Peralkaline acid volcanic rocks of oceanic islands

Occurrences of peralkaline acid volcanic rocks on oceanic islands are reviewed. Peralkaline differentiates are usually associated with mildly alkaline or transitional basalts and often with related sodic intermediate rocks. A compositional gap between basaltic and salic rocks is not invariably present. Although a comenditic end member is more usual in the oceanic suites, pantellerites are particularly well developed on Socorro Island and also on Gran Canaria where they form extensive ignimbrite sheets. There may be a genetic distinction between peralkaline rocks of islands which lie near the crests of oceanic rises and those which are built on broad submarine plateaux.     

Crystal fractionation of the basalt comendite series of the mount Edziza volcanic complex, British Columbia: Major and trace elements

The Mount Edziza Volcanic Complex in north-central British Columbia includes a group of overlapping basaltic shields, salic composite volcanoes, domes and small calderas that range in age from 7.5 Ma to less than 2000 years B.P. The volcanic assemblage is chemically bimodal, comprising voluminous alkali olivine basalt and hawaiite, a salic suite of mainly peralkaline trachyte and comendite, plus a relatively small volume of intermediate rocks (trachybasalt, tristanite, mugearite, benmoreite). The complex is the product of five cycles of magmatic activity, each of which began with alkali olivine basalt and culminated with the eruption of salic magma. The regular chemical variation shown by almost 100 major- and trace-element analyses suggests a genetic lineage between the basic and salic members of each cycle. Least-squares mathematical modelling, indicates that the salic rocks (trachyte and comendite) have formed by fractionation of observed phenocryst and cumulate nodule mineral phases from a common alkali olivine basalt parent magma.Hawaiite is thought to be a cumulate rock, formed by partial fractionation and feldspar accumulation within rising columns of primary alkali olivine basalt. Fractionation leading from alkali olivine basalt through trachybasalt and trachyte to comendite is believed to have taken place where primary basalt became trapped in large crustal reservoirs. The early removal of olivine, clinopyroxene and plagioclase, leading to a trachytic residuum, and subsequent fractionation of mainly alkali feldspar, leading to the peralkaline end members, is consistent with major- and trace-element variation and with isotopic and REE data.The chemical diversity of the complex is attributed to its location over a zone of crustal extension where mantle-derived basalt, trapped in large high-level reservoirs, underwent prolonged fractionation.     

Silicic peralkaline volcanic rocks of the afar depression (Ethiopia)

Three main types of recent volcanism may be distinguished in the Afar Depression: 1) oceanic volcanism of the axial ranges; 2) volcanism along the margins where an attenuated sialic crust probably occurs; 3) mainly fissural volcanism of Central-Southern Afar, with associated central volcanoes, similar as a whole to the volcanism of the Ethiopian Rift Valley. Peralkaline silicic volcanic rocks are found in all the three groups but showing some different characteristics which seem related to their geological location and which probably reflect different sources. Moreover emplacement of peralkaline granitic bodies, associated with volcanics of the same composition, marks the first stage of formation of the Afar Depression, in the Early Miocene. Axial Ranges: Erta’Ale and Boina volcanic ranges indicate that peralkaline rocks are the final liquids produced by fractionation of basalt in shallow magma chambers of central volcanoes. The parental magma is a transitional type of basalt with a mildly alkalic affinity, which fractionated under lowpH₂O-pO₂ conditions. Transition to peralkaline liquids is realized without passing a «true» trachytic (low silica) stage. The first peralkaline liquid is a low silica comendite and evidence exists that «plagioclase effect» was active in determining the first peralkalinity. Within the peralkaline field a fractionation mainly controlled by alkali feldspar progressively increases the peralkalinity and silica oversaturation of residual liquids (transition from comendites to pantellerites). The most peralkaline pantellerites of Boina are produced by fractionation of an alkali feldspar of constant composition (Ab_65–68 Or_35–32) suggesting that these liquids lie on a «low temperature zone» of the peralkaline oversaturated system. Marginal Units: On the borders of the depression peralkaline silicics are found in volcanic massifs mainly made of metaluminous silicic products. Petrology and geochemistry suggest a complex origin. Crystal fractionation, contamination with sialic crust and chemical changes related to a volatile rich phase, all these processes probably played a role in the genesis of these peralkaline silicic rocks. Central-Southern Afar Fissural Volcanism: Mildly alkaline basalts are associated with peralkaline and metaluminous silicics; intermediate rocks are very scanty. Fractionation from deep seated magmatic bodies with selective eruptivity and partial melting at depth of associated basalts or of a common source material are possible genetic mechanisms.     

Silicic volcanism in Iceland: Composition and distribution within the active volcanic zones

Silicic volcanic rocks within the active volcanic zones of Iceland are mainly confined to central volcanoes. The volcanic zones of Iceland can be divided into rift zones and flank zones. Each of these zones contains several central volcanoes, most of which have produced minor amounts of silicic rocks. The silicic rocks occur as lavas and domes or as tephra layers, welded tuffs and ignimbrites, formed both in effusive and explosive eruptions. They tend to be glassy or very fine-grained, containing small amounts of phenocrysts. Plagioclase (andesine–oligoclase), anorthoclase or occasionally sanidine coexist with minerals such as augite, fayalite, pigeonite, orthopyroxene and magnetite. Quartz phenocrysts are exceedingly rare. Zoning of phenocrysts is limited and the pattern is variable. A set of 90 samples representing all active central volcanoes that have erupted silicic rocks was analysed for major- and trace-elements. The silicic rocks can be classified as dacites, trachytes, low-alkali rhyolites and alkalic rhyolites. Some of the trachytes and alkalic rhyolites are peralkaline (mostly comenditic). Trachytes and alkalic rhyolites are only found within the flank zones, while dacites and low-alkali rhyolites are mostly confined to the rift zones. The Icelandic rhyolites plot close to the thermal minimum in the “granite” system, while dacites and trachytes plot within the plagioclase field and towards the alkali feldspar temperature minimum. The silicic rocks are relatively Fe-rich and Ca-poor indicating low water pressure in the source. Trace element concentrations follow similar patterns in most central volcanoes. Exceptions are Torfajökull where silicic rocks display a negative correlation of Ba to Th and unusually high Th-contents, and the western flank zone where Ba-concentrations are highly variable. The ratios of different high field-strength elements are generally similar within each central volcano or region, which probably reflects different ratios in the source materials. Isotope systematics indicate that the silicic rocks are derived from older basaltic rocks similar to those from the same volcano, and that meteoric water has played a role in the genesis of the silicic rocks. Traditionally, the petrogenesis of silicic rocks in Iceland has been explained by various models of fractional crystallization or partial melting. The available data seems to be better explained by near-solidus differentiation than by near-liquidus differentiation. The silicic minimum melts can be extracted from the rigid framework of the near-solidus source by the process of solidification front instability or by deformation-assisted melt segregation. The source of the silicic rocks is within the intrusive complex beneath a central volcano rather than in a large, long-lived magma chamber.     

Petrography and mineralogy of the peralkaline silicic rocks

The mineralogy and textures of quartz-normative peralkaline trachytes and the comendites and pantellerites are described in detail, using published examples and new information on rocks from the Naivasha area, central Kenya. Mineral assemblages are tabulated for each rock type, and attention is drawn to the mineralogical differences between glassy and crystalline varieties. The petrography of peralkaline ignimbrites is briefly discussed. The chemical compositions and optical properties of the major mineral phases in peralkaline silicic rocks are tabulated, with a brief discussion on their stability relationships.     

Volcanological aspects of peralkaline silicic welded ash-flow tuffs

Peralkaline silicic welded ash-flow tuffs differ characteristically in a number of properties from most calc-alkaline welded tuffs, due to their generally lower viscosity and higher temperatures. For example, individual cooling units are relatively small (less than 30 m thick, less than 5 km³ in volume); rocks can be thoroughly welded and crystallized to feldspar, quartz, and mafic minerals; several primary deformational structures (e.g. lineations, stretching of pumice, folds, ramp structures) indicate late stage laminar creep, resulting from the low yield strength of the nearly homogeneous glass of very low viscosity. Theoretical considerations also suggest that peralkaline melts are of low viscosity and high temperature, as inferred from,e.g., their chemical composition (high iron- and alkali-, and low alumina-concentrations). The low viscosity may also be due to trapping of volatiles. Absence or paucity of OH-bearing phenocryst phases, paucity of pyroclastic rocks, other than ash flow tuffs, formed from highly explosive eruptions, and apparently high crystallization temperatures, indicate that peralkaline silicic magmas are comparatively dry. The common occurrence of peralkaline ash-flow tuffs may be due to an increased water content of the magmas, resulting also in amphibole phenocrysts in some welded tuffs, or to specific volcanotectonic conditions. Ash flows of peralkaline composition move as particularly dense particulate flows. This type of flowage and the very rapid welding of the low viscosity glass lead to a high degree of homogenization of the fine glass shards. This in turn inhibits complete degassing of the collapsing ash flow. Semiclosed systems result where gas overpressures can develop and where volatiles play an important role by fluxing crystallization and transporting dissolved matter. Several types of vesicles can form under these conditions: (a) Spherical vesicles within collapsed ash and pumice particles formed after deposition of the ash flow. (b) Round or irregular vesicles transsecting pyroclastic particles, vesicle sheets, and large cavities, several m in diameter, may form in a largely homogenized ash-flow tuff beneath tightly welded layers. (c) Lensoid cavities formed during granophyric crystallization of large pumice particles. Small ash particles of peralkaline composition may assume spherical shapes due to their low viscosity and in some cases, expansion of bubbles. They form during transport and are preserved under low load pressure in the top part of cooling units. Globule lavas and most froth flows are interpreted as welded ash-flow tuffs, some of their unusual features being due to their peralkaline composition.     

Distribution of Ca in highly fractionated peralkaline magmas

Many peralkaline rhyolites and granites contain less than 0.15 wt.% CaO. In contrast, strongly fractionated peralkaline nepheline syenites and phonolites usually contain greater than 0.5 wt.% CaO. Consideration of known distributions of Ca between crystals and liquid in conjunction with crystal fractionation does not provide an adequate explanation of the contrasting levels of Ca depletion observed. Examination of the suites of late-crystallizing accessory phases in peralkaline rocks suggests that Ca is more soluble in undersaturated magmas than in over-saturated magmas. Activities for CaO in silicic and phonolitic rocks are calculated and the latter have higher CaO activities than the former and this may manifest itself in the different suites of accessory phases and levels of Ca depletion noted in natural rocks.     

The acid rocks of the ocean basins

The occurrences of acid rocks in the ocean basins are reviewed, and their genesis is discussed. It is concluded that fractional crystallisation of basalt magma at low pressure or at high pressure could adequately explain the dominant chemical features of most occurrences. For those cases where anomalously high initial⁸⁷Sr/⁸⁶Sr ratios compared to the basalts are observed, formation by repeated remobilisation of early formed acid material associated with the volcanic pile is suggested. The Daly gap is likely to be a distortion introduced by the limitations of subaerial sampling, and subsequent bias during analysis and interpretation. Although volatile transfer could be a subsidiary factor in the evolution of the acid rocks, there is insufficient data available in any single case to demonstrate the effectiveness of such a process.     

The quaternary caldera volcano emuruangogolak,kenya rift,and the petrology of a bimodal ferrobasalt-pantelleritic trachyte association

Emuruangogolak is a Quaternary basalitrachyte volcano situated in the Suguta graben of the northern Kenva rift, and probably erupted last early in this century. Following the construction of an early trachytic shield volcano, two episodes of caldera collapse occurred. each preceded by explosive pvroclastic activity. Post-calelera volcanism consisted of alternating phases of basalt and trachyte eruption. The basic lavas are high-Ti ferrobasalts of a mildly alkaline ‘transitional’ composition and the trachytes are peralkaline and oversaturated. A distinct compositional bimodality exists and no rocks in the range 49–59°. SiO, have been found. Major and trace element analyses suggest that the trachytes are genetically related to the basalts. Associations of almost identical lavas occur in Ethiopia. Pantelleria and the Azores but with the presence of intermediate terms Fractional crystallization is the mechanism currently preferred to account for the origin of the trachytes. The ‘Daly gap’ may be a consequence of a crystallization process which limits the volume of intermediate magma available at any time. In addition, the physical properties and spatial distribution of the different magmas probably discriminate against the cruption of lavas of intermediate composition.     

Compositional changes during crystallization of some peralkaline silicic lavas of the Kenya rift valley

Comparison of crystalline and glassy comendites shows that compositional changes during solidification involve losses of Na, Cs, Cl, F, gain of Sr, and gains and losses of REE at near-solidus temperatures. Variation in a thick trachyte lava suggests deuteric mobility of the same elements, while variation in a trachyte-comendite suite shows that Fe loss can also occur. Al, Si, Zr, Hf, Nb, Ta, Th, U, and Rb are not significantly affected, nor are the proportions of the REE, and these elements can be relied upon in petrochemical studies of crystalline peralkaline silicic rocks.     

Role of water in pantellerite genesis

Sub-solidus, mass transfer experiments along a temperature gradient, in which water reacted with alkali olivine basalt, trachybasalt, and peridotite, have yielded acidic extracts directly from basic material. It is proposed that the process may also occur under magmatic conditions by progressive and efficient enrichment of the upper portions of a magma reservoir in the constituents of petrogeny’s residua via an aqueous fluid. This hypothesis could explain the production of silica-oversaturated rocks from nepheline-normative parent material, and their association with basalts of alkaline affinity but without volumetrically significant compositional intermediates. In the case of the trachybasalt experiment the transfer products were also peralkaline and the hypothesis may thus be extended to peralkaline oversaturated rocks.     

Large-scale silicic alkalic magmatism associated with the Buckhorn Caldera, Trans-Pecos Texas, USA: comparison with Pantelleria, Italy

Three major rhyolite systems in the northeastern Davis and adjacent Barrilla Mountains include lava units that bracketed a large pantelleritic ignimbrite (Gomez Tuff) in rapid eruptions spanning 300,000 years. Extensive silicic lavas formed the shields of the Star Mountain Formation (37.2 Ma-K/Ar; 36.84 Ma ³⁹Ar/⁴⁰Ar), and the Adobe Canyon Formation (37.1 Ma-K/Ar; 36.51-³⁹Ar/⁴⁰Ar). The Gomez Tuff (36.6 Ma-K/Ar; 36.74-³⁹Ar/⁴⁰Ar) blanketed a large region around the 18×24 km diameter Buckhorn caldera, within which it ponded, forming sections up to 500 m thick. Gomez eruption was preceded by pantelleritic rhyolite domes (36.87, 36.91 Ma-³⁹Ar/⁴⁰Ar), some of which blocked movement of Star Mountain lava flows. Following collapse, the Buckhorn caldera was filled by trachyte lava. Adobe Canyon rhyolite lavas then covered much of the region. Star Mountain Formation (~220 km³) is composed of multiple flows ranging from quartz trachyte to mildly peralkalic rhyolite; three major types form a total of at least six major flows in the northeastern Davis Mountains. Adobe Canyon Formation (~125 km³) contains fewer flows, some up to 180 m thick, of chemically homogenous, mildly peralkalic comendite, extending up to 40 km. Gomez Tuff (~220 km³) may represent the largest known pantellerite. It is typically less than 100 m thick in extra-caldera sections, where it shows a pyroclastic base and top, although interiors are commonly rheomorphic, containing flow banding and ramp structures. Most sections contain one cooling unit; two sections contain a smaller, upper cooling unit. Chemically, the tuff is fairly homogeneous, but is more evolved than early pantelleritic domes. Overall, although Davis Mountains silicic units were generated through open system processes, the pantellerites appear to have evolved by processes dominated by extensive fractional crystallization from parental trachytes similar to that erupted in pre- and post-caldera lavas. Comparison with the Pantelleria volcano suggests that the most likely parental magma for the Buckhorn series is transitional basalt, similar to that erupted in minor, younger Basin and Range volcanism after about 24 Ma. Roughly contemporaneous mafic lavas associated with the Buckhorn caldera appear to have assimilated or mixed with crustal melts, and, generally, may not be regarded as mafic precursors of the Buckhorn silicic rocks, They thus form a false Daly Gap as opposed to the true basalt/trachyte Daly gap of Pantelleria. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users. This paper constitutes part of a special issue dedicated to Bill Bonnichsen on the petrogenesis and volcanology of anorogenic rhyolites.     

Le complexe filonien de Fuerteventura (Iles Canaries)

Two sectors in the island have been distinguished which depend on the intensity of the penetration of the dyke swarm: the central-western sector, where the proportion of dykes in relation to the host rocks is — in some cases — more than 90% and all the rest of the island, where the dykes are in less proportion. In the central-western sector the predominant direction varies from N-10^o-E to N-20^o-E, whilst in the rest of the island a clear constant direction does not exist, although there are some directional trends more or less well defined. Most dykes are of basaltic composition, but there are also salic dykes (trachytic and phonolitic) and camptonites. The dykes from the western sector have suffered intense transformation processes (albitization and amphibolization), so that it is frequent to find dykes that do not preserve practically any one of their primary minerals. As the albitization process also affects the host rocks and has a regional development; we can speak of a regional metamorphism equivalent in its intensity to the Green Schist. Three main generations of dykes have been established: 1st) those of an approximately N-S direction, forming the older generation, constituted by multiple injections, 2nd) those genetically related to the Basaltic Series I, 3rd) those of salic composition, closely related with the syenitic-trachytic intrusions.     

Trans-Atlantic correlation of eruptive sequences and individual silicic volcanic units within the Paraná-Etendeka igneous province

The Mesozoic volcanic rocks of the Serra Geral Formation in the Paraná Basin, South America, and of the Etendeka Group in northwestern Namibia were erupted shortly before the opening of the South Atlantic. The major widespread silicic volcanic units in the Etendeka Group are interpreted as rheoignimbrites (Milner et al., 1992) and are interbedded with tholeiitic basalts and basaltic andesites.The southern portion of the Etendeka Group is subdivided into a basal Awahab Formation which is overlain disconformably by the Tafelberg Formation. Both formations contain silicic and mafic units. Bulk composition, initial ⁸⁷Sr/⁸⁶Sr ratios, phenocryst assemblages and mineral compositions are used to correlate silicic units of the Awahab Formation with the basal units of the Palmas silicic volcanic rocks in the southern Paraná Basin. Silicic units of the Tafelberg Formation can similarly be correlated with silicic units in the upper portion of the Palmas succession, which are also disconformable on the units below them. Not all silicic units in these successions are present in both the Etendeka and Paraná areas, but where correlation of individual units is possible, then this is found to be consistent with the overall stratigraphic sequence.Silicic units in the Awahab Formation were erupted from the Messum Igneous Complex in Namibia and their correlation into Brazil indicates that individual eruptive units must have travelled over 340 km from their source. Serial changes in the composition of silicic units in the Awahab Formation and their correlatives indicates that they were erupted from a single magma system from which a total of ˜ 8600 km³ of material was erupted.     

Some data on the comendite type area of S. Pietro and S. Antioco islands,Sardinia

A summary of the available data on the peralkaline rocks of S. Pietro and S. Antioco islands, together with, new chemical analyses and some preliminary K-Ar ages are reported. Peralkaline rocks occur as ignimbrites, lava flows and domes usually deeply affected by hydrothermal alteration. Pantelleritic varieties are found within the dominantly comenditic association, which display K₂O contents higher than Na₂O ones. K-Ar data indicate that these peralkaline rocks have a middle Miocene age (? 15 m.y.). They occur in close field association with coheval andesitic and subalkaline acid volcanics belonging to the final products of the Tertiary calc-alkaline volcanic cicle of Sardinia.     

The influence of trace amounts of water on the viscosity of rhyolites

 As a major volatile in volcanic systems, water has a significant influence on the rheological properties of silicic magmas. This is especially so at minor water contents relevant to the emplacement of silicic lavas. To investigate the influence of water on the viscosity of natural rhyolitic obsidians, a novel strategy has been adopted employing parallel-plate and micropenetration techniques. Viscosities have been determined on three types of material: (a) raw water-bearing obsidians; (b) remelted (1650  °C, 1 atm) degassed glasses of the obsidians; and (c) hydrothermally hydrated (1300  °C, 3 kbar) obsidians. Ten natural rhyolitic obsidians (peraluminous, calc-alkaline and peralkaline) were employed: seven originated from lava flows and contained <0.2 wt.% H₂O, two samples were F-rich from pyroclastic successions, and one was an obsidian cobble with 1.5 wt.% water also associated with pyroclastic units. Melt compositions and water contents were stable during viscometry. The measured decreases in activation energies of viscous flow and viscosity with small amounts of water are much greater than the Shaw calculation scheme predicts. In addition, a marked non-linear decrease in η exists with increasing water content. In contrast to the case for peralkaline rhyolites, 0.1–0.2 wt.% water decreases activation energies significantly (up to 30%) for calc-alkaline compositions. These results have important implications for the ease of near-surface degassing of silicic magmas during emplacement and permit the testing of calculational models for viscosity, largely based on synthetic systems. Received: 7 July 1997 / Accepted: 6 April 1998     

A review of Karoo vulcanicity in Southern Africa

Vocanic rocks of Karoo age which today cover more than 140,000 km² of the southern African sub-continent occur as scattered outliers representing eroded remnants of an originally more extensive volcanic province. The rocks are best preserved in central southern Africa including Lesotho, and the continental margin areas of Namibia in the west and Mozambique. Zambabwe, Swaziland and South Africa in the east. Extensive lava fields (yet few volcanoes) dykes, sills, layered intrusions and at least two major dyke swarms characterise the region. Volcano-stratigraphic and geochemical mapping have been used to subdivide the volcanic successions found in the different areas and recently adopted nomenclature is presented. Considerable more variability and complexity occurs in the volcanic succession than was previously recognised: geochemical variations and stratigraphic relationships indicating that four major provinces can be recognised. Rocks from the central Karoo areas are primarily of basaltic composition whereas those from the western and eastern marginal areas include mafic basic, intermediate and acid types. Emplacement of rocks such as carbonatites, nephelinites, and picrite basalts enriched in incompatible elements, indicate that derivation from a heterogeneously enriched source played a significant role in the petrogenesis of a large proportion of the Karoo mafic and basic rocks. Age relationships of the volcanic rocks reveal that vulcanicity extended over a period of 130 m.y. from mid-Triassic to early Cretaceous time.     

Variation trends in the alkaline salic rocks of tenerife

The ratio salic rocks/basalts is higher in Tenerife than in other Atlantic islands. It is also surprising the number of intermediate types between phonolites and trachytes and the basalts. The salic rocks of Tenerife have been grouped in to two large units, one related to the edifice of « Las Cañadas » and the other to « Teide-Pico Viejo ». The top of the former collapsed and the latter was built in the Caldera thus formed. Both units belong to a middle atlantic series, but the atlantic character of « Teide-Pico Viejo » is stronger. A clear alkalinitization can be observed during the whole evolution. Most of the materials which are related to the Cañadas edifice are near the saturation line, and they must be classed as phonolites and Na-trachytes. In these rocks a variation trend related to that of the former alkaline basalts can be observed. In the latest episodes of their evolution cutaxite and pumice emissions appeared with great intensity. The « Teide-Pico Viejo » lava-flows are always of phonolite types with high amounts of normative nepheline. These materials also represented the end of the differentiation trend of an alkaline basaltic series, which started after the Cañadas edifice was built. This second trend ended in less silica-rich rocks than those of the Cañadas series.     

Silicic differentiates of abyssal oceanic magmas: Evidence for late-magmatic vapor transport of potassium

Massive, nearly holocrystalline dolerites from DSDP Hole 417D contain from 0.5 to 1.5% of granophyric patches composed mainly of Na-plagioclase and quartz. These patches are compositionally similar to other crystalline silicic rocks from oceanic spreading centers and differ from rarer abyssal silicic glasses. Crystalline varieties withSiO₂60wt.% generally haveNa/K>10, whereas silicic glasses have Na/K in the range 3–6. While crystal fractionation readily accounts for the Na₂O and K₂O contents of abyssal silicic glasses, both the 417D granophyres and other crystalline abyssal silicic rocks have much lower K₂O than that predicted by any reasonable crystal-liquid fractionation model. We propose that high-temperature vapor phase transport is responsible for removal of potassium during late-stage crystallization of these rocks. This allows for the formation of cogenetic silicic glassy and crystalline rocks with greatly different Na/K ratios. These observations and interpretations lead to a more confident assignment of high Na/K silicic rocks of oceanic and ophiolitic environments to a cogenetic origin with basaltic oceanic crust.