal of Paleolimnology. Epub ahead of print 1 March. DOI: 10.1007/s10933-012-9648-8.

31.

Oldfield

Richardson

(1990) Lake sediment magnetism and atmospheric deposition. Philosophical Transactions of the Royal Society of London Series B – Biological Sciences 327: 325–330.

32.

Oldfield

Statham

(1963) Pollen analytical data from Urswick Tarn and Ellerside Moss, North Lancashire. New Phytologist 62: 53–66.

33.

Oldfield

Richardson

Appleby

(1995) Radiometric dating (Pb-210, Cs-137, Am-241) of recent ombrotrophic peat accumulation and evidence for changes in mass-balance. The Holocene 5: 141–148.

34.

Oldfield

Thompson

Barber

(1978) Changing atmospheric fallout of magnetic particles recorded in recent ombrotrophic peat sections. Science 199: 679–680.

35.

Oldfield

Tolonen

Thompson

(1981) History of particulate atmospheric-pollution from magnetic measurements in dated Finnish peat profiles. Ambio 10: 185–188.

36.

Pakarinen

Tolonen

(1977) On the growth rate and dating of surface peat. Suo 28: 19–24.

37.

Patrick

Flower

Appleby

. (1989) Palaeoecological Evaluation of the Recent Acidification of Lochnagar, Scotland (PRU research papers, issue 34). London: Palaeoecology Research Unit, University College London.

38.

Renberg

Brodin

Cronberg

. (1990) Recent acidification and biological changes in Lilla-Oresjon, Southwest Sweden, and the relation to atmospheric-pollution and land-use history. Philosophical Transactions of the Royal Society of London Series B – Biological Sciences 327: 391–396.

39.

Richardson

(1986) The mineral magnetic record in recent ombrotrophic peat synchronized by fine resolution pollen analysis. Physics of the Earth and Planetary Interiors 42: 48–56.

40.

Rose

(1990) A method for the selective removal of inorganic ash particles from lake sediments. Journal of Paleolimnology 4: 61–67.

41.

Rose

(1994) Characterization of carbonaceous particles from lake sediments. Hydrobiologia 274: 127–132.

42.

Rothwell

Lindsay

(2007) Mapping contemporary magnetic mineral concentrations in peat soils using fine resolution digital terrain data. Catena 70: 465–474.

43.

Sahota

JTS

Mayewski

Oldfield

. (1996) Magnetic measurements of Greenland and Himalayan ice-core samples. The Holocene 6: 477–480.

44.

Sharonova

Anshits

Oruzhenkov

. (2003) Composition and morphology on magnetic microspheres in power-plant fly ash of coal from the Ekibastuz and Kuznetsk basins. Chemistry for Sustainable Development 11: 639–648.

45.

Snowball

Hounslow

Nilsson

(2013) Geomagnetic and mineral magnetic characterization of the Anthropocene. Geological Society: Special Publications. Epub ahead of print 19 December. DOI: 10.1144/SP395.13.

46.

Steffen

Crutzen

McNeill

(2007) The Anthropocene: Are humans now overwhelming the great forces of nature. Ambio 36: 614–621.

47.

Thompson

Oldfield

(1986) Environmental Magnetism. London: Allen & Unwin, 227 pp.

48.

Thompson

Bloemendal

Dearing

. (1980) Environmental applications of magnetic measurements. Science 207: 481–486.

49.

Tishmack

Burns

(2004) The chemistry and mineralogy of coal and coal combustion products. In: Gieré

Stille

(eds) Energy, Waste and the Environment: A Geochemical Perspective. London: Geological Society of London (Special Publications 236), pp. 223–246.

50.

Tolonen

Oldfield

(1986) The record of magnetic-mineral and heavy-metal deposition at Regent-Street Bog, Fredricton, New-Brunswick, Canada. Physics of the Earth and Planetary Interiors 42: 57–66.

51.

Van der Post

Oldfield

Haworth

. (1997) A record of accelerated erosion in the recent sediments of Blelham Tarn in the English Lake district. Journal of Paleolimnology 18: 103–120.

52.

Williams

(1992) Evidence for the dissolution of magnetite in recent Scottish peats. Quaternary Research 37: 171–182.

53.

Winburn

Lerach

McCarthy

. (2000) Quantification of ferrite spinel and haematite in magnetically enriched fractions. Advances in X-ray Analysis (JCPDS-International Centre for Diffraction Data 2000) 43: 350–255.