Petrology and structure of the Esja Quarternary volcanic region, southwest Iceland

• Main Author: Friðleifsson, Ingvar Birgir
• Published: University of Oxford 1973
• Subjects: 551; Geology
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.255892

The stratigraphy of Esja is described and the chronology of the succession established by means of the geomagnetic time scale. The Esja volcanic succession is about 2.4km thick, and comprises olivine tholeiites (25%), tholeiites (68%), basaltic andesites (5%), icelandites and rhyolites (2%). Volcanism was active in the Esja region for just over one million years, and during this time span, .at least ten glaciations occurred in the region. The stratigraphic succession is therefore characterized by sequences of lava flows intercalated, at intervals, by thick subglacial hyaloclastite units. Two central volcanoes were active in the Esja region; the Kjalarnes volcano was active for about 0.6 million years and was succeeded after a short interval by the Stardalur volcano, which remained active for about 0.3 million years. Flood-basalt volcanism was concomitant with the central volcanism, and most of the olivine tholeiites are considered to have been erupted in fissures and shield volcanoes unrelated to the central volcanoes. Igneous activity apparently migrated eastwards with time, reflecting the westward crustal drift away from the active volcanic zone, which is a subaerial extension of the mid-Atlantic ridge. The volcanics are tilted and downfaulted towards the east. The irregular topography created by the glaciations in Esja repeatedly prevented lavas erupted in the active volcanic zone from spreading over the tectonically less active neighbour- hood, thus producing angular unconformities in the stratigraphic succession from which the tectonic history of the region can be read. The Esja evidence suggests that tectonic activity is chiefly restricted to the active volcanic zone, and that the crust becomes tectonically inactive soon after it has drifted away from the active zone. Intrusive activity in Esja can be divided into three phases. The oldest dykes in the region trend N 25°E and contemporaneous sheets dip towards the Kjalarnes peninsula, where the intrusive activity culminated in the formation of a multiple dolerite sheet. This intrusion may have been preceded by a caldera collapse in the Kjalarnes area. After the intrusion of the Kjalarnes dolerites the regional trend of dykes changed to N 40°E, and a narrow dyke swarm (representing up to 20% dilation) cut across the Kjalarnes central volcano. The dyke swarm was succeeded by cone sheets focussing to the south of Leidhamrar, and the second phase culminated in the intrusion of large dolerite sheets in Þverfell and Lauganípa. Following a brief interval, during which flood-basalt volcanism was dominant in Esja, the Stardalur central volcano became active and, during its life span, minor intrusions were predominantly in sheet form. Caldera collapse in the Stardalur volcano was followed by the intrusion of basic cone sheets, large dolerite sheets, a sill and finally a laccolith within the caldera. Long after the caldera had been filled the caldera fault zone dominated over the regional fault pattern at depth so that basic and acid volcanics alike were erupted concentrically with and parasitically to the Stardalur caldera. Large dolerite intrusions in Esja are found chiefly within or at the boundaries of the thick hyaloclastite units, and there is evidence of dykes cutting straight through lava successions, but spreading out laterally to form sill-like bodies once they enter the less coherent hyaloclastites. A survey of the literature shows that the majority of large basic intrusions .in Iceland are accommodated in relatively soft and "structureless" host rocks, such as tuffaceous hyaloclastites, sediments, vent and caldera agglomerates, hydrothermally propylitized lavas, and "hot" and still partly liquid acid intrusive material. The majority of the large intrusions are in the form of inclined sheets, but sills and laccoliths are formed when the intrusions are emplaced at shallow levels (perhaps less than 1km) in the crust. The coincidence of central volcanoes having a great bulk of shallow level intrusions, with positive gravity anomalies, and the sites of shallow depth to layer 3 in Iceland strongly suggests that crustal layer 3 consists mostly of basic intrusions. A comparison of the densities of primary and secondary minerals of tholeiitic rocks suggests that infilling of vesicles of porous basalt lavas by secondary minerals will not make the rock as dense as a non-porous rock of the same composition. The estimated density difference of 0.2g/cm³ between crustal layers 2 and 3 can apparently not be ascribed to secondary alteration of subaerial lavas, but can readily be explained by a transition from altered lavas to non-porous intrusives. It is proposed that the sharp boundary between layers 2 and 3 results from the lavas at the base of layer 2 reaching a degree of alteration at which the rock becomes sufficiently incoherent to accommodate large basic intrusions. The "metamorphic boundary" proposed by Pálmason (1971) to explain the correlation between the thermal gradient and depth to layer 3 in Iceland is not therefore primarily a density boundary, but a boundary at which the lavas loose their strength as a result of alteration and host voluminous dense intrusives. The large scale features of crustal layer 3 in Iceland can be explained within the framework of this model. There is a complete range in composition from olivine tholeiites to rhyolites in the Esja volcanic succession, and the majority of the rocks contain some phenocrysts. Crystal fractionation appears to be a feasible mechanism to explain the chemical variation within at least the basaltic rocks in Esja, but whether the intermediate and acid rocks are formed by extensive fractionation or by partial melting of crustal material cannot be answered. The apparent coincidence in time of the emplacement of large basic intrusives and the commencement of intermediate and acid volcanism in eastern Esja may suggest that the rise of voluminous basic magmas has raised the thermal gradient sufficiently to produce the intermediate and acid rocks by partial melting at the base of the crust. Positive gravity anomalies associated with the Kjalarnes and Stardalur central volcanoes are attributed to high level intrusives in the core regions of the two centres. Specific gravity measurements of the chemically analysed rocks from Esja show a range of densities from about 2.5 (rhyolite) to about 3.15g/cm³ (olivine tholeiite). Local gravity anomalies commonly found associated with central volcanoes are probably due both to local concentrations of rocks of different chemistry and to a high percentage of intrusives. A comparison of the average density of the crust in eastern Iceland and that of the crust on the Iceland-Faeroe ridge suggests that a considerable part of the negative Bouguer gravity anomaly of Iceland (Einarsson 1954) can be explained in terms of geochemical differences between the volcanics of Iceland and those of the surrounding areas. The similarity of the gravity profiles from the aseismic Iceland-Faeroe ridge and from the active Reykjanes ridge to the centre of Iceland suggests that, if a hot spot contributes to the bowl-shaped gravity anomaly of Iceland (Bott et al 1971), then it is not connected with layer 4 under the Reykjanes ridge. A positive magnetic anomaly associated with the Stardalur caldera is explained in terms of a thick pile of normal polarity eruptives within the caldera being surrounded essentially by reverse polarity eruptives. The combined effects of a high magnetite content (which may be caused by unusually high partial pressure of oxygen in the melt) and a high palaeofield strength may cause the very high magnetic intensity of lavas, which give rise to a sharp maximum within the Stardalur magnetic anomaly. Assuming a common cause for three other strong magnetic anomalies, which, with Stardalur, lie on a straight line and are separated regularly in space and time, the possibility of a mantle controlled "high partial pressure of oxygen spot" migrating at half the spreading speed along the spreading axis is discussed.