NE Iceland Research Articles

Temporal variation of chemical and mechanical weathering in NE Iceland: Evaluation of a steady-state model of erosion

This study critically assesses the temporal sensitivity of the steady-state model of erosion that has been applied to chemical and mechanical weathering studies of volcanic islands and the continents, using only one sample from each catchment. The model assumes a geochemical mass balance between the initially unweathered rock of a drainage basin and the dissolved and solid loads of the river. Chemical composition of 178 samples of suspended and dissolved inorganic river constituents, collected in 1998–2002, were studied from five basaltic river catchments in NE Iceland. The Hydrological Service in Iceland has monitored the discharge and the total suspended inorganic matter concentration (SIM) of the glacial rivers for ~ four decades, making it possible to compare modelled and measured SIM fluxes. Concentration of SIM and grain size increased with discharge. As proportion of clay size particles in the SIM samples increased, concentrations of insoluble elements increased and of soluble decreased. The highest proportion of altered basaltic glass was in the clay size particles. The concentration ratio of insoluble elements in the SIM was used along with data on chemical composition of unweathered rocks (high-Mg basalts, tholeiites, rhyolites) to calculate the pristine composition of the original catchment rocks. The calculated rhyolite proportions compare nicely with area-weighted average proportions, from geological maps of these catchments. The calculated composition of the unweathered bedrock was used in the steady-state model, together with the chemical composition of the suspended and dissolved constituents of the river. Seasonal changes in dissolved constituent concentrations resulted in too low modelled concentrations of SIM mod at high discharge (and too high SIM mod at low discharge). Samples collected at annual average river dissolved load yielded SIM mod concentrations close to the measured ones. According to the model, the studied rivers had specific mechanical denudation rates of 1.3–3.0 kg/m 2/yr whereas the average measured rates were 0.8–3.5 kg/m 2/yr which are among the highest on Earth. This study validates the use of a steady-state model of erosion to estimate mechanical weathering rates at the scale of a river catchment when the collected riverine dissolved load represents the average chemical composition over a mean hydrological year.

Constraints on the dynamics of subglacial basalt eruptions from geological and geochemical observations at Kverkfjöll, NE-Iceland

The Kverkfjoll area, NE Iceland is characterised by subglacial basalt pillow lavas erupted under thick ice during the last major glaciation in Iceland. The water contents of slightly vesiculated glassy rims of pillows in six localities range from 0.85±0.03 to 1.04±0.03 wt %. The water content measurements allow the ice thickness to be estimated at between 1.2 and 1.6 km, with the range reflecting the uncertainty in the CO2 and water contents of the melt. The upper estimates agree with other observations and models that the ice thickness in the centre of Iceland was 1.5–2.0 km at the time of the last glacial maximum. Many of the pillows in the Kverkfjoll area are characterised by vesiculated cores (40–60% vesicles) surrounded by a thick outer zone of moderately vesicular basalt (15–20% vesicles). The core contains ∼1 mm diameter spherical vesicles distributed uniformly. This observation suggests a sudden decompression and vesiculation of the still molten core followed by rapid cooling. The cores are attributed to a jokulhlaup in which melt water created by the eruption is suddenly released reducing the environmental pressure. Mass balance and solubility relationships for water allow a pressure decrease to be calculated from the observed change of vesicularity of between 4.4 and 4.7 MPa depressurization equivalent to a drop in the water level in the range 440–470 m. Consideration of the thickness of solid crust around the molten cores at the time of the jokulhlaup indicates an interval of 1–3 days between pillow emplacement and the jokulhlaup. Upper limits for ice melting rates of order 10−3 m/s are indicated. This interpretation suggests that jokulhlaups can reactivate eruptions.

Reconstruction of the largest Holocene jökulhlaup within Jökulsá á Fjöllum, NE Iceland

Glacial outburst floods (jökulhlaups) have a significant role for landscape evolution in NE Iceland. A number of jökulhlaups have routed from the northern margin of Vatnajökull during the Holocene. In this study, reconstruction of the largest Holocene jökulhlaup along Jökulsá á Fjöllum, NE Iceland was undertaken using the HEC-RAS hydraulic modelling and HEC-GeoRAS flood mapping techniques with a Digital Elevation Model (DEM) derived from ERS-InSAR data and field-based wash limit evidence. The largest jökulhlaup produced extensive erosional and depositional landforms across an inundated area of ∼1390 km 2 and is calculated to have had a peak discharge of 0.9×10 6 m 3 s −1. Power per unit area within this jökulhlaup varied from 6 to 46,000 W m −2. Jökulhlaup hydraulics are related to geomorphogical evidence at three key sites: in Vaðalda, Upptyppingar and Möðrudalur sub-areas in order to explain the abrupt spatial variation of the flood characteristics on a regional scale and to relate erosional and depositional features to spatial variations in jökulhlaup hydraulics. These process-form relationships of the largest jökulhlaup along the Jökulsá á Fjöllum are compared with large outburst floods elsewhere. The largest Jökulsá á Fjöllum jökulhlaup had a factor of ∼20 times smaller discharge and ∼a factor of 20 times lower power per unit area than Altai palaeoflood—the largest known flood on the Earth.

Reply to the: Comment on “Pb isotopic analysis of standards and samples using a 207Pb– 204Pb double spike and thallium to correct for mass bias with a double focusing MC-ICP-MS” by Baker et al.

The following Comment addresses the issueconcerning the accuracy of the thallium-corrected Pbisotope data from Theistareykir, NE Iceland (Strackeet al., 2003), and the drill core recovered by theHawaii Scientific Drilling Project (HSDP; Blichert-Toft et al., 2003). Baker et al. (2004) claim that,because of improper mass-spectrometry techniques,these data sets are inaccurate to a degree that by farexceeds the errors reported in these two papers.Baker et al. (2004) dismiss the Theistareykir(Stracke et al., 2003) and HSDP data (Blichert-Toftet al., 2003) and assert that the Pb isotope ratiosproduced in Lyon are more than 3000 ppm offcompared to the claimed reproducibility of less thanabout 500 ppm (Stracke et al., 2003; Blichert-Toft etal., 2003; Albare`de et al., 2004). Baker et al’s. (2004)arguments are based on a comparison of Copenhagenand Lyon data for solution and rock standards and onebcommonQ sample. The Baker et al. (2004) compar-ison is, however, erroneous because,1. The only solution standard (98B) analyzed by boththe Copenhagen and Lyon groups shows anexcellent agreement (103, 109, 174 ppm on

Source enrichment processes responsible for isotopic anomalies in oceanic island basalts

A new method has been developed to separate the compositional variations in ocean island basalts into those that result from variations in source composition and from the melting process itself. The approach depends on correlations between isotope ratios, which can only come from source inhomogeneities, and elemental concentrations. Analysis of three data sets shows that the inhomogeneities beneath Theistareykir, in NE Iceland, Kilauea and Pitcairn can be produced by subduction of oceanic islands and volcanic ridges. The thicknesses of the lithosphere on which such islands were constructed and potential temperatures of the plumes that produced them can be estimated from the geochemical observations. Model ages are harder to determine, though simple assumptions give about 400 Ma for the Theistareykir source and 1.2 Ga for Kilauea. The model may also provide a physical explanation for the commonly used isotopic classification of ocean island basalts, with the isotopic composition changing from HIMU through EMII to EMI as the melt fraction increases. These results have been obtained from a small number of data sets obtained from ocean island basalts erupted in small areas during short time intervals. More such observations are needed to discover whether geochemical observations from other islands are consistent with the same model.

Sedimentology of cold-climate aeolian sandsheet deposits in the Askja region of northeast Iceland

Active aeolian sandsheets cover an area of 270 km 2 in the Askja region of NE Iceland. The sandsheet is divided into (i) a deflationary upwind part that is characterised by deflation lags, ventifacted boulder fields and restricted aeolian sand patches in the lee of obstacles, (ii) an accumulating central part that is characterised by continuous aeolian sand cover up to 10 m thick, where three scales of aeolian ripples are evident along with larger scale forms including longitudinal zibars, and (iii) a downwind part that is currently subject to aeolian bypass with only localised accumulation occurring in topographic hollows between basaltic lava fields. Trains of aeolian granule megaripples are a characteristic landform of the downwind part of the sandsheet. These sinuous-crested, oblique bedforms attain heights of 0.35 m and downwind spacings of 6 m. They are composed principally of low-density, pebble-grade pumice clasts and other reworked volcaniclastic detritus derived from the nearby Askja volcanic centre. The distribution of landforms across the sandsheet reflects a regional aeolian sediment budget that is controlled by dynamic interactions between glacial, ice-margin fluvial, aeolian and volcaniclastic processes under the influence of a distinctive climatic regime. Sediment supply at the upwind sandsheet margin, adjacent to the Dyngjujökull glacier, increases during summer periods when rates of glacial ablation are at a maximum. Longer term controls on sediment supply rate are determined by episodes of glacier surging and retreat, and by periodic eruptions of volcaniclastic debris from the Askja volcanic centre. The availability of sediment for aeolian transport is controlled by the level of the groundwater table, the depth and continuity of the permafrost layer in the proglacial area, the degree of development of armoured pebble lags and freezing of the sandsheet surface during winter months. The transport capacity of the wind varies on a seasonal basis with mean wind velocities and wind gustiness reaching a maximum during winter periods. Both winter and summer mean wind velocities have been steadily increasing since formal records began in 1969. The suite of aeolian landforms developed within the Askja region, as defined by sediment texture and composition, and bedform scale, morphology and spatial distribution, reflects a special set of controlling circumstances, which mean that the Askja Sandsheet deposits differ significantly from those described from elsewhere.

Melt mixing and crystallization under Theistareykir, northeast Iceland

Analysis of the compositions of crystals and melt inclusions from a suite of 40 gabbroic and wehrlitic nodules in a single eruptive body provides a record of concurrent mixing and crystallization of melts under NE Iceland. The crystals in the nodules have a similar range of compositions to those found as phenocrysts in the flow, and many of the nodules may have been generated by crystallization of a magma with a similar composition to that of the host flow. While plagioclase is only present in nodules where the average forsterite content of olivines is <Fo88.5, clinopyroxene is found in nodules with average olivine forsterite contents of up to Fo91.6. The order of crystallization ol → ol + cpx → ol + cpx + plg is consistent with generation of the nodules at pressures >0.8 GPa and is in agreement with estimates of crystallization pressures for the host basalt. The relationship between the compositional variability of melt inclusions and the forsterite content of the host olivine is revealed by REE analyses of over 120 melt inclusions. The degree of variability in REE concentrations and REE/Yb ratios decreases with falling forsterite content of the host olivine, as expected if melt mixing and fractional crystallization are operating together. The standard deviation of the REEs falls by a factor of ∼4 between Fo90 and Fo87. This change in olivine composition can be produced by crystallization of 20% which occurs on cooling of ∼50°C. The relative rates of mixing, cooling and crystallization may provide constraints upon the dynamics of magma bodies. The oxygen isotopic composition of olivines from the nodules and phenocrysts is highly variable (δ18O from 3.3–5.2 per mil) and shows little correlation with the forsterite content of the olivine. The full range of oxygen isotope variation is present in olivines with Fo89–90, and the low δ18O signal is associated with melts of high Mg# and La/Yb. Such geochemical relationships cannot be produced by assimilation of low Mg# crustal materials alone, and may reflect oxygen isotopic variation within the mantle source. The geochemistry of the melt inclusions and their host crystals can be accounted for by fractional melting of a mantle source with variable composition, followed by concurrent mixing and crystallization beneath the Moho.

Land cover characteristics in ne iceland with special reference to jökulhlaup geomorphology

ABSTRACT Subglacial eruptions in Vatnajbkull have accounted for several jökulhlaups (glacial outburst floods) in the Northern Volcanic Zone (NVZ). These events and aeolian processes have had a considerable impact on the landscape evolution of the area. Most of this area is occupied by barren land cover; the northern margin of the barren land cover is advancing northwards, burying vegetation under wind‐blown sediment. This paper presents a land‐cover classification based on a supervised Landsat TM image classification with pre‐processing and extensive field observations. Four land cover categories were identified: (a) lava cover (34.8%); (b) barren sediment cover (39.0%); (c) vegetation (25.1%); and (d) water and snow (1.1%). The mapping of sand transport routes demonstrates that a major aeolian sand transportation pathway is situated in the western part of the study area. The sedimentary formation elongated towards the northeast is evidence of active and continuous aeolian sand transportation towards the north. Interpretation of the satellite image suggests that four main areas are affected by jökulhlaups along the Jökulsáá Fjöllum: Ásbyrgi, Grímsstaðir, Herðubreið–Möðrudalur, and the Dyngjujökull sandur. In addition, jökulhlaup‐related sediment cover (8%) in the study area, together with erosional features, are evidence of a severe and extensive jökulhlaup‐induced process of land degradation.

Geochemical variability in a single flow from northern Iceland

Compositional variability is present for almost every element analyzed in 70 whole rock samples and 40 olivine‐hosted melt inclusions from the Borgarhraun lava flow in NE Iceland. The variation of incompatible element concentrations can be produced by incomplete mixing of fractional melts, while the compatible element variations in the whole rock samples can be explained by addition/removal of crystals found in the flow. The melt inclusion incompatible element compositions are more variable than the whole rock compositions, and the magnitude of variability of the whole rock samples can be matched if each hand specimen is made of magma that was formed by mixing of 20–30 batches of melt with the composition of the inclusions. Clinopyroxene barometry and the major element composition of the whole rock samples suggest that mixing took place at pressures of ∼0.9 GPa in sub‐Moho magma chambers. The spatial distribution of the concentrations of incompatible elements is not random in Borgarhraun and samples with a separation of <4 km have concentrations that are more similar than expected from a random distribution of geochemistry. This separation distance corresponds to a lava volume of 0.014–0.14 km3. This volume may be controlled by magma mixing and episodic eruption of melts from a sub‐Moho chamber. The concentration variations of compatible elements with position in the flow are little different to those expected for random variation. It is likely that the crystal removal and addition processes that control major element compositions generate variability on short length scales.

Geomorphology of the Ódáðahraun semi-desert, NE Iceland; a Landsat TM-based land cover mapping

Geomorphology of the Ódáðahraun semi-desert, NE Iceland; a Landsat TM-based land cover mapping

Farmers' perception of land-cover changes in NE Iceland

Land degradation is considered to be a major environmental problem by scientific and political institutions in Iceland. In past years, land degradation problems related to sheep and horse grazing have gradually come to the fore in public discussion of environmental questions. However, land degradation is often evaluated in different ways by different groups of people. This has led to a lack of consensus regarding rates and severity of land degradation. Therefore, it is important to study differences in environmental perception, and how these differences can be dealt with in the context of sustainable management. This paper aims to highlight farmers' perception of land degradation in Iceland, as perceptions of the environment determine the basis for human activities related to the land. A total of 100 farmers in NE Iceland were questioned and interviewed for their opinions on land degradation, its history, causes and severity. The results indicate that farmers consider land degradation to be a slow process, and not a catastrophic phenomenon. The understanding of the general ecological processes by the users affects their practices and concern for their environment. This feeds back on the land-use system and ultimately causes changes in land-cover. Therefore, if we are to increase our understanding of the processes that lead to land-cover changes we have to integrate the perceptions of land users and technical experts. In this way we may improve conservation practices and land-resource management policies. Copyright © 2000 John Wiley & Sons, Ltd.

Effective tension-shear relationships in extensional fissure swarms, axial rift zone of northeastern Iceland

The geometry of fracture systems in selected areas of the active Krafla fissure swarm, mid-Atlantic ridge, northeastern Iceland, is analysed. Based on geodetic analysis of the present-day topography at the top of Holocene basaltic lava flows which fill the axial rift zone, the deformation of this initially horizontal surface can be reconstructed. Extensional deformation is localised at all scales and block tilting, though present, remains minor. Using simple models of the surface expression of normal faults, the geometrical characteristics of the topographic features related to active deformation during tectonic-volcanic events are quantitatively analysed. At crustal depths of about 1 km, normal faults are present and have an average 70 ° dip. Comparison with the dip data of older normal faults observed in the uplifted and eroded shoulders of the rift zone, at palaeodepths of 1–2 km, indicates that this dip determination is valid. Comparisons between the local case study and structural analyses of active fissure swarms on a larger scale suggest that normal faulting plays a major role in the middle section of the thin, newly formed brittle crust of the rift zone. In the axial oceanic rift zone of NE Iceland, the extensional deformation in the upper crust is dominated by horizontal tension and shear of normal sense, their relative importance depending on depth. Absolute tension dominates in the uppermost several hundred metres of the crust, resulting in the development of fissure swarms. Effective tension plays an important role at a deeper level (2–5 km), because of the presence of magmatic fluid pressure from magma chambers which feed dyke injections. At crustal depths of about 1 km, normal shear prevails along fault planes which dip 60 °–75 °. This importance of normal shear at moderate depth, between upper and lower crustal levels where tension prevails, is pointed out. Within the extensional context of rifting, these variations of tectonic behaviour with depth are controlled by both the lithostatic pressure and the effective tension induced by the presence of magmatic fluid pressure.

The Krafla spreading segment, Iceland: 1. Three‐dimensional crustal structure and the spatial and temporal distribution of local earthquakes

The geothermal seismicity of the Krafla Volcanic System, NE Iceland, was monitored for 3 months in 1985 using a dense, local seismometer network. The seismicity was continuous, and the spatial and temporal distributions were roughly known prior to monitoring. The instruments could thereby be deployed in a well‐positioned array. A total of 489 locatable events were recorded within the network, and 1771 arrival times were inverted to calculate the three‐dimensional P‐wave velocity structure and hypocentral locations. Low‐velocity volumes were imaged beneath the Krafla and Námafjall geothermal areas and indicate zones of hydrothermal alteration. High‐velocity bodies beneath the Krafla caldera rim are interpreted as gabbroic intrusions. Using a three‐dimensional velocity structure instead of a refraction‐based one‐dimensional model to locate the hypocenters significantly improves their location quality and illustrates the shortcomings of using refraction‐based models to locate earthquakes in local, anomalous areas. Seismic activity was concentrated within the Krafla and Námafjall geothermal areas and in a narrow zone where dike injections had occurred 8 and 5 years earlier. The activity occurred in the depth range 0–3 km. The seismic rate for the whole area was one magnitude 3.2 event per year and the b value was 0.77±0.10. Most of the seismicity appears to result from geothermal processes in the manner proposed for other Icelandic areas (Foulger and Long, 1984; Foulger, 1988b). Seismicity directly beneath the Bjarnarflag well field within the Námafjall area is probably induced partly by geothermal exploitation. Considerable seismicity also occurred immediately below Leirhnjúkur, a site of intense geothermal activity in the center of the Krafla caldera that overlies a roof pendant in the magma chamber below. This volume of high seismicity is probably highly fractured and may provide a conduit for magma escaping from the magma chamber.

The Krafla spreading segment, Iceland: 2. The accretionary stress cycle and nonshear earthquake focal mechanisms

The continuous geothermal seismicity of the Krafla volcanic system, NE Iceland, was monitored in 1985, immediately following a major spreading episode (Foulger et al., 1989; Arnott, 1990; Arnott and Foulger, this issue). Focal mechanisms of 153 well‐located, shallow earthquakes were determined using P wave polarity data. Errors resulting from poorly modeled crustal structure were reduced by calculating hypocentral locations using a three‐dimensional crustal model derived from a simultaneous inversion and by deriving takeoff angles and azimuths using three‐dimensional ray tracing. The final data are thus of exceptionally high quality. Most of the events display radiation patterns consistent with a double‐couple source model. The source orientations exhibit no systematic pattern beneath the Bjarnarflag well field, a little coherence in the Krafla caldera, and a general tendency for the greatest principal stress to be oriented normal to the rift zone in a zone of recent dike injection. Normal, thrust, and strike‐slip shear events are mixed together, and five events have radiation patterns inconsistent with a double‐couple source model and are attributed to nonshear source processes. Variable non‐double‐couple events were observed, with implosive and explosive/volume‐conserving mechanisms and no clear systematic pattern of type or orientation for the system as a whole, though patterns were discernable in some subsets of events. The nonshear suite, like the shear suite, was therefore in general heterogeneous. The results indicate that a systematic deviatoric stress field was absent in the Krafla spreading segment in 1985, in contrast with the Reykjanes segment in 1977 and the Hengill segment in 1981 (Klein et al., 1977; Foulger, 1988b). The observations are consistent with a model where a strong, systematic deviatoric stress field characterizes the accretionary plate boundary during the interrifting and prerifting phases of the rifting cycle. This systematic stress is partially or wholly released during episodic rifting and spreading episodes, such as that experienced by the Krafla system 1975–1984, and is absent in the immediate postrifting phase of the rifting cycle. Nonshear geothermal earthquakes have been reported from all of the three Icelandic spreading segments studied in detail to date.

Plate dynamics near divergent boundaries: Geophysical implications of postrifting crustal deformation in NE Iceland

The bulk of a tectonic plate is thought to move continously at a rate consistent with the geologic average. On the other hand, movements are highly episodic at plate boundaries. We study the plate dynamics that relate to these two different modes by modelling the displacements observed using the global positioning system in Northeast Iceland 1987–1990. These observations were made about 10 years after an episodic divergent movement between the North American and Eurasian plates 1975–1981. The horizontal displacement field fits well a two‐dimensional model of postrupture stress relaxation assuming a thin elastic layer overlying a layer of Newtonian viscosity. This analysis indicates values of about 10 m2/s for the stress diffusivity and 0.3–2 × 1018 Pa s for the Newtonian viscosity of the lower layer. However, no significant correlation exists between the observed and modeled vertical displacements probably because of the relative inaccuracy of the vertical component observations. Assuming that contemporary plate motion is the sum of many displacements that have diffused from boundaries where episodic displacements occur periodically, we simulate the spatial transition from episodic to continuous plate movements. The plate “boundary zone” where movements are episodic or quasi‐episodic is of the order of 100 km wide, depending on the stress diffusivity (which may be fairly uniform throughout the world) and the frequency of episodic movements.

Evidence for subsidence in the 1989 Arctic winter stratosphere from airborne infrared composition measurements

Simultaneous measurements of the stratospheric burdens of CO2, HCN, N2O, CH4, OCS, CF2Cl2, CFCl3, CHF2Cl and HF were made by the Jet Propulsion Laboratory MkIV interferometer on board the NASA DC‐8 aircraft during January and early February 1989 as part of the Airborne Arctic Stratosphere Experiment. Data were acquired on 11 flights at altitudes of up to 12 km over a geographic region covering the NE Atlantic Ocean, Iceland and Greenland. The results obtained show large variations in the burdens of these tracers due to the effects of transport. The tropospheric source gas burdens were reduced inside the polar vortex, suggesting that the air had subsided with respect to the surrounding mid‐latitude air. Increased HF burdens inside the vortex support this interpretation. The results obtained from the different tracers are highly consistent with each other and indicate that in the 15‐ to 20‐km altitude range inside the vortex, surfaces of constant volume mixing ratio were located some 5–6 km lower in absolute altitude than outside the vortex. The results also indicate that the magnitude of this subsidence increases with altitude. These conclusions are consistent with other measurements.

Composition measurements of the 1989 Arctic winter stratosphere by airborne infrared solar absorption spectroscopy

Simultaneous measurements of the stratospheric burdens of H2O, HDO, OCS, CO2, O3, N2O, CO, CH4, CF2Cl2, CFCl3, CHF2Cl, C2H6, HCN, NO, NO2, HNO3, ClNO3, HOCl, HCl, and HF were made by the Jet Propulsion Laboratory MkIV interferometer on board the NASA DC‐8 aircraft during January and early February 1989 as part of the Airborne Arctic Stratosphere Experiment (AASE). Data were acquired on 11 flights at altitudes of up to 12 km over a geographic region covering the NE Atlantic Ocean, Iceland, and Greenland. Analyses of the chemically active gases reveal highly perturbed conditions within the vortex. The ClNO3 abundance was chemically enhanced near the edge of the vortex but was then depleted inside. HCl was chemically depleted near the vortex edge and became even more depleted inside. In fact, by late January deep inside the vortex, HCl was either completely removed up to 27‐km altitude, or partially removed to an even greater altitude. NO2 was also severely depleted inside the vortex. In contrast to Antarctica, H2O and HNO3 were both more abundant inside the vortex than outside. While for H2O this is solely a consequence of descent (without accompanying dehydration), HNO3 additionally shows evidence for chemical enhancement inside the vortex. One exception to the high HNO3 abundances inside the vortex occurred on January 31 when stratospheric temperatures above the aircraft fell below 190 K. However, following this event, HNO3 burdens fully recovered, suggesting that if the loss on January 31 was due to temporary freeze‐out of HNO3, the resulting particles reevaporated above 12 km. Taken together, these results suggest that although the Arctic vortex did not get cold enough to produce any dehydration, nor as vertically extensive denitrification as occurred in Antarctica, nevertheless, enough heterogeneous chemistry still occurred to convert over 90% of the inorganic chlorine to active forms in the 14‐ to 27‐km altitude range by early February 1989.

Elastic deformation models of Krafla Volcano, Iceland, for the decade 1975 through 1985

The dynamics of the shallow magma reservoir at Krafla Volcano in NE Iceland have been analyzed with three types of elastic models based on over 70 surveys of tilt and displacement made from 1975 to 1985, a period of continuous volcano-tectonic activity. Modeling results are integrated with geophysical and geological information to estimate the position, geometry, and volume change of magma reservoir domains subjected to periodic inflation and deflation. Dominating influences on magma reservoir dynamics are examined in the context of activity in Krafla's associated rift system and the deformation of its caldera from 1975–1985. Rather than the fluid-filled cavity concept of Mogi (1958) and most recent workers, our models are idealized as strained regions with spherical, double-spherical, or general ellipsoidal symmetry. The models are mathematically generalized from that of Mogi (1958) and are derived by inversion of displacements. Model results from different displacement components are remarkably consistent, although models based on vertical displacements typically have errors-of-fit much closer to expected measurement errors than those based on tilt or horizontal displacements. About one half of the double-sphere and ellipsoid models have significantly better fits than single-sphere models. Double-sphere model results are consistent with a 2 to 3 km center-to-center separation of magma storage zones from at least 2 km to 4 km depth by portions of the fissure system, as implied by S-wave attenuation patterns for 1976–1977. However, all models suggest pressurization zones of more limited extent than possible domains of storage inferred from S-wave attenuation. Ellipsoid models typically implied unrealistically shallow depths of magma storage. Caldera inflation rates decreased after January 1978 when the caldera periodically reinflated to its level prior to the initial December 1975 deflation event. From 1975–1980 the volume and duration of inflation between deflation events was strongly correlated with the volume of the previous deflation. After 1980 there was a significant increase in the duration of inflation periods and a decrease in rates of caldera inflation and fissure system widening. Consistent with these results, the magma reservoir is conceptualized as a hot rock mass containing numerous magma chambers and pressure-sensitive conduits connecting the chambers and deep magma sources. Magma is injected into the fissure system at critical pressures determined by the confining stress and rock mass strength. The duration and volume of inflation required to reach a critical pressure threshold is largely dependent on the volume of magma released in the previous deflation. Reduction of the extension rate across the Krafla fissure system after 1980 suggests that extensional forces were also reduced. A resultant rise of confining pressure on the magma reservoir and a reduced capacity of the fissure system to accommodate dike injection in combination would have increased critical stress levels for reservoir deflation and reduced the pressure gradient driving magma supply from deep sources.

The Fjallgardar volcanic ridge in NE Iceland: an aborted early stage plate boundary or a volcanically dormant zone?

Summary The extinct Fjallgardar volcanic ridge (FVR) in NE Iceland is approximately 15–20 km wide and extends for about 190 km from the Vatnajökull glacier toward the Slétta peninsula in the far N. The FVR was mainly formed subglacially and strikingly parallels the curvature of the present median line (Askja to Theistareykir centres) of the neovolcanic zone (NVZ) which lies about 40 km to the W. The basalts from the Fjallgardar and Slétta regions are dominantly evolved tholeiites as shown by major- and trace-element compositions of 42 samples. The average magnesium number is 47 (range 37–58). 90 per cent of the samples can be grouped into either ‘low-’ potassium (0.1%–0.4% K 2 O) or ‘high-’ potassium (0.5%–0.7% K 2 O) units. The high-potassium lavas in the FVR are chemically similar to Tertiary tholeiites of eastern Iceland. A test for temporal geochemical variations for three stratigraphic formations was carried out for the Mödrudalur area of the FVR. The oldest of the three formations (M3) consists of low-potassium basalts only, whereas the two younger formations (M4 and M5) consist predominantly of more evolved high-potassium basalts. Formation M5 is most probably fed into Mödrudalur from the Kverkfjöll volcanic system some 75 km further S. Volcanism in the FVR S of Slétta was short-lived and lasted for perhaps 200 000–300 000 years. Subsurface lateral flow of magma from the NVZ is probably responsible for the Recent subaerial volcanism in Slétta. In the Slétta area the older subglacially formed ridges have only a low-potassium basalt composition. The FVR may be viewed as ‘abortive’ and as the precursor to the NVZ in northeastern Iceland. During the Brunhes magnetic polarity chron (less than 0.7 Ma) a rift jump is postulated to have occurred from the FVR to the present NVZ, approximately 40 km to the W. E.-W lateral rift jumps (20–40 km) may have been common in the eastern half of Iceland in the vicinity of the Iceland hot-spot located 150–200 km E of the Reykjanes and Kolbeinsey mid-ocean ridges.

Geodetic measurements of surface deformations during the present rifting episode in NE Iceland

Iceland is part of the accreting plate boundary between the American and European plates. Since 1965 a number of repeated triangulation‐trilateration surveys spanning the neovolcanic zone in NE Iceland were made in order to detect crustal movements. All the measurements for a single observation epoch are separately adjusted by the method of least squares. Statistical tests are carried out to check the observations and adjustment models for outliers and model errors. To detect significant movements in repeatedly measured geodetic networks with multivariate design, a testing rule is described. After unexpected horizontal contractions of the surface of approximately 0.5 m from 1965 to 1971, enormous deformations have been found since the beginning of a major rifting episode in December 1975. During the period from 1971 to 1980 a horizontal expansion of about 7.5 m and vertical displacements of 3 m were observed across the approximately 3‐km‐wide active rifting zone of the Krafla fissure swarm. In adjacent zones, compressional strains of up to 2×10−4 occurred, rapidly decreasing to a value of 1×10−5 at a distance of 40 km from the rifting axis. After 1980 the pattern of ground deformation has altered. The east‐west shifts have decreased and are superposed by a radial component of motion around the Krafla caldera. West of the Krafla fissure swarm a shear zone is found.