163

Journal of Paleolimnology 20: 163–173, 1998. c 1998 Kluwer Academic Publishers. Printed in the Netherlands.

Magnetic properties of recent sediments in Lake Baikal, Siberia  J. A. Dearing1 , J. F. Boyle1 , P. G. Appleby2, A. W. Mackay3 & R. J. Flower3 1

Environmental Magnetism Laboratory, Department of Geography, University of Liverpool, Liverpool L69 3BX, UK (e-mail: [email protected]) 2 Department of Applied Mathematics and Theoretical Physics, University of Liverpool, Liverpool L69 3BX. UK 3 Environmental Change Research Centre, University College London, 26 Bedford Way, London, WC1H 0AP, UK Received 23 June 1997; accepted 15 December 1997

Key words: Lake Baikal, mineral magnetism, 210 Pb, reductive diagenesis, erosion

Abstract Mineral magnetic measurements of six 210 Pb-dated surface cores from different basins of Lake Baikal, Siberia, show temporal records controlled by a range of internal and external processes. With the exception of sediments on the Academician Ridge, there is clear evidence for widespread reductive diagenesis effects on the ferrimagnetic component coupled with neo-formation of paramagnetic iron minerals. Greigite formation, bacterial magnetosome accumulation and turbidite layers may affect the properties of some sediment levels. Concentrations of canted antiferromagnetic minerals (eg. haematite) appear to increase from the 19th century onwards. These minerals are less affected by dissolution processes and probably represent detrital minerals delivered by catchment fluvial processes. The magnetic evidence for recent atmospheric pollution by fossil-fuel combustion processes is weak in all the cores, and supports the findings from studies of spherical carbonaceous particles (SCPs) and heavy metals that pollution is largely restricted to the southern basin. Correlations between recent sediments based on magnetic data may be insecure over long distances or between basins. Introduction Mineral magnetic measurements (Thompson & Oldfield, 1986) have been applied to the study of lake sediments for more than two decades. Thompson et al.’s (1975) study of magnetic susceptibility profiles from the recent sediments of Lough Neagh, Northern Ireland, was the first to demonstrate links between high concentrations of detrital ferrimagnetic minerals and catchment disturbance, but since then numerous studies have shown the use of magnetic measurements to correlate cores (e.g. Dearing, 1986) and to reconstruct histories of erosion (e.g. Thompson & Morton, 1978; Oldfield et al., 1986; Higgitt et al., 1990; Snowball & Thompson, 1990; Dearing, 1992; Foster et al., 1996) and atmospheric pollution (e.g. Oldfield & Richard This is the fifth in a series of seven papers published in this special issue dedicated to the paleolimnology of Lake Baikal. Dr. Roger Flower collected these papers.

son, 1990; Williams, 1992). These different aspects of sediment study are all important to the present Lake Baikal research programme (Flower, 1998). Large distances require core correlation techniques to improve the confidence of accepting regional rather than local interpretation of sediment records and to help identify anomalous sedimentation rates and patterns caused by turbidites (Lees et al., 1998b). Cultural activities which are most likely to have altered the recent Lake Baikal ecosystem are accelerated erosion from agricultural and deforested landscapes and atmospheric pollution from various industrial point sources. Mineral magnetic measurements also give information about climatic change over glacial-interglacial timescales. Peck et al. (1994) have reconstructed palaeoclimates for the past 250 kyr at Lake Baikal from slowly accumulating sediments sampled on the Academician Ridge (Figure 1). Interglacial periods are characterised by lower density diatomaceous sediment with small concentrations of

164

Figure 1. Lake Baikal, site and location, showing coring locations of dated cores used in this study and other cores.

165 low coercivity minerals whereas glacial periods show higher clay contents and concentrations of all magnetic minerals especially high coercivity minerals, such as haematite, which are attributed to increased aeolian inputs during arid phases. Validation of such proposed links between magnetic mineralogy and environmental processes is often easier to achieve within recent sediments where environmental conditions and processes are more easily identified and more sharply defined. This is particularly important in the case of Lake Baikal where exceptionally deep water and slow accumulation rates could provide the conditions for reductive diagenesis of magnetic minerals (Anderson & Rippey, 1988; Snowball, 1993), authigenic formation of ferrimagnetic iron sulphides (Hilton & Lishman, 1985; Hilton et al., 1986; Snowball & Thompson, 1988) and magnetotactic bacteria (Oldfield, 1994; Snowball, 1994) to play major roles in controlling the mineral magnetic characteristics of sediment. Therefore the main aim of this paper is to evaluate alternative controls on sediment magnetism in Lake Baikal.

Site and sampling Cores of recent sediment were sampled between 1992 and 1994 from a large number of locations using a purpose-designed box corer (Flower et al., 1995). The present study presents results from the six 210 Pb-dated cores (Figure 1) extending to sediment depths ranging between 12 and 40 cm taken from the southern basin (BAIK 6 and 38), the distal part of the Selenga delta (BAIK 19), the central basin (BAIK 22), the Academician Ridge (BAIK 25) and the northern basin (BAIK 29), which are complemented by results from analyses of geochemistry (Boyle et al., 1998), microfossils (Flower et al., 1998) and spherical carbonaceous particles (Rose et al., 1998). Preliminary results from BAIK 6 have been published previously (Flower et al., 1995). All the cores comprise diatomaceous muds with a colour change from grey to brown within the depth range 4–12 cm assumed to indicate the contemporary oxic-anoxic boundary. Cores BAIK 19, 22 and 29 show evidence for the presence of discreet fine-grained and diatom-rich turbidite layers.

Magnetic measurements and dating The sediments were extruded at either 1 cm or 0.5 cm intervals and oven-dried (40  C). Each sample of oven-

dried material was gently ground from which an aliquot of  0:3 g was taken for measurement in a Molspin vibrating sample magnetometer (VSM). The VSM measures the magnetisation of the sample through a pre-set sequence of field strengths ranging from zero to 1000 mT. Isothermal remanence is measured in zero field (0:1 mT) after each magnetisation step. The VSM is calibrated to a palladium standard (moment at 1000 mT = 31.23 mA m2) and field strengths are repeatable to 0:1 mT at zero field and to 1 mT at high fields (> 100 mT). The precision of the equipment as shown by the coefficient of variation for repeated measurements of the calibration sample is  0:1%. The detection limit is 0.01 mA m2 and the nylon sample holders have a diamagnetic moment of  ,0:2 mA m2 at 1000 mT. Small sample masses ruled out measurements of low field AC susceptibility, frequencydependent susceptibility and anhysteretic remanent magnetisation. Magnetisation and isothermal remanence data were used to calculate the following mass specific or ratio mineral magnetic parameters: low field DC susceptibility (LF :  @ 5 mT) high field paramagnetic susceptibility (para :  @ 800 mT) ferrimagnetic susceptibility (ferri : LF , para ) percent paramagnetic susceptibility (% para : para = LF  100) percent ferrimagnetic susceptibility (% ferri% : ferri = LF  100) high field remanence (HIRM: IRM @ 1000 mT – IRM @ 100 mT) (saturation) remanence (Mrs : IRM @ 1000 mT) (saturation) magnetisation (Ms : M @ 1000 mT corrected for paramagnetism) ratio of saturation remanence to saturation magnetisation (Mrs /Ms ) ratio of low remanence to saturation remanence (S ratio : IRM @ 100 mT/IRM @ 1000 mT). Sediment chronologies were obtained from analyses of 210 Pb, 226 Ra, 137 Cs and 241 Am using gamma spectrometry and both CIC and CRS models: results shown in this paper are described and evaluated in detail by Appleby et al. (1998). In summary, the last  100 years is contained in the uppermost 15 cm and sedimentation rates rise in all cores during the 20th century. Dates of  1900 and  1950 are shown in the magnetic records.

166

Figure 2. Mass specific concentration magnetic parameters plotted against depth for the six dated cores using spline fitting, showing 210 Pb-derived (Appleby et al., 1998) sediment dates 1900 and 1950 as horizontal lines.





167

Figure 3. Interparametric ratios% ferri =para , Mrs /Ms , S-ratio plotted against depth for the six dated cores, showing 210 Pb-derived (Appleby et al., 1998) sediment dates 1900 and 1950 as horizontal lines.





168 Results In broad terms, the profiles of percent dry weight, mass specific concentration parameters (Figure 2) and interparametric ratios (Figure 3) for the set BAIK 6, 19, 22 and 29 show some similarities in terms of trends and data variance, but a lack of correlation between the details of dated sections indicates that the role of local or in situ processes is probably strong. Decreasing upcore percent dry weight values in the lower parts of BAIK 25 and 29 are paralleled by decreasing concentrations of magnetic minerals (Figure 2) but this relationship breaks down in the upper parts of cores where low densities in all cores probably reflect relatively high concentrations of labile organic matter rather than real reductions in minerogenic matter. Profiles for LF (Figure 2) and ferri (not shown) correlate well in each core showing that LF is controlled largely by the ferrimagnetic component. However, the paramagnetic (para ) contribution to LF (Figure 3) is significant across the lake, ranging from  10– 15% in BAIK 6 to  20–80% in BAIK 38. Though the general trends in mass specific ferrimagnetic components (LF , Ms , Mrs ) are often parallel in each core, there are many examples of non-parallel peaks suggesting that ferrimagnetic minerals and grain-sizes may differ between samples in a core. With the exceptions of BAIK 22 and BAIK 25, there is a tendency for the concentrations of ferrimagnetic minerals to rise towards the sediment surface, and all cores show at least one ferimagnetic ‘spike’ in the upper 4 cm. Within sediments accumulated since 1900 the ferrimagnetic concentrations are highest in BAIK 6, followed by BAIK 22 and 19, and then by BAIK 25, 29 and 38. S-ratios (Figure 3) decrease in the upper 10 cm in BAIK 6, 19, 22 and 38 indicating a gradual increase in the proportion of canted antiferromagnetic minerals compared with ferrimagnetic minerals. This trend is seen only in the upper 2 cm of BAIK 29 and is reversed in BAIK 25; in both of these cores lower S-ratios suggest a higher proportion of canted antiferromagnetic minerals than in the other cores. Mrs /Ms ratios (Figure 3) are indicative of the relative importance of ‘hard’ magnetic phases, such as stable single domain magnetite, greigite and haematite, and show trends in BAIK 6, 19, 29 and 38, which are generally the inverse of those for S-ratios. In general, canted antiferromagnetic minerals, such as haematite, seem to dominate changes in the ‘hard’ components more than either stable single domain magnetite or greigite. Increasing trends in the HIRM data (Figure 2)

since 1900, indicative of the concentration of canted antiferromagnetic minerals, are strongest for BAIK 6 followed by BAIK 19, 22 and 29, but the trend is decreasing in BAIK 25 and ill-defined in BAIK 38. Trends for paramagnetic minerals show no common trends with sediment depth (Figures 2 and 3) for either mass concentrations (para ) or relative proportions (% para ) or clear correlations with other parameters. In BAIK 6 and 22 significant peaks in the concentrations of paramagnetic minerals occur just below the mud-water interface (0–3 cm) and in BAIK 22 at 8–12 cm.

Discussion 1. Post-depositional changes There are four types of post-depositional effects on the magnetic properties of detrital minerals: reductive diagenesis of ferrimagnetic minerals; authigenic formation of ferrimagnetic greigite; authigenic formation of paramagnetic/canted antiferromagnetic oxyhydroxides, such as ferrihydrite and goethite; and incorporation of ferrimagnetic magnetosomes from magnetotactic bacteria. Values for ferri are not reduced to zero in any of the cores, indicating that reductive diagenesis acting on ferrimagnets, if present, is partial. The strongest evidence that reductive diagenesis has affected some sections of sediment is in the relationship between the ferrimagnetic, paramagnetic and total Fe records (Boyle et al., 1998). In BAIK 6 and 22, para and total Fe are linearly associated (Figure 4) suggesting that peaks in total Fe concentration in these cores are largely controlled by the presence of paramagnetic iron oxyhydroxides. The source of paramagnetic oxyhydroxides may be either detrital or authigenic, but in Lake Baikal the evidence suggests an authigenic source derived from the partial reduction of previously deposited iron oxides. In both cores and to a lesser extent others, the peaks in total Fe and para are preceded (i.e. lower in the core) by reductions in the ferrimagnetic component and followed (i.e. higher in the core) by peaks in Mn (Boyle et al., 1998; Figure 1). This is consistent with a reducing front in the upper few centimetres causing the partial dissolution of ferrimagnetic minerals and the subsequent migration of Fe and Mn upwards to the oxic zone and precipitation as paramagnetic minerals. The one exception is BAIK 25 (Academician Ridge) which shows no evidence for Fe and Mn mobility in

169

Figure 4. Mass specific paramagnetic susceptibility (para ) plotted against total Fe (Boyle et al., 1998) for the six dated cores.

170 the uppermost sediments, and the weakest association between para and total Fe (Figure 4) suggesting that the coring site lies in a deep oxidised zone of the lake bed where para is controlled by the concentration of detrital clay minerals. The sharp peaks in ferri in the uppermost sediments of all cores except BAIK 6 could be evidence for the presence of either greigite (Fe3 S4 ) or magnetite in the form of bacterial magnetosomes. Each peak is associated with a sharp rise in the S-ratio confirming a shift towards dominantly low coercivity (soft) ferrimagnetic properties. Magnetic discrimination between the two minerals is through their remanence properties and the redox conditions of their formation. Greigite exhibits high coercivity (hard) ferrimagnetic properties with characteristically (very) high Mrs values relative to LF but is a product of strictly reducing conditions in organic and sulphur-rich sediments. Magnetotactic bacteria produce stable single domain magnetite grains with relatively hard magnetic properties, but the live bacteria are restricted to microaerophilic environments. The apparently soft magnetic behaviour coupled with the continuous presence of well-oxygenated bottom water over a large part of the Lake Baikal bed, allowing the widespread precipitation of Fe and Mn in surface oxidised sediments, suggests that the sharp peaks of ferrimagnetic concentrations are possibly zones of magnetosome production and accumulation rather than greigite formation. This argument is supported by three further observations: the absence of ferrimagnetic peaks in more strongly reducing deeper sediments; the lack of very high Mrs values diagnostic of greigite; and the lack of evidence for greigite in or near those sediment zones where reductive diagenetic processes are believed to have operated. None of the lower sections of cores shows peaks in ferri of comparable magnitude to those at or near the surface which implies that acceptance of a magnetosome explanation has also to include total or partial magnetosome dissolution in sediments lying below 5 cm or alternatively that the existence of magnetotactic bacteria is a recent phenomenon. Snowball (1994) confirmed that magnetosomes dominated the ferrimagnetic properties of the upper  50 cm of some Swedish lake sediments, but that these were subject to reductive diagenesis as they became buried to leave coarser multidomain grains dominating the ferrimagnetic properties of old sediment. Microscopic or molecular methods are required to confirm these alternative explanations.

2. Turbidites Turbidites are rapidly deposited layers derived from extreme events such as floods, or the reworking and slumping of upslope sediments. In some cores from Lake Baikal they are characterised by minerogenic and diatomaceous layers fining upwards from sandy to silty, with an initial rise in LF values which decline up-core (Lees et al., 1998b). Turbidite sedimentation between 1.2–1.4 cm and 5.75–6.75 cm in BAIK 22 is indicated from sediment accumulation rates (Appleby et al., 1998) and lithology descriptions (Mackay et al., 1998). LF values show a decreasing trend over the depth range 1–7 cm and are matched by similar decreasing trends in ferri , Ms and Mrs , low values of paramagnetic parameters (para and % para ), but broadly constant values of HIRM. Close inspection of the density record of BAIK 22 show that some peaks in density are associated with small minima in magnetic concentration parameters, and a similar association can be found in the turbidite layer of BAIK 29 (0.6– 0.8 cm). The evidence suggests that turbidites in these cores are composed of slumped diatomaceous sediments rather than reworking of minerogenic sediment from the margins with higher density detrital minerals (cf. Lees et al., 1998b) and that they may dilute the concentration of catchment-derived detrital minerals. 3. Catchment erosion Apart from the sand particles in flood and shallow water slump turbidites, catchment-derived material which reaches the deep water zones is likely to consist of clay-silt sized minerogenic particles eroded from soils and terrestrial sediments. An influx of minerogenic particles within this size range could be expected to give a variety of magnetic signals depending upon the mineralogy of the geological source, the soil type and particle-size. However, most clay and silt-sized minerogenic sediments, unless dominated by quartz will include ferrimagnetic, paramagnetic and canted antiferromagnetic minerals. All the lake sediment samples contain measurable concentrations of ferrimagnetic, paramagnetic and canted antiferromagnetic minerals, indicating that a consistent detrital mineral contribution from catchment sources is a possibility. However the evidence for widespread postdepositional changes makes interpretation in terms of detrital sources of ferrimagnetic and paramagnetic data problematical. Canted antiferromagnetic minerals (haematite and goethite) are the iron oxyhydroxides

171 most resistant to reductive dissolution which suggests that any record of long term (i.e. excluding extreme flood events) shifts in the delivery of eroded catchment material will be best recorded in the HIRM profiles. The consistency in the trends of HIRM, increasing since at least 1900 in the three deep water cores (BAIK 6, 22 and 29) and BAIK 19 opposite the Selenga inflow suggests that catchment erosion particularly from igneous rock sources has increased since the 19th century; Lees et al.’s (1998a) magnetic data for sediment sources in the catchment show that igneous rocks have at least 2 the HIRM value of sedimentary rocks and topsoil. The trigger for increased erosion is not known, but could include changes in farming and forestry practices, increased construction or climate changes leading to, for example, higher flood frequencies or faster snowmelt. The post-1950 increase in sedimentation rates (Appleby et al., 1998) coupled with a rising trend of HIRM values argues for a recent acceleration in erosion, but the dilution effects of biogenic silica and organic matter on the magnetic records need to be evaluated before this argument can be accepted. The S-ratio profiles support an interpretation of higher HIRM in terms of increased canted antiferromagnetic mineral concentrations, but caution is needed because the S-ratio profiles are clearly influenced by the absolute changes in the concentrations of paramagnetic and ferrimagnetic concentrations brought about by dissolution, authigenesis and magnetosome production. 4. Atmospheric pollution Magnetic spheres produced in predominately coalfired combustion processes contain both magnetite and haematite and are therefore detectable by ferrimagnetic and canted antiferromagnetic parameters. Postdepositional changes to the ferrimagnets would indicate that, as with eroded catchment material, the optimum magnetic parameter is HIRM. Values of HIRM in catchment samples are as high for modern polluted snow samples as for igneous rock samples (Lees et al., 1998a). Discrimination of the effects of atmospheric pollution from erosion on the magnetic record is therefore difficult but the weight of evidence points to the latter. Studies of heavy metals (Boyle et al., 1998) and SCPs (Rose et al., 1998) in these cores concludes that the major control on heavy metal concentration is sediment supply and only in the southern basin is atmospheric pollution detected. The highest post-1950 HIRM values are recorded in BAIK 6 from the south-

ern basin and the trend parallels the curve for SCPs from the 1930s when they are first recorded in the sediments, but the rising HIRM values start in this and in other cores during the 19th century. 5. Glacial-interglacial mineral magnetic records In their study of glacial-interglacial sediments from the Academician Ridge, Peck et al. (1994) used HIRM to infer aeolian inputs of iron-stained grains during glacial periods. The present study shows that during interglacial times, HIRM may indicate fluviallytransported soils and sediment, raising the question of how records of HIRM during glacial periods may be confirmed as showing the effects of long distance dust transport. The dominance of glacial sediments by angular quartz grains (Flower pers. comm.) supports a ‘mountain loess’ origin (Smalley, 1995), but fluvial transport within a periglacial catchment over relatively short distances could also deliver sediment with high concentrations of canted antiferromagnetic minerals. Long interglacial mineral magnetic records of detrital minerals in cores from the southern basin may provide information about the growth and function of the Selenga Delta as it responds to major lake level changes (cf. Romashkin & Williams, 1997). 6. Core correlations The recent magnetic records from Lake Baikal demonstrate the array of potential mineral sources, processes and transformations encompassing the catchment, the lake and the sediments. Intuitively, the likelihood of tracing synchronous layers between cores will become increasingly difficult as the distance increases. Lees et al. (1998b) show convincing evidence for core correlations based on magnetic signatures of turbidites over distances of up to tens of kilometres, but on the evidence of the present study it must remain doubtful whether magnetic-based correlations in recent sediments between basins or even sub-basins are possible everywhere. Indeed, given the strong morphological and hydrodynamic contrasts between basins, the large catchment size and the numerous inflows, it is unlikely that short term influxes of fluvially-derived detrital sediments to the lake are driven by synchronous and catchment-wide processes. For studies of recent sediments, the best means for core correlation is likely to be based on atmospherically-derived properties (hence the success of 210 Pb dating) such as SCPs (Rose et al., 1998). In older sediments, signals of autochthonous

172 processes such as the widespread  18th century rise in one species of diatom (Mackay et al., 1998) may represent a sounder basis for core correlation between basins. In long glacial-interglacial sediment sequences, major climate-mediated shifts in the whole lake’s sediment system will be useful for correlation purposes as demonstrated by Julius et al.’s (1997) analysis of diatom-rich interglacial and diatom-poor glacial sediments in the southern basin and Peck et al.’s (1994) use of susceptibility to identify glacial-interglacial boundaries across large distances on the Academician Ridge.

teith who helped to collected most of the sediment cores, to Bob Jude who measured some of the samples and produced the diagrams, and to Dr J. A. Lees for offering comments on a first draft. The Lake Baikal programme received financial support from the Royal Society (BICER), the Leverhulme Trust (Project F134AZ), and ENSIS Ltd (University College London). Thanks are due to both reviewers for making constructive comments on the text.

References Conclusions Mineral magnetic measurements of six recent sediment cores from across Lake Baikal show the possible effects of a variety of controls including reductive diagenesis, authigenesis of paramagnetic minerals, bacterial magnetosome production, atmosphericallyderived magnetic particulates and catchment-derived detrital minerals. With the exception of a core from the Academician Ridge, the evidence is strongest for the presence of widespread reductive diagenesis of iron-containing minerals including ferrimagnets with upward migration of Fe to form paramagnetic minerals in the seasonally oxic zone beneath the mud-water interface. Sediments on the Academician Ridge show evidence for magnetosome accumulation but signals of reductive diagenesis are weak. There is evidence from HIRM data that there has been a widespread increase in the delivery of catchment-derived minerogenic material to the deep parts of the northern and southern basins since the 19th century. The opposite trend found on the Academician Ridge may be explained by its isolation from sediment-carrying bottom currents and the diluting effect of diatomaceous silica on magnetic concentration parameters. Further diagnostic tests of mineralogy and sediment composition are required to confirm these findings.

Acknowledgements The work presented here has been carried out in collaboration with other colleagues in the Lake Baikal Programme. Field work facilities at Irkutsk and on Lake Baikal were kindly provided by Prof. Grachev, Director of the Limnological Institute in Irkutsk. Thanks are due to Dr Ye. V. Lihkoshway who organised the main sediment collection expedition, to Don Mon-

Anderson, N. J. & B. Rippey, 1988. Diagenesis of magnetic minerals in the recent sediments of a eutrophic lake. Limnol. Oceanogr. 33: 1476–1492. Appleby, P. G., R. J. Flower, A. W. Mackay & N. L. Rose, 1998. Paleolimnological assessment of recent environmental change in Lake Baikal: sediment chronology. J. Paleolimn. 20: 119–133. Boyle, J. F., A. W. Mackay, N. L. Rose, R. J. Flower & P. G. Appleby, 1998. Sediment heavy metal record in Lake Baikal: natural and anthropogenic sources. J. Palaeolimn. 20: 135–150. Dearing, J. A., 1986. Core correlation and total sediment influx. In: B. E. Berglund (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology, John Wiley & Sons Ltd.: 247–265. Dearing, J. A., 1992. Sediment yields and sources in a Welsh lake catchment during the past 800 years. Earth Surf. Proc. Landf. 17: 1–22. Flower, R. J., 1998. Paleolimnology and recent environmental change in Lake Baikal: an introduction and overview of interrelated concurrent studies. J. Paleolimn. 20: 107–117. Flower, R. J., A. W. Mackay, N. L. Rose, J. F. Boyle, J. A. Dearing, P. A. Appleby, A. E. Kuzmina & L. Z. Granina, 1995. Sedimentary records of recent environmental change in Lake Baikal, Siberia. The Holocene 5: 323–327. Foster, I. D. L., P. N. Owens & D. E. Walling, 1996. Sediment yields and sediment delivery in the catchments of Slapton Lower Ley, South Devon, UK. Field Studies 8: 629–661. Granina, L. Z., 1992. Ferromanganese formations in Lake Baikal: formation conditions, composition, distribution. IPPCCE Newsletter 6: 41–47. Granina, L. Z., E. B. Karabanov & E. Callender, 1993. Relics of oxidized ferromanganese formations in the bottom sediments of Lake Baikal. IPPCCE Newsletter 7: 32–39. Higgitt, S. R., F. Oldfield & P. G. Appleby, 1990. The records of land use change and soil erosion in the late-Holocene sediments in the Petit Lac d’Annecy, eastern France. Holocene 1: 14–28. Hilton, J. & J. P. Lishman, 1985. The effect of redox changes on the magnetic susceptibility of sediments from a seasonally anoxic lake. Limnol. Oceaongr. 30: 907–909. Hilton, J., Lishman, J. P. & Chapman, J. S., 1986. Magnetic and chemical characterisation of a diagenetic magnetic mineral formed in the sediments of productive lakes. Chem. Geol. 56: 325–333. Julius, M. L., E. F. Stoermer, S. M. Colman, & T. C. Moore, 1997. A preliminary investigation of siliceous microfossil succession in late Quaternary sediments from Lake Baikal, Siberia. J. Paleolimn. 18: 187–204. Lees, J. A., R. J. Flower & P. G. Appleby, 1998a. Mineral magnetic and physical properties of surficial sediments and onshore sam-

173 ples from the southern basin of Lake Baikal, Siberia. J. Paleolimn. 20: 175–186. Lees, J. A., R. J. Flower, D. Ryves, D. Vologina & M. Sturm, 1998b. Identifying sedimentation patterns in Lake Baikal using whole core and surface scanning magnetic susceptibility. J. Palaeolimn. 20: 187–202. Mackay, A. W., R. J. Flower, A. E. Kuzmina, L. Z. Granina, N. L. Rose, P. G. Appleby, J. F. Boyle & R. W. Battarbee, 1998. Diatom succession trends in recent sediments from Lake Baikal and their relationship to pollution and to climate change. Phil. Trans. Roy. Soc. Lond., in press. Oldfield, F., 1994. Towards the discrimination of fine-grained ferrimagnets by magnetic measurements in lake and near-shore marine sediments. J. Geophys. Res. 99: 9045–9050. Oldfield, F., C. Barnosky, E. B. Leopold & J. P. Smith, 1986. Mineral magnetic studies of lake sediments. Hydrobiologia 103: 37–44. Oldfield, F. & N. Richardson, 1990. Lake sediment magnetism and atmospheric deposition. Phil. Trans. Roy. Soc. Lond. B327: 325– 330. Peck, J. A., J. W. King, S. M. Colman & V. A. Kravchinsky, 1994. A rock-magnetic record from lake Baikal, Siberia: Evidence for Late Quaternary climate change. Earth Planet Sci. Lett. 12: 221– 238. Romashkin, P. A. & D. F. Williams, 1997. Sedimentation history of the Selenga Delta, Lake Baikal: simulation and interpretation, J. Paleolimn. 18: 179–186. Rose, N. L., P. G. Appleby, J. F. Boyle, A. W. Mackay & R. J. Flower, 1998. The spatial and temporal distribution of fossil-fuel derived

pollutants in the sediment record of Lake Baikal, eastern Siberia. J. Paleolimn. 20: 151–162. Smalley, I., 1995. Making the material: the formation of silt-sized primary mineral particles for loess deposits. Quat. Sci. Rev. 14: 645–652. Snowball, I. F., 1993. Geochemical control of magnetite dissolution in subarctic lake sediments and the implications for environmental magnetism. J. Quat. Sci. 8: 339–346. Snowball, I. F., 1994. Bacterial magnetite and the magnetic properties of sediments in a Swedish lake. Earth Planet. Sci. Lett. 126: 129–142. Snowball, I. F. & R. Thompson, 1988. The occurrence of greigite in sediments from Loch Lomond, J. Quat. Sci. 3: 121–125. Snowball, I. & R. Thompson, 1990. A mineral magnetic study of Holocene sedimentation in Lough Catherine, Northern Ireland. Boreas 19: 127–146. Thompson, R., R. W. Battarbee, P. E. O’Sullivan & F. Oldfield, 1975. Magnetic susceptibility of lake sediments. Limnol. Oceanogr. 20: 687–698. Thompson, R. & F. Oldfield, 1986. Environmental Magnetism. Allen & Unwin, London, 227 pp. Thompson, R. & D. J. Morton, 1978. Magnetic susceptibility and particle-size distribution in recent sediments of the Loch Lomond drainage basin, Scotland. J. Sed. Petrol. 49: 801–811. Williams, T. M., 1992. A sedimentary record of the deposition of heavy metals and magnetic oxides in the Loch Dee basin, Galloway, Scotland, since c.AD 1500. Holocene 1: 142–150.