Abstract

The nature of the alteration of basaltic, andesitic and rhyolitic glass of Holocene and Pleistocene age and their physical and chemical environments have been investigated in the ash layers within the cryogenic soils associated with the volcanoes in the central depression of Kamchatka. One of the main factors controlling the alteration of the volcanic glass is their initial chemistry with those of andesitic (SiO2 = 53–65 wt.%) and basaltic (SiO2 < 53 wt.%) compositions being characterized by the presence of allophane, whereas volcanic glass of rhyolitic composition (SiO2>65 wt.%) are characterized by opal. Variations in the age of eruption of individual ashes, the amount and nature of the soil water, the depth of the active annual freeze-thawing layer, the thermal conductivity of the weathering soils, do not play a controlling role in the type of weathering products of the ashes but may affect their rates of alteration.

The highly active volcanic area of Kamchatka in eastern Russia is part of the circum-Pacific belt of andesitic volcanism. It is situated north of the 49th parallel of latitude and is characterized by a severely cold climate with permafrost widespread above an altitude of 700 m. The Kamchatka peninsula is blanketed by a soil-pyroclastic cover (Melekestsev et al., 1969) consisting of loose, pyroclastic material from explosive eruptions and buried soil beds. Ashes from major eruptions form clear marker horizons and can be traced over wide areas. This has allowed a well defined tephrostratigraphy to be established that is based upon the chemical signature of the glass (Braitseva et al., 1997). Between the key ash marker horizons are other pyroclastic deposits derived from both minor local and remote eruptions and from the secondary re-deposition of ash (Bazanova et al., 2005).

Considerable research has been carried out on the weathering of volcanic glass. For volcanic ash soils in warm humid climates Wada & Wada (1977) suggested that the sequence of weathering is volcanic glass → allophane → halloysite or volcanic glass → opal → allophane → halloysite. For the rhyolitic glass in the humid conditions in Japan, the allophane → halloysite transformation occurs after 6000 years (Wada & Wada, 1977; Nagasawa, 1978; Wada, 1989). However, other studies have shown that, depending on their different glass compositions, volcanic ash may weather directly to halloysite as well as to allophane (Parfitt et al., 1983; Parfitt & Wilson, 1985). Basaltic glass (SiO2:Al2O3 < 4) has an intrinsically smaller molar Si:Al ratio and a soft and porous structure that favours weathering at a much higher rate than rhyolitic glass (SiO2:Al2O3 = 5 to 6) (Lowe, 1986).

The influence of climatic variation on the weathering of volcanic glass is still to be clarified, particularly in the extremely cold climates associated with high precipitation. In this paper we report on our study of the breakdown of volcanic glass in the soil-pyroclastic cover associated with the Klyuchevskaya group of volcanoes situated in the Central Depression of Kamchatka (55–56°N, 160–161°E). The group consists of several active volcanoes (Klyuchevskoy (∼4800 m), Bezymianny (∼2900 m), Ushkovsky (∼3900 m), Plosky Tolbachik (∼3100 m) plus ten others that are not active today, together with numerous small cinder cones, extrusive domes and other volcanic features (Braitseva et al., 1995). It is the location where the Large Fissure Tolbachik Eruption took place in 1975 and 1976, during which about 500 km2 were covered by scoria and ash and 3 new cinder cones and lava fields were formed (Fedotov & Markhinin, 1983).

The climate of the study area is characterized by cold winters, as well as mild and wet summers that are mainly influenced by cyclonic activity. No permanent meteorological station exists in the mountainous areas, but the Klyuchi station (WMO 32389), situated in the Kamchatka river valley at an elevation of 29 m and about 30–40 km from the central part of volcanic area, provides a long-term climatic record for the region. The mean annual average temperature in Klyuchi is –0.7°C for 1910–2007 and ∼0°C for 1970–2007. The total annual precipitation in Klyuchi is ∼500 mm, whereas in the mountains it is estimated to be ∼1000 mm (Muravyev, 1999). In Klyuchi the snow cover usually lasts from October to June with a snow cover thickness of ∼1.5–3 m in forested areas and 50–80 cm or less in open areas due to redistribution by strong winds (Abramov et al., 2008).

Permafrost and periglacial processes are wide-spread. The lower boundary of the permafrost is at 750–900 m above sea level for the northern slopes and 650–800 m for the southern ones. Numerous solifluction lobes, clay cryoturbation spots, polygonal structures, and areas of sorted ground occur between altitudes of 1000 and 1700 m. The thickness of the seasonal freezing-thawing active layer decreases with increasing altitude and may vary slightly from year to year. The maximum seasonal melting of permafrost is up to 2.5 m at 900 m and 50 cm at 2500 m (Abramov et al., 2008).

There are from six to eight soil profiles developed in the soil-pyroclastic cover associated with these volcanoes. In each there is an organic horizon that overlies volcanic ash. The soils have a small humus content ∼ 1 to 2.4%, which is not typical for most volcanic soils of Kamchatka (Zakharikhina, 2009); this reflects the interplay of permafrost, relatively low rainfall and the abundant deposition of airborne volcanic material.

Analytical Methods.

A total of 24 samples (Table 1) were collected at elevations between 130 and 1630 m. Samples 1-6 were collected from the active layer in the thawed state, sample 7 in the frozen state from a depth of 2.50 m in a drill hole and other samples from the frozen active layer.

The moisture content (ratio of the mass of water to the mass of dry sample) was determined by drying the wet samples at 105°C for 12 h. Adsorbed water was determined by drying samples to constant weight from room temperature to 105°C.

Thermal studies (TG, DTG, DTA analyses) were carried out using a Q-1500D derivatograph over the range from room temperature to 1000°C with a heating rate of 20°C/min (Topor et al., 1987). The loss of absorbed (low) water over the range from room temperature to 110°C was observed on the heating curves of the samples (3, 4, 8, 9, 13, 15, 17, 24), and the endothermic effects recorded over the range from 110 to 750°C corresponded to the loss of structural (high) water by the ash material. The total mass loss (structural water + absorbed water) is designated as Wterm and its values are shown in Table 6.

From the weight difference between the wet and dried sample and the volumes of the sample, the dry bulk density and water content were calculated. Solid particle density (mineral density of rock) was measured for rock powder on an exclusive device ELA (Kalachev et al., 1997). The underlying principle of this device is the law of Boil-Moriotte (pressure×volume = constant in an isothermal environment); the accuracy is 0.02 g/cm3.

Particle-size distribution patterns were determined for most samples by a combination of sieving and the pipette method after dispersion with sodium hexametaphosphate.

Chemical analysis was restricted to several representative samples based on their similar stratigraphical positions in comparative marker layers (e.g. between layer Sh2 and Sh3) or in relation to existing published data.

The chemical compositions of the volcanic ash samples 17, 14, 15 and 2224 were determined by wet chemical analyses in the chemical laboratory of the Petrology Department of Moscow State University. SiO2 and H2O were determined gravimetrically; CaO, MgO, Al2O3, TiO2 and Fe2O3, complexometrically; and Na2O and K2O by flame photometry (Jeffery & Hutchison, 1983).

The salinity of the soil water was determined using a salinometer based on electrical conductivity (EC) measurements of soil-water extract.

The mineral composition of the ashes was determined by infrared absorption spectroscopy using a LOMO FSM-1201 infrared Fourier spectrometer. The absorption spectra were recorded over the range from 400 to 4000 cm–1 at room temperature with an accuracy of frequency determination of ±1 cm–1. The samples were prepared using two procedures: powdered samples (2 mg per 250 mg of KBr) were pressed into transparent tablets, or as suspensions in petroleum oil (Plyusnina, 1977). Interpretation of the spectra was made by comparison with standard mineral spectra published by Peng Wenshi (1982). SEM images of the samples and chemical composition of the volcanic glass and some minerals were obtained at the Petrology Department of Moscow State University by EPMA using a Jeol JSM-6480LV instrument with an INCA-Energy 350 spectrometer for compositional analysis. The samples were prepared as polished thin sections and SEM images presented in this paper are backscatter images.

Transmission Electron Microscopy (TEM) analysis was performed in the NTNU/SINTEF Gemini Centre using a JEOL 2010F instrument equipped with an Oxford Instruments SDD X-ray detector for chemical analysis. Samples were prepared by first dispersing a small sample of ash in propanol in a glass vial, assisted by placing the suspension in an ultrasonic bath. A significant proportion of the ash settled within a few seconds. A drop of the remaining suspension, expected to contain a higher proportion of the smaller component particles, was placed onto a perforated carbon film supported on a standard Cu-mesh TEM grid and allowed to dry before examination.

Samples 17 were radiocarbon dated at the laboratory of the Department of Geography, University of Zurich (GIUZ) using AMS (accelerator mass spectrometry) with a tandem accelerator at the Institute of Particle Physics based at the Swiss Federal Institute of Technology Zurich (ETH).

The dating of other samples was made by comparison to the well established scheme of tephrochronology developed for Kamchatka and based on a combination of the varying chemistry of the marker ash layers and radiocarbon dating of the organic soils, combined in sample 24 with pollen and diatom analysis (Braitseva et al., 1968). Marker ashes are dated and have their own index accompanied by a number or letter. For example, the tephra of the Shiveluch volcano from eruption ‘Sh2’ are 950 years old, from eruption ‘Sh3’ – 1400 years, from eruption ‘Sh5’ – 2550 years, and from eruption ‘Sh8300’ – 8300 years. The Khangar volcano tephra horizon with the index ‘Khg’ are 6850 years old, and the Ksudach volcano ashes with the index ‘KS1’ are 1800 years old (Ponomareva et al., 2007a, b). The dating of the samples between marker layers was made by calculation.

Determinative methods for thermal properties and unfrozen water content in frozen and thawed grounds are described in Motenko & Kuznetsova (2009a), Kuznetsova et al. (2011) and Kuznetsova & Motenko (2011b, 2012).

Results

Sample ages

The ashes are mainly of Holocene age (0–12000 years), the only exception being sample 24 which has been dated by its sporo-pollen and diatom assemblages as belonging to the upper part of the early Pleistocene (Q12) (Braitseva et al., 1968) (Table 1).

Physical properties

The physical properties of the soils are summarized in Table 1. The water contents range from 13 to 69%, the adsorbed water ranges from 0 to 7.3%, the dry density from 0.9 to 1.6 g/cm3 and the solid particle density from 2.1 to 2.8 g/cm3. Particle size distribution showed that nearly all the ashes are of fine sand size. The predominant fraction in most samples is 0.1–0.05 mm, although for samples 24 it is 0.05–0.01 mm, for sample 17 it is 0.25–0.1 mm and sample 24 is predominantly of silt size with 60% of its total being less than 56 μm (Kiryanov, 1981). The particle size limits applied for sand, silt and clay are based on standard procedures used in Russia, but these are similar to those used in Europe and USA.

Chemical properties

The major primary components in fresh tephras include volcanic glass, plagioclase and pyroxenes as well as other minerals (Nanzyo, 2002). The most abundant component is glass which can be divided into noncoloured and coloured varieties. Rhyolitic, dacitic and andesitic tephras contain non-coloured volcanic glass while basaltic tephra contains coloured volcanic glass (Yamada & Shoji, 1983).

The chemical compositions of ash and volcanic glass in several samples are shown in Tables 2 and 3. In addition reference is also made to chemical data for ash and glass samples used in this study that have previously been published by Braitseva et al. (1997) and Ponomareva et al. (2007a,b). The volcanic glasses are generally either andesitic (SiO2 < 65%) or rhyolititic (SiO2 > 65%), although the glass from samples 8, 20 and 21 is basaltic (SiO2 < 53%) (Braitseva et al., 1997). The bulk composition of the fresh tephra reflects the overall bulk chemical composition of its mineral and glass components and not necessarily the composition of the glass components. For example, combinations of andesitic ash with rhyolitic glass or basaltic ash with andesitic glass are not uncommon.

The soil water salinity measurements indicate that all ashes, except sample 24, have low salinities; their total soluble salt content is about 0.02–0.03%. Sample 24 has a high salinity, the total soluble salt content is 1.815 % and the main component is sulfate (SO42 1.242%). The hydrochloric acid extract from this sample shows that the content of SO42 is 0.47%, and Fe2+ 0.44%. Assuming that Ca2+ is associated with the sulfate ion, the content of CaSO4 is 0.67%. The pH of this saline water sample was 3.4 (Table 4).

The organic carbon content of the samples range from 0 to 0.03 wt.%.

Infra-red analyses

Yamada & Shoji (1975) found that the composition of tephra is dependent on particle size; the mineral content tends to be high in the fraction coarser than 0.1 mm in diameter, whereas volcanic glass tends to be abundant in the <0.1 mm fraction. Infra-red analysis was carried out therefore on the volcanic glass-rich fine fraction (<0.1 mm size fraction) of the tephra.

Peng Wenshi (1982) showed that the typical frequency of the Si–O absorption bands for opal occur at 472, 790 and 1100 cm–1, and for allophane at 430, 575 and 965 cm–1.

The IR absorption spectra from samples of the glass samples are shown in Fig. 1. The spectrum from sample 2 indicates the presence of “anhydrous opal” (bands at 446, 786 and 1080 cm–1), whereas the spectra from samples 9 (15), 13, 32 (Fig. 1b,c,d), indicate opal (SiO2·nH2O) with a varying water content. The Si–O absorption bands occur at 465–471, 781, 1050–1060 cm–1. The spectrum shown in Fig. 1e is for sample 8 (35 years old) which is essentially an aluminosilicate glass containing “anhydrous allophane” (IR absorption bands at 440, 553 and 987 cm–1), whereas samples 3, 7, 18 and 20 (Fig. 1f,g,h,i) contain allophane (mSiO2·kAl2O3·nH2O) (Si–O absorption bands at 434–457, 565–567, 980–995 cm–1).

The occurrence of water in the form of molecular H2O was identified by the band near 1630 cm–1 in the IR absorption spectra. This corresponds to deformation vibrations (vibrations with angular alterations) of the H2O molecule, and the simultaneous presence of a broad absorption band at 3100–3700 cm–1 in the spectra related to valence vibrations on O–H bonds. An increase in water content for each glass type was observed in the IR absorption spectra with increasing age of the ash. The absorption spectra also demonstrated the presence of traces of crystalline phases which we were unable to identify in our investigation.

Comparing IR results with glass composition indicates that allophane is associated with andesitic and basaltic glass, and opal with rhyolitic glass.

SEM and TEM analyses

Allophane. Backscatter electron micrographs from SEM analyses are shown in the Figs 2 and 3.

Figure 2 shows the image of a glass particle with an alteration rim that has been interpreted as allophane. The accompanying elemental distribution maps suggest a relative enrichment of Fe and Al in the allophane coating accompanied by desilicification of the glass and possible leaching of Mg, Na, K and Ca along the contact with the allophane.

Figure 3 shows backscatter images of two ash samples; no. 5–1500 years old (Fig. 3a) and no. 18–9000 years old (Fig. 3b). In the younger sample with less weathering, allophane forms a rim along the contact between vesicles and glass, but in the older sample with greater age and weathering, the leaching and desilicification of glass has resulted in very fine-grained mixtures of short-range-order aluminosilicates (allophane) (Fig. 3b), with Al/Si ratios of ∼ 1.4–1.6 (Table 5, A1 and A2).

Opaline silica. Opaline silica was present as very fine particles and these could only be seen clearly at high magnification using transmission electron microscopy. TEM micrographs of sample 24 are shown in the Figs 4, 5, 6. There is some evidence that the opaline silica particles are spherical in form and are closely packed in microaggregates (Fig. 4a and b), as described by Shoji & Masui (1971) and by Pollard & Weaver (1973). The forces and processes involved in the aggregation of the particles are not entirely understood, but there are suggestions that the long-range electrostatic repulsion force, that is between similarly charged objects, and the short-range van der Waal’s forces are the major forces that predetermined the configuration of the aggregate made up of silica spheres (Deveson, 2004). The opaline silica particles are extremely thin; the size of each particle can be up to 10–20 nm.

Figure 5a shows a corroded edge of the opaline silica which suggests that dissolution occurs during weathering (Henmi & Parfitt, 1980). The completely altered soft zone is filled with spherical bodies (Fig. 5b) and these are characteristic of opaline silica (Pollard & Weaver, 1973). These particles differ from glass fragments which are more angular (Fig. 6). The elemental compositions of the opaline silica and volcanic glass are shown in Table 5 with the analysed areas marked on the micrographs.

Gaillou et al. (2008) has emphasized that opal is not a pure form of silica as it contains water as a component, and various impurities and trace elements can also enter its structure. The most common impurity is aluminium, which substitutes for silicon (Bartoli et al., 1990). The investigation of Australian opals, representing 95% of world production (Horton, 2002), as well as opals from USA and Brazil, shows that the ‘impurities’ (i.e. elements present at concentration above 500 ppm) include Al, Ca, K, Mg, Fe and Na. Elements at concentrations below 500 ppm are referred to as ‘trace elements’. These are mainly Ba, Sr, Rb, Mn and Ti (Bayliss & Males, 1965; Bartoli et al., 1990; McOrist et al., 1994; McOrist & Smallwood, 1995, 1997). Some studies link the chemistry of opals to that of the host rock, even if modified by weathering processes (Gaillou et al., 2008).

Interpretation

Henmi & Parfitt (1980) concluded that opaline silica is characteristic of the early stages of weathering of volcanic ash in the humid temperate climatic zone. Our observations of volcanic ash from Kamchatka show that cold humid climates and permafrost can preserve opaline silica.

Figure 7 is a crossplot of the total and adsorbed water contents (Wterm and Wg) of volcanic ash samples (Table 6) vs. sample age. This shows a correlation between the total mass loss of water (Wterm) in the ashes and the age and chemical composition of the glass contained in those ashes. Ashes containing glass with andesitic composition have twice as much water as ashes of a similar age with glass of rhyolite composition (Fig. 7a). Similar correlation between the age of the ash and the amount of adsorbed water (Wg) is evident when the samples were dried (Table 6; Fig. 7b). The water content of the ashes containing opal is considerably lower than for those containing allophane. This suggests a higher rate of transformation of the ashes containing allophane.

Samples 9 and 15, both containing rhyolitic glass and of similar age (Table 6), contain opal but display considerable differences in the total water content of their host ashes (Table 6). This might be explained by the different rates of alteration. Sample 15 comes from the forest zone at an altitude of 290 m, whereas sample 9 is from the tundra zone at 695 m. The higher water content of the ash in sample 15 could reflect the enhanced weathering associated with the higher average temperature at 290 m.

Unfrozen water content

The unfrozen water content of soils under permafrost conditions is an essential part of their thermo-physical characteristics. It alters the thermal conductivity and heat capacity of volcanic ashes and soils and the rate at which volcanic glass and minerals may weather under permafrost conditions (Motenko & Kuznetsova, 2009a,b; Kuznetsova et al., 2011).

Figure 8 is a crossplot showing the unfrozen water content, Ww, and temperature for samples containing either allophane or opal. For temperature below –3°C (where the variation of unfrozen water content is insignificant) the value of Ww changes from 2 to 13% for ashes containing allophane (area 1) and from 0 to 3% for ashes containing opal (area 2).

According to many investigations (e.g. Henmi & Wada, 1976; Theng et al., 1982), allophane consists of hollow spheres with an outer diameter of about 4–5 nm, with defects in the wall texture which consequently form micro-pores (0.3–2.0 nm) and with an active specific surface area of 800 m2/g. The unfrozen water retention capacity of the ashes containing allophane could be related to the presence of these very small pores and the extensive active specific surface areas. This is in contrast to opals with a somewhat similar basic physical structure of more or less close-packed silica spheres of amorphous silica. However their active specific surface area is much lower (<0.5 m2/g; Jones & Segnit, 1969), which may be one of the explanations of the smaller amount of unfrozen water content in samples with opal.

Sample 24 is a rhyolitic ash from the exposure Krutoy on the bank of the Kamchata River at an elevation of 130 m. The ash has a high-salinity pore water content (see above) and is considerably older than the other samples (upper Lower Pleistocene cf. Holocene) we have investigated. The high salinity is thought to be related to the incorporation of sulfates (and sometimes chlorides) during eruptions and not to post-magmatic alteration (Girina, 1998). The sample is exceptional, as it is sandwiched between two layers of unsalted clayey loam (Fig. 9a, samples 24′ and 24′). In spite of its age the IR and TEM analyses indicated only the presence of opaline silica and no halloysite or other clay minerals from the alteration of the volcanic glass. The unfrozen water content is high, nearly 27% in the temperature range below the phase transition of water (< –3°C), whereas in the overlying and underlying samples (24′, 24′) Ww is only 2–3%. This exceptional high value may reflect the high salinity of sample 24. Figure 9b shows the unfrozen water content Wwvs. the negative temperature for these three samples. The limited alteration of sample 24 suggests its preservation by palaeopermafrost conditions persisting in the profile in spite of its low altitude (130 m).

Figure 10 shows the relationship between the unfrozen water content (Ww) and the age of all the investigated samples. There are four groups depending on sampling altitude: 1600 m, 1000–1300 m, 700–800 m, and 290 m and below. These are discussed below.

Group 1. The samples 1, 3, 5 and 7 are of similar age, ∼1500 years old. They are from the tundra zone on the Tolbachinskiy Pass at an altitude of 1600 m where permafrost is widespread. Samples 1, 3 and 5 were collected from the seasonal melting layer (the active layer) at the top of the soil, which is the site of cryogenic weathering. These soils experienced radical temperature fluctuations with repeated freezing and thawing cycles which are a controlling factor in the composition and structural changes taking place in this layer. All the samples contain andesitic glass and are associated with allophane but contain varying amounts of unfrozen water. These samples show differences in their colour and granulometric composition. Samples 1, 5 and 7 are brown in colour and have a similar grain size, whereas sample 3 is bright orange and is of finer grain size. The active freeze-thaw boundary at the time of observation was along a bright yellow, very fine grained ash layer. Drawing an analogy between the ash from this existing active boundary and sample 3, we conclude that an earlier freeze-thaw boundary was situated higher in the profile along the horizon where sample 3 was collected.

Two extreme weathering situations occur in this area; they are represented by two samples. Sample 7 is from a borehole in the permafrost. It has been permanently in the frozen state and was never involved in the active layer. Sample 3 is from a freeze-thaw boundary. Samples 1 and 5 were collected from the upper part of the active layer where there is no repeated temperature transition during the warm season. The highest content of unfrozen water was found in sample 3, and the smallest in samples 1 and 5, with a difference of 8% between them.

Group II. This group of samples (2, 4, 6 and 8) are from the tundra zone at an altitude of 1000–1300 m in the proximity of the Large Fissure Eruption at Tolbachik and near the Kamen volcano area. Their age is nearly 35 years. The glass from samples 2, 4 and 6 is of rhyolitic composition and is associated with opal; sample 8 has a basaltic composition (Girina, 1998) and is associated with allophane. The content of unfrozen water is low (0–0.5%).

Group III. These samples (913, 19–21 and 23) are from an altitude of 700–800 m from an area within the tundra zone with seasonal permafrost. The glass composition of samples 9, 11–13, 19 and 23 is rhyolitic and is associated with opal, sample 10 is andesitic and is associated with allophane, and samples 20 and 21 are basaltic and are associated with allophane (see Table 3, Braitseva et al., 1997; Ponomareva et al., 2007a,b). There is some correlation between the unfrozen water content, sample age and the presence of allophane or opal. The unfrozen water contents tend to be higher in the older samples or those containing allophane.

Group IV. The samples 14–18 were from an altitude of 290 m. There is only seasonal freezing at this altitude, but it is already within the upper part of the forest zone.

Samples 15 and 17 have rhyolitic glass and contain opal; samples 14, 16 and 18 have andesitic glass and contain allophane. The unfrozen water content is higher in samples with allophane than those with opal.

Thermal properties

Measurements of thermal conductivity were made for volcanic ashes of different age, mineralogy, chemistry and grain size, and have been reported in previous publications (Motenko et al., 2008; Kuznetsova & Motenko, 2011a,b).

These data show that when the density (ρd) and moisture content (W) of the ash samples change from 0.7 to 1.65 g/cm3 and from 10 to 80% respectively the thermal conductivity (λ) increases from 0.37 to 1.0 W/(m·K) in the thawed state and from 0.41 to 1.27 W/(m·K) in the frozen state. All physical properties of frozen grounds are controlled by unfrozen water (Ww) and ice content (W–Ww), because the thermal conductivity (λ) of ice (2.2 W/(m·K)) is much greater than that of water (0.56 W/(m·K)). The influence of the presence of allophane on Ww/(WWw) ratio and therefore on the thermal conductivity in frozen volcanic ashes is seen in Fig. 11. This shows the thermal conductivity of frozen ashes versus the unfrozen water and ice ratio at temperatures below –3°C. In the ashes containing opal (lines I and II), there is little unfrozen water with the combined unfrozen water and ice content ratio varying from 0 to 0.08. On line I there are data for samples 2, 4 and 19 in which there is practically no unfrozen water; on line II there are data for samples 6, 9, 12 and 17, with values of Ww ranging from 0.5 to 1.1%.

In the ashes containing allophane (lines III, IV and V) with unfrozen water contents ranging from 2 to 11%, the ratio Ww/(WWw) varies from 0.08 to 0.58. The line III averages data for samples 10, 14, 15, 20 and 21, where the unfrozen water content is 0.8–3%, The line IV for samples 1, 5 and 7 has values of 2–4%, and line V for samples 3 and 18 has an unfrozen water content of 10–11%.

A small change in the water to ice ratio, for example from 0.02 to 0.04 in the ashes containing opal, causes a decrease of thermal conductivity by a factor of 2. For ashes containing allophane there is a change in thermal conductivity by a factor of 3 due to a greater variation in this ratio from 0.19 to 0.58.

Conclusions

  1. The mineralogical analysis showed that allophane was found in ash samples with andesitic and basaltic glass, and opal in samples with rhyolitic glass.

  2. Even in the oldest ash deposits such as an example from the second half of early Pleistocene (Q12) there was no evidence of halloysite or other clays which could be linked with palaeopermafrost conditions.

  3. The amount of structural and adsorbed water depends on sample age and glass composition. The content is higher in the samples with basaltic and andesitic glass containing allophane, and lower in the samples with rhyolitic glass and opal.

  4. The content of unfrozen water is higher in volcanic ashes containing allophane than in those containing opal.

  5. The highest proportion of unfrozen water occurred in an ash with high soil water salinity.

  6. There are four distinct groups of ash samples related to altitude that can be distinguished by their volcanic glass alteration and unfrozen water contents. There are two high-altitude groups above 1600 m and 1100 m asl, an intermediate region consisting of samples from 700–800 m, and a low altitude region with samples from below 290 m. The first three regions are in tundra zones and the last one in the forest zone. The greatest alteration takes place in the high-altitude region in samples from the seasonally melting layer, especially in the interface between frozen and thawed grounds.

  7. Thermal conductivity in frozen soils depends on the unfrozen water and ice content and the mineralogical composition of the samples. Even a small change in the water to ice ratio in the ashes containing opal causes a decrease in thermal conductivity by a factor of 2. For ashes containing allophane, this change is by a factor of 3, reflecting the greater variation in the water to ice ratios of these ashes.

Experimental work was cattied out at the Department of Geocryology of Lomonosov Moscow State University as a part of the first author’s PhD work.

We wish to thank the following: S.W. Danielsen from SINTEF Building and Infrastructure for help in preparing the manuscript; J. Walmsley from SINTEF Materials and Chemistry for TEM analysis; M.F. Vigasina and L.V. Melchakova from Lomonosov Moscow State University for IR and thermal studies; Muravyev Ya.D., Deputy Director in Volcanology and Seismology Institute of the RAS FEB, for the field works in Kamchatka in 2009-2011, and for valuable consultations; C.V. Jeans, H.F.Shaw and M.J.Wilson for their constructive comments and editorial help.