Geochemistry, mineralogy, and geochemical mass balance on major elements in two peat bog profiles (Jura Mountains, Switzerland)

Geochemistry, mineralogy, and geochemical mass balance on major elements in two peat bog profiles (Jura Mountains, Switzerland)

INCKIDIW ELSEVIER ISOTOPE GEOSCIENCE Chemical Geology 138 (1997) 25-53 Geochemistry, mineralogy, and geochemical mass balance on major elements in ...

3MB Sizes 0 Downloads 0 Views

INCKIDIW

ELSEVIER

ISOTOPE GEOSCIENCE Chemical Geology 138 (1997) 25-53

Geochemistry, mineralogy, and geochemical mass balance on major elements in two peat bog profiles (Jura Mountains, Switzerland) Philipp Steinmann ‘, William Shotyk * Geologisches Institut, Uniuersitiit Bern, Baltzerstr. I, CH-3012, Bern, Switzerland Received 15 August 1995; accepted 5 December 1995

Abstract

The mineralogical and chemical composition of peats from two Sphagnumbogs in the Franches Montagnes region (Jura Mountains, Switzerland) were compared. The peats in the top 102 cm at Etang de la Gr&re (EGr) represent 2110 years of peat formation compared with 1730 years for the first 84 cm at Tomb&e de Genevez (TGe). Scandium was used as a conservative tracer to distinguish between the primary sources of major elements to the bogs: atmospheric deposition of soil-derived aerosols (EGr and TGe); physical incorporation of elements along with mineral matter from the underlying sediments (TGe); and adsorption or complexation by the peat following diffusion from groundwater (TGe). These results support previous findings which showed that the EGr core (102 cm) is exclusively ombrogenic compared with TGe where the peats below approx. 30 cm become predominantly minerogenic. The modem rates of SC accumulation (past 100 years) are similar in the two cores: 39 pg mm2 a- ’ at EGr versus 52 pg m-2 ,-1 at TGe. However, comparison with deeper sections of the EGr core reveals that the present-day rates of atmospheric SC deposition are nearly 3 times greater than the long-term average rate, reflecting the higher concentrations of soil-derived atmospheric aerosols today. The rates of SC accumulation in the TGe core are slightly higher than at EGr; this reflects the incorporation of mineral matter from the underlying sediments. In the deeper, minerogenic peats at TGe, the long term rates of Fe and Ca accumulation are 5 and 7 times higher, respectively, than in the EGr core. These fluxes greatly exceed the difference in SC flux, and cannot be explained by the differences in amount of mineral matter in the peats. These enrichments require an independent explanation, and are most likely the result of adsorption and/or complexation of cationic Ca and Fe species diffusing from deeper layers in the profile. The mineralogical composition of the ash is mainly quartz @O-90%), with feldspar (5-15%) and muscovite (5-15%), although various other minerals (5-10%) are commonly present. Biogenic Si represents another important fraction of ash and is abundant (30-‘70%) at discrete depths. The vertical profiles showed no significant changes in mineralogy with depth: assuming a constant composition of source materials, the lack of progressive mineralogical change suggests that the fine-grained silicates supplied by soil dust have not been measurably weathered during the past two millennia. A number of

Correspondingauthor. Phone: +41 31 631-8770. Fax: +41 31 631-4843. E-mail: [email protected] ’Institut F.-A. Forel, 10, route de Suisse, C&1290, Versoix, Switzerland.

??

0009-2541/97/$17.00 Copyright 6 1997 Elsevier Science B.V. All rights reserved. PII SOOOS-2541(96)00171-4

26 peaks

P. Steinmann, W. Shotyk/Chemical

Geology 138 (1997) 25-53

in Fe/SC are found in both cores above peaks in ash content and density. These appear to correspond with temporal

fluctuations

in wet/dry

conditions:

accumulation

of Fe-oxides

during dry periods and leaching during wetter conditions.

Keywords: Peat bog; Ombotrophic; Atmospheric deposition; SC; Fe; Ca

1. Introduction Little is known about the rates and mechanisms of mineral-fluid interactions in the organic-rich, anaerobic, acidic, pore waters of peat bogs. The precipitation of secondary Fe minerals such as siderite and vivianite in fens and swamps has been studied on a regional scale in some detail (Van Bemmelen, 1899; Kivinen, 1936; Bushinskii, 1946; Puustj’rirvi, 1952; Lukashev and Kovalev, 1969; Postma, 1977, 1981, 1982, 1983; Selmer-Olsen and Lie, 1983; Virtanen, 19941, but little is known about the reductive dissolution of primary Fe-containing phases in bogs. Some recent studies of atmospheric pollution have included measurements of isothermal remanent magnetization (IRM) in peat cores as an index of the concentrations and chronology of deposition of ferric particles produced during fossil fuel combustion (Oldfield et al., 1978, 1981). The validity of these records of magnetic minerals depends on the assumption that these particles are unaffected by such post-depositional processes as dissolution. Using peat cores from various bogs, Richardson (1986) found that cores taken from hummocks (predominantly aerobic microenvironments in which oxidizing conditions are presumed to dominate) and hollows (water-saturated, anoxic sites) revealed synchronous changes in the abundance and distribution of magnetic particles. The profiles presented by Richardson (1986) support the argument that Fe mineral dissolution must have been taking place sufficiently slowly that the magnetic mineral record was effectively preserved. More recently, however, Williams (1992) showed that the declining concentrations of magnetic particles with increasing depth in peat cores coincided with the level of the water table at the time of sample collection; Eh measurements revealed a sharp redoxcline at these depths. SEM analyses of magnetic particles isolated from one of the peat cores revealed Fe oxide surface coatings on a sample taken from a depth of 1 cm; at this depth the pore waters are assumed to be oxic most of the time, and the Fe oxides were preserved. In contrast, in a sample collected below

the water table (at a depth of 24 cm), there was no detectable Fe oxide surface layer and the sphere appeared fractured. Williams (1992) argued that these particles are measurably reactive in acidic, anoxic peats and has questioned the validity of using the occurrence of magnetic particles in peat cores to construct historical records of industrial pollution. Similarly, there have been few quantitative studies concerning the dissolution of silicate minerals in peat bog environments, probably at least partly due to the difficulty of performing quantitative mineralogical analyses in peat samples which are only 1 to 2% by weight mineral matter (Finney and Farnham, 1968). The few reports which have been published have presented some conflicting views on the rates of mineral dissolution in peat bogs. Lukashev et al. (1974), for example, compared rates of atmospheric deposition of silicates (dust) to Belorussian peat bogs with the rates of ash accumulation in the peats and estimated that 90% of the mineral matter supplied to the bogs had dissolved in situ. Limited support for this interpretation has been published by Hodder et al. (1991) who have reported comparatively rapid dissolution of ferromagnesian minerals in volcanic ash buried in some acidic New Zealand peat bogs. Biotite in particular was extensively attacked and from one tephra layer it was completely removed in less than 800 years of burial. In contrast, individual shards of Icelandic volcanic glass which had been deposited in Scottish blanket bogs almost 4000 years BP (Hekla 4 eruption) revealed remarkable preservation (Dugmore et al., 1992). Electron microprobe analyses of these shards yielded major element compositions which were not significantly different from the corresponding tephra layers collected from loessial soils in Iceland. Dugmore et al. (1992) concluded that the fine silicic glass shards retain their characteristic chemical composition in the weathering environments of both bogs and soils. With respect to pH neutral-alkaline fens and swamps, there have been reports suggesting that organic acids may promote the dissolution of quartz (Bennet and Siegel, 1987) and aluminosilicates (Be-

P. Steinmann, W. Shotyk/ ChemicalGeology138 (1997)25-53

nnet et al., 1991); this process is restricted to minerogenie mires where pH values are in the range 6-8 (i.e. in fens and in the deeper peat layers underlying acidic bogs). For example, at a depth of 100 cm below the surface of the Red Lake Peatland (with a pore water pH of 6.91, SEM analyses showed that many of the Fe-bearing silicates were clearly etched. Similarly, etched feldspars were found at a depth of 200 cm where the pore waters had a pH of 7.2. Bennet et al. (1991) argued further that organic complexation of Si may be responsible for the complete absence of diatoms in humified peat, and may help explain the formation of Si-depleted underclays commonly associated with coal seams. In contrast to the weathering process in neutral-alkaline fen peats, in acidic bog peats (pH < 5) Bennet et al. (1991) found only minor weathering of Fe silicates and Al silicates, and no recognizable weathering of quartz. One of the main objectives of the present study was to determine the sources of inorganic solids to two contrasting peat bogs in close proximity: Etang de la Grubre (EGr) and La Tourbibre de Genevez (TGe). Previous work has shown that the top 1 m at EGr consists exclusively of ombrogenic peat whereas at TGe, below approximately 30 cm the profile is essentially minerogenic (Shotyk and Steinmann, 1994; Shotyk, 1996a,b; Steinmann and Shotyk, 1997). In other words, at EGr, atmospheric deposition is the sole sourc:e of inorganic solids supplied to the peat whereas at TGe, the inorganic fraction is derived from both atmospheric and terrestrial sources. The terrestrial sources of inorganic solids to the peats at TGe may in’clude both mineral matter physically incorporated from the underlying sediments as well as elements which were added to the peats by mineral-water interactions; one of the goals of this study was to estimat,e the relative importance of each of these processes. A second objective was to identify the dominant mineral phases present in the ash fraction of the peats and to quantify any changes in mineralogy with increasing depth (and therefore, increasing age) in the profiles. Because the EGr core is ombrogenic, it provides a unique opportunity to study possible mineralogical and chemical changes to soil-derived aerosols as a function of time since they were deposited on tlhe surface of the bog. The base of the 2f core from EGr (96-102 cm) has been dated at 2110 f 30 years BP (Shotyk, 1996b). Thus, this

21

core should provide a record of possible mineralogical transformations on the scale of two millennia. 2. Materials and methods 2.1. Description of the study sites, peat stratigraphy, and trophic status Etang de la Gruhre (EGr) and Tourbibre de Genevez (TGe) are two of a number of bogs occurring in the Franches Montagnes region of the Jura Mountains, Switzerland (Fig. la). The Franches Montagnes is a calcareous plateau at an altitude of approximately 1000 m (a.s.1.). EGr is located 4 km west of TGe which lies just 2 km north of the village Tramelan. The bogs are 22.5 ha (EGr) and 7.2 ha (TGe) in extent. At EGr, peat formation began in the late Glacial approximately 12000 years BP (Joray, 1942) compared with 5000 years BP at TGe (Reille, 1991). EGr and TGe are registered in the Inventory

a

b

1OOm

“NNW



4

TGe

il‘ 1002 d E .F ;u I

999 996 1000 994 1 i

I_” 100m



Fig. 1. (a) Arrow showing approximate location of the bogs in the Franches Montagnes of the Jura Mountains, Switzerland. (b) Cross-section of Tourbike de Genevez (TGe) after Welten (1964). The peat is clearly elevated above the surrounding dolines. Dark layers in the peat profiles indicate more decomposed, less permeable peat layers. Arrows indicate major flow paths of pore waters. (c) Cross-section of Etang de la Gru&re (EGr) after Joray (1942).

28

P. Steinmann, W. Shotyk/Chemical Geology 138 (1997) 25-53

of Raised and Transitional Mires of Switzerland as bogs No. 2 and No. 4, respectively (G&rig et al., 1984). The peat bog at EGr is strongly domed, with more than 6 m of peat accumulation in the centre where the cores were taken. The surface vegetation in the treeless central parts of the peat dome is dominated by Sphagnum (mainly S. magellanicum) and Eriophorum (see also Feldmeyer-Christe, 1990 and Griinig et al., 1984). The field description of the peat stratigraphy at the coring site is as follows, with the degree of decomposition on the IO-point scale of von Post (von Post and Granlund, 1925) given in brackets (see e.g. Clymo, 1983; Landva et al., 1986; Hobbs, 1986; Shotyk, 1988): O-25 cm Sphagnumdominated peats with Eriophorum (H2); 25-50 cm Sphagnum peat with few Carex and Eriophorum remains (H4); 50-60 cm Sphagnum-Eriophorum peats (H4); 60- 130 cm Sphagnum-Eriophorum peats (H5); 130-250 cm Sphagnum-Eriophorum peats (H8); 250-420 cm Sphagnum-dominated peat with few Eriophorum (H8); 420-550 cm Carex peat with Eriophorum (H4); 550-650 cm Carex peat with few Eriophorum (H3). At 650 cm there is a sharp boundary with the underlying grey silty clay. EGr has developed on Oxfordian clays and marls consisting essentially of 40% clay and 40% quartz, with < 5% of each of plagioclase feldspar and potassium feldspar, and approximately 3% calcite and 2% dolomite (Steinmann, 1995). The topography of the site combined with the presence of a 2 m thick layer of highly decomposed (H8) peat suggests that the upper 2 to 3 m of this profile is beyond the influence of laterally penetrating waters. A comparison of the major element chemistry of the pore waters with local rainwater indicated that the entire core collected at EGr (core 2f representing O-102 cm at EGr) is ombrogenic (Shotyk and Steinmann, 1994; Steinmann and Shotyk, 1997). Independent support for this interpretation was obtained from the Ca/Mg molar ratio of the peats which revealed values equal to or less than the corresponding rainwater average (Shotyk, 1996a). The stratigraphy indicates that the uppermost 420 cm at EGr consists of Sphagnum-dominated peat; the botanical composition of this peat suggests that ombrogenic conditions may have prevailed at the time that these plants were growing. However, it has been clearly

shown that during burial, decomposition, and compaction peats which were originally ombrogenic may be penetrated by mineral soil water or groundwater and become overprinted to some extent with a minerogenic chemical signature (see the review by Shotyk, 1996b). In addition to these processes which are continually active in the peatlands themselves, a rise in the local water table external to the mire could have a comparable effect. Thus, the botanical composition of peats is no reliable indication of their trophic status: the presence of Carex peats helow 420 cm certainly indicates minerogenic conditions at the time of their formation, but the presence of Sphagnum peats above does not necessarily mean that the overlying peats are ombrogenic today. The abundance and vertical distribution of ash is also consistent with the possibility that all of the Sphagnum-Eriophorum peats at these sites may be ombrogenie: the peats above 400 cm contain < 3% ash, values which are in the range generally given for Sphagnum bog peats (Naucke, 19901, whereas the peats below this depth contain more than 3% ash. However, at EGr the Ca/Mg ratios are equal to or lower than the value measured in local rainwater

L.

*.._.. ‘..

1

. . . . . . . . . ca#Jg

-.._ 1:. -.....*

2

EGr

‘:.

.c

e

:ql CaMg TGe

*.

:

__................ ...

. ..--

..::I:.. *:.. .:_.: . . ..-

.:I . . . . .._.___.___~~~

si E 7



0

‘.: :

/..’ .-*

’j ’ 5

“,



I

!



10 15 20 camg ratlo (molar)

*’ 25

Fig. 2. Complete vertical profiles of Ca/Mg at EGr and TGe. Vertical line indicates the average Ca/Mg molar ratio of precipitation in the Jura (Shotyk and Steinmann, 1994). which is used here to distinguish ombrotrophic peats from minerotrophic peats.

P. Steinmann, W. Shoiyk/Chemical

only down to a depth of 2.5 m (Fig. 2), indicating that ombrogenic peat does no longer extend below this depth. The very low Ca/Mg vales near the surface are caused by a relative enrichment of Mg (see below). Below 2.5 m the increasing ratio of Ca to Mg indicates that rainwater alone cannot explain the composition of Ihe peats, and that an additional source of Ca must be invoked to explain the profile. The Ca/Mg molar ratio of the peats, therefore, indicates that atmospheric supply has been the exclusive source of inorganic solids to the peats in the top 250 cm of the bog, ,whereas deeper peat layers have additionally been affected by mineral soil water and groundwater. The profile at EGr highlights the important difference between the bog/fen limit which is strictly a biological parameter revealed by the botanical composition of the peats (and located at 420 cm at EGr) and the ombrogenic/minerogenic boundary which is a geochemical parameter indicated by the Ca/Mg molar ratio of the peats (and found at 250 cm at EGr). The peat bog ‘La Tourbi&re de Genevez’ (TGe) has developed on an outcrop of Tertiary clays and marls (up to 1 m thick) which is elevated above and surrounded by a ring of karstic sink holes (Fig. lb). Penetration of calcareous water from the surrounding area into the bog is, therefore, prevented by (1) the generally flat topography of the area, (2) elevation of the bog above the surrounding terrain, and (3) the presence of the dolines surrounding the bog (Gerber and Monbaron, 1990). This elliptical bog extends no more than 590 m frclm west to east, and 170 m from north to south. TGe also has a pronounced domed structure (Welten, 1964): from north to south the increase in elevation from the edge of the peatland to the centre of the dome is 2.5 m (Fig. lb). All peat cores were taken in the treeless centre of the dome. The surface vegetation is dominated by Sphagnum (S. j&urn> and Eriophorum (cf. Gtinig et al., 1984; Feldmeyer-Christe, 1990). The peat stratigraphy at the coring site is: O-25 cm Sphagnumdominated peat (1~3); 25-70 cm SphagnumEriophorum peat with few Curex remains (H4-H5); 70-100 cm Carex-dominated peats with Eriophorum (H4). At 140 cm the mineral soil is reached. A brownish to greenish peaty clay (at 140-160 cm) with ash content increasing from 75% to 85% is followed by a greyis’h to greenish clayey silt between

Geology 138 (1997) 25-53

29

160 and 200 cm containing roots and yellowish spbts. The mineral material at the base of this peat core consists essentially of 60% clay, 30% quartz, < 5% of each of plagioclase feldspar and potassium feldspar, < 2% calcite and dolomite (Steinmann, 1995). At TGe, the ash contents of the peats below approximately 20 cm increase progressively with depth, achieving values (5-7%) which are well in the range for minerogenic Curex fen peats (Naucke, 1990). In other words, atmospheric inputs alone cannot account for the relatively high ash contents, and below 20 cm or so, terrestrial inputs must have been increasingly important with depth. The chemistry of pore waters in core li (taken from 0 to 84 cm at TGe) indicated that no more than 20-30 cm of peat at TGe is ombrogenic (Shotyk and Steinmann, 1994; Steinmann and Shotyk, 1997). Similarly, the Ca/Mg molar ratio in the peats (Fig. 2) reveal an ombrogenic to minerogenic transition taking place within approximately 30 cm of the surface (Shotyk, 1996a). 2.2. Chronology of peat accumulation Both the EGr 2f and the TGe li cores were age-dated using 2’0Pb (Appleby et al., 1997), and the chronology evaluated using independent pollen chronostratigraphic markers in replicate cores. The *“Pb chronology dates sample 2fll (27-30 cm) at AD 1905 + 6 and sample 2f12 (30-33 cm) at AD 1879 + 11. The depth of 30 cm in the EGr core 2f is therefore estimated to correspond to the year 1892 f 8. Given that the core was taken in the year 1991 the uppermost 36 cm represent approximately 100 (99 f 8) years of peat accumulation. At TGe sample li13 (33-36 cm) dates at 1900 + 11 years and sample li14 (36-39 cm) dates at 1886 f 15. The depth of 36 cm at TGe is therefore dated at AD 1886 + 15, i.e. the uppermost 36 cm of core li embraces 98 f 13 years. The depth of the modem peats accumulated over the last 100 years (30 cm at EGr; 36 cm at TGe) is indicated in all chemical profiles as a horizontal line. Cannabis pollen was found in the EGr 2h core (replicate of the 2f core) only in samples deeper than 24 cm (Fankhauser, 1995). In the 2f core, this depth was dated at AD 1929 rt 3. Similarly, at TGe, Cannabis was only found in a replicate core (11)

30

P. Steinmunn, W. Shotyk/ Chemical Geology 138 (1997) 25.53

below 30 cm; this depth was dated at AD 1923 f 6 in core li. Cannabis was grown commercially in the Jura until approximately 1930 and the pollen chronostratigraphic markers provide, therefore, an independent confirmation of the 210Pb chronologies for both the EGr and the TGe bogs. Using t4C, the following radiocarbon dates were obtained for basal samples from these two cores: sample 294/35 at EGr (96-102 cm, average 99 cm), 2110 f 30 BP (radiocarbon sample B-6459); sample li29 at TGe (81-84 cm, average 82.5 cm>, 1730 f 30 BP (~-6458). 2.3. Sample collection and preparation Peat cores from Etang de la Grubre (EGr) and Tourbibre de Genevez (TGe) were taken in August 1991 using a Wardenaar peat profile sampler (Wardenaar, 1987). The cores were rectangular, 10 X 10 cm in cross-section with lengths of 102 cm (EGr) and 84 cm (TGe), respectively. The cores were wrapped in plastic foil and brought to the lab, where they were sectioned into 3-cm slices using a stainless steel knife with serrated blade. Each slice was put into a plastic bag, weighed, and the pore waters squeezed by hand. The pore waters were analyzed for pH, major cations and anions and DOC (Shotyk and Steinmann, 1994). After squeezing pore waters the peats were dried at 105°C in Teflon bowls, and macerated in a centrifugal mill equipped with a Ti rotor and 0.25-mm sieve (Ultracentrifugal Mill ZM l-T, F.K. Retsch GmbH and Co., Haan, Germany). The milling was carried out in a Class 100 laminar flow clean air cabinet. The samples were put in acid-washed polyethylene beakers with screw cap. They were homogenized with an end over end shaker overnight and stored. The density was calculated from the volume of the Wardenaar core slices (300 cm3) and the dry weight. 2.4. Chemical analyses of peats A 5-g aliquot of each sample was analyzed using instrumental neutron activation (INAA) for the elements Ba, Br, Ca, Fe, Na, and SC (Activation Laboratories Ltd., Ancaster, Ontario, Canada). The application of INAA for analyzing peat samples has been described by several authors (Njlstad et al., 1987,

Steinnes and Njistad, 1995; Ntahokaja and Zikovsky, 1995). A series of reference standard materials was analyzed as blind standards in triplicate. These standard reference materials included coals (NBS-SRM 1632b, NBS-SRM 1635, SARM 19, SARM 20) and plant materials (NBS-SRM 1575, NBS-SRM 1547, NBS-SRM 1515, BCR Trace Elements in Rye Grass, IRNAT Lucerne), plus an in-house peat reference material (Ontario Geological Survey OGS 1878P). Standard deviations (in ppm) estimated as pooled variances from these measurements are: SC (0.0111, Br (0.4), Ca (701, Ba (21, Na (6.51, Fe (17). For those elements for which certified values exist (Fe, Ca, Na, Ba) the measured values were within 90% and 110% of the certified values. The elements Si, Ti, Al, Fe, Mn, Mg, Ca, Na, K, P, S and Cl were analyzed by wavelength-dispersive X-ray fluorescence spectroscopy (Geological Institute, Ukrainian Academy of Sciences, Kiev). Pressed pellets were prepared adding boric acid as a supporting matrix. Standardization was accomplished using certified reference materials. XRF has also been used in previous studies of peats (Metso et al., 1976; Giirres, 1991). Ash was determined following the ASTM D 2974 procedure (Andrejko et al., 1983). One gram of previously dried, milled and homogenized peat was placed in a preheated, acid-washed porcelain crucible and dried again at 105°C for 4 h. After 10 min cooling in a desiccator, dry weight was determined to 1 mg. Ashing was accomplished by heating at 550°C overnight. After cooling in the desiccator for 15 min the crucible was weighed again, and the ash content (weight percent) calculated. The (absolute) standard deviation of the method is 0.074% ash (estimated from 70 duplicate and triplicate analyses). Acid-insoluble ash (AIA) was determined by digesting a known amount of 550°C ash for 15 min in 6 N HCl while heating. The digest was filtered through an ash-free paper filter, which was then ashed in an acid-washed porcelain crucible and the amount of AIA weighed. The small amount of 550°C ash available for the determination of AIA led to errors > 10% for samples with low AIA contents. 2.5. Mineralogical analyses For mineralogical analyses the residue obtained from the ash determination was cleaned with 0.1 N

P. Steinmann,

W. [email protected]/

Chemical

HCl. Prior to examinations under the optical microscope the very fine fraction (less than about 5 km) was removed. Clove oil was used as an immersion liquid. The refraction index of clove oil is close to that of quartz, thus it is useful for identifying other minerals when quartz is so abundant. For scanning electron microscopy the oil from the specimen used under the optical microscope was washed off with methanol and the glass slide was coated with aluminum. Thus, both analytical methods could be used to study the same grains. Energy-dispersive X-ray analyses (SEM-EDAX) helped to identify very small or opaque grains. 2.6. Normalization oj‘the chemical profiles to scandium In order to compensate for the natural variation in the amount of dust and rain supplied to the bog, the chemical profiles were normalized to SC. Scandium is a conservative (very insoluble) element hosted in a variety of ferromagnesian minerals and is widely dispersed in crustal rocks (Frondl, 1970). Scandium has been found to behave as conservatively as Ti in the peat profiles studied (Shotyk, 1996b). The rea-

Geology

31

138 (1997125-53

sons for choosing SC over other conservative elements like Ti or Zr are the low detection limits for SC afforded by INAA (the SC concentrations in the two peat cores were always at least 10 times higher than the detection limit of 0.01 ppm), and the wide distribution of SC in soil-derived aerosols (in contrast Ti and Zr are mainly concentrated in ilmenite and zircon, respectively). Here the concentration of SC in the peats is used as a quantitative indicator of the natural variation in lithogenic inputs to the bogs. For a given element X any changes in the X/SC ratio in the profile indicates either a change in the relative amount of element X supplied to the peat (e.g. an additional source), or an enrichment or depletion due to postdepositional redistribution of X.

3. Results 3.1. Relative abundance of the elements Fig. 3 shows the concentrations of elements in the ombrogenic peat layers normalized to the average composition of shale (Krauskopf, 1982). Shale, a

/

??

EGr

I

ki

i[ /-

I

I””



1

I””

"

t

Element Fig. 3. Enrichment factors (EF) for various elements in ombrogenic peat (i.e. O-40 cm at TGe and O-100 cm at EGr) versus average shale. Enrichment factors were calculated by dividing the element/%-ratio of peat by the corresponding element/&-ratio of shale; i.e. (X,/X,)/&,/SC,), where X, and X, are the concentration of a given element and SC, and SC, are the scandium concentrations of peat and shale, respectively.

P. Steinmann, W. Shotyk/

32

Chemical Geology 138 (1997) 25-53

weathering product of crustal rocks, was chosen for comparison because much of the inorganic input to ombrogenic peats is derived from the weathering of crustal materials. The same general trend is seen if the concentrations in the peats are normalized to typical crustal abundance. Elements of the main groups VB to VIIB (i.e. P, As, Sb, S, Se, Cl and Br) plus Zn are strongly enriched in peat versus average shale; this list includes many of Goldschmidt’s ‘biophile’ elements and may reflect the affinity of living organisms for these elements. Manganese appears to be enriched relative to iron, but this may be due to overestimation of Mn which was generally close to the detection limits by INAA.

Table 1 Average composition of peat core TGe li (O-84 cm, dominantly minerotrophic peats) and peat core EGr 2f (O-102 cm, entirely ombrotrophic peats) Element

Minerotrophic TGe

Na K

@pm)

Mg Ca Sr Ba

125 487 314 3750 16.5 18.7

Si Al Fe

5430 2200 1920

Ti SC P S Cl Br AS co Cr cs Hf La Sb Se Zn Ash

93 0.44 598 1640 357 10.2 5.00 0.96 3.5 0.14 0.21 1.73 0.28 0.29 14.1 3.9%

peat,

Ombrotrophic EGr (ppm) 127 392 170 879 7.6 11.6 5790 2600 497 96 0.39 420 1410 287 11.0 0.96 0.54 2.5 0.12 0.18 1.54 0.24 0.28 13.7 3.00%

peat,

+EGr --t--TGe

F

Fig. 4. Distribution of several elements in ombrogenic peat and pore water, calculated from the composition of the ombrogenic peat layers of each bog and the typical composition of a corresponding volume of pore water (cf. Steinmann and Shotyk, 1997).

The following elements are significantly more abundant at TGe compared to EGr: Ca, Fe, As, Sb, Zn, Cr, Mg, Sr, Ba, Co, P, S, Cl, Br. The difference is especially pronounced with respect to Ca, and probably reflects the contributions from non-atmospheric processes at TGe even in the uppermost 40 cm of Sphagnum-dominated peat. At EGr, Fe is clearly depleted relative to average shale. Sodium is also depleted relative to average shale, possibly because of the ease by which it can be leached from the peat in exchange for H+ (Shotyk and Steinmann, 1994). The average concentrations of inorganic elements in the two analyzed peat cores (integrated over the entire core) are given in Table 1. The major difference between the exclusively ombrogenic peats at EGr and the dominantly minerogenic peats at TGe is the higher abundance of Ca and Fe at TGe. Calcium and Fe are also the dominant dissolved metals in the deeper pore waters at TGe (Steinmann and Shotyk, 19971, suggesting that both of these elements are supplied to the peat subsequent to the reaction of acidic pore fluids with their host minerals in the ash fraction; this process is discussed in more detail later. In addition to Ca and Fe, the average concentrations of Sr and Ba are higher in the peats at TGe compared with EGr. Arsenic is 5 times more abundant in the TGe core compared to EGr and the natural enrichment of As at TGe is described in more detail elsewhere (Shotyk, 1996a). The distribution ratio (concentration in peat [mg/kg]/concentration in pore water [mg/ll> for a

P. Steinmann, W. Shotyk/ Chemical Geology 138 (1997) 25-53

variety of elements is shown in Fig. 4 and reveals a wide range (almost 3 orders of magnitude) in preference for the solid or iaqueous phase. Elements which prefer the solid phase are expected to be less mobile in the peat profiles compared with elements which prefer the aqueous phase. Aluminum, despite the low pH of the waters and the abundance of dissolved organic ligands, exhibits the highest ratio, i.e. the highest preference fclr the solid phase. At the other extreme, the Na concentrations in the solid phase are only 100 times higher than the average pore water concentrations, suggesting much greater potential for post-depositional migration.

and showing corresponding peaks. In other words, the peaks at these depths were found using both physical (ash content) and chemical (SC by INAA and Si by XRF, respectively) analytical methods on three different subsamples. Moreover, a second core (21) from the same bog shows a very similar ash profile: all but the peak at 11 cm are found in both ash profiles. The largest peaks in the ash profile (35, 50 and 89 cm) correlate with an increased peat density (Fig. 5). The higher degree of decomposition of the peat in these layers is shown not only by the density, but also the lower rubbed fibre content and the lower percentage of extractable pore water (Shotyk, unpublished data). Note that the ash contents are too low throughout the profile to have had a significant effect on the measured peat densities. In other words, variations in the amount of mineral matter in the profile could not have contributed to the measured differences in peat density. The ash contents at TGe are less than 3% only to a depth of 30 cm. Beyond this depth the ash content increases to reach more than 5% at depths > 70 cm (Fig. 5). Ash contents > 5% are characteristic of minerogenic peats (Naucke, 1990). The ash profile at TGe, therefore, is consistent with the view that the peat layers deeper than approximately 30 cm are

3.2. Ash and density Most of the EGr samples showed ash contents ranging from 1 to 2.5%, values which are characteristic of Sphagnum peats (Vuorela, 1983; Tolonen, 1984; Naucke, 1990). The ash profile of peat core 2f shows four pronounced peaks at depths of 11, 35, 50, and 89 cm with ash contents up to 8% (Fig. 5). Minor peaks are recognized at 62, 68, and 80 cm; evidence supporting the identification of these minor peaks in ash content are given by the SC and Si profiles, measured using two independent methods, EGr 2f

EGr 2f

0.5

SC (mm)

1

3 6 9 AIA, ash (%)

TGe Ii

EGr 21

0.1

33

0.2

P (g cm-7

0.5 SC (rwm)

TGe li

TGe li

1

3

6

/

9

AIA, ash (%)

1

0.1

I

0.2

P (9 cm- 7

Fig. 5. Profiles of SC, ash and AIA (dotted line), and density of dry peat for Et3 and TGe. The horizontal lines indicate the peat layer accumulated over the last 100 years. Open triangles show the most significant ash peaks which have corresponding SC peaks. Error bars show + 1 standard deviatl!on.

34

P. Steinmann, W. Shotyk/Chemical

essentially minerogenic (Shotyk and Steinmann, 1994; Shotyk, 1996a; this report). At TGe the ash profile shows less pronounced peaks than at EGr. The clearest peaks, found in the SC profile as well, are around 8, 41, and 57 cm. Again, these peaks correlate with increased peat density and more advanced decomposition (as indicated by differences in the amount of expressible pore water). 3.3. Mineralogical

composition

of the ash

Identified silicate minerals and biogenic particles are listed in Table 2. Among the mineral grains quartz by far is dominant. Potassium feldspar, albite and micas occur as minor minerals, whereas many other minerals are found in trace amounts. Several shards of volcanic glass were found, but no distinct volcanic ash layers were recognized. Point counting

Table 2 Minerals and biogenic particles found in ombrotrophic

Geology 138 (1997) 25-53

of grain mounts of ash from the ombrogenic layer at TGe (O-30 cm) showed an important contribution (30-70%) of biogenic grains, with a high abundance of diatom tests near 30 cm. The mineral fraction is composed of 60-90% quartz, 5-15% feldspar, 515% muscovite, and 5-10% others (mainly biotite, epidote, pyroxenes, and opaques, but with a few grains of tourmaline and zircon). Most other samples, including those with high ash content at EGr 35 cm and 50 cm, showed similar mineralogical compositions. Due to the generally small but varying grain size and due to the overabundance of quartz and opaline silica it was not possible to quantify small changes in mineralogy by point counting. Powder X-ray diffraction analyses on samples of 50-100 mg showed tiny peaks due to feldspars and micas, but accurate quantification of these minerals versus quartz was not

peats: identification

based on optical microscopy

and SEM (EDS-spectra)

Mineral or biogenic particle

Description

Quartz Diatoms Dhrysophyta Rhizopod tests Opaline phytoliths Volcanic glass K-feldspar Albite Epidote Zoisite (?) Augite (?) Diopside Na-pyroxene Aegirinaugite (?) Tourmaline Muscovite Biotite Chlorite? Zircon Rutile (?) Anatase (?) Brookite (?) Ilmenite (?) Titanite Fe-oxides Cr-spine1 (?) Andalusite (?) REE-mineral (cerite?) Carbonaceous particles

clear, bluish or faintly clouded elongate forms (Pint&aria? Eunotia?) 50 pm remains; small oval to round flakes; l-5 km spheres; 20-25 pm various forms; e.g. dumb-bell shaped brown or greyish, irregular shards with inclusions of tiny greenish pyroxenesl?); often clouded grains of slight carnation color; 20-50 pm colourless (bluish), rounded, sometimes twinned; 10 pm greenish, transparent, weak pleochroism colourless irregular grains green irregular grains pale green, elongate green irregular grains green irregular grains elongate shape, often with terminal faces: greenish: 4-10-50 pm rel. large platy grains; greenish; often gray rims; 35 pm yellowish to reddish-brown platy grains; 30 pm greenish-brown, platy short prismatic rounded grains with bipyramidal terminations reddish grains brownish euhedral grain; tabular, tetragonal fresh-looking olive green grain opaque grain greenish opaque to dark reddish small grains; 10 pm; Fe-rich spheres; 3 pm opaque colourless grain subround colourless grain with high R.I. opaque grain

K > Ca

P. Steinmann, W. Shotyk/Chemical

of ash in both profiles. At EGr, AIA accounts for 23-86% of the ash (mean: 49%), whereas at TGe, the range is 1l-56% (mean: 41%). Thus, the inorganic solids which must have been present originally in the organic (biological) fraction of the peat account for an average of 50% of the ash at EGr and 60% at TGe. At TGe the proportion of non-mineral inorganic solids is higher than at EGr and increases in importance with depth, probably reflecting the addition of dissolved solids (primarily Ca and Fe) to the increasingly minerogenic peats by diffusion and adsorption; this process is discussed further later. Note that the peaks at EGr 35, 50, and 89 cm are largely due to increased concentrations of AIA (Fig. 5). These peaks, therefore, could not have been caused by biological processes operating within the bog. Instead, they represent higher concentrations of soil-derived silicate dust, either the result of an increase in the rate of supply, changes in the rates of peat accumulation or both.

possible. The mineral proportions observed microscopically are essentially uniform with depth. In other words, there are no discernible changes in relative mineral abundance with depth in the profiles. The ash remaining after burning the peat at 550°C consists of oxides, hydroxides, carbonates, sulphates, mineral grains, and particles of biogenic origin. The presence of oxides/hydroxides like CaO or KOH was inferred from the high pH (ca. 13) measured in aqueous extracts of these ashes. Carbonates were revealed by effervescence upon adding dilute acid. Iron oxides imparted a reddish colour to some ashes, and hematite was found by X-ray diffraction analysis of reddish peat ash. The presence of sulphate was indicated by the sulphate concentration in aqueous extracts of peat ashI. Much of the oxide, hydroxide, carbonate and sulphate fractions may have formed during ashing from cations which were originally held in the organic fraction, and are therefore artifacts of the combustion process. The combined contribution of silicate minerals and biogenic silica particles to the ash content is given by the acid-insoluble ash (AIA) fraction. As shown in Fig. 5, AIA makes up an important portion

3.4. Major element cations: Al, Si, Ca, and Fe Aluminum and Si (Fig. 6) are the most abundant cations in the peat samples. The Al and Si concentra-

TGe li

TGe li

EGr 2f

EGr 2f

35

Geology 138 (1997) 25-53

100~ f, / /I”, , , /,1 0.2

0.4

Al (%) Fig. 6. Aluminum diatoms.

and Si concentration

0.6

1

2

si (%)

0.2

0.4

Al (%)

0.6

0.2

0.6

1

si (%)

profiles of the peat core at ECir and TGe. The peak indicated (d) corresponds

to an abundance

of

36

P. Steinmann, W. Shotyk/

Chemical Geology 138 (1997) 25-53

tion ranges at EGr and TGe are similar, but the distribution of these elements differs: at TGe there is a progressive increase with depth for both elements. Both Al and Si correlate with ash content and with SC (correlation coefficient 0.98) in both profiles. The Si peak near 30 cm in the TGe core is exceptional in that it has no corresponding ash peak. However, it does coincide with an abundance of diatom tests seen in the microscope. The Si/Sc ratio (Fig. 12) provides additional evidence that the peats at this depth have an atypical source of Si. Moreover, the Si concentrations in the pore waters between 30 and 40 cm typically are higher than at other depths, suggesting that the diatom tests represent a more soluble form of Si. At EGr iron concentrations range from 200 to 400 ppm at the very top of the profile and in the samples below 39 cm, while higher concentrations (up to 900 ppm) are found between 3 cm and 39 cm. Five peaks are identified by dark arrows in Fig. 7. These peaks in the Fe profiles were found using two independent analytical methods (INAA and XRF) on a different set of subsamples. All of these Fe peaks correspond

to relative minima in the SC profile and they are emphasized when the Fe profile is normalized to SC. In contrast to the EGr profile, at TGe the Fe increases tenfold from the top (300 ppm) to the bottom (0.3%). Two more pronounced peaks (also identified independently in the XRF profile) are at 4 and 52 cm (Fig. 7). Below 30 cm at EGr the Ca concentrations are relatively constant (- 600 ppm), while above 30 cm they are 2-3 times higher (Fig. 7). At TGe Ca increases from a subsurface minimum of approximately 900 ppm to almost 6000 ppm at 82 cm (Fig. 7). The steady increase of Ca towards the bottom indicates that Ca is an important component of terrestrial inputs to the minerogenic peats at TGe. The Ca profile in the solid phase at TGe is paralleled by the Ca concentrations in the pore waters which increase from 2 mg/l at around 20 cm to 100 mg/l at a depth of 1.8 m (Steinmann and Shotyk, 1997). 3.5. Less abundant cations: Mg, Ba, Na, and K Below the 35 cm ash peak at EGr, the Mg concentrations are approximately 100 ppm and are

EGr 2f

EGr 2f

TGe li

TGe li

a

1 a

0.05

Fe (%)

0.1

0.1

0.2

Ca (%)

0.3

0.2 Fe (%)

0.4

0.2

0.4

0.6

Ca (%)

Fig. 7. Iron and Ca concentration profiles for EGr and TGe. The filled triangles show the most significant observed in independent XRF measurements. The open triangles indicate the ash peaks (cf. Fig. 3).

Fe peaks, which were

P. Steinmann, W. Shotyk/Chemical Geology 138 (1997) 25-53 EGr 2f

EGr 2f

TGe li

EGr 2f

I

’j ’

37

TGe li

I

I I

TGe Ii

I ,

1

/

10

20

._

L

w 500

h

1000

5

(fwm)

10

15

IO

Sr (rwm)

Fig. 8. Magnesium,

20

30

400

Ba (wm)

Mg (wm)

Sr and Ba concentration

effectively constant with depth (Fig. 8). Above 15 cm the concentrations of Mg increase progressively towards the top, with the maximum concentration approaching 900 ppm. Normalizing to SC (profile not

EGr 2f

800

w 10

20

30

Ba (wm)

profiles at EGr and TGe.

shown) emphasizes the strong Mg enrichment in the uppermost 10 cm which is apparently due to plant uptake of this essential nutrient. The gradual changes which are seen in the uppermost part of the Mg

EGr 2f

TGe li

TGe ii

I

/

I

I

I

_

20

g s. 2. $j.

40

-O

60

80

I

r

200 Na (mm)

400

1

I

1000

30

Sr (wm)

I

-

2000

K (pm)

Fig. 9. Sodium and K concentration

100

200

500

Na bwm) profiles at EGr and TGe.

T

1500

K hvm)

38

P. Steinmann, W. Shotyk/ Chemical Geology 138 (1997) 25-53

profiles contrast with the more erratic distribution of Ca; this suggests that either the metals are being supplied independently, or more likely, that they behave differently within the profile. The decreasing Ca/Mg ratio toward the top of the EGr profile (Fig. 2) is a consequence of this strong Mg enrichment at the surface of the bog. At TGe the Mg concentrations increase both upwards and downwards from a minimum of about 200 ppm at 30 cm (Fig. 8). While the Mg/Sc profile suggests strong bioaccumulation at TGe also, the concentrations of Mg in the peats at the top of the TGe profile are only one-half the EGr values; this may indicate that Mg is more available to plants growing at TGe, and therefore it is concentrated to a much smaller extent. The concentrations of Ba are higher in the deeper sections of TGe compared with EGr (Fig. 81, but Ba generally follows SC in these sections of both cores. Barium concentrations are higher at the top of both profiles. Normalizing to SC reveals relatively constant values in the lower half of the cores but increasing values above approximately 30 cm, indicating a Ba enrichment in the surface layers. Normaliz-

ing the Ba profiles to Ca (profile not shown) emphasizes the most prominent Ba peaks (EGr 35 cm, 50 cm; TGe 41 cm, 57 cm). Assuming that atmospheric inputs had a constant ratio of Ba/Ca, the peaks in Ba/Ca indicates depletion of Ca relative to Ba in these zones. In contrast, towards the top of the profiles Ba is slightly enriched relative to Ca. Regarding the ‘near-surface-enrichment’ of the alkaline earths the sequence Mg > Ba > Ca > Sr is observed. This order is also apparent from nearsurface peat profiles of these elements from other bogs (Market-t and Thornton, 1990). The Na concentrations at both bogs roughly follow the ash profiles with the exception of a large additional peak at the surface at TGe (Fig. 9). The Na/Sc ratios at EGr are constant from 40 to 100 cm but increase fivefold from 40 cm to the surface with two peaks near 15 cm and 0 cm. At TGe Na/Sc shows a strong increase from 10 cm to the surface. The potassium profiles are very similar to the Na profiles: K parallels the ash profiles with the exception of a strong near-surface maximum at both bogs (Fig. 9). For example, below 21 cm at EGr (i.e. below the biologically active surface layer) there is a

EGr 21

EGr 2f

TGe li

TGe li

---e

i



I



I

t

500 15

300 Cf (pm)

10

15

20

Br O-wm)

300

500

Cf (wm)

Fig. 10. Chlorine and Br profiles for EGr and TGe.

5

15 Br (fwm)

P. Steinmann, W. Shotyk/

EGr 2f

I

39

Chemical Geology 138 (1997) 25-53

TGe li

EGr 2f

TGe li

I

5

80

200

600 P (ppm)

Fig. 11. Phosphorus

s

P @pm)

(%)

and sulphur concentration

strong correlation between K and Ti (r2 = 0.95, n = 27) which suggests that most of this K resides in the mineral fraction.

very

3.6. Cl, Br, S, and .P Chlorine is rather similarly distributed in the two peat profiles, with Cl concentrations ranging from 200 to 400 ppm at :EGr and from 300 to 500 ppm at TGe (Fig. 10). At both bogs the highest Cl concentrations are found near the top with a minimum about lo-15 cm b’eneath, apparently reflecting the uptake of Cl by the living plants. Bromine concentrations at EGr range from 6 to 19 ppm (average 11 ppm); at TGe the variation is smaller with concentrations ranging from 7 to 14 ppm (average 10 ppm) (Fig. 10). The Br profiles of the two bogs have some common features distinct from the Cl profiles. First, a subsurface Br maximum is found below the Cl maximum in both cores. Second, deeper in the profiles at both bogs (but more pronounced at EGr) the Br profiles follow the density profiles. In other words, Br is enriched in zones of very low pH (subsurface) and in highly decomposed peats. This is consistent with a fixation of Br in organic matter, as it is generally observed in soils

s

(%)

profiles for EGr and TGe.

(Kabata-Pendias and Pendias, 1992). The efficiency of Br retention in a peat bog seems to depend on the pH and the availability of decomposed peat, two factors which may vary during the development of a bog. The most significant input of Cl and Br is rainwater because substantial inputs of particulate sea salts are unlikely in these continental sites. Furthermore, input of Br from car exhausts is less than 15% of total Br (calculated using a Br/Pb ratio of car exhaust of 0.39 (Foltescu et al., 1994) and the lead concentrations in the peats of EGr and TGe (Shotyk, 1996b). The S and P profiles for the two bogs are given in Fig. 11, with concentrations generally higher in the deeper layers of TGe compared to EGr, and maximum concentrations found in the upper 40 cm.

4. Discussion 4.1. Processes controlling ash content and peat bulk density

Both cores show some prominent ash peaks and corresponding peaks in the profiles of related ele-

40

P. Steinmann, W. Shotyk/Chemicul

Geology 138 (1997) 25-53

ments. At EGr these are most notably at 35 cm and 50 cm, and at TGe at 41 cm and 57 cm. These peaks occur in a zone of increased density which exists in the middle of both cores (Fig. 5). There are several possible explanations for the ash peaks. The ash contents of peats at the time of their formation are a function of both peat accumulation rate and rate of dust input. The original ash concentration may be increased by the decomposition of organic matter or precipitation of new, diagenetic solid phases. Alternatively, mineral dissolution can decrease the ash content. Extraordinary events like forest- or moorfires, deposition of volcanic ash (Persson, 1971; Dugmore, 1989; Zoltai, 1989; Hodder et al., 1991; Dugmore et al., 1992; Blackford et al., 1992) or Saharan dust may result in discrete peaks in ash content. The mineralogical analyses, however, failed to reveal any volcanic ash layers. Also, no charcoal layers (indicating nearby fire events) were found (Andreas Fankhauser, pers. commun.). Thus, the peaks in ash content appear to have resulted primarily from discrete periods of reduced rates of peat accumulation and possibly higher rates of dust input. This interpretation is supported by the presence of denser peats at the corresponding depths. Dense, more decomposed peat layers form during relatively dry periods when peat growth is slow and

decomposition (oxidation) enhanced (Ikonen, 1993). Such dry periods may be regional effects such as climate change, or local influences, such as peatland drainage for peat cutting or forestry. The ash peaks are then caused by slow rates of peat accumulation, with more time available for dust input; enhanced rates of dust deposition may also be a consequence of dry climate. Following a dry phase, renewed moist conditions will again lead to increased growth rates of peat-forming plants and to the formation of a less decomposed and less dense peat, with lower ash contents. The most prominent ash peaks in each core, EGr 35 cm and TGe 57 cm, agree rather well with the increase in Pinus pollen. The peat samples in which Pinus begins to increase are EGr 2h14 and TGe 1120; these samples have since been dated using radiocarbon at 560 f 60 years BP (B-6601) and 680 f 60 years BP (B-66001, respectively. It appears, therefore, that afforestation with Pinus began in this area as long ago as the 13th or 14th century. The reduced peat accumulation rates and higher ash contents may reflect a temporary drying of the bog surfaces in response to local land use practices (such as extensive forest clearing), as well as higher dust inputs from local human activity. While much remains to be discovered about the details of local land

Table 3 Net accumulation

10%

rates at EGr and TGe; errors (1 a) are approximately EGr

Years: Depth:

Dry

Ash SC

Al

peat(a)

AD 1892-1991 O-30 cm

Rain water

TGe

(mgm-2 a-I 1

2110 BP-AD 1892 30- 102 cm (mg mm2 a-‘)

AD 1893-1991 O-36 cm (mg m

240 4715 0.039 248

32 1085 0.015 103

271 6914 0.052 263

-2

a-l)

1730 BP-AD 1893 36-84 cm (mg m-’ a-‘)

(mg m

33 1545 0.019 93

Si Na K Mg Ca

785 29 89 74 372

220 4.2 13 3.9 21

1145 22 126 75 801

201 4.9 17 11 139

175 113 50 450

Fe P S Cl

126 78 290 65

16 15 48 10

225 158 530 94

82 20 50 12

898 388

-2

,-I

)

P. Steinmann,W. Shotyk/ ChemicalGeology 138 (1997) 25-53

use activities, the ash peaks at EGr (35 cm) and TGe (57 cm) as well as the increase in Pinus appear to be the result of local environmental changes. Other Central European studies have also found ash peaks at depths corresponding to increases of Pinus or Picea due to afforestation (Menke, 1987; Hijlzer and Hijlzer, 1987, 1988a,b; Giirres and Bludau, 1992). 4.2. Peat bog inventories

of atmospheric

deposition

Age dating of the cores allows an inventory to be calculated for the major elements which, in turn, may be compared with the estimated total inputs to the bogs. The available age dates (cf. above) allow two time frames to be compared: the average annual rate of accumulation over the long term (2110 years at EGr and 1730 years at TGe) and the average annual rate of accumulation during the past century (30 cm at EGr and 36 cm at TGe). Fluxes calculated for these modem (past century) and ancient (pre1900) accumulations are given in Table 3. 4.2.1. Peat accumu,!ation rates The rate of organic matter accumulation during these time frames is very similar at the two sites (Table 3). At EGr, for example, the net rates of dry matter accumulation are 240 g mP2 a-’ (averaged over the past 100 years) and 32 g me2 a-’ (averaged over the previous 2050 years). The corresponding values for TGe are 271 g mW2 a-’ and 33 g me2 a-’ for the past 100 and previous 1670 years, respectively. The average annual rate of peat accumulation is much higher during tbe past century compared to the past two millennia because the former rate represents mainly fresh, poorly decomposed peat and includes plant material living in the bogs at the time of sample collection. 4.2.2. Rates of atmospheric dust deposition Because plants and animals have no biological requirement for SC and because SC behaves conservatively in peat profiles, this metal is a reliable indicator of the recent increases in soil dust inputs. In the modem peats (accumulated during tbe past century) the SC fluxes are in good agreement: 39 + 4 p,g mW2 a-l at EGr compared with 52 f 5.5 p.g m-2 a-1 at TGe (Table 3). Also the long-term SC

41

flux is similar at both sites: 15 f 1.5 kg me2 a-’ at EGr versus 19 f 2 p,g rnM2 a-l at TGe. At EGr where the entire peat core (102 cm) is exclusively ombrogenic (as shown by independent assessment of the trophic status of the cores; cf. Shotyk and Steinmann, 1994; Shotyk, 1996a), the SC inventory, therefore, reflects atmospheric deposition only. Since the two bogs are in close vicinity in a relatively open, flat terrain, atmospheric inputs can reasonably be expected to be essentially identical at the two sites. The slightly higher SC flux at TGe shows, therefore, a small but measurable non-atmospheric source of solid inputs at TGe. Because the TGe profile is physically isolated from groundwater sources external to the bog, the most likely source of this extra SC to the peats at TGe is mineral matter incorporated from the underlying sediments by terrestrial processes. Comparison of modem accumulation rates with long-term accumulation shows that modem SC fluxes are 2.6 times (EGr) and 2.7 times (TGe) higher than the annual fluxes averaged over the long-term (Table 3). The increased modem SC flux can be attributed to land use changes such as large-scale forest clearing and enhanced soil erosion due to agriculture (Vuorela, 1983; Tolonen, 1984; Damman et al., 1992). 4.2.3. Apparent accumulation

Ca and Fe accumulation of ash

rates and

In contrast to SC, the fluxes of Fe and Ca are considerably higher at TGe than at EGr, both in the upper part (past century) and, especially, in the older peats. The long-term Ca flux at TGe exceeds that at EGr by nearly 7 times, and the Fe flux by more than 5 times. In the older peats the SC accumulation rates show the presence of a non-atmospheric particulate fraction which was most likely incorporated into the peat (at TGe) by a terrestrial process. Given that the long-term SC accumulation rates at TGe are only 30% higher than at EGr, the physical incorporation of mineral matter could at most explain only a small fraction of the excess Ca and Fe in the peats at TGe. Some other process, therefore, must be active and allows Ca and Fe to become enriched in the TGe core far out of proportion with the amount of mineral matter present in the peat. The simplest possible explanation for the enrichments is chemical diagene-

P. Steinmann, W. Shotyk/Chemical

42

sis subsequent to peat formation; this would most likely include chemical reaction of the basal mineral sediment at TGe with the pore fluids, release of Ca and Fe to the pore waters, molecular diffusion of Ca and Fe upwards in response to the concentration gradient, and accumulation of Ca and Fe in the peats by cation exchange and surface complexation, respectively. Diffusional input of Fe and Ca at TGe is consistent with the pore water profiles of the two elements which show increasing concentrations with depth. At TGe the pore water concentrations typically increase from approximately 1 ppm Ca near the surface to 3 ppm Ca at a depth of 30 cm; up to 6 ppm are reached at 60 cm and concentrations as high as 70 ppm and more are found in the underlying layers (140 cm) (cf. Steinmann and Shotyk, 1997). Similarly, the Fe concentrations rise from about 0.3 ppm near the surface to 0.8 ppm Fe at 30 cm, 1.7 ppm Fe at 60 cm and 24 ppm Fe at 140 cm. In contrast at EGr the concentrations of Ca and Fe are much lower (approx. 1 ppm Ca and 0.1 ppm Fe) and rather constant down to a depth of about 2.5 m. During the past century, the average rates of Ca and Fe accumulation in the peat at TGe have been approximately twice those for the same time period at EGr. Even in the uppermost 36 cm of the TGe profile, therefore, there has been a significant influx of these two metals from underlying peat layers. 4.2.4. Comparison of ombrogenic bog inventory with rainwater

inputs

In contrast to Sc, some of the major elements are retained in the peat profiles to a very limited extent. For example, consider Na and Cl whose rates of supply can be estimated using available data for average annual precipitation and the 8-year average rainwater composition in the Franches Montagnes region (Shotyk and Steinmann, 1994). Assuming that the concentrations of Na and Cl in rainwater have not changed significantly, the long-term (past 2 millennia) rates of Na and Cl accumulation in the two profiles (Table 3) amount to only 2.4% of the Na and Cl supplied to the bogs. Even if modern rainwater contains higher concentrations of Na and Cl, it is still clear that most of the Cl- and Na+ has not been retained by the peat; instead, these ions leave the bog surface by laterally flowing waters in the poorly decomposed surface layers. For the other major ele-

Geology 138 (1997125-53

ment cations, long-term retention is estimated to represent 4.6% of the Ca, 7.8% of the Mg, and 11.6% of the K. In contrast to the long-term average, the amounts of K, Mg, and Ca supplied to the ombrogenic bog by rainwater are comparable to the amounts retained in the younger section of the bog (Table 3); this section of the core (top 30 cm) includes the biologically active layer. 4.3. Geochemistry

of major elements

4.3.1. Al and Si

Normalized to SC the Al profiles are relatively flat at TGe (Fig. 12) with no apparent change with depth (time). This relatively constant Al/St profile reflects the common origin (soil-derived mineral dust) of Al and SC, the generally slow rate of aluminosilicate mineral dissolution (Casey et al., 19911, even at pH 4, and the low solubility of Al (Baes and Mesmer, 1976). At EGr, the AI/SC ratio is slightly higher in the deeper layers compared to the surface layers, and the values are generally higher than at TGe (Fig. 12). Assuming that the mineralogical compositions are the same at the two sites, some Al appears to have been lost from the TGe profile. One possible explanation is the higher average depth to the water table at TGe compared to EGr. At TGe there may be more fluid flow, providing more opportunity for leaching the dissolved forms of Al which are generated during mineral weathering. Second, the higher concentrations of Ca and Fe in pore waters of TGe might help to displace Al bound at exchange sites at TGe. The Si/Sc ratio is nearly 6 times higher in the upper parts than in the lower parts of both profiles. The higher values in the upper sections of the profiles is due to increased input in biogenic opaline silica from the opal phytoliths of agricultural plants (e.g. cereals) in the last 100 years or so; these represent a substantial addition of a SC-poor Si-fraction to the peats. The Si/Sc profiles also show peaks corresponding to the ash peaks at EGr 35 cm, 50 cm and 89 cm, as well as (less pronounced) at TGe 57 cm. The peaks in ash content, therefore, are relatively rich in Si. At TGe a peak in the Si/Sc profile found at 30 cm can be attributed to an abundance of diatoms (as revealed by optical microscopy and SEM images of the ash fraction).

P. Steinmann, W. Shotyk/

43

Chemical Geology 138 (1997) 25-53

4.3.2. Ca and Fe in the ombrogenic layers The distribution of Ca and Fe in the two peat profiles differs markedly from the SC distribution as is shown in Fig. 12. Departure from the SC profile may be either due to changes in composition of inputs or post-depositional redistribution. Here evidence is provided which indicates that the distribution of Ca and Fe in these cores has undergone significant post-depositional redistribution. At EGr the Fe profile shows no correlation with the SC or the ash profile. In fact, in detail the SC and Fe profiles are negatively correlated; consequently the lowest Fe/SC ratios are found at depths corresponding to peaks in ash and density (e.g. 11 cm, 35 cm, 50 cm, and 89 cm; Fig. 5). It has already been argued that episodic dry events must have given rise to the increased bttlk density and ash contents at discrete depths. The regular variations in Fe/SC which follow these changes, therefore, are probably also the product of the alternating dry and wet cycles. Iron has become depleted relative to SC in the more decomposed peat layers, and subsequently enriched relative to SC in overlying layers. The most likely geochemical explanation for the apparent Fe dynamics is the effect of alternating oxic-anoxic

conditions on the redox state, solubility, and speciation of Fe (e.g. Hem, 1972). During the dry phase with slower peat accumulation, Fe solids at the oxygenated bog surface are expected to be stable and, together with SC and other lithogenic elements are relatively enriched in the peat (Fig. 13a). When conditions become wetter again, this well decomposed layer will become strongly anoxic, allowing Fe to be released preferentially from the original (inorganic) solid phase, perhaps by reductive dissolution. As this ferrous Fe diffuses upward in response to the pore water concentration gradient which has developed, it will subsequently be oxidized near the surface of the bog and become strongly adsorbed to the peat (Fig. 13b). The reduction and oxidation of Fe in relation to the pH and Eh of mire waters was studied in detail by F’uustj%rvi(1952) who noted that ‘oily Fe films’ only precipitate in oxygenated peatland surface waters when the pH is neutral to alkaline. Given the low pH of peat bog surface waters, precipitation of new inorganic Fe solids is an unlikely removal mechanism for the diffusing Fe(U). Instead, given the great stability of Fe(II1) complexes with natural organic ligands (e.g. Steinmann and Shotyk, 1997 and references therein), adsorption of

20 1

100-j

i

I

0.2 AI/SC

I

I

0.6

I

t

1

i

I

1

(x1000)

Fig. 12. Sc-normalized profiles for TGe and EGr: Al/%, 100 years at both bogs.

I

I

2 SilSc

/

'

t

3

I’,‘,’ /

2 Fe/SC

4

6

6

(x1000)

Si/Sc, Ca/Sc, Fe/SC. The horizontal lines show the peataccumulated in the last

44

P. Steinmann, W. Shotyk/Chemical

c ;;

d

time

Fig. 13. Conceptual model for post-depositional redistribution of Fe in a bog profile. (a) During a dry phase, the rates of bog growth and peat accumulation are limited, and peat decomposition is promoted: a relatively humified peat layer develops at the surface of the bog (shaded zone) in which both SC (solid line) and Fe (dotted line) become residually enriched. (b) Later, as peat accumulation begins again during renewed wetter conditions, Fe oxides in the humified layer may be reductively dissolved and Fe diffuses upward where it becomes organically bound in the oxidizing zone. (c) Owing to the stability of organic complexes of Fe, organic fixation is effectively permanent, and Fe is not remobilized when conditions again become reducing as bog growth continues.

Fe(II1) by the peat is the most likely sink for this solute. The sequence of events hypothesized here explains why the peaks in Fe/SC always slightly overlie (and therefore formed subsequent to) the peaks in ash content which are found in well decomposed peat layers and represent dry periods in the development cycle. This series of processes is not a continuous phenomenon, but rather is restricted to the dry, stationary phases of bog development. These patterns of Fe depletion and redeposition are stable and are remarkably well preserved as palaeo-horizons in the peat core, attesting to the great stability of Fe(II1) organic complexes (Fig. 13~). In addition to the two cores described here, peaks

in Fe overlying peaks in ash can be observed in other studies. Hiilzer and Holzer (198&a) present a short profile (1 cm resolution) from the Blindensee bog in the Black Forest. Their profile shows an ash peak around 14 cm where the peat is slightly denser and more decomposed, as revealed by the higher C content; an Fe peak occurs only 6 cm above this zone. A similar profile (Seemisse, also in the Black Forest) was described by the same authors (Hiilzer and Holzer, 1988b); here, an ash peak in a denser peat

Geology 138 (1997) 25-53

layer at 18 cm is followed by a Fe maximum 3 cm higher. In a third profile (Homisgrinde; Hiilzer and Holzer, 1987) an Fe peak is seen l-2 cm above an ash peak at 7 cm. Menke (1987) found an Fe peak at about 5 cm over an ash peak at 11 cm, and a large Fe maxima at 60 cm above an Al maximum at 87 cm. Gorres (1991) analyzed with 5 or 10 cm resolution, so that small-scale dislocation of Fe could not be seen. However, even in his profiles the correlation between Fe and ash is often negative. All of these profiles indicate that Fe relocation is both significant and frequent, with release of dissolved forms from ash-rich layers and subsequent fixation in nearby, overlying layers. The vertical extent of relocation is small, operating on a scale of cm to dm. In contrast to Fe, Mn is typically removed from the entire peat profile, with elevated concentrations restricted to the uppermost layers of the core which are both oxic and biologically active. An additional feature of the Fe/SC profile at EGr is the much higher values in the upper part of the profile, corresponding to the present century (see Table 3). Since redistribution of Fe from the lower half of the core toward the upper part is unlikely (this distance is beyond the typical range of Fe-relocation), the Fe enrichment was most likely caused by an additional input of a SC-poor Fe-bearing phase. One possible explanation is the increase in dust from road construction; limestone containing Fe as an abundant trace element could help explain the Fe enrichment. Industrial sources of Fe-containing dust also may have been involved. The Ca distribution at EGr is remarkably similar to that of Fe. The lowest Ca/Sc ratios are found at depths corresponding to ash-rich layers. Again, this indicates dissolution of Ca-bearing minerals and the redistribution and fixation of Ca in organic forms in surrounding peat layers. The similarities between Fe/SC and Ca/Sc argue in favour of adsorption/complexation as the removal mechanism for Fe(II1) in the oxic surface layer of the bog, as opposed to precipitation of an inorganic phase. Like Fe/SC, the Ca/Sc ratios are generally higher in the uppermost layers, corresponding to the last 100 years of deposition. It is unlikely that this represents a large-scale redistribution of Ca because the surface peat layers are acidic (pH 4); this may limit Ca adsorption (Shotyk and Steinmann, 1994). Again, the

P. Steinmann, W. Shotyk/Chemical

relatively high metal/St ratios in this section of the bog probably reflect an increase in the rate of deposition of calcareous dust during the present century. Given the imperfect retention of Ca (Table 31, it is clear that the increase in the ratio of Ca/Sc may not reflect quantitatively the increase in Ca deposition. In summary, within the ombrogenic peat core at EGr, significant redistribution of Fe and Ca has taken place on a small scale (cm to dm), but the more pronounced enrichments (Fe/SC and Ca/Sc) in the uppermost 35 cm are largely the result of increasing anthropogenic inputs. 4.3.3. Fe and Ca in the minerogenic layers At TGe which is predominantly minerogenic, two Fe/SC peaks as well as Ca/Sc peaks occur above peaks in the ash profile, corresponding to the minima in the Fe/SC and Ca/Sc profiles. Again, these changes indicate redistribution of both metals taking place over small distances. However, in contrast to the ombrogenic core at EGr, the Fe/SC and Ca/Sc ratios in general are much higher at TGe, especially in the lower half of the profile (Fig. 12). As discussed above higher contents of Fe and Ca at TGe are due to binding of Fe and Ca ions which entered the peat column as dissolved species via diffusion

Geology 138 (1997) 2.5-53

45

from the underlying sediments. Hill and Siegel ( 199 1) report similar enrichments of Fe and Ca, but in their case the peats are predominantly minerotrophic, and the enrichments clearly a response to groundwater flow patterns. The influence of the fluid-mineral reactions in the basal sediments on the chemistry of the overlying peat layers is more clearly seen when the peat profiles are viewed in their entirety. The TGe profile, for example, shows peaks in Fe/Al and Ca/Al at 100 cm (Fig. 14a); below this depth the amount of mineral matter continues to increase as the underlying sediments are approached, but the Fe/Al and Ca/Al ratios decrease (Al is used here as a conversative metal instead of SC because SC analyses are not available for the deeper samples). Thus, both Fe and Ca have become enriched in the peats, well out of proportion with their abundances in the mineral sediments. These profiles are consistent with the view that inorganically bound Fe and Ca in the underlying sediments and in ash-rich peat layers have been released to the pore waters and subsequently become fixed in the peats. Additional evidence for the diffusion of these cations is given by the pore water profiles which show the gradients of dissolved Ca and Fe at TGe and in the lower part of

mCa/AI 0

5

10 15 20 25 30 35 40

160 0

20

40

60

100

60

0

5

10

15

ash

ash I

20

I

Fig. 14. Complete profiles of Fe/Al and Ca/Al at (a) TGe and (b) EKk

46

P. Steinmann, W. Shotyk/ Chemic:a1Geology 138

EGr (Steinmann and Shotyk, 1997). The upward diffusion of Ca at EGr has overprinted 2 m of previously ombtogenic Sphagnum bog peat with a minerogenic signature (Fig. 14b). This process illustrates the great danger in using the botanical composition of peats as an indicator of their trophic status: peats which have a botanical composition typical of bogs may easily have a geochemical composition characteristic of minerogenic fens.

(1997) 25-53

4.4. Mineralogy profdes

and mineral stability

in the peat

4.4.1. Calculating the essential mineralogy Analyses of the ash fraction using optical microscopy revealed that quartz, opaline silica, albite and K-feldspar are the dominant minerals in the peat ash, and this explains at least qualitatively the strong correlation between the ash content and the elements

25.5 z

37.5

S 5

49.5

u8

61.5 73.5 65.5 96

0

1

2

3

4 5 % of dry weight

6

7

6

b

1.5 10.5 19.5 28.5 37.5 46.5 55.5 67.5 79.5 88.5

EGr 2f

97.5

0

20

40

60

80

100

% of AIA

Fig. 15. Mineralogy of EGr as calculated from the elemental composition and AIA as the sum of insoluble silicates. AAIA is indicated when the calculated amount of silicates did not account for the total AL4. (a) Calculated amounts of each mineral present and organic or oxide fractions of Al, Fe, Ca, Mg, and K (percent by weight of dry peat). (b) Calculated mineralogical composition of AU.

P. Steinmann, W. Shotyk/ Chemical Geology I38 (1997) 25-53

Si, Al, Na, and K beneath the zone of biological cycling at the surface of the bog. In order to gain .a better idea of the chemical forms of elements in the peat, a hypothetical mineralogical composition ,of the ash was calculated from the chemical data. This allows an estimate to be made of the portion of each element residing in specific minerals and to assess possible changes in mineral abundances with depth. The calculated mineralogy employed the measured amount of acid-insoluble ash (AIA) as a quantitative indicator of the total amount of mineral material in each sample, and the relative abundance of minerals obtained from optical microscopy was used to constrain the calculated mineral abundances. The constraints for the calculations were expressed in terms of the equations shown in Table 4. The calculated mineralogy was assumed to consist exclusively of the following phases: SiO, (including quartz and opaline silica), albite, K-feldspar (Kfs), kaolinite, Al-hydroxide, and iron oxide; Ca, K, Nal hydroxides were used as well in these calculations. Albite and K-feldspar were assumed to be the only important minerals containing Na and K, respectively (Eq. (1)). The amount of albite and K-feldspar were restricted to less than 15% of AIA, in accordance with the results from optical microscopy (:Eq. (2)). The sum of silicate minerals (quartz, albite, K feldspar, and kaolinite) was taken to be equal to AIA (Eq. (3)) and to contain all of the measured Si (Eq. (4)). Quartz (which for the purpose of the calculations included biogenic silica) had to account for more than 60% of AIA (Eq. (5)), in accordance with the results from optical microscopy. Iron was assigned to Fe-oxide (Eq. (6)). Calcium, Na, and K in excess of the feldspar-bound fraction (i.e. originally present in the organic fraction prior to ashing) were calculated as hydroxides (Eq. (6)). The results of these calculations are shown in Figs. 15 and 16. If all non-feldspar silica were calculated as quartz/SiO, (as a surrogate for quartz + biogenic silica), the total weight of feldspar and quartz would be less than AIA for most samples; this implies that another important silica-bearing phase with a lower Si-content must have been present in the ash. Because the calculated amount of feldspar used up only a small portion of the Al present, the missing silicate is presumably rich in Al. Also, the phase should

41

contain low K or Na because these two cations are mainly contained in the feldspars present. Furthermore, the missing silicate should be a clay mineral because greater amounts of other minerals would have been observed in the microscope. Kaolinite is the most likely candidate because it is a common weathering product in soils and sediments, and should be an important component in soil-derived atmospheric aerosols. Moreover, kaolinite is a clay mineral which has been found in ombrogenic peats (Finney and Famham, 1968). Kaolinite was included in the calculations as shown in Eq. (4) (Table 4). Even with the inclusion of the aluminum-rich phase kaolinite, however, a considerable amount of Al is left over, and this must have belonged to the soluble ash fraction. In the (550°C) ash this fraction of Al is most likely present as AI(O whereas in the original peat it may have occurred either as gibbsite (another common weathering product in soils) or was organically bound. After calculating the mineralogical composition of AIA not only Al remained in excess, but also Fe, Ca and Mg. These elements must therefore have originally been bound in the acid soluble fraction of the ash. As mentioned above in the 550°C ash a part of these cations are present as oxides and hydroxides. The sum of the calculated ash fractions (AIA plus the calculated amounts of oxides and hydroxides of

Table 4 Constraints used to calculate the mineralogical ash fraction from the chemical composition

0.14Kfs

composition

5 K; 0.088 albite 5 Na

albite,Kfs

< 0.15AIA

quartz + albite + Kfs + kaolinite = AIA

of the

(1) (2) (3)

0.301 albite + 0.321 Kfs - 0.217 kaolinite + 0.467 quartz = Si (4) quartz > 0.6 AL4

(5)

Fe,O,

(6)

= 1.43Fe

Naexcess = Na - 0.088 albite; KeICeSS= K - 0.14Kfs

(7)

AlexCelS= Al - 0.103 albite - 0.097 Kfs - 0.209 kaolinite

(8)

48

P. Steinmann, W. Shotyk/ Chemical Geology 138 (1997) 25-53

Ca, Mg, Fe, and excess Al) yields an ash content which is still 10 to 50% below the measured ash contents. This difference reveals that some of the Ca, Mg, Fe and Al originally present in the organic fraction of the peat were converted to sulphates, phosphates or carbonates during combustion. The calculated mineralogy reveals that the observed amounts of both feldspars can account for most of the K and Na present in the profile below the biologically active zone. The relatively constant proportions of plagioclase and K feldspar seen in the microscope, and the lack of a vertical trend in the

concentrations of these two elements, argues against measurable dissolution of the feldspar fraction of peat ash over time. The calculations show further that kaolinite is likely to be an important constituent in the peat ash, and that a significant pool of soluble Al exists in the peat profile. Because burning at 550°C and dissolution in dilute HCl are involved during treatment of the ash fraction, it was not possible to confirm or quantify such phases as Ca-carbonate or Fe-hydroxides in the fresh peat. We therefore cannot distinguish between Fe and Ca which may have been organically bound

a I

? ? MIA ? ? kaol ? ? albits

1.5

2 2.5 % of dry weight

3

3.5

b

? ?kaol ? ?albite

TGe li ,

0

20

40

60

80

100

% of AIA

Fig. 16. Calculated

mineralogy

for TGe (cf. caption Fig. 15)

P. Steinmann, W. Shotyk/Chemical Geology 138 (1997) 25-53

versus CaCO, or Fe,O, which also may have been present. Furthermore, the presence of traces of Sipoor silicates containing cations such as Ca and Fe (e.g. epidote or biotite which were sometimes seen in the microscope), was neglected in the calculations and this helps to explain why the calculated AIA is smaller than the measured AIA (indicated as A AIA in Figs. 15 and 16) in some samples. For example, in the EGr profile Mg shows the most prominent peak at the depth of the highest AIA peak at 35 cm. Therefore, it is likely that some of the Mg too is bound in silicate minerals. For comparison with the results presented here, Finney and Famham (1968) found the following composition of particles > 2 km in peats from raised bogs in Minnesota: 30% biogenic opal; 30% quartz; 18% orthoclase; 6% micas; 2% hornblende; 1% opaques; and 1% others. Finney’s results indicate more feldspar and less muscovite than found in this study. Indeed, the lo-15% muscovite overestimates the volumetric contribution of this mineral because of the grain shape of micas. The contribution of minerals (including biogenic opal) to ash was 5070% and again 50-70% of the mineral fraction was > 2 p,m. No signilicant changes in the relative amounts of quartz, feldspar and mica were found within the low ash peats of these Minnesota bogs, an observation which argues against significant dissolution of these minerals. In the clay mineral fraction Finney and Far&am (1968) found kaolinite, illite, vermiculite, and montmorillonite but no quantification of the clay minerals was possible. Clay minerals are likely to be present in the TGe and EGr peats, too. 4.4.2. Dissolution of Fe oxides Given the very srnall size of the mineral grains (soil-derived aerosols) and the low pH of the waters, there is no obvious reason why mineral matter decomposes so slowly in the peat profiles. With respect to the Fe oxides, the apparent stability of these solids in the anoxic waters is inconsistent with the relatively rapid rates of reductive dissolution of Fe oxides. The rate of reductive dissolution of hematite at pH 3 in the presence of ascorbate, for example, is on the order of micromoles per hour (Stumm, 1992). Above we have shown evidence that many of the iron oxides have been dissolved and subsequently

49

were fixed by binding to organic substances. It might also be possible that some of the original iron oxides were not dissolved due to the formation of a prohibitive organic coating of the minerals. Experiments have shown that the adsorption of large molecular weight organic acids on iron oxide surfaces may lead to the formation of organic, polymeric coatings which significantly reduce mineral dissolution rates (Blaser, 1974; Blaser et al., 1981). Consider the dissolution of Fe,O,, for example: the pH of the point of zero charge for o-Fe,O, is 8.5 (Stumm, 1992). At pH 4, therefore, extensive adsorption of anionic organic molecules on the oxide surface can be expected. The adsorption and electron transfer reactions are relatively fast, and the rate-limit step of the dissolution process is the de-sorption of the organic molecule (Stumm, 1992). Given the large average size of the molecules, the number of bonds which need to be broken per molecule adsorbed may be prohibitively large. In these circumstances the organic molecules effectively form protective surface coatings which protect the oxide from attack by the acid solution. 4.4.3. Dissolution of plagioclase feldspar The relative abundances of albite and K-feldspar observed in the peat samples (5 to 15% of mineral ash) can account for effectively all of the Na and K in the profile beneath the zone of active biological enrichment. Furthermore, the relative abundance of albite and K-feldspar from 40 to 100 cm is more or less constant (Fig. 16b), providing no indication of significant dissolution of either of these phases during the past 2000 years. Assuming that the relative abundance of these two minerals in the dust deposited on the surface of the bog has not changed significantly over time, these findings suggest that these two mineral phases have not measurably reacted with the organic-rich, acidic-pore fluids. The result presented here can be further examined in relation to the known rate of dissolution of plagioclase feldspar. At pH 4, this rate is on the order of lo-i5 mol cm-’ s-l at 25°C (Casey et al., 1991). In other words, under laboratory conditions with the solution being far from equilibrium and no organic material present, at pH 4 it takes one year to dissolve 0.0085 mg of feldspar per cm* of mineral surface. Assuming that the average dimensions of the feldspar grains in the ombrogenic peats at EGr are 10 X 10 X

50

P. Steinmann, W. Shotyk/

Chemical Geology 138 (1997) 25-53

1 pm (this is typical of the feldspar grains observed using SEMI, there would be 0.113 mg of mineral per cm2 of feldspar surface available for reaction. Given the dissolution rate for plagioclase at pH 4, this amount of feldspar would be expected to dissolve within approximately 10 years. The mineralogical and geochemical data presented here, however, indicate that this has not happened. In fact, more than 1000 years has elapsed and essentially all of the feldspar initially present remains in the peat profile. Even if 10% of the feldspar has reacted, the apparent dissolution rate of feldspar in acidic peat bogs is no greater than approx. lo-‘* mol cm-’ s-l. In other words, the rate of dissolution of feldspar in these pore fluids may easily be 3 orders of magnitude lower than the experimentally derived rates in HCl at pH 4. There are several possible reasons for the lack of reactivity in the bog. First, the average annual temperature is around 6°C which would reduce the dissolution rate relative to the rate at 25°C. Second, the pore waters in the peat bog might be closer to equilibrium with feldspar than the HCl solution used in the experiments. However, these pore waters are rather dilute, and some of the cations are complexed with humic acids (Steinmann and Shotyk, 1997). The saturation indices (IAP/K,,) calculated for the feldspar minerals in the ombrogenic peats are all less than lo-” (Steinmann, 1995). Thus, the saturation state of the pore waters with respect to feldspar cannot explain the reduced dissolution rates. While the elevated concentrations of Al in solution reported by Bennet et al. (1991) and also measured by us (Steinmann and Shotyk, 1997) certainly result from the dissolution of some Al-bearing solid phase, these Al concentrations themselves do not necessarily indicate a high rate of mineral dissolution because the fluid flow rate in the peats is very low. Also, the Al present in solution may be derived from the dissolution of an Al oxide phase, a common constituent in soil-derived atmospheric aerosols, rather than a silicate mineral. With respect to the possible role of organic acids in the weathering of feldspars, the literature evidence is controversial. For example, Ochs et al. (1993) found experimentally that humic acids and exudates from tree roots may inhibit the dissolution of minerals even under acidic conditions. In contrast, in other

experiments large molecular weight humic acids neither inhibited nor promoted the rate of plagioclase feldspar dissolution, compared to the dissolution rate in HCl at pH 4 (Eggenberger, 1995). Finally, another factor which may contribute to the low reactivity of feldspar minerals in bog environments at pH 4 is the development of a siliceous residual surface layer on the mineral grains due to preferential removal of Al and cations such as Na and Ca (Shotyk and Metson, 1994). Experimental studies have shown that during the early stages of feldspar dissolution (days to Oweeks)these layers may extend to several thousand Angstroms (Casey et al., 1989a,b; Shotyk and Nesbitt, 1992). These layers are generally assumed to have chemical properties similar to that of amorphous silica. Due to the relatively high Si concentrations in the pore waters (Steinmann and Shotyk, 19971, however, amorphous silica is much closer to equilibrium (saturation index approaching lo- ’) than is either of the feldspars. Thus, the slow rate of reaction of these siliceous layers at pH 4 may control the overall rate of feldspar dissolution in acidic bogs.

5. Summary and conclusions The two contrasting peat profiles studied indicate three main sources of major elements to the peats: (a) atmospheric deposition of soil-derived aerosols; (b) mineral matter in sediments which may become physically incorporated in the peats; (c) diffusion from mineral soil water and groundwater. At EGr, all of the evidence from this and previous studies (peat stratigraphy, ash contents, pore water chemistry, and major element chemistry of the peats) indicates that all of the inorganic solids in the 2f core (102 cm) were supplied via the atmosphere; this includes both precipitation and aerosols derived from the weathering of crustal rocks. Using SC as a conservative tracer, the rates of atmospheric dust deposition to the recent peats (past 100 years) at EGr and TGe were found to be similar. The ombrogenic EGr core reveals modem rates of atmospheric dust deposition which are nearly 3 times greater than the average rates over the long term (past two millennia), a reflection of the greater abundance of aerosols today compared with the historical past.

P. Steinmann, W. Shotyk/Chemical Geology 138 (1997) 25-53

At TGe the slightly higher rate of SC accumulation cannot be explained by atmospheric inputs only, and an additional non-atmospheric particulate source is needed to explain the profile of this and other elements. The metal/SC ratios indicated that the rates of Ca and Fe accumulation are higher than can be explained by differences in mineral matter alone. The most likely cause of these enrichments is adsorption and/or complexation of dissolved metals from solution, following their diffusion from deeper layers. The relatively constant proportions of aluminosilicates with depth (time) suggests that the rates of mineral dissolution in bogs is generally slow. Support for this interpretation is provided by the chemical data which shows that essentially all of the Na and K below the biologically active zone can be explained by two mineral phases: plagioclase and K feldspar, respectively. The lack of a trend in Na/Sc or K/SC with depth suggests that these two mineral phases have not undergone measurable dissolution during the past two millennia.

Acknowledgements We thank Dr. [email protected] Cheburkin, Geological Institute Ukrainian Academy of Sciences, Kiev for XRF analyses and Dr. Peter Appleby of the University of Liverpool for the *l’Pb age dating. This study was financially supported by the Canton of Bern @EVA Lottofonds) and the Swiss National Science Foundation (Grants 21-30207.90 and 20-36371.92). The preparation, processing, and age dating of the basal peat samples were carried out by the Radiocarbon Laboratory of the Physics Institute of the University of Beme. The manuscript was improved with the helpful comments of Drs. Don Siegel and Kimmo Virtanen.

References Andrejko, M.J., Fiene, F. and Cohen, A.D., 1983. Comparison of ashing techniques for detexmination of the inorganic content of peats. In: P.M. Jarrett (Editor), Testing of Peats and Organic Soils. ASTM, Philadelphia, pp. 5-20.

51

Appleby, PG., Shotyk, W. and Fankbauser, A., 1997. “‘Pb age dating of three peat cores in the Jura Mountains, Switzerland. Water Air Soil Pollution, Special issue: Peat Bog Archives of Atmospheric Metal Deposition, in press. Baes, C.F.J. and Mesmer, R.E., 1976. The Hydrolysis of Cations. John Wiley and Sons, New York. Bennet, P. and Siegel, D.I., 1987. Increased solubility of quartz in water due to complexing by organic compounds. Nature, 326: 684-686. Bennet, P.C., Siegel, D.I., Hill, B.M. and Glaser, P.H., 1991. Fate of silicate minerals in a peat bog. Geology, 19: 328-331. Blackford, J.J., Edwards, KJ., Dugmore, AJ., COOL,G.T. and Buckland, P.C., 1992. Ice.landic volcanic ash and the mid-Holocene Scats pine (Pinus sylvestris) pollen decline in northern Scotland. Holocene, 2: 260-265. Blaser, P., 1974. Mechanismen der Eisenaufnahme und -ve.rlagerung durch wasserliisliche Strcusubstanzen in podsoligen B&n. Mitt. Dtsch. Bode&d. Ges., 20: 447-457 (in GelltXUl). Blaser, P., Flier, H. and Sidler, T., 1981. Riickkoppehmgsmechanismen bcim Stofftransport: Verlagerung vou Fe(III) durch wasserkislichc Streusubstanzen. Mitt. Dtsch. Bodenknd. Ges., 30: 111-122 (in German). Bushinskii, G.I., 1946. The conditions in the formation of siderites, vivianites, and brown iron ores in the peat bogs of White Russia. Byull. Moskov. Obschchestva Ispytat. Prirody, Otdel. Geol., 21: 65-80. Casey, W.H., West&h, H.R., Arnold, G.W. and Banfield, J.F., 1989a. The surface chemistry of dissolved labradorite feldspar. Geochim. Cosmochim. Acta, 53: 821-832. Casey, W.H., Westrich, H.R., Massis, T., Banfield, J.F. and Arnold, G.W., 1989b. The surface of labradorite feldspar after acid hydrolysis. Chem. Geol., 78: 205-218. Casey, W.H., Westrich, H.R. and Holdma, G.R., 1991. Dissolution rates of plagioclase at pH = 2 and 3. Am. Mineral., 76: 211-217. Clymo, R.S., 1983. Peat. In: A.J.P. Gore (Editor), Mires: Swamp, Bog, Fen and Moor. Ecosystems of the World, 4A. General Studies. Elsevier, Amsterdam, pp. 159-224. Dammau, A.W.H., Tolonen, K. and Sallantaus, T., 1992. Element retention and removal in ombmtropbic peat of H&de.tkeidas,a boreal Finnish peat bog. Suo, 43: 137-145. Dugmore, A., 1989. Icelandic volcanic ash in Scotland. Scott. Geogr. Mag., 105: 168-172. Dugmore, A.J., Newton, A.J., Sugden, D.E. and Larsen, G., 1992. Gcochemical stability of fine-grained siliiic Holocene tcphra in Iceland and Scotland. J. Quat. Sci., 7: 173-183. Eggenbcrger, U., 1995. Mineral Weathering in Soils: Experiments, Field Studies, and Modeling. Ph.D. Thesis, University of Berue, unpubl. Fankhauser, A., 1995. Pollenanalytische Untersuchungen zur jiingsten Vegetationsgeschichte der Franches Montagnes. Diploma Thesis, Geobotanical Institute, University of Beme, unpubl. (in German). Feldmeyer-Chris& E., 199O.&ude phyto&ologique de tourbi&res de Franches-Montagnes (canton du Jura et de Beme, Suisse). Mat. Lev6 G.%bot. Suisse, 66: 1-163 (in French).

52

P. Steinmann, W. Shotyk/ Chemical Geology 138 (1997) 25-53

Finney, H.R. and Famham, R.S., 1968. Mineralogy of the inorganic fraction of peat from two raised bogs in northern Minnesota. Thii Int. Peat Congr., pp. 102-108. Foltescu, V.L., Isakson, J., Selin, E. and Stikans, M., 1994. Measured fluxes of sulphur, chlorine and some antbropogenic metals to the Swedish west coast. Atmos. Environ., 28: 26392649. Frondl, C., 1970. Scandium. In: K.H. Wedepohl (Editor), Handbook of Geochemistry. Springer-Verlag, Berlin. Gerber, E. and Monbaron, M., 1990. Die Torfmoore des niirdlichen Faltenjuras in iluem morpho-strukturalen Zusammenhang. UKPIK, 7: 31-44 (in German). G&es, M., 1991. iiber den Eintrag anorganischer N%hrstoffe in ombrogene Moore als Indikator der ehemaligen Aerosolbelastung. Ph.D. Thesis, J. Cramer, Berlin, 179 pp. (in German). G&es, M. and Bludau, W., 1992. Der Zusammenhang zwischen pollen- und 14Canalytisch ermittelten Siedlungsphasen und e&ihten Mineralstoffgehalten in Profilen des Weidfilzes (Stamberger See). Tehna, 22: 123-144 (in German). Grhnig, A., Vetterli, L. and Wildi, O., 1984. Inventaire de hautsmarais et marais de transition de Suisse. Swiss Fed. Inst. For. Res., Birmensdorf, ca. 2100 pp. (in French). Hem, J.D., 1972. Chemical factors that influence the availability of iron and manganese in aqueous systems. Geol. Sot. Am., Spec. Pap., 140: 17-24. Hill, B.M. and Siegel, D.I., 1991. Groundwater flow and the metal content of peat. J. Hydrol., 123: 21 l-224. Hobbs, N.B., 1986. Mire morphology and the properties and behaviour of some British and foreign peats. Q. J. Eng. Geol. London, 19: 7-80. Hodder, Q.P.W., de Lange, P.J. and Lowe, D.J., 1991. Dissolution and depletion of ferromagnesian minerals from Holocene tephra layers in an acid bog, New Zealand, and implications for tephra correlation. J. Quat. Sci., 6: 195-208. Hiilzer, A. and HSlzer, A., 1987. Pahaookologische Moor-Untersuchungen an der Homisgrinde im Nordschwarzwald. Carolinea, 45: 43-50. (in German). Hiilzer, A. and I-Ii?lzer, A., 1988a. Untersuchungen zur jiingeren Vegetationsund Siedlungsgeschichte im Blindensee-Moor (Mhtlerer Schwarzwald). Carolinea, 46: 23-30 (in German). Hijlzer, A. and Hiilzer, A., 1988b. Untersuchungen zur jiingeren Vegetationsund Siedlungsgeschichte in der Seemisse am Ruhestein. Tehna, 18: 157-174 (in German). Ikonen, L., 1993. Holocene development and peat growth of the raised bog Pesiinsuo in southwestern Finland. Geol. Surv. Finl. Bull., 370: l-58. Joray, M., 1942. L’etang de la Gmyere. Etude pollenanalytique et stratigraphique de la tomb&e. Mat. Le.& Geobot. Suisse, 25: 1-117 (in French). Kabata-Pendias, A. and Pendias, H., 1992. Trace Elements in Soils and Plants. CRC Press, Boca Raton, 2nd ed., 365 pp. Kivinen, E., 1936. Zur Kenntnis der Eisenkarbonatvorkommnisse in der Mooren Finnlands. Z. Pflanzenem’ti. Ding. Bcdenkd., 45: 96-104 (in Germane. Krauskopf, K.B., 1982. Introduction to Geochemistry. McGrawHill, New York, 2nd ed., 617 pp.

Landva, A.O., Korpijaakko, E.O. and Pheeney, P.E.. 1986. Notes on the original von Post peat and peatland classification system. National Research Council Canada, Ottawa, Gnt. Lukashev, K.I. and Kovalev, V.A., 1969. Mineralogical-geochemical systems of iron in recent peat bogs. Dokl. Akad. Nauk SSSR, 187: 1390-1393. Lukashev, K.I., Kovalev, V.A. and Generalova, V.A., 1974. Peat bogs as a medium for the secondary weathering of alumina silicates. Dokl. Akad. Nat& Beloruss. SSR, 18: 357-360. Markert, B. and Thornton, I., 1990. Multi-element analysis of an English peat bog soil. Water Air Soil Pollut., 49: 113-123. Met&e, B., 1987. Geobotanische und geochemische Untersuchungen an einem Torfprotil zur Frage nathrlicher und anthropogener Elementverfrachtung. Geol. Jahrb., 95A: 3-l 12 (in German). Metso, J., Niinisto, L. and Yliruokanen, I., 1976. Inorganic constituents in Finnish fuel peat ash. Kemia-Kemi, 3: 409-412. Naucke, W., 1990. Chemie von Moor und Torf. In: K. Giittlich (Editor), Moor und Torfkunde. E. Schweizerbart, Stuttgart, pp. 237-261 (in German). Njistad, O., Naeumann, R. and Steinnes, E., 1987. Variations in atmospheric trace element deposition studied by lNAA of peat cores from ombrogenic bogs. J. Radioanal. Nucl. Chem., Articles, 114: 69-74. Ntahokaja, W. and Ziiovsky, L., 1995. A systematic neutron activation analysis of peat from Burundi. J. Radioanal. Nucl. Chem. Lett., 199: 431-441. Ckhs, M., Brunner, I., Stumm, W. and Cosovic, B., 1993. Effects of root exudates and humic substances on weathering kinetics. Water Air Soil Pollut., 68: 213-229. Oldfield, F., Thompson, R. and Barber, K.E., 1978. Changing atmospheric fallout of magnetic particles recorded in recent ombrotrophic peat sections. Science, 199: 679-680. Oldfield, F., Tolonen, K. and Thompson, R., 1981. History of particulate atmospheric pollution from magnetic measurements in dated Finnish peat profiles. Ambio, 10: 185-188. Persson, C., 1971. Tephrochronological investigation of peat deposits in Scandinavia and on the Faeroe Islands. Sver. Geol. Unders., 65: 3-34. Postma, D., 1977. The occurrence and chemical composition of recent Fe-rich mixed carbonates in a river bog. J. Sediment. Petrol., 47: 1089-1098. Postma, D., 1981. Formation of siderite and vivianite and the pore-water composition of a recent bog sediment in Denmark. Chem. Geol., 31: 225-244. Postma, D., 1982. Pyrite and siderite formation in brackish and freshwater sediments. Am. J. Sci., 282: 1151-1183. Postma, D., 1983. Pyrite and siderite oxidation in swamp sediments. J. Soil Sci., 34: 163-182. Puustj%rvi, V., 1952. The precipitation of iron in peat soils. Acta Agral. Fenn., 78: 5-72. Reille, M., 1991. L’origine de la station de pin B crochets de la tourbi&e de Pinet (Aude) et de quelque stations isolees de cet arbre dam les Voges et le Jura. Bull. Sot. Bot. Fr., 138 (L&t. Bot. 2): 123-148 (in French). Richardson, N., 1986. The mineral magnetic record in recent

P. Steinmann, W. Shotyk/Chemical

ombrotrophic peat synchronized by fine resolution pollen analysis. Phys. Earth Planet. Inter., 42: 48-56. Selmer-Olsen, A.R. and. Lie, 0.. 1983. Jemkarbonatmfelling (side&t) y myr. Saertryk av Jord og myr, 4: 131-140 (in Norwegian with Englrsh summary). Shotyk, W., 1988. Review of the inorganic geochemistry of peats and peatland waters. Earth Sci. Rev., 25: 95-176. Shotyk, W., 1996a. Peat bog archives of atmospheric metal deposition: geochemical evaluation of peat profiles, natural variations in metal concentrations, and metal enrichment factors. Environ. Rev., 4: 149-183. Shotyk, W., 1996b. Natural and anthmpogenic enrichments of As, Cu. Pb, Sb, and Zn in ombrotrophic versus minerotmphic peat bog profiles, Jura Mountains, Switzerland. Water, Air, Soil Pollut., 90: 375-405. Shotyk, W. and Metson, J.B., 1994. Secondary ion mass spectrometry (SIMS) and its application to chemical weathering. Rev. Geophys., 32: 1!)7-220. Shotyk, W. and Nesbitt, H.W., 1992. Incongruent and congruent dissolution of plagioclase feldspar: effect of feldspar composition and ligand complexation. Geoderma, 55: 55-78. Shotyk, W. and Steinmann, P., 1994. Pore-water indicator of rainwater-dominated versus groundwater-dominated peat bog profiles, Jura Mountains, Switzerland. Chem. Geol., 116: 137-146. Steinmann, P., 1995. Geochemistry of atmospheric inputs and pore water genesis in two contrasting Sphagnum bog profiles, Jura Mountains, Switzerland. Ph.D. Thesis, Geol. Inst., Univ. Beme, unpubl. Steinmann, P. and Shotyk, W., 1997. Chemical composition, pH, and redox state of S and Fe in complete vertical pore water

Geology 138 (1997) 25-53

53

profiles from two Sphagnum peat bogs, Jura Mountains, Switzerland. Geochim. Cosmochim. Acta, in press. Steinnes, E. and Njhtad, O., 1995. Ombrotrophic peat bogs as monitors of trends in atmospheric deposition of pollutants: role of neutron activation analysis in studies of peat samples. J. Radioanal. Nucl. Chem., Articles, 192: 205-213. Stumm, W., 1992. Chemistry of the Solid-Water Interface. John Wiley and Sons, New York, 428 pp. Tolonen, K., 1984. Interpretation of changes in the ash content of ombrotmphic peat layers. Bull. Geol. Sot. Finl., 56: 207-219. Van Bemmelen, J.M., 1899. iiber das Vorkommen, die Zusammensetzung und die Bildung von Eisenanhaufungen in und unter Mooren. Z. Anal. Chem., 12: 313-379 (in German). Virtanen, K., 1994. Geological control of iron and phosphorous precipitates in mires of the Ruukki-Vihanti area, central Finland. Bull. Geol. Surv. Finl., 375: l-69. von Post, L. and Granlund, E., 1925. Sodra Sveriges tortvtillgangar. Sver. Geol. Unders., Ser. C, 335, A&ok, 19(2), 127 pp (in Swedish). Vuorela, I., 1983. Field erosion by wind as indicated by fluctuations in the ash content of Sphagnum peat. Bull. Geol. Sot. Finl., 55: 25-33. Wardenaar, E.C.P., 1987. A new hand tool for cutting peat profiles. Can. J. Bot., 65: 1772-1773. Welten, M., 1964. La Tourbibre de Genevez. Mitt. Nat. Ges. Bern, 21: 67-73. Williams, M., 1992. Evidence for the dissolution of magnetite in recent Scottish peats. Quat. Res., 171-182. Zoltai, S.C., 1989. Late Quatemary volcanic ash in the peatlands of central Alberta. Can. J. Earth Sci., 26: 207-214.