3
The Use of Tephra in Linking European Sequences of Terrestrial, Lacustrine, Marine and Ice Palaeoenvironmental Records
Abstract
Tephra provides a valuable additional chronological and stratigraphical tool for the palaeoecologist and environmental archaeologist. It is proving particularly useful in north-west Europe where well-dated Icelandic tephras provide good chronological control. Improved concentration and detection methods now permit the use of these valuable chronological markers in a wide range of Quaternary deposits. Chronologies established using clean ombrogenous peats can now be applied to lacustrine and marine sediments of low organic content. Study of tephra layers in well-dated ice cores will lead to much improved dating for tephras from prehistoric times.
Introduction
Tephra is the air-fall component of ejecta from a volcano. Strictly speaking, it includes glassy material and crystalline material from the magma and also any broken volcanic cone rock that is carried into the atmosphere. The glassy component is liquid magma that cools rapidly as it is shot into the atmosphere and forms a glass without crystallising. The quantities produced can be huge–geologists describe eruptions in terms of the number of cubic kilometres of ash produced–and it is shot many kilometres into the atmosphere. Thus, it is not really surprising that it can travel long distances from its source.
Because eruptions are generally short-lived, tephra layers provide very precise time markers in sediments and deposits. They have the potential to provide an additional time control to archaeologists to supplement radiocarbon dating. Tephras (in this case visible layers) have been used extensively in the study of the early settlement of Iceland and the Faroe islands (Dugmore and Newton 1998; Wastegaard et al. 2001). Using the methods described here to identify much smaller amounts of tephra, the technique can now be extended to Ireland and Britain and other parts of north-west Europe.
Geographical Distribution and Temporal Range
Tephras may be found in many areas of the world, and in fact there is probably no part of the globe where tephras are unlikely, albeit in low concentrations. It has been claimed that tephra from an eruption in AD 1259 has been found in both the Arctic and the Antarctic ice sheets (Ram and Gayley 1991; Palais et al. 1992). Their use in stratigraphic studies and as a dating tool is well developed in north-west Europe using Icelandic tephra (Dugmore et al. 1995; Pilcher et al. 1996; Hall and Pilcher 2002; van den Bogaard and Schmincke 2002), in North Island New Zealand (Froggatt and Lowe 1990) and in northern USA (Beget and Keskinen 1991; Beget et al. 1991). Similar work is in progress in many other areas such as Mexico, the Antarctic Peninsula (Bjork et al. 1991), the Falkland Islands (Hall et al. 2001) and Kamchatka (Russian Federation) (Braitseva et al. 1993) etc. Most of these studies concentrate on the Late Quaternary and Holocene.
Practicalities
Most tephra studies so far have been on continuous sequences such as peat bogs and lake sediments. Samples for archaeological dating would ideally use similar sequences–ditch fill deposits, for example. Where single samples from sealed archaeological contexts are investigated the greatest care must be taken to ensure stratigraphic integrity. A contaminated tephra sample is just as useless as a contaminated radiocarbon sample! It is difficult to give guidelines for sample size that are generally applicable. For peat and lake sediments a sample of about 5 cm2 surface area would be used. Because extraction from mineral soils is more complex, twice this surface area might be a good target when sampling these types of soils and sediments.
How do we Detect Tephra in a Sediment Core or Sample?
Tephra particles are typically less than 100 microns (commonly 20–60µ) in size and are glass. The ideal medium for detection is in a highly organic matrix such as peat. All we need to do is burn the organic material, dissolve the peat ash in dilute hydrochloric acid (10% HCl) and wash the tephra and other minerals clean. The clean material can then be mounted in resin and examined under a microscope at about x100 magnification. This process is described by Pilcher and Hall (1992). Where the matrix has more than about 1% mineral content the separation of the tephra is more difficult. Sieving between mesh sizes of 26 and 75µ will retain most of the tephra and will remove all the fine clay and silt particles. If this is not adequate, then we resort to separation using a heavy liquid (Lowe and Turney 1997). Sodium polytungstate dissolved in water provides densities up to about 3.2. As the glass of the tephra is relatively light it can be floated off most other minerals using a density of 2.5. Density separations are quite time consuming and the heavy liquid is very expensive and must be recovered and re-used.
Preparation of samples from lacustrine sediments may need additional treatment to remove diatoms. These have a similar density to the tephra and follow the tephra in the heavy liquid separation. Prolonged treatment (2–4 hours) with warm potassium hydroxide at a concentration less than 10% will dissolve most diatoms and has been demonstrated not to affect the tephra chemistry.
Preparation for Geochemistry
The simple preparation by burning is not suitable when chemical analysis is required as the burning changes the nature of the glass chemistry. In this case the organic matter is dissolved in a chemical oxidising mixture of concentrated sulphuric acid and concentrated nitric acid as described in Hall and Pilcher (2002). The tephra recovered from the acid treatment is washed on a 26µ sieve mesh. At this point microscopic assessment will determine whether either or both of the diatom treatment or the heavy liquid separation, described above, will be needed. Finally, the clean tephra is dried onto a ground glass slide. The tephra is covered in a layer of epoxy resin (Araldite). When this has set the preparation is ground using 12µ alumina on a glass plate until the tephra is positioned just at the surface. The slide is then polished using diamond or alumina polishing powders until the surface is smooth, and flat surfaces of the tephra exposed. The difference between a perfect slide and one on which all the tephra has been ground or polished away is only a few microns!
Recognition of Tephra
There is a wide range of tephra morphologies and colours. While these can be some guide to origins they are not capable of providing a definitive identification. Examples (a) to (d) in Figure 3.1 show some of the different morphologies and colours seen in Icelandic tephras. Van den Bogaard and Schmincke (2002) highlight the diagnostic value of crystalline inclusions in some tephra particles. One of the greatest problems facing those starting tephra studies is that of distinguishing tephra from biogenic silica. Biogenic silica comes in many forms and some will survive the alkaline treatment mentioned above. Morphologically distinct forms such as diatoms and many plant phytoliths are not a problem, but other types that have formed inside higher plant cells can often mimic the shape of large tephra bubbles. Example (e) in Figure 3.1 shows various biogenic silica bodies that have been found in tephra preparations. With experience, the biogenic silica can be distinguished by its different refractive index and often by a very fine sculpturing of the surface that can be seen at high magnification of the light microscope. The effect of the refractive index depends on the refractive index of the mounting medium. With the histomount medium used by the author, the biogenic silica tends to have a bolder outline than the tephra.
The results of microscope identification can be expressed graphically alongside other stratigraphic information. Figure 3.2 shows a stratigraphic sequence from Garry Bog in Co. Antrim, Northern Ireland. The peat monolith was sampled at 1 cm intervals and prepared by the burning technique. Selected samples were prepared by wet chemistry and the tephras identified by microprobe analysis (see below).
Figure 3.1: Different morphologies and colours seen in some Icelandic historic tephras. a) Hekla AD 1510, b) Oraefajokull AD 1362, c) Hekla AD 1104 (‘Hekla 1’), d) Unknown origin c. AD 1250 and e) various biogenic silica bodies that have been found in tephra preparations.
Chemical Fingerprinting of Tephras
Volcanic systems differ in their chemistry, individual volcanoes may also differ and in some cases, such as in Iceland, many individual eruptions are distinct. This distinction is much clearer in some places than others. For example, in South America many eruptions share the same chemistry. In the distal tephra studies described here it is not possible to collect enough tephra for X-ray fluorescence (XRF) analysis, so we use electron microprobe analysis of single shards of tephra. Because the differences between eruptions are quite subtle it is normal to use the more sensitive wavelength dispersive microprobe analysis rather than the simpler and more commonly available energy dispersive analysis. The wavelength dispersive analysis requires a flat surface, hence the need to mount the tephra in resin and grind and polish the preparation. This has the added advantage that it exposes a fresh surface of the tephra. It is clear from the pitted surface sometimes seen in prehistoric tephras that some degradation of the surface occurs with age (Dugmore et al. 1992).
Figure 3.2: Stratigraphic tephra sequence from Garry Bog in Co. Antrim, Northern Ireland.
Both sodium and potassium are important diagnostic elements in volcanic glass. However, it is a feature of glasses that the sodium and to a lesser extent the potassium in the glass is driven out by the heat from the electron beam during microprobe analysis (Hunt and Hill 1993). Sodium must be analysed first to reduce this problem and it is usual to de-focus the beam to about 8µ diameter to reduce the heating. The operating conditions of the microprobe may also affect the extent of sodium loss. We operate the Belfast Jeol 733 Superprobe at 15 kv and 10 na beam current. It is usual to analyse somewhere between ten and 20 individual tephra fragments. For some glasses as few as five analyses may be diagnostic, but where a particular volcano has produced tephra of varying composition during the course of an eruption many analyses may be needed to obtain a representative population (Dugmore et al. 1992).
One of the main reasons for the rapidly increasing success of tephra studies in north-west Europe has been the ability to locate and chemically identify very low concentrations of tephra. Hall and Pilcher (2002) describe the identification of concentrations as low as 1–2 tephra fragments per cm3 of peat. This is achieved by using a light microscope with x–y mechanical stage to record the positions of individual tephra particles on a slide and then converting these co-ordinates to the x and y stage co-ordinates of the microprobe. This allows the microprobe to be driven directly to the tephra particle. Many modern microprobes have no optical imaging capability. This eliminates the very valuable use of polarised light to distinguish crystalline minerals from the glassy tephra. Even where there is an optical system these are never of the quality of a good research light microscope and the separation of tephra from biogenic silica using such optics stretches the capabilities of the most experienced tephra analyst!
Type Material and Matching
Type material comes from profiles examined in Iceland. Many have been analysed, many more remain to be done. Tephra stratigraphies close to the source volcanoes are extremely complex, with much re-working and wind- and water-generated movement of tephras. Much of the work on type material has been carried out on clearly stratified sequences in the peat bogs of northern Iceland (Larsen and Thorarinsson 1997; Larsen et al. 1999).
Work in a new area benefits from the construction of a detailed tephra stratigraphy using the best possible ombrogenous peat sequence before moving on to work on more difficult series from lacustrine or near-shore marine deposits. Analyses based on microprobe analysis as described above are stored in an international tephra data bank (TEPHRABASE) and are available to all researchers at the following web address: http://www.geo.ed.ac.uk/tephra/tbasehom.html
Why is Tephra Useful?
A typical eruption lasts for days or perhaps even months, but a very short time in relation to typical archaeological timescales and to bog and lake stratigraphy. Thus, an ash layer forms a highly defined time marker. Apart from tephra, such isochrons are rare in lake sediments and almost unknown in peat bogs. In an archaeological context, one might hope to find a known-age tephra below or above a cultural context that would provide a bracketing date. Such dates can be very useful when combined, using Bayesian statistics, with a series of radiocarbon measurements. There is considerable potential for using tephra in wetland archaeology where tephras could potentially provide a link between cultural layers in the archaeological site and palaeoenvironmental contexts nearby.
Tephra for Correlation Between Cores, Between Sample Sites and Between Regions
One of the perennial problems of working in bog and lake deposits is correlation from one core to another. Magnetic susceptibility is quick and sometimes works well but with highly organic sediments will often not provide a definitive link. Tephra is ideal for this as the markers are so sharp. It is a mistake to believe that it is a quick technique, however. It is normally assumed that parallel cores taken a few metres apart on the same day are going to be comparable. If one core is used for pollen analysis and the other to provide material for AMS dating or for chemical analysis then tephra may provide an insight into critical offsets between the two cores.
Tephra as a Taphonomic Guide
This I believe to be one of the most valuable things that tephra can tell us and so far has been very little used. As the tephra fall is assumed to be at least within a single year we can make a good assessment of post-depositional processes by looking at the vertical spread of the tephra in a sediment core. As the particle size of the typical tephra layer is of the same order as that of pollen, for example, it can suggest the extent to which pollen may have been moved vertically in the sediment profile. If the tephra is spread over 5 cm, there would be little point in undertaking a pollen study at 1 cm resolution! The upper graph in Figure 3.3 shows fine resolution sampling over a tephra layer in Sphagnum peat. In this medium there appears to be very little tephra movement and 80% of the tephra is restricted to 8 mm. In contrast, the lower graph in Figure 3.3 shows sampling in an upland blanket peat where there has clearly been considerable tephra movement. Such material would not be appropriate for a high resolution pollen study, and the findings also suggest that other studies such as radiocarbon dating need to be carried out with caution.
Tephra as a Chronology
This depends, of course, on knowing when eruptions took place and then of tying a tephra fall to a particular eruption. Iceland has been settled since about 850 AD and the inhabitants have kept detailed records of eruptions since that time. From about 1100 AD, we have a well recorded chronology of calendrical accuracy (N. Ogilvie 2005, pers. comm.). Before settlement in Iceland we are dependent on dating tephra layers by independent means. Radiocarbon dating is clearly a possibility but this raises the question of would it not be just as good to date the sediment core directly rather than date tephra somewhere else then relate it to the sediment core. There are actually several good reasons for doing this and these centre on radiocarbon’s ability to provide accurate dates.
We can pick an ideal material for radiocarbon dating, such as fast-grown Sphagnum peat, and then apply the date to material such as low-organic lake sediment or marine or brackish water sediment that is unsuitable for precision radiocarbon dating. Even using high precision radiocarbon dating will not provide adequate precision in some time periods because of the nature of natural radiocarbon variations (Pilcher 1993; see Barratt and Reimer, this volume). In some time periods, a radiocarbon measurement may relate to a range of real ages–for instance, samples of several different real ages may all have the same radiocarbon date. One way round this is to use a technique called ‘wiggle matching’. A series of samples (typically five or more samples) covering the event in question are measured using the highest available radiocarbon precision. These measurements can then be compared with the ‘wiggles’ in the radiocarbon calibration curve. This technique was tested on the known-age eruption of Hekla in AD 1104. The best estimate of the date of the eruption based on the wiggle match was 1088 ± 20 BP. The application to an eruption of unknown date is illustrated in Figure 3.4, which shows the wiggle match dating of the Icelandic Hekla 4 eruption (Pilcher et al. 1996). We used a Sphagnum-rich ombrogenous peat in which the tephra was sharply defined. Five radiocarbon samples were measured with a precision of ± 18 years and the sequence matched to the calibration curve. The technique has been used on several other eruptions such as Hekla 3 (van den Bogaard and Schmincke 2002) and the mid-Holocene ‘Lairg’ tephras, dated to 4774–4677 cal. BC and 4997–4902 cal. BC (Pilcher et al. 1996). It is thus clearly easier to obtain accurate dates on individual tephras, than it is on individual horizons of interest at different sites of varying sediment and stratigraphy.
Figure 3.3: Upper graph–Fine resolution sampling over a tephra layer in a Sphagnum peat. In this medium there appears to be very little tephra movement and 80% of the tephra is restricted to 8 mm. Lower graph–Sampling in an upland blanket peat where there has clearly been considerable tephra movement.
Figure 3.4: Wiggle match dating of the Icelandic Hekla 4 eruption (Pilcher et al. 1996).
Tephra from Ice Cores
The detection of tephra from ice cores opens up the possibility of using the superior chronology of the most recent ice cores to date prehistoric tephras and thus assist in the correlation of the huge climate archive in the ice with those in terrestrial and marine sediments. In the most recent Greenland ice core–North Grip (NGRIP)–an automated chemical analysis system was implemented in the drill-site laboratory. The melted ice for analysis was filtered through micropore filters and these filters were kept for future analysis of the particulate content. We have prepared a selection of these filters for microprobe analysis by embedding slices of filter in Araldite resin rather than trying to wash particles off the filter. Successful light microscope detection of tephra was achieved and the same slides were then ground and polished for microprobe analysis.
Samples of the archived older ice cores such as GRIP are being melted and the water centrifuged to concentrate particulate matter for tephra analysis.
An Example of the Use of Tephra in a Palaeoenvironmental Investigation
Recently, a team from the University of Massachusetts at Amherst and Queen’s University of Belfast went to the Lofoten Islands off the north-west coast of Norway, just inside the Arctic Circle, to look for sediment that might give information on the past position of the Polar Front during the Holocene. Because we were working in a range of terrestrial, brackish and marine sediments we hoped to use tephra as an additional dating method. As no tephra work had been done in this area before, we selected several peat profiles to develop a local tephra stratigraphy before trying to apply this to more difficult lake sediments. One profile has been fully analysed.
The stratigraphy contains identified Icelandic tephras of AD 1362, AD 1158 and AD 1104 eruptions and also the big eruption of Hekla in 2310 BC. A more recent layer may be the AD 1510 eruption of Hekla and the profile extends back to the late glacial where the well-known Vedda tephra was detected. In all some 24 district tephra layers were isolated (Pilcher et al. 2005).
Once we know where the key dating layers are we can start to look for these in the lake cores. It is too time-consuming to process each lake sediment and brackish water core so we use AMS dating of terrestrial macrofossils to roughly define time periods of likely interest. Once tephras have been found we can use them to assess the bioturbation in the cores and can correlate from core to core and lake basin to lake basin. Eventually, we will be able to correlate our findings with other work on a transect along the Atlantic seaboard of Europe, focusing on key dates such as the 2310 BC eruption.
Problems
— It sounds easy! Identification of the tephra requires a lot of very careful microscopy. Experience with a technique such as pollen or diatom analysis is a distinct asset.
— Microprobe analysis requires access to a microprobe capable of critical wavelength dispersive analysis.
— Not all areas receive tephra. Experience with the fall-out of the Chernobyl disaster shows how patchy aerosol fall out can be.
— Not all tephras are chemically distinct. For example the AD 1947 and AD 1510 tephras from Hekla are very similar, as are those of Hekla 4 (2310 BC) and Hekla 5 (5990 BC). Normally, other stratigraphic information allows these to be separated.
References
Beget, J. and Keskinen, M. 1991. The Stampede tephra: a middle Pleistocene marker bed in glacial and aeolian deposits of central Alaska. Canadian Journal of Earth Sciences 28, 991–1002.
Beget, J., Edwards, M., Hopkins, D., Keskinen, M. and Kukla, G. 1991. Old Crow tephra found at the Palisades of the Yukon, Alaska. Quaternary Research 35, 291–7.
Bjork, S., Sandgren, P. and Zale, R. 1991. Late Holocene tephrochronology of the northern Antarctic peninsula. Quaternary Research 36, 322–8.
Bogaard, C. van den and Schmincke, H-U. 2002. Linking the North Atlantic to central Europe: a high resolution Holocene tephrochronological record from Northern Germany. Journal of Quaternary Science 17, 3–20.
Braitseva, O. A., Sulerzhitsky, L. D., Litasova, S. N., Melekestsev, I. V. and Ponomareva, V. V. 1993. Radiocarbon dating and tephrochronology in Kamchatka. Radiocarbon 35, 363–476.
Dugmore, A. J. and Newton, A. J. 1998. Holocene tephra layers in the Faroe Islands. Frodskaparrit 46, 191–205.
Dugmore, A. J., Larsen, G. and Newton, A. J. 1995. Seven isochrones in Scotland. The Holocene 5, 257–66.
Dugmore, A. J., Newton, A. J. and Sugden, D. E. 1992. Geochemical stability of fine-grained silicic Holocene tephra in Iceland and Scotland. Journal of Quaternary Science 7, 173–83.
Froggatt, P. C. and Lowe, D. J. 1990. A review of late Quaternary silicic and some other tephra formations from New Zealand: their stratigraphy, nomenclature, distribution, volume, and age. New Zealand Journal of Geology and Geophysics 33, 89–109.
Hall, V. A. and Pilcher, J. R. 2002. Late Quaternary Icelandic tephras in Ireland and Great Britain: detection, characterization and usefulness. The Holocene 12, 223–30.
Hall, V. A., Holmes, J. and Wilson, P. 2001. Holocene tephrochronological studies in the Falkland Islands, pp. 39–44 in Juvigne, E. and Raynal, J. P. (eds.), Tephras; Chronology and Archaeology (Les dossiers de l’Archeo-logis No 1). Goudet: Conseil-General de Haute Loire.
Hunt, J. B. and Hill, P. G. 1993. Tephra chemistry: a discussion of some persistent analytical problems. The Holocene 3, 271–8.
Larsen, G. and Thorarinsson, S. 1977. H4 and other acid tephra layers. Jokull 27, 29–47.
Larsen, G., Dugmore, A. J. and Newton, A. J. 1999. Geochemistry of historical age silicic tephras in Iceland. The Holocene 9, 463–71.
Lowe, J. J. and Turney, C. S. M. 1997. Vedde ash discovered in a small lake basin on the Scottish mainland. Journal of the Geological Society 154, 605–12.
Palais, J. M., Germani, M. S. and Zielinski, G. A. 1992. Inter-hemispheric transport of volcanic ash from a 1259 A.D. volcanic eruption to the Greenland and Antarctic ice sheets. Geophysical Research Letters 19, 801–4.
Pilcher, J. R. 1993. Radiocarbon dating and the palynologist: a realistic approach to precision and accuracy, pp. 23–32 in Chambers, F. M. (ed.), Climate Change and Human Impact on the Landscape. London: Chapman and Hall.
Pilcher, J. R and Hall, V. A. 1992. Towards a tephrochronology for the Holocene of the north of Ireland. The Holocene 2, 255–9.
Pilcher, J. R., Hall, V. A. and McCormac, F. G. 1995. Dates of Icelandic volcanic eruptions from tephra layers in Irish peats. The Holocene 5, 103–10.
Pilcher, J. R., Hall, V. A. and McCormac, F. G. 1996. An outline tephrochronology for the north of Ireland. Journal of Quaternary Science 11, 485–94.
Pilcher, J. R., Bradley, R. S., Francus, P. and Anderson, L. 2005. A Holocene tephra record from the Lofoten Islands, Arctic Norway. Boreas 34, 1–21.
Ram, M. and Gayley, R. I. 1991. Long-range transport of volcanic ash to the Greenland ice sheet. Nature 349, 401–4.
Wastegaard, S., Bjorck, S., Grauert, M. and Hannon, G. E. 2001. The Mjauvotn tephra and other Holocene tephra horizons from Faroe Islands: a link between the Icelandic source region, the Nordic Seas and the European continent. The Holocene 11, 101–9.