By comparing high-quality volcanic gas and whole rock compositions, we calculated the apparent (observed) mass partition coefficients Kd* for 58 elements on six basaltic volcanoes located in arc and rift/hotspot settings. The inferred Kd* vary from � 1100 for sulfur to 0.0001 for zirconium, i.e., within seven orders of magnitude. Only 14 elements have Kd* > 1, including highly volatile S, Se, Te and halogens, as well as Tl, Re, Os, Bi, Cd, Au, In and As. Alkali metals have Kd* in the rangefrom 0.1 for Cs to 0.01 for Na. Partition coefficients of other rock-forming elements are <0.001. The partition coefficients for elements depend on element speciation and concentrations of ligand-forming elements in the gas such as sulfur and chlorine.
Elements transported in the gas predominantly as halides have higher partition coefficients in HCl-rich arc gases, whereas elements preferably forming sulfides, hydrides and free atoms, have higher Kd* in sulfur-rich, HCl-poor and reduced rift/hot-spot gases. Degassing directly from the free melt surface is negligible; deep gas passing through the erupting vent is quickly overwhelmed by the signal of low-pressure degassing. Equilibration of rising bubbles with the surrounding melt almost eliminates the difference between Kd* calculated for degassing lava flows (no connection with deep magma) and for lava lakes and open-vent volcanoes (convective mass exchange with deep magma takes place). Diffusion does not strongly affect the apparent partitioning of magmas degassing at surface. Gas bubbles growing in near-surface silicate melts at atmospheric pressure have a large density difference compared to the surrounding melt of 12–15 thousand times. This leads to the rapid expansion of such bubbles and a decrease in the thickness of the diffusion boundary layer in the melt due to its stretching around the growing bubble, which sharply decreases diffusion fractionation. As a result, the apparent partition coefficients (Kd*) for degassing basaltic volcanoes are close to the equilibrium ones (Kd) for most of the elements. The partition coefficients of volatile elements (S and Cl) calculated from the comparison of volcanic gas and rock compositions are in agreement with the values determined previously via experiments or theoretical modeling.

Pb isotope data are used to constrain the chemical contribution of the subducted components in the recycling beneath Klyuchevskoy volcano, the most active volcano in the Kamchatkan arc. The Pb isotope ratios of Klyuchevskoy basalts (206Pb/204Pb= 18.26–18.30, 207/Pb204Pb= 15.45–15.48, 208/Pb204Pb= 37.83–37.91) define a narrow range that falls within the Pacific mid-ocean ridge basalt (MORB) field and are among the least radiogenic island arc basalts measured to date. These data are similar to data from three other Quaternary Kamchatkan volcanoes: Tolbachik, Kumroch-Shish, and Maly Semiachik. In contrast, North Pacific sediments (primarily siliceous oozes) collected parallel to the Kamchatkan trench during Ocean Drilling Program Leg 145, have Pb isotope ratios (206Pb/204Pb= 18.51–18.78, 207Pb/204Pb= 15.56–15.64, 208Pb/204Pb= 38.49–38.75) that are more radiogenic than either the Klyuchevskoy basalts or Pacific MORB. Incorporation of even a small amount of sediment in the source of the Klyuchevskoy magmas would shift the Pb isotope ratios of the erupted basalts from the MORB field to more radiogenic values. The absence of 10Be and elevated Pb isotope ratios in the Kamchatkan volcanic lavas, despite the presence of distinctively radiogenic Pb in the North Pacific sediments makes it unlikely that sediments or sediment-derived fluids are involved in the source magmas beneath Kamchatka. The Kamchatkan arc thus represents an “end-member” whereby little or no sediment is involved in terms of elemental recycling and arc magma genesis. The major and trace elements, Pb, Sr and Nd isotope data of the Kamchatkan basalts are most consistently explained if derived from a fluid-fluxed, peridotitic mantle wedge source, wherein the fluid composition is dominantly controlled by dehydration of altered oceanic crust, imparting a radiogenic 87Sr/86Sr, and MORB-like Pb isotope signature to the mantle source. The erupted Klyuchevskoy lavas preserve a slab signature derived from incompatible elements that are strongly partitioned into the fluid. The 30 km of arc crust through which the Klyuchevskoy magmas traverse prior to eruption is not composed of older crust, but must be juvenile, similar in isotopic composition to MORB.

Detailed studies of volcanic tremor envelopes with frequencies ranging from 5.5⋅10-6 to 2.5⋅10-2 Hz (50 hrs - 40 sec), recorded during the Klyuchevskoi volcano eruptions of 1983 and 1984, revealed five major frequencies: 1.1⋅10-2 Hz (T1 = 1 min 34 sec), 2.5⋅10-3 Hz (T2 = 6 min 10 sec), 4.2⋅10-4 Hz (T3 = 40 min), 5.1⋅10-5 Hz (T4 = 5 hrs 30 min), 7.7⋅10-6 Hz (T5 = 36 hrs), as well as superpositions of their harmonics. In the 1993 eruption, fluctuations in the volcanic tremor envelopes have frequencies of TI = 2 hrs 48 min and TII = 6 hrs 12 min, which correspond to periodicities in the dynamics of eruptions identified by visual observations since 1932. The distribution of peak amplitudes has been found to vary in relation to eruption intensity—increasing eruption strength correlates with an increase in the amplitude of low frequency peaks, and vice versa. It is concluded that volcanic tremor allows monitoring of eruption dynamics. Possible reasons for the occurrence of periodicities are discussed, but a comprehensive model for this phenomenon has not yet been developed.

Data on the geology, petrography, and geochemistry of Middle–Late-Pleistocene rocks from the Tolbachik volcanic massif (Kamchatka, Klyuchevskaya group of volcanoes) are presented and compared with rocks from the neighboring Mount Povorotnaya, Klyuchevskaya group basement, and Holocene–historical Tolbachik monogenetic cones. Two volcanic series of lavas, middle-K and high-K, are found in the Tolbachik massif. The results of our data analysis and computer modeling of crystallization at different P–T–H2O–fO2 conditions allow us to reconstruct the geochemical history of the massif. The Tolbachik volcanic massif started to form earlier than 86 ka based on K–Ar dating. During the formation of the pedestal and the lower parts of the stratovolcanoes, the middle-K melts, depleted relative to NMORB, fractionated in water-rich conditions (about 3 of H2O). At the Late Pleistocene–Holocene boundary, a large fissure zone was initiated and the geodynamical regime changed. Upwelling associated with intra-arc rifting generated melting from the same mantle source that produced magmas more enriched in incompatible trace elements and subduction components; these magmas are high-K, not depleted relative to N-MORB melts with island arc signatures and rift-like characteristics. The fissure opening caused degassing during magma ascent, and the high-K melts fractionated at anhydrous conditions. These high-K rocks contributed to the formation of the upper parts of stratovolcanoes. At the beginning of Holocene, the high-K rocks became prevalent and formed cinder cones and associated lava fields along the fissure zone. However, some features, including 1975–1976 Northern Breakthrough, are represented by middle-K high-Mg rocks, suggesting that both middle-K and high-K melts still exist in the Tolbachik system. Our results show that fractional crystallization at different water conditions and a variably depleted upper mantle source are responsible for all observed variations in rocks within the Tolbachik volcanic massif. Sr–Nd isotopes are consistent with 2–4 crustal assimilation during formation of the pedestal and stratovolcanoes, while the young lava fields do not show evidence of crustal assimilation. Major and trace element data coupled with K–Ar dating provide strong evidence that Mount Povorotnaya, located in 8 km northeast of Plosky Tolbachik, is an old block of the Tolbachik massif pedestal and for the moment it is the oldest (306 ka) known object in Klyuchevskaya group of volcanoes.