The AD 1362 Öræfajökull eruption, S.E. Iceland: Physical volcanology andvolatile release
Kirti Sharma a,⁎, Stephen Self a, Stephen Blake a, Thorvaldur Thordarson b, Gudrun Larsen c
a Department of Earth Sciences, The Open University, Milton Keynes, MK7 6AA, UKb School of Geosciences, University of Edinburgh, Edinburgh, EH9 3JW, UKc Science Institute, University of Iceland, Dunhagi 3, IS-107 Reykjavik, Iceland
ption of Öræfajökull volcano, Iceland, in AD 1362 is described and interpretedbased on the sequence of pyroclastic fall and flow deposits at 10 proximal locations around the south side ofthe volcano. Öræfajökull is an ice-clad stratovolcano in south central Iceland which has an ice-filled caldera(4–5 km diameter) of uncertain origin. The main phase of the eruption took place over a few days in June andproceeded in three main phases that produced widely dispersed fallout deposits and a pyroclastic flowdeposit. An initial phase of phreatomagmatic eruptive activity produced a volumetrically minor, coarse ashfall deposit (unit A) with a bi-lobate dispersal. This was followed by a second phreatomagmatic, possiblyphreatoplinian, phase that deposited more fine ash beds (unit B), dispersed to the SSE. Phases A and B werefollowed by an intense, climactic Plinian phase that lasted ∼8–12 h and produced unit C, a coarse-lapilli,pumice-clast-dominated fall deposit in the proximal region. At the end of Plinian activity, pyroclastic flowsformed a poorly-sorted deposit, unit D, presently of very limited thickness and exposed distribution. Much ofEastern Iceland is covered with a very fine distal ash layer, dispersed to the NE. This was probably depositedfrom an umbrella cloud and is the distal representation of the Plinian fallout. A total bulk fall deposit volumeof ∼2.3 km3 is calculated (∼1.2 km3 DRE). Pyroclastic flow deposit volumes have been crudely estimated tobe b0.1 km3. Maximum clast size data interpreted by 1-D models suggests an eruption column ∼30 km highand mass discharge rates of ∼108 kg s−1. Ash fall may have taken place from heights around 15 km, above thelocal tropopause (∼10 km), with coarser clasts dispersed below that under a different wind regime. Analysesof glass inclusions and matrix glasses suggest that the syn-eruptive SO2 release was only ∼1 Mt. This result issupported by published Greenland ice-core acidity peak data that also suggest very minor sulphatedeposition and thus SO2 release. The small sulphur release reflects the low sulphur solubility in the 1362rhyolitic melt. The low tropopause over Iceland and the 30-km-high eruption column certainly led tostratospheric injection of gas and ash but little sulphate aerosol was generated. Moreover, pre-eruptive anddegassed halogen concentrations (Cl, F) indicate that these volatiles were not efficiently released during theeruption. Besides the local pyroclastic flow (and related lahar) hazard, the impact of the Öræfajökull 1362eruption was perhaps restricted to widespread ash fall across Eastern Iceland and parts of northern Europe.
Öræfajökull is an ice-capped stratovolcano located in south-eastIceland (64.00° N, 16.65° W), on the southern margin of theVatnajökull ice cap. The volcano rises ∼2000 m above sea level, thesummit being ∼10 km NW of the Atlantic Ocean shoreline. It isprimarily composed of sub-glacial pillow lavas, tuffs and breccias,basaltic and andesitic lava flows and rhyolite intrusives and extrusives,intercalated with interglacial sediments (Thorarinsson, 1958; Steven-son et al., 2006). The summit is truncated by a 4 to 5 km diameter ice-filled caldera of unknown origin and age. Several peaks (nunataks),
l rights reserved.
most of them remnants of old lava domes, rise above the ice and markthe margins of the caldera wall. The highest nunatak (and the highestpoint in Iceland), Hvannadalshnjúkur at 2110 m, is a pristine rhyolitelava dome that possibly occupies the 1362 eruptive vent site. Therehave been two explosive eruptions from Öræfajökull in historicaltimes, a small basaltic-andesite (benmoreitic) eruption in AD 1727–1728 and the larger rhyolitic eruption of AD 1362 (Thorarinsson, 1958;Prestvik, 1979, 1982).
Öræfajökull erupted violently in 1362, with historical evidencesuggesting that activity started around June 15 (Thorarinsson, 1958).By Thorarinsson's account this was a large explosive eruption thatproduced much ash and devastated the local area. Explosive activitywas accompanied and followed bymajor jökulhlaups (glacier outburstfloods), carrying pumice and gravel that demolished the remaining
Fig. 1.Map indicating locations where ash fall from AD 1362 event has been found. Volcanic glass shards correlated to the 1362 eruption have been discovered in south Ireland (Halland Pilcher 2002), Norway, and Scandinavia (J Pilcher, pers. comm. 2004), and Greenland (Palais et al., 1991). Inset map shows location of Öræfajökull in Iceland. Triangles mark activesilicic volcanoes of Eastern Volcanic Zone (EVZ): Ö = Öræfajökull, K = Katla, T = Torfajökull, H = Hekla, E = Eyjafjallajökull, A = Askja. Light-grey shaded areas denote main ice caps.
720 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
settlements on the lowlands and caused further fatalities. Contem-porary reports, summarized and translated in the classic work byThorarinsson (1958), describe “volcanic eruption kept burning withsuch monstrous fury as to lay waste to the whole of Litlaherad1,causing desolation for some 100 miles. On even ground one sank inthe sand up to the middle of the leg, and winds swept ash into suchdrifts that buildings were almost obliterated. Volcanic eruption withdarkness so intense that roads could not be seen at midday”. Wherethere had previously been an expanse of water on the south side of thevolcano, glacier bursts triggered by the activity carried large quantitiesof gravel, and pyroclastic material, forming an unstable braid–plain(sandur).
The 1362 eruption is thought to be the largest rhyolite eruption tohave occurred in Iceland during historic times (Thorarinsson, 1958).Fine ash was dispersed across the Atlantic Ocean at least as far afieldas Greenland, ∼1300 km from source, and Scandinavia, ∼1220 kmfrom source, and parts of the UK, as shown in Fig. 1 (Persson, 1971;Palais et al., 1991; Hall and Pilcher, 2002; Pilcher et al., 2005).Thorarinsson (1958) presented an isopach map of the widespreadash fall deposit across Eastern Iceland, and Larsen et al. (1999)revised the dispersal axis of the deposit slightly (their Fig. 3).Although the main explosive phases apparently lasted for only a fewdays, historical accounts indicate that eruptive activity probably
1 The coastal area around Öræfajökull.
continued until the autumn of 1362 (Thorarinsson, 1958). This latteractivity possibly involved the growth of the rhyolite dome Hvanna-dalshnjúkur at the 1362 vent site.
In this paper we report observations on the pyroclastic deposits inthe medial to proximal area around the volcano and use them tointerpret the eruption sequence and physical parameters of this majorIcelandic eruption. Whole-rock, mineral and glass inclusion analysesare also reported and used to infer themagma properties and estimatemagma volatile release. This paper adds to the recent contribution ofSelbekk and Trønnes (2006) about the 1362 magma body. We con-centrate on the type and dispersal of the deposits, calculate a newvolume for the ejecta, and quantify the release of volatiles during thiseruption.
2. Deposit stratigraphy and characteristics
The 1362 deposits are only patchily preserved in the proximalregion, as most of them have been stripped away from the steepglacial valleys and ridges on the flanks of the volcano. The volcanosummit, including the vent area, is ice-capped, precluding thepreservation of any very proximal sections. The northern to easternproximal–medial fallout dispersal was also on ice and the depositshave presumably been removed or modified by runoff and glaciermovements. However, tephrochronological studies have found the1362 layer (commonly used as a marker horizon) in outlet glaciers onnorth-western Vatnajökull (Larsen et al., 1998) and widely across
721K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
Eastern Iceland (Thorarinsson, 1958). Nevertheless, there are severalmoderately well-preserved, accessible exposures of 1362 depositslocated on the lower, ice-free southern flanks of the volcano. Tensections that provide control on the proximal deposit stratigraphywere studied and logged in detail (Table 1, locations also marked inFig. 2 inset map); extra details are provided by occasional incompleteexposures (Sharma, 2006). We have not attempted to re-map thedistal deposit across Eastern Iceland.
We identify threemain fall units (A, B and C; see Figs. 2 and 3a) anda pyroclastic flow unit (D). Each unit contains one or more sub-units,but it must be noted that the sub-unit divisions, in A and B especially,represent a simplification of a more complex set of deposits. Due tothe patchy preservation of the fall units, a more detailed subdivisionwas not made. Stratigraphic sections in Fig. 2 show that the fall unitsdecrease in thickness at various rates with distance from source,reflecting varying eruption intensities. All major fall units thin morerapidly to the west than towards easterly directions. The mappeddispersal of the individual fallout units A–C are shown in Fig. 4a, andwill be discussed in a later section. In this work, grain size termsused (e.g. lapilli, ash) follow the definitions in McPhie et al. (1993),see Fig. 2.
2.1. Unit A (sub-units A1, A2)
The lowermost depositional unit of the 1362 eruption is composedof a fine- to coarse-grained ash bed, up to 14 cm thick in total (Fig. 4a).Two sub-units can be identified, A1 and A2. Sub-unit A1 is a locallydistributed (found at two localities only), thin (∼1 cm), very finewhite ash layer. Where present it forms the basal layer of the 1362sequence. A2 is dominantly a medium- to coarse-grained ash andlapilli bed, mainly brown to grey in colour, but often displaying ayellowish tinge, containing small (1 to 10 mm) intercalated pumiceand lithic lapilli; unit A2 is also often stratified into thinner layers(Fig.3b). Pumice clasts in unit A are small, elongate and have tubularvesicles. Lithic clasts are mostly chips of fragmented, altered obsidian,which are assumed to be non-juvenile. The top of unit A is often
Ö-1 Old ruined farm next toHighway 1; 2–3 steep patchesof deposit on hillside.
9.5 231
Ö-2 Bank on north side of Hvalvörðugil(also Öldur), at stream ∼0.7 kmnorth of Highway 1 (NW of Hof).
10.9 220
Ö-3 Small section south of Hof–reworkeddeposit.
14.9 187
Ö-4 Gljúfurá (small stream) northof Highway 1 near Hofnes farm.
15.2 179
Ö-5 Mulagljúfur Valley north-west ofHighway 1.
11.3 090
Ö-6 Partial section in peat bog nextto Highway 1.
14.1 120
Ö-7 On path to Svartifoss waterfallabove Skeidara.
14.5 283
Ö-8 Patch of deposit on hilltopabove Svartifoss.
16.2 282
Ö-9 Bölti, above Svínafellsjökull. 8.9 259Ö-10 Hnappavellir, small pit next to
Highway 1.13.8 159
Ö-11 Ingólfshöfði. 24.2 176Ö-12 Small section at start of Laki
road, near Hunkubakkar.43 251
Ö-13 Near Stigarjökull, north of location 10. 8.1 156
eroded and missing in proximal sections. However, at localities wherethe unit top is preserved, the upper 1–2 cm is often a distinct “millet-seed” layer of extremely well-sorted fine-grained, rounded, pumice(0.5 to 2.5 mm).
2.2. Unit B (sub-units B1, B2)
Unit B is distinguished from unit A by its differing grain size andcolour; two sub-units are recognised — B1 and B2. B1 is a fine-grained, grey ash bed, up to 8 cm thick. Lenses containing lapilli-sizeobsidian chips and small crystals can be found distributed irregularlywithin the bed. Overlying B1 in most proximal sections is sub-unitB2, a medium-grained white to grey ash bed, up to 30 cm thick,containing scattered small (5–10 mm) pumice and lithic lapilli. Thebase of B2 contains extremely fine-grained, white or grey, thin (0.1 to0.5-cm thick) ash horizons interbedded within the coarser ash beds(Fig. 3b).
In the proximal region we also identify additional layers that arepresent at very few localities and cannot be correlated with other sub-units. These layers (B3 and B4) are grouped as part of the unit Bsequence and are most likely the result of minor pyroclastic flowactivity occurring prior to, or perhaps during, the beginning of activitythat produced unit C.
2.3. Layer B3
Layer B3 consists of accretionary-lapilli-bearing fine- to medium-grained ash, and is found only at locations 10,11 and 13. At location 13,the most proximal exposure studied (∼8 km from the vent), B3 is avery fine-grained dark grey ash, ∼30-cm thick, containing accre-tionary lapilli “ghosts” and ash aggregates. At location 10, ∼15 kmaway from the vent, this ash layer is thinner (∼15 cm) and containsabundant accretionary lapilli with amaximumdiameter of 10–15mm;they are extremely fragile and fall apart when touched (Fig. 3c). At thebase of this layer, a sandy, crystal-rich, 2 to 3-cm-thick sub-layer isoften found.
2.4. Layer B4
Layer B4 is preserved only at the most proximal location (13). It isa 30-cm-thick, medium- to-coarse-grained ash deposit with somesmall (∼1 to 5 mm) pumice lapilli. The poorly-sorted nature of thisbed and the ash-dominated matrix suggest that it is a proximallocal pyroclastic flow deposit, possibly derived from small columncollapses before the climactic stage of the eruption (cf. June 13–14pre-climactic events during the 1991 Pinatubo eruption, Hoblittet al., 1996).
2.5. Unit C
Unit C consists of a massive to crudely layered unit that forms athinning sheet, extending ∼20–25 km out to the south coast. At themost distal exposure found (location 11, ∼24 km from the vent), thetrue thickness of the deposit is truncated by erosion. Based on itsfines-poor, relatively well-sorted nature, this unit is interpreted tobe a medium- to coarse-grained, clast-supported, fall deposit dom-inated by coarse pumice lapilli and bombs, with minor lithic clasts.Where fully preserved, the unit varies in thickness from 60 to180 cm in the sections examined and shows both normal and re-verse grading. In most sections, the middle of the deposit is verycoarse-grained (Fig. 3d); above this, however, coarse pumice lapillilayers alternate with beds of finer pumice lapilli. The pumice clastsare rounded to sub-angular (Fig. 3e), although in some sectionsclasts are more angular and elongate, parallel to the long-axis ofstretched vesicles. Large pumice clasts are extremely fragile andhave an exceptionally low density, 280 (±10) kg m−3, calculated by
722 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
measuring the volume and weight of cut cubes of pumice. In-dividual clasts commonly range in diameter from 10 to 300 mm,however, occasional clasts up to 600 mm in diameter are also found.Breakage upon impact must partly control the observed clast size.The pumice is light grey to white in colour with a platy texture,and is extremely vesicular and crystal-poor in hand specimen. Atlocation 10, larger pumice clasts are sometimes coated with fineash, suggesting that they fell through a wet, fines-rich eruptioncloud.
At locations 2 and 10, pumice clasts containing dark grey bands, upto 2–3 cmwide, occur (Fig. 3f), and at location 4, small scattered darkgrey pumice clasts are found within the deposit together with thewhite clasts. The dark bands and grey pumice clasts are both crystal-poor (in hand specimen). The grey pumices are small in maximumdiameter (up to 60 mm) and denser than the accompanying whitepumice clasts.
There are very few juvenile lithic clasts in unit C (b3 wt.%). Wherepresent they are typically small (maximum diameter ∼10–30 mm)weathered fragments of basalt and rhyolite, most likely derived fromolder sub-glacially eruptedmaterial that forms the volcano core. Lithicclasts often display a red–brown staining, indicative of hydrothermalalteration. Occasionally, in the upper parts of unit C, 1-cm thick lithic-rich horizons are found consisting of 10–20-mm diameter obsidianclasts (e.g., locations 10 and 13, Fig. 2). In some proximal locations(e.g., 2, 4, and 10) a thin discontinuousmedium-grained grey ash layer,with scarce small pumice fragments is found on top of the mainpumice fall unit and below unit D.
2.6. Unit D (sub-units D1, D2)
Overlying unit C at most locations is a poorly preserved unit,∼70–110 cm thick in total, which probably has a pyroclastic floworigin. We give primary deposit thickness values (Fig. 4d), and do notinclude reworked material (which is present at locations 5, 7 and 9;Fig. 2). In the proximal sections this deposit can be divided into twodistinct sub-units (D1 and D2). D1 is a poorly-sorted, sometimesstratified, matrix-rich, thin bed, with medium- to coarse-sizedpumice and lithic lapilli, supported in a vesicular ash matrix. AboveD1, unit D2 is a very poorly-sorted bed, consisting of a fine, vesicularash matrix and small (20–30 mm diameter) dense pumice clasts(Fig. 3.3a). Occasionally accretionary lapilli “ghosts” and fragile ashaggregates are present towards the top of the deposit. Crude normalgrading can sometimes be recognised, along with a fine ashy top.In sections where the pyroclastic flow deposit is better-preserved(e.g. at location 5), a fine-grained ash fall layer sits on top of unitD2; it is usually mixed into the soil above. Thin horizons of fine-grained dark ash originating from later explosive activity in theEastern Volcanic Zone (EVZ), including Öræfajökull 1727 and Laki1783 are also found intercalated within the soil.
3. Deposit dispersal and clast sizes
3.1. Dispersal
Owing to the different dispersal and thinning characteristics ofeach fall unit in the 1362 deposit sequence, separate isopach mapswere constructed to assess the dispersal characteristics of each unit inproximal–medial areas around the volcano (Fig. 4). Isopach maps arepresented on the basis of control provided by the 10 locations loggedand occasional other observations.
Deposits of the opening phase of the eruption have a complex in-ternal stratigraphy and different-appearing units occur at the differentlocations logged. These are grouped as unit A. Sub-unit A1 can be seenonly in two locations, is small, and distributed to the southwest. Theisopach map for fall unit A2 (Fig. 4a) suggests a bi-lobate dispersal,with one lobe extending directly to the west, and a second lobe dis-
persed SSE of the vent. These two separate lobes may reflect changingwind conditions, but the exact relationship between the two lobes isdifficult to define based on the available exposures. Thus, for thewhole of unit A, multiple lobes suggest at least three explosive events,each with a different dispersal axis (and, thus wind field). There mayhave been more sub-units dispersed in northerly directions, onto theice cap.
Isopachs for the total fall unit B define a narrow south-easterlydispersal (Fig. 4b), similar to the distribution of the main fall unit (C);(Fig. 4c). The 1 cm isopach is the smallest mapped isopach, and isestimated (based on contouring of the isopachs and isopach extra-polation) to extend ∼15–20 km off the south coast. Insufficient dataexist to produce separate maps for sub-units B1–B3, but layer B3 isthought to have a very narrow distribution, extending directly southfrom the vent for ∼20 km (Fig. 4b).
The main fall deposit (C) is more widely dispersed with an axis tothe SSE of the vent area (Fig. 4c). Isopachs are regularly spaced andindicate thinning occurs at an exponential rate. The 5-cm isopach isthe smallest mapped and is estimated to extend to ∼25 km offshore tothe south but does not extend as far to the NW (i.e. the deposit thinsupwind to the NW). Larsen et al. (1999) also produced a distributionmap for the 1362 fall deposit on which the dispersal axis extends tothe south-east, consistent with the dispersal direction of the proximalunits determined in this study.
Owing to the poor preservation of the pyroclastic flow deposits(unit D) it has not been possible to accurately constrain the originalextent and distribution, and calculate a volume for them. Fig. 4d showsthe proximal locationswhere primary flow sequences weremeasured.A contemporary description of offshore pumice rafts from Öræfajökull(given by Thorarinsson, 1958) suggests that pyroclastic flows mayhave travelled at least 25–30 km from the vent and into the ocean.(Note that the location and configuration of the coastline in 1362 mayhave been different from that of today). Units C and D were producedby the climactic phase of the 1362 eruption. Unit C, where mapped,(Fig. 4c) is considered to be part of, and formed simultaneously with,the widespread pyroclastic fall unit across eastern Iceland (see laterdiscussion).
3.2. Clast sizes
Maximum pumice and lithic sizes in unit Cwere determined in thefield by measuring and averaging the long axes of the five largestclasts found in the deposit at each location. Care was taken to selectclasts that were in situ and, in the case of the maximum pumicemeasurements, had not undergone any breakage. The dense nature ofjuvenile lithic clasts means that they are unlikely to break duringtransport, therefore their distribution is often a better indicator oferuption column dynamics. However, in the case of Öræfajökull,maximum lithic data were difficult to obtain as lithic clasts are smalland few in number even in the proximal sections. Althoughthe pumice clasts are more fragile in nature, measurements in situ atleast present a complete size range in the proximal area. Pumice clasts25–30 cm in length can be found ∼10 km from the vent, along thedispersal axis; clasts up to 3 cm long can be found 15–20 km awayfrom the vent area, indicating energetic eruption conditions. This isconfirmed by the maximum lithic data; lithic clasts up to 1 cm inlength can be found 20–25 km from source. Isopleth maps for averagemaximum pumice and maximum lithic data are plotted in Fig. 5. Bothmaximum lithic and pumice isopleths are ellipsoidal in shape andfollow the same distribution as the isopachs for unit C.
4. Deposit interpretation
Based on our descriptions of the proximal 1362 deposits, and thedistribution of the eruptive units, we conclude that three main unitswere produced as a result of changing styles of explosive activity. The
Fig. 2. Stratigraphic sections through Öræfajökull 1362 (Ö1362) deposits (soil below not shown). Sections are all from proximal locations on south flank of the volcano. Index map shows approximate location of each section (black circles with location number); red triangle (labelled Ö) marks volcano summit, grey shaded area marks Vatnajökull ice cap.Thickness of stratigraphic sections given in cm; grain size variations (horizontal axis) measured using grain size scale of McPhie et al. (1993). McPhie et al. (1993) definitions for ash are: fine-grained ash (fg)=b0.1 mm, medium-grained ash (mg)=0.1–0.5 mm, coarse-grained ash (cg)=0.5–2mm; for lapilli (pumice and lithic clasts): fine grained (fg)=2–4mm,medium grained (mg)=4–64mm, coarse grained (cg)=64–256mm, very coarse grained (vcg)=N256mm. Ash/lapilli grain size boundary is marked on sections 5 and 8 for reference. Dashed tie lines mark the top of fall unit A, and the base and top of Plinian fall deposit unit C. Unit designations are listed alongside sections 13, 10, and 5. Sections from locations3 and 6 are not shown, as 1362 units are largely reworked at these localities.
pp. 723–726K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
Fig. 3. Photographs showing Öræfajökull 1362 deposits and detailed clast/unit features. Scale on photographs is 10 cm with 1 cm bars. (a) Proximal 1362 stratigraphic sequence;∼55 cm thick, shows different deposits of main eruption phases. Sequence is from location 9, 9 km to NW and upwind from vent (see also log of deposit sequence at location 9,Fig. 3.2). Soil above section contains thin horizons of fine-grained basaltic ash. (b) Detail of base at location 9; sub-unit A2 forms base of sequence; A1 missing at this location; B3 alsomissing. Stratification, resulting from changes in grain size in unit A can be seen in the base of this section. Above A, unit B forms thin bands of fine ash fallout. (c) Sample of delicateaccretionary-lapilli-bearing ash sub-unit B3 from location 10. (d) Unit C at location 4, totalling 70 cm thick; note horizon enriched in coarse pumice and lithic clasts towards middle.(e) Typical large, platy, fractured pumice clast N20 cm long from middle of unit C at location 4. (f) Cut section through pumice clast with a dark band from unit C, collected frommiddle of deposit at location 10.
727K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
start of the eruption was characterised by phreatomagmatic activity;this was followed by a climactic Plinian phase, then, finally, columncollapse and pyroclastic flow-forming activity ensued.
In the lower part of the 1362 sequence (i.e. sub-units A1, A2, B1, B2),the presence of finely fragmented juvenile clasts suggests a highdegree of magma–water interaction at the vent. The low degreeof sorting may also indicate that this opening phreatomagmaticphase involved abundant external water (Houghton et al., 2000)with fine wet ash deposited along with coarser clasts. Stratified bedswithin unit A2 imply that this early activity was pulsating (e.g. Luhr,2000). The presence of both ash and pumice lapilli in sub-unit A2
indicates that water may have been entrained into the eruptioncolumn below the vent, enhancing fragmentation to produce ash,eventually resulting in the premature deposition of ash aggregates
(which were held together by the surface tension of water in thehigher parts of the eruption cloud) alongside coarser ash lapilli. Fallinglapilli can also flush out finer ash from the column, and rain seededfrom the eruption cloud can possibly do the same, resulting in wet,cohesive deposits. (Sparks et al., 1997; Houghton et al., 2004). Unit Adeposits are interpreted as locally dispersed phreatomagmatic ashbeds, consistent with the vent opening up under ice.
Unit B deposit shows features that are more typical than unit Aof phreatoplinian deposits, such as a fine grain size near source, poorsorting and a widespread dispersal, most likely caused by magmafragmentation resulting from magma–water interaction and highermagma flux rates during the transition to a drier Plinian phaseof activity (e.g. Sparks et al., 1981). The origin of layer B3 may beattributable to co-ignimbrite ash clouds associated with early,
Fig. 4. Thickness distribution maps for Öræfajökull 1362 proximal fall and flow units (isopach and individual location data in cm). Only locations where primary unit thickness waspreserved are shown. Isopachs on unit A map are based on unit A2 thicknesses; isopachs on unit B map are based on unit B1+B2 thickness. Stippled areas on maps a and b definedistribution area of proximal fall sub-units A1 and B3, respectively. Map d shows locations where primary pyroclastic flowmaterial was found; total pyroclastic flow thickness is unitD1+D2; dotted line on map d shows approximate distribution area of unit D. Locations where possible pyroclastic flow bed is largely reworked are marked with an asterisk. Shadedgrey area on all maps represents Vatnajökull ice cap.
728 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
localized pyroclastic flows, emplaced prior to the main explosiveactivity that produced unit C. Alternatively, layer B3 may simplyrepresent localized accretionary-lapilli-bearing phreatoplinian ashfall, although this is considered more unlikely. Despite the distinctnature of this unit, Thorarinsson (1958) does not mention anaccretionary-lapilli-bearing ash bed in his 1362 deposit stratigraphy,leading us to conclude that this unit is locally distributed within theproximal region only. Thus, we think that unit B was the result of ahigher intensity magma flux, with water (from ice) still having freeaccess to the vent.
Unit C is interpreted to have formed via a steady, sustained,high Plinian eruption column linked to a stable, established vent,and represents the climactic phase. Relatively ‘lithic-rich’ intervals(∼3–6 wt.% juvenile lithic clasts) within this deposit perhaps reflectepisodes of increased vent erosion and possible vent-wall collapse(cf. Wilson et al., 1980; Adams et al., 2001). Grading patterns withinthe deposit, together with the absence of well-defined fall units,suggests constant-rate activity that perhaps occasionally exhibitedfluctuations in eruption intensity. For example, the finer-grained,discontinuous horizons found in unit C most likely indicate a reduc-tion in the eruptive power during the Plinian event (cf. 1991 Hudson
eruption, Chile, Scasso et al., 1994). The highly vesicular nature of thepumice bombs and lapilli in unit C suggests that they formed mainlyby magmatic vesiculation and fragmentation. The much lowerpercentage of ash-grade fallout in the unit suggests that water nolonger had free access to the vent. Possibly, by the onset of phase Cthere were sufficient deposits around the vent, and/or most of the icehad been removed, enabling the upper conduit and vent to be more-or-less dry.
As stated earlier, the pyroclastic flow deposits (unit D) areespecially poorly preserved. In proximal sections they often showevidence of fluvial reworking, including what is interpreted assecondary cross-stratification at the top. It is therefore impossibleto accurately constrain the original extent of the deposits and thuscalculate their exact volume. It is possible that the ash bed on topof the thin pyroclastic flow deposit seen at location 5 (Fig. 2) isfine co-ignimbrite fallout generated by pyroclastic flows travellingover the very steep, glacially-eroded slopes of the volcano andgenerating fine ash clouds. However, as exposures of the uppermostlayers are often truncated, reworked, and much reduced by winderosion, it was not possible to identify co-ignimbrite fallout withconfidence.
Fig. 5. Maximum clast size maps for Öræfajökull 1362 unit C. (a) Isopleth map based on average maximum pumice (MP) data. (b) Average maximum lithic data (ML) isopleth map.Maximum clast sizes and isopleths are shown in cm. Both maximum clast size isopleths are dispersed to the SE.
729K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
At present, most of the identified, mapped units show quite alarge degree of change in thickness and character between theexposures. This indicates that there may have been more variabilityin the units than can be presently resolved; certainly there wassome syn-eruptive, intra-fall unit erosion occurring in this dynamicenvironment. Units A–B were most likely produced in a series ofshort-lived early explosive episodes, with a hiatus between eachone permitting erosion, building up to the climactic eruptiveepisodes and the deposition of units C and D (similar to theopening phases of the Pinatubo 1991 eruption; Hoblitt et al., 1996).
5. Distal ash fall distribution
Thorarinsson's (1958) map of the 1362 deposit across EasternIceland indicates that a widespread fine ash fall deposit exists.He first recognised it in many soil profiles and described it asmedium- to coarse-grained ash, containing silt size pumice clasts.Thorarinsson did not divide the mapped deposit into individual fallunits; however, he did comment on various layers in some proximalexposures. In some sections the deposit is described as havingprimary stratification, resulting from “changes in the intensity ofvolcanic paroxysms”. His map is based on total deposit thickness andhis measurements may include some reworked material at the topof the primary 1362 deposit. This could account for the differencesin deposit thickness between his map and where we have logged thedeposit in the same near-source areas. We only observed andsampled the distal ash fall (where it shows no discernable layering)∼43 km west of the volcano (location 12), where it is a few mil-limetres thick.
The isopach map produced by Thorarinsson (1958) showsa different distribution pattern (Fig. 6a) to the general trend ofthe proximal fall units determined in this study (Fig. 4). His isopachsare regularly spaced, ellipsoidal in shape and extend offshore out tothe east, defining an east-south-east dispersal axis; the smallestmapped isopach is 0.1 cm. On his map, the 0.1 cm isopach extends∼100 km to the west and ∼180 km to the north of the volcano.Locations where no 1362 ash layer is present are also shown.Larsen (unpublished data 2006) suggests that the deposit, at tracethicknesses, does exist west and north of the areas where it isshown by Thorarinsson (Fig. 6a) to be absent. Isopach patternsshown in Fig. 4 suggest that individual fall units B and C have asouth-easterly dispersal axis in both thickness and grain size. These
vary from thinner and finer on the southwest side of Öræfajökull tothicker and coarser south of the summit crater, and thinner andfiner on the south-east side. By comparison, the proximal isopachshape of the total 1362 deposit on Thorarinsson's map seems to belargely controlled by his interpretation of the distal dispersal acrossEastern Iceland.
To further assess the total deposit distribution and estimate thefall deposit volume, we have constructed a new total fall depositisopach map (Fig. 6b and c). On this we plot our proximal data andselected distal thickness data from Thorarinsson's map, togetherwith new thickness data from exposures to the northeast (R. Carey,unpublished data 2006) and west of the volcano. Thorarinsson(1958) thickness data are taken to represent primary falloutthicknesses, i.e. not reworked or thinned by secondary processes.Fig. 6b shows the land-based distribution only and Fig. 6c shows apossible wider dispersal based on extrapolated isopachs off Iceland.These distribution maps define a general north-easterly dispersalaxis, with the deposit thinning more rapidly towards the southwestthan the northeast. The isopachs are drawn with a bulge towardsthe south to accommodate the distribution of proximal fall units A,B, and C, which have a dominant dispersal towards the SSE. The0.1-cm isopach is the thinnest isopach that can be reasonablycompleted and its extrapolated extent (covering ∼195,000 km2)falls ∼100 km offshore from the easternmost point of Iceland and125 km offshore from the south coast. The isopach shape anddistribution for this total deposit distribution map is still largelycontrolled by the widespread fine-grained distal ash fallout acrossEastern Iceland mapped by Thorarinsson (1958). The relationshipbetween the proximal and distal fallout from the 1362 eruption isdiscussed in Section 10.
6. Volume estimate
Volumes of explosive eruption deposits that occur in oceanisland settings are commonly difficult to estimate because a largeproportion of pyroclastic material is often dispersed offshore(Walker, 1981). This fact, coupled with the poor preservationand lack of complete exposures of the 1362 deposit in the harshproximal environment, means that any calculated volume will havea large associated uncertainty. We use the extrapolated total depositisopach map (Fig. 6c) to estimate a bulk fall deposit volume. Basedupon the proximal sections where primary 1362 pyroclastic
Fig. 6. Isopachmaps for thewhole Öræfajökull 1362 (Ö1362) fall deposits. Map (a) is Thorarinsson's (1958) isopachmap. Black circles mark locations where primary deposit thicknesswas measured, crosses mark locations where the 1362 deposit was not found. Map (b) is a total fall deposit map showing land-based distribution only. Map is based on proximalthickness data from this study (grey circles), selected data points from Thorarinsson's map (black circles) and new data from the west of the volcano (open circles). Proximalthicknesses (data from this study) are cumulative fall unit thicknesses (i.e. unit A+unit B+unit C). Map (c) is an extrapolated version of map (b); on this map isopachs wereextrapolated to give likely isopach areas and calculate deposit volume. On all maps, individual thicknesses and isopach values shown in cm.
730 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
Fig. 7. Log thickness versus A1/2 plots for Ö1362 fall units. (a) Shows data for individualfall units, A to C, based uponproximal thickness data (shown on isopachmaps, Fig. 4a–c).Fall units A and B are represented by a single line segment; unit C by two line segmentswith different slopes (see text for discussion). West (black squares) and east (whitesquares) lobes for fall unit A are plotted separately. Volumes are calculated by integrationof area under each of these line segments; data is extrapolated to infinity from thesmallest mapped isopach. (b) Represents total fall deposit thickness (based uponextrapolated isopach map in Fig. 6c). Ö1362 total fall deposit data can be divided intothree distinct line segments. Number labels on each point correspond to point numbersused in defining the three segments (see Table 3); k is slope of line segments. Calculatingthe area under the three segments gives a bulk fall deposit volume of ∼2.2 km3. The twoinflectionpoints (at 25 kmand 75 km)may represent changes in nature of the depositingcolumn (refer to text for more discussion). For comparison the unit C proximal segmentis also plotted on (b) as a solid line. Vertical dashed linesmark segment inflection points.
731K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
flow deposits were measured, we estimate that these deposits havean areal distribution of ∼80 km2 (Fig. 4d). Using a minimum averagethickness of 1 m, a minimum bulk volume of ∼0.1 km3 is indicated.
For the fall deposits, adopting the approach of Pyle (1989, 1995),the square root of the area (A1/2) enclosed by each isopach isplotted against thickness on a log scale (Fig. 7). On such a plot,thickness data of pyroclastic fall units tend to follow a general
Table 2Summary of Öræfajökull 1362 associated volume data for fall units A–C. Note all volumes calcpyroclastic flow deposit is estimated to be ∼0.1 km3, refer to text for further discussion
exponential decrease, producing one or more straight line seg-ments. Many Plinian fall deposits also show a third segment whichrepresents the distal portion of the fall deposit (Pyle, 1995) formedunder a different fluid dynamical regime (Bonadonna et al., 1998).The area beneath each segment can be integrated to yield aminimum bulk deposit volume (including the volume outsideof the smallest mapped isopach). On the A1/2 versus thicknessdiagram for the proximal fall units (A–C), data for fall units A (bothlobes plotted separately) and B are represented by single linesegments, and the coarse pumice fall unit C (as shown in Fig. 4c)is defined by two segments (Fig. 7a). Integrating the area underthese line segments to infinity gives fall unit deposit volumes of0.11 km3, 0.12 km3, and 0.58 km3 for units A, B, and C respectively(Table 2).
Fig. 7b shows the A1/2 versus thickness plot for the totalÖræfajökull 1362 tephra fall deposit, including the distal ash fall,based on Fig. 6c. Three discrete line segments can be recognised.Following the method of Fierstein and Nathenson (1992), the areaunder each of these three line segments was integrated to obtainthe deposit volume, including the volume beyond the smallestmapped isopach (0.1 cm) to infinity (Pyle, 1989). Using thisapproach we completed several versions of the volume calculation,each time changing the points used to create the three segments.We also made estimates by fitting only one or two line segments tothe data. Each variation produced similar values for the totalvolume (Table 3). Averaging the results for three line segmentsyields a bulk fall deposit volume of 2.2±0.05 km3 which, when weinclude the small pyroclastic flow volume, yields a total eruptionbulk volume of ∼2.3 km3. This is equivalent to ∼1.2 km3 dense-rockequivalent (DRE) based on a magma density of 2470 kg m−3 and anaverage compacted bulk deposit density of 1250 kg m−3, or a massof ∼3.0×1012 kg. These figures are significantly smaller than the10 km3 bulk deposit volume and 2 km3 magma volume estimatedby Thorarinsson (1958) based on his isopach map. Note that theformer volume is Thorarinsson's estimate of the amount of freshlyfallen, un-compacted ash. The thickness–A1/2 plot was also used toobtain a theoretical maximum thickness (T0) of 350 cm for the totaldeposit; this T0 estimate approximately agrees with the fall depositthicknesses (unit A+unit B+unit C) measured in the field at themost proximal locality (290 cm at location 13)∼8 km from the vent,Fig. 2.
7. Eruptive parameters
To quantify and compare fall deposits, Pyle (1989, 1995)proposed several parameters to describe thinning rates for falldeposits and changes in clast size with distance from the vent. Thethickness half-distance parameter, bt, defines the A1/2 distance overwhich the deposit thickness halves; tephra deposit plots often havemore than one segment with different bt values. For Pliniandeposits, bt typically ranges between 1 and 10 km for the proximalportion, and can range from 20 to ≥100 km over the distal portion
ulated to∝. Dense rock equivalent (DRE) volume is shown in brackets. Volume of Unit D,
proximal) bt (medial) bt (distal) bc (MP) bc (ML)) (km) (km) (km) (km)
– – –
– – –
– – –
– – 3.86 4.74
Table 4Summary of eruptive parameters for the Öræfajökull 1362 eruption and its deposits
bt (proximal) total deposit 4.8 kmbt (medial) total deposit 13.4 kmbt (distal) total deposit 29.2 km
(b) Unit CUnit C bulk volume 0.58 km3 (0.29 km3 DREa)Minimum eruption durationb 8–12 hEruption column height (HT)c 30–34 kmMDRd 1.0×108 kg s−1
VDRe 4.0×104 m3 s−1
bc (MP)f 3.9 kmbc (ML) 4.7 kmbt (proximal) unit Cg 3.0 kmbt (medial) unit Cg 4.8 kmbc/bt (unit C)h 1.3
a Based on a rhyolite magma density of 2470 g/cm3 and an average compacted bulkdeposit density of 1250 g/cm3.
b Minimum eruption duration of Plinian phase, estimated using accumulation ratesby column height (Wilson and Hildreth, 1997).
c Calculated using maximum clast size data, MP (maximum pumice), ML (maximumlithic); HB ∼21–24 km.
d Mass discharge rate (MDR) estimated from Fig. 6, Sparks (1986).e Volumetric discharge rate (VDR) estimated from Fig. 6, Sparks (1986).f Clast half-distance (Pyle, 1989).g Thickness half-distance (Pyle, 1989).h Half-distance ratio, used to classify deposit dispersal and fragmentation (Pyle,
1989).
Table 3Summary of thickness versus A1/2 dataset for Öræfajökull 1362 total fall deposit. Note allvolumes calculated to ∞. Points 1–9 correspond to numbered labels in Fig. 7
Data for combined total deposit isopach map (Fig. 6)
732 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
(Houghton et al., 2000). Using isopach area and thickness data, btfor the Öræfajökull 1362 Plinian pumice fall unit (C) is estimated tobe 3.0 km in the proximal region and 4.8 km in the medial portion.The maximum clast half-distance, bc, defines the average distanceacross which the maximum clast size halves. Using maximum lithic(ML) and maximum pumice (MP) data, bc for MP is ∼3.9 km andfor ML ∼4.7 km (Fig. 8). Pyle (1989) also used the slope of the bestfit line on an A1/2–maximum clast size plot (Fig. 8) to determineHB, the neutral buoyancy height of an eruption column. HB canbe related to the total column height (HT) by the approximationHB/HT=0.7. Using these relationships, HB for the Öræfajökull 1362climactic Plinian phase (i.e. unit C) is 21 km (MP) and 24 km (ML);therefore HT is 30 km (MP) to 34 km (ML).
Fig. 8. Log maximum clast size versus A1/2 plot for maximum pumice (MP) andmaximum lithic (ML) data from unit C, shown in Fig. 3.5. The slope of the best fit line, k,can be used to estimate clast half-distance (bc), which can then be used to calculateneutral buoyancy height (HB), see text for further discussion.
The total eruption column height (HT) can be used in conjunctionwith the temperature of the erupting mixture to estimate dense-rockmass discharge rate (MDR) and volumetric discharge rate (VDR),after Sparks (1986). For Öræfajökull the magmatic temperature(obtained from Fe–Ti oxide geothermometry) is 828 °C, whichwe assume was the temperature of the erupting mixture. Combiningthis temperature with a column height (HT) estimate of ∼30 km,and assuming temperate atmospheric conditions, yields a MDR of1×108 kg s−1 or a VDR of 4×104 m3 s−1 (Table 4). Mass discharge ratesalso provide a measure of eruptive intensity (Carey and Sigurdsson1989). The Öræfajökull 1362 fall units were emplaced under con-ditions similar in intensity to Plinian events such as the eruptions ofFogo A (Agua de Pau volcano, Azores), Santa María (1902), and ElChichón (1982) unit A.
Pyle (2000; after Walker, 1981) proposes eruption magnitudeand intensity scales to characterise the size and power of a volcaniceruption. The eruption magnitude (M) is defined as log10 [eruptedmass (kg)]−7; and the eruption intensity (I) is defined as log10[mass eruption rate (kg s−1)]+3. Based on the total mass of fall andpyroclastic flow units (∼3.0×1012 kg), the magnitude for Öræfajö-kull 1362 is M5.3, smaller than those of the 1991 Pinatubo climaticeruption (M6.1) and the 1912 Katmai-Novarupta eruption (M6.3).The eruption intensity is 11, calculated using a mass discharge rateof 1×108 kg s−1. Using these parameters the 1362 eruption ofÖræfajökull was similar in magnitude and intensity to the 1956Bezimianny eruption, and lower in intensity than the Pinatubo andKatmai-Novarupta events.
The above estimates are based on calibration with 1-D models oferuption columns and fitting these to fallout deposit parameters. Theyare presented to permit comparison with previous, similarly-basedestimates of eruptive parameters for other events. More recent 2-D and3-D-basedmodelling of convectively driven eruption columns (e.g., Neriet al., 2003; Suzuki et al., 2005; Herzog and Graf, 2007) predict lowerHB, and thus ashplume fallout heights, formaintained columns. AtMDRsof 107–108 kg s−1, neutral buoyancy heights of 15–17 km are indicated
733K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
Fig. 9. Classification scheme for tephra fall deposits (after Pyle, 1989). The half-distanceratio (bc/bt) represents the total grain size population and the thickness half-distance(bt) represents the dispersal. The diagram is contoured for clast half-distance (bc) andtotal column height, HT (km). Öræfajökull 1362 (grey triangle), together with mostPlinian deposits, plots between the bc=3 km and bc=8 km contours.
(HT 21–24 km) and this altitude may be the maximum from whichdownwind dispersal of ash occurred.
As details of the timing and total duration of the 1362 eruptionare unclear from historical accounts, we attempt to constrainthe duration of the main phase using known accumulation rates offall deposits based upon eruption column heights (Wilson andHildreth, 1997). Using an estimate of 30 km for HT, the Wilson andHildreth compilation suggests appropriate accumulation rates of0.023 mm s−1 at location 10 (∼14 km along the dispersal axis; Unit Cthickness 97 cm) and 0.030 mm s−1 at location 13 (8 km along thedispersal axis; Unit C thickness 143 cm), respectively. Based on theserates and the unit C thicknesses at these locations south of the vent,the duration for the main Plinian phase is estimated to be 11–13 h.These estimated durations are longer than the eruption durationdetermined semi-independently using the estimated mass dischargerate (1×108 kg s−1) and the unit Cmass (7×1011 kg), which is ∼2–4 h,or the total deposit mass (3×1012 kg), which is ∼8 h. Estimates of8 to N12 h duration for the period of most voluminous output arein general agreement with an overall reported eruption durationof 1–2 days (Thorarinsson, 1958).
The reasonably high dispersive power for the 1362 event is alsoreflected in the clast half-distance parameters (bc) and half-distance ratio (bc/bt). The bc/bt ratio is principally controlled bythe grain size characteristics of the erupting mixture at the vent.Thus there is a wide variation in bc/bt values, with lower ratiosexpected in deposits that are dominated by finer-grained clasts.Most deposits, however, exhibit bc/bt ratios between 0.5 and 1.5(Houghton et al., 2000). A small bc/bt ratio thus implies greaterdispersal, as the deposit thins more slowly. The 1362 Plinianphase has a large bc value (∼3.9 to 4.7 km) but a relatively small bc/bt ratio (1.27 km). These eruption parameters rank the 1362Öræfajökull eruption alongside other intense Plinian events suchas Quizapu (1932), Fogo A and El Chichón A (1982) (Carey andSigurdsson, 1989).
The dispersive power of the 1362 Plinian event and the higheruption plume (24–30 km), combined with the northerly location ofthe volcano and the low height of the tropopause (∼10 km) during thesummer months, imply that ash and gas were able to easily reachlevels where strong lower stratospheric winds are more likely to
transport lapilli and ash for greater distances. During early June,stratospheric winds above Iceland blow from an approximate SSWdirection (Lacasse, 2001) dispersing tephra in a general NNE to NEdirection, consistent with our interpretation of the 1362 fine ashfallout. This is generally supported by dispersal of the fine ash com-ponent from heights of ∼18–20 km, whereas the coarser, faster-fallingparticles (forming the preserved proximal part of unit C) were dis-persed from heights just above the tropopause (10–16 km) in an ESEdirection (Lacasse, 2001).
A small bc/bt ratio (1.27) on Pyle's (1989) classification scheme(Fig. 9) also suggests that there was efficient fragmentation ofmagma during the 1362 eruption, resulting in a large volume offine ash being produced during the Plinian phase, which con-tributed to the considerable amount of distal fallout. This rein-forces the interpretation that a relatively small proportion ofthe magma erupted during the Plinian phase forms the coarseproximal pumice fall layer, while the rest of the magma washighly fragmented and entered the umbrella cloud region of theeruption plume to be widely dispersed both up- and downwind(Fig. 9).
8. Petrology and whole-rock geochemistry
Major and trace element bulk compositions of the Öræfajökull1362 ejecta were determined by analysing selected pumiceclasts using X-ray Fluorescence Spectroscopy (XRF). The results(Appendix A (Tables A1–A3)) show that the 1362 white pumicesare all low-silica, high-K rhyolites, with approximately 70 wt.% SiO2
and ∼9 wt.% Na2O+K2O, and notably high concentrations ofiron (mean Fe2O3 ∼3.8 wt.%) in agreement with previous studies(Prestvik, 1982; Larsen et al., 1999). There is no discernible dif-ference in the whole-rock composition of rhyolite juvenile clastsbetween the various eruptive units (Table A1). The Öræfajökull1362 rhyolite can be distinguished from other Icelandic siliciceruption deposits by its low MgO and CaO content (b0.1 wt.% andb1.2 wt.% respectively) and high Na2O (∼5 wt.%) values (e.g., Larsenet al., 1999).
Glass from the Öræfajökull 1362 pumices was analysed by electronmicroprobe and is rhyolitic in composition (Tables A2 and A3; fulldetails of the analytical setup are outlined in Appendix A). Matrix glasscompositions from each of the eruptive units (A–D) and glass inclusioncompositions overlap both with each other and with the whole-rockpumice composition (Tables A2 and A3; see also Selbekk and Trønnes(2006), although their study largely features analyses of whole tephrasamples).
White pumice clasts from the 1362 deposits are relatively crystal-poor. They contain ∼1–3% (determined via point counting of thinsections and back-scatter electron images) phenocrysts (defined hereas ≥0.3 mm) set in a glassy, microlite-poor matrix. The mineralassemblage consists of clinopyroxene (En4.3Fs51.5Wo44.2), fayaliticolivine (Fo1.8), and plagioclase (Ab79.8An15.2Or5.0) with minoramounts of Fe–Ti oxides and very occasional apatite crystals. Thecomposition of the main mineral phases is similar in the differenteruptive units, consistent with both the constant bulk whole-rockand glass chemistry (Tables A1 and A2). Geochemical data presentedhere indicate that the bulk of the 1362 magma was homogenous andthe composition did not apparently change as the eruptionprogressed, which is also reflected in the glass andmineral chemistry.Selbekk and Trønnes (2006) have also reported a similar interpreta-tion, however Unit C also contains sparse grey pumice clasts anda few rhyolite, white pumice clasts containing bands of dark greymaterial (banded pumice) (Fig. 3.3f).
Both the grey pumice and the bands of darker material in therhyolite pumice are less finely vesicular than the rhyolite pumice;where present vesicles are sub-rounded to angular in shape and areoften elongated or coalesced. Phenocryst contents of these bands are
734 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
small, b0.5%. However, the main difference between the grey andbandedpumice and the rhyolite pumice is the groundmass crystallinity;with both types of ‘mafic’ samples containing a large amount of feldsparand pyroxenemicrolites.Whole-rock analyses (XRF) of the grey pumiceand of thematerial in the dark bands (Table A1) show that it is andesitic(high-Si) to dacitic in composition (63 to 66 wt.% SiO2 respectively).Feldspar and clinopyroxene are the mainmineral phases present in thegrey pumice, although minor amounts of Fe–Ti oxides and apatite arealso present. Apatite occurs both as small microlite-size crystals andalso as small inclusionswithin clinopyroxene crystals. Pyroxene crystalsexhibit a narrow range in compositionwith all analysed crystals havinga mean composition of approximately En30.8Fs28.9Wo40.3. In contrast,feldspar crystals show a wide range in composition, with anorthitecontents varying from An21.5–An91.7. Pyroxene is the dominantmineral phase in grey bands, with feldspar, apatite and Fe–Ti oxidespresent as accessory phases. Occasional crystals of olivine are alsopresent. Clinopyroxene compositions vary from En16.0Fs41.0Wo43.0 toEn29.5Fs30.9Wo39.6. Plagioclase is rare in the intermediate compositionband; a few crystals present are similar in composition to those in therhyolite pumice (Ab72.2An25.1Or2.7). However, occasional crystals differfrom this mean composition, showing higher anorthite and lowerorthoclase contents (Ab26.7An73.1Or0.2), implying that these crystalspossibly formed in a more mafic/intermediate melt. Olivine isuncommon in the banded material; a single crystal was analysedand shown to have a higher forsterite content (Fo11.9) than that of theolivine analysed in the rhyolite pumice (Fo1.8). The presence of the greypumice and the dark-banded rhyolitic pumice provides tantalisingevidence for magma mixing. However, more data is required tounderstand the origin or extent of this mixing and its significanceduring the 1362 eruption.
8.1. Magmatic intensive parameters
Magmatic temperature and oxygen fugacity for the Öræfajökull1362 rhyolite magma were determined using co-existing magnetiteand ilmenite compositions, following the method of Anderson et al.(1993). As iron–titanium (Fe–Ti) oxides re-equilibrate faster thansilicates following changes in P–T–X conditions (Gardner et al.,1995), they are the crystal phases most likely to record pre-eruptiveP–T conditions. Equilibrium pairs of Fe–Ti oxide crystals wereidentified by the Mg/Mn partitioning criterion of Bacon andHirschmann (1988), with Mg/Mn ratios for ilmenite and magnetiteranging from 0.04–0.07 and 0.01–0.12 respectively. Oxide mineralcompositions were entered into the QUILF 4.1 software (Andersonet al., 1993). QUILF calculations yielded a magmatic temperatureof ∼828 °C, and a log fO2 of −15.6, ΔNNO=−2.0 (approximatelytwo log units below the Ni–NiO buffer). These values indicatethat the 1362 Öræfajökull magma was comparatively reduced;they are similar to those found by other workers using differentformulations of the magnetite–ilmenite geothermometer (Carmi-chael, 1967; Ghiorso and Sack, 1991).
Four glass inclusions and two matrix glass shards (bubblewall fragments) from unit C pumice clasts were analysed usingmicro-FTIR in transmission mode following the method out-lined in Appendix A. The glass inclusions contained 1.7–2.0 wt.%H2O, whereas degassed matrix glasses record H2O contents of0.2–0.3 wt.%. Using the microprobe analytical total (differencefrom 100%) as an indicator of unanalysed volatiles (e.g. Devineet al., 1995), the inclusion H2O values determined via FTIR are inbroad agreement with the microprobe totals of ∼30 rhyolitic glassinclusions (Table A3).
9. Estimate of SO2 release and atmospheric mass loading
Sulphur concentrations in matrix glasses (Cmatrix) and glass inclu-sions in pyroxene and feldspar crystals (Cinc) were determined by
electronmicroprobeanalysis fromunitCpumice clasts as this representsthe largest volume andmost intense part of the eruption. Matrix glass Scontents range from 0.008 to 0.011 wt.% S. In contrast, compositionallysimilar glass inclusions show a range of pre-eruptive sulphur concen-trations, from∼0.006 to∼0.024wt.% S (Fig.10; Table A3). Asmany of thematrix glass values are close to the microprobe detection limit forsulphur (3σ limit∼78ppm), avalueof 80ppmwasassumed forCmatrix inanalyses with low concentrations. This value is in good agreement withion chromatography analyses of the crystal-poor pumice clasts whichgives a bulk S content of 72 ppm (A. DiMuro, pers. comm., 2004). Thesevalues are slightly higher than the ones shown by Palais and Sigurdsson(1989); however, this differencemay simply be due to improvedmodernmicroprobe analytical capabilities. The difference between pre-eruptive(Cinc in wt.%) and degassed sulphur contents (Cmatrix in wt.%) multipliedby the mass of magma erupted (MV in kg) and the magma glass (liquid)fraction (1−Wxtls) yields themass of sulphur released to the atmosphere(MS):
MS ¼ MV 1 −Wxtlsð Þ Cinc − Cmatrix½ �100
:
Multiplying the mass of S by a factor of 2 (2.Ms) gives themass of SO2 emitted (Sharma et al., 2004). Only glass inclusionswith the same major element composition as the matrix glasswere considered for use in this calculation, thus ensuring thatthe inclusions used represent non-degassed equivalents of thedegassed matrix liquid. Based on a statistical selection process,six glass inclusions are sufficiently similar to the degassed ma-trix glass composition. The sulphur content of these inclusionsranges from 0.012 to 0.024 wt.% (average 0.020 wt.% S). The highestrecorded S concentration (0.024 wt.% S), was used for the pre-eruptive S (Cinc) to calculate a S degassing maximum (Table A3).We estimate that ∼1 Mt (0.94±0.2) SO2 was injected via the30 km high Plinian eruption column into the upper atmosphereduring the 1362 eruption, assuming a total magma mass (MV) of3×1012 kg, and a crystal mass fraction (Wxtls) of 0.02 (liquid fraction0.98).
Pre-eruptive and degassed chlorine and fluorine concentra-tions were also measured in the 1362 glass inclusions and ma-trix glasses (Tables A2 and A3). Pre-eruptive chlorine contents havea narrow range varying from 0.191–0.219 wt.% Cl. Chlorine valuesin matrix glasses show almost identical concentrations, rangingbetween 0.195 and 0.215 wt.% Cl. In contrast, fluorine contentsin both inclusions and matrix glasses show significant variations.Pre-eruptive fluorine values range between 0.043 and 0.190 wt.%F; however, the majority of inclusions exhibit values between0.071 and 0.090 wt.% F. Fluorine in the matrix glasses rangesfrom 0.096 to 0.164 wt.% F. These values suggest that halogensdid not efficiently degas from the melt during eruption and themass released was not large enough to cause significant atmo-spheric perturbations to the background concentrations of thesespecies.
10. Discussion
10.1. Interpretation of total 1362 fall deposit dispersal and origin of thefine distal ash
There is an apparent misfit between the dispersal of the wholedeposit and the dispersal of the individual proximal fall units (A–C) asdetermined in this study; in this section we explain possible reasonsfor this. The isopach shapes and resulting easterly dispersal onThorarinsson's (1958) isopach map may have resulted from theassumption of a uniform deposit dispersal pattern, leading him to‘fold’ the on-land thickness data over to the south on an E–W axis toobtain the dispersal over the sea. For each on-land thickness
735K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
measurement he plotted, another point of the same thickness occursat the same distance from the volcano directly to the south of the on-land point. This approach could account for his deposit isopachs withan almost easterly dispersal axis.
When isopachs on the total deposit map are drawn from on-landdata alone, without consideration of the expected shape of thedispersal pattern, the deposit dispersal has a north-easterly trendingaxis (Fig. 6b). At Öræfajökull this is consistent with the prevailingdirection of strong stratospheric winds during the early summermonths in Iceland. At elevations of 24–30 km, early summerstratospheric winds will blow from an approximate W to SWdirection (Lacasse, 2001) therefore dispersing fine ash from highPlinian eruption columns in a general ENE direction. This is alsoconsistent with the dispersal direction of fine ash from other majorPlinian eruptions in Iceland (e.g., Askja, 1875; R. Carey, pers. comm.,2006).
We consider three possibilities for the origin of the distal fine ashfall.
10.1.1. Fallout from an umbrella cloudThe shape of the new total deposit isopach map (Fig. 6c) suggests
deposition from an umbrella cloud during the main C–D phases,including upwind dispersal to the WSW to distances of 100 km.This scenario is similar to that of the 1991 Pinatubo eruption,where most of the fine ash rose in one convective eruption col-umn and was dispersed as a single umbrella cloud, with fineash fallout occurring to a distance of 200 km upwind (Koyaguchi,1996).
The 1815 eruption of Tambora also involved a major Plinian phasethat produced several fall units, with a subordinate amount ofpyroclastic flow deposits (Sigurdsson and Carey, 1989; Self et al.,2004). Approximately 90% of the erupted volume at Tambora wasdistal fallout and the proximal equivalent of the distal ash is not aspecific unit. Rather, the distal fallout is believed to be the distalequivalent of the proximal Plinian unit F4 and other fall units inter-calated within the pyroclastic flow deposits, plus some co-ignimbriteash (Self et al., 2004). It is possible that a similar situation occurred
Fig. 10. S versus FeO plot for all unit Cmatrix glasses and glass inclusions. Grey circles = glass(1989), grey square = glass inclusion, white square = matrix glass. Dashed line indicates 3σreliable and are therefore excluded from the petrologic method calculation (see text for furtherror bar shown for reference.
during the 1362 Öræfajökull eruption, with a large proportion of thedominantly fine-grained fallout from phases C and D being depositedin distal regions.
The three straight line segments on the Öræfajökull 1362thickness versus A1/2 plot (Fig. 7b) are similar to those seen onplots for other Plinian deposits, e.g. Quizapu, 1932 (Hildreth andDrake, 1992) and Huaynaputina 1600 AD (Adams et al., 2001)interpreted to be deposited from umbrella clouds. At Öræfajökull,the first inflection point at A1/2 ∼25–28 km possibly represents themaximum spread of the convecting column and the decrease inballistic contributions, thus signalling a change to umbrella clouddeposition, as seen during the 1932 eruption of Volcan Quizapu(Hildreth and Drake, 1992). Note that thickness data for unit C,when plotted on the A1/2–thickness diagram for the total deposit(Fig. 7), has a similar slope and segment length to segment 1 for thetotal deposit, consistent with the fact that segment 1 for the totaldeposit represents coarse, clast dominated fallout in the proximalregion. The second inflection point, at A1/2 ∼75 km, reflects thetransition from coarse-grained fallout, dominated by pumice clasts,to fine-grained, slower-settling fallout, rich in glass shards. Thus, thesecond, distal inflection point at Öræfajökull in Fig. 7b representschanges in the particle settling behaviour of the co-Plinian ash,similar to that described for the Huaynaputina 1600 AD deposit(Adams et al., 2001) and Novarupta 1912 (Fierstein and Hildreth,1992). In these cases, fine-grained ash is interpreted to have de-coupled from the more coarse-grained fallout and was held aloft byturbulence in the higher atmosphere before fallout from theumbrella cloud.
10.1.2. Wind-shift during the main Plinian phaseThe proximal, coarse pumice fall unit C appears to represent
several hours of deposition and forms the mapped dispersal pattern(Figs. 4c and 5), which has an axis to the SSE. Fine-grained, distalash fall from this part of C would have fallen into the sea. However,perhaps phase C continued and a wind shift occurred that resultedin the dispersal of fall deposits on a more north-easterly dispersalaxis. The fine, distal part of this later unit C ash fell out over
inclusions, white circles = matrix glass. Also shown are data from Palais and Sigurdssonmicroprobe detection limit for S; analyses that lie below this line cannot be considereder information). Note that the Palais and Sigurdsson (1989) data fall below this line. 1σ
736 K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
the eastern part of the island and formed the distal unit mappedby Thorarinsson. If this occurred, much of the coarser, proximalfallout of the later unit C would have fallen onto Vatnajökull icecap and may not be readily observable today, and, further, thetotal volume of the eruption may be somewhat underestimated.Contemporary accounts do not provide sufficient constraints toevaluate whether the climatic phase of the eruption lasted muchlonger than 8–12 h.
10.1.3. Origin by co-ignimbrite ash fallSome studies (e.g. Hildreth and Drake, 1992) have also shown
that the distal segment inflection on the thickness versus A1/2 plot(Fig. 7b) can be attributed to the input of large amounts of co-ignimbrite ash to distal fallout. Could this possibly be a thirdexplanation for the distal 1362 ash fallout observed? Our estimateof pyroclastic flow volume appears too small to have generatedsignificant amounts of co-ignimbrite fallout in the distal region, butthere may have been more pyroclastic flow deposits that wereeroded away or buried since 1362. Based on crystal concentrationstudies, approximately 2–3 km3 of pyroclastic flow deposit isrequired to generate ∼1 km3 (bulk volume) of associated co-ignimbrite ash (Walker, 1972). This study suggests that it is unlikelythat such large volumes of pyroclastic flow deposit existed, andwe consider this the least likely option for the origin of the distalash cloud.
10.2. Mass of S released and potential atmospheric impact
The small estimated mass of SO2 released during the Öræfajö-kull 1362 eruption can be explained by low initial sulphurconcentration in the 1362 magma. Earlier studies (Wallace, 2001,2004) have noted a discrepancy between petrologic estimatesof SO2 release and satellite measurements of volcanic SO2 in arceruptions (e.g., Pinatubo 1991). This discrepancy has been attrib-uted to the presence of a sulphur-rich gas phase: in arc magmassulphur is sequestered from the melt into a separate S-bearing fluidphase, thus depleting the melt with respect to sulphur. Sharmaet al. (2004) report that excess sulphur does not seem to be asignificant factor in non-arc basaltic volcanic systems and that ifany excess fluid phase was to form in these systems, then it is likelyto be sulphur poor. Studies suggest that sulphur solubility is afunction of oxygen fugacity and temperature in rhyolite melts (e.g.Scaillet et al., 1998); low sulphur solubility in rhyolites is attributedto a combination of low melt FeO content and low temperatures.The comparatively low log fO2 (ΔNNO −2.0) for the 1362 magmameans that the development of a separate, pre-eruption S-rich fluidphase is unlikely. At these low ΔNNO values, the magma is reducedand thus stabilises formation of an immiscible sulphide phaserather than sequestering sulphur into a SO4
2− sulphate-bearingphase. Experimental work by Scaillet et al. (1998) also shows thatin more reduced silicic magmas (i.e., those co-existing withpyrrhotite), fluid/melt partition coefficients are drastically reduced(values around 1 compared with 50–2600 for oxidised magmas), aspyrrhotite locks up nearly all of the sulphur in the magma. Thusexplosive eruptions involving reduced, cool silicic magmas, such asÖræfajökull 1362, tend to release minor amounts of SO2 into theatmosphere. In the case of Öræfajökull, independent estimates ofthe SO2 release (see below) also support a small eruptive SO2 yield.
Small shards of glass (∼30 μm in diameter) attributed to the1362 eruption have been identified in the GISP-2 Greenland ice core(Palais et al., 1991). Associated with these shards is a minor sulphatepeak (SO4
2−). Following the method outlined by Zielinski (1995),acidity peaks in ice cores can be used to infer the mass of SO2
released. For the 1362 eruption, the estimate of SO2 mass derivedin this way is on the order of 0.4 to 0.9 Mt, in agreement with ourpetrologic estimate.
As discussed earlier, we estimate that the 1362 Plinian eruptioncolumn had a maximum height of 30 km. This, coupled with the factthat during the summer months the tropopause height in Iceland isat an altitude of 9–10 km, suggests that most of the SO2 releasedwas injected into the stratosphere. The amount of SO2 releasedduring the 1362 eruption is sufficient only to have generated∼1.7 Mt of sulphate aerosol. This assumes a gas (SO2) to particle(H2SO4+H2O) conversion efficiency of 86%, based on the satellite-observed stratospheric aerosol conversion efficiency during the1991 Pinatubo eruption (McCormick et al., 1995). A small strato-spheric aerosol mass such as this would have a negligible atmo-spheric impact. Therefore, considering the small syn-eruptivevolatile release and taking into account the low tropopause aboveIceland, the high eruption columns generated during the Plinianphase, and the wide dispersal area of the fine distal ash, perhaps thegreatest atmospheric or environmental impact from the 1362Öræfajökull eruption resulted from the injection of volcanic ashand fine dust particles into the stratosphere.
11. Conclusions
The 1362 eruption of Öræfajökull produced three main fallunits during phreatomagmatic and Plinian activity. Initial phreato-magmatic activity produced fine- to coarse-grained ash beds (unitsA and B) with a SSE dispersal. The climactic phase of the eruptionwas an intense Plinian eruption lasting ∼8–12 h, that gener-ated a 24–30 km high eruption plume and produced a pumice-dominated, coarse fall deposit (unit C), dispersed to the south, inproximal areas, and a widespread fine, distal ash that coveredmuch of northeast Iceland. The differing dispersals possiblyresulted from decoupling of fines material from the coarserpumice-dominated material, with the fines held at higher levelsin a dynamic umbrella cloud system and the coarse proximalpumice fallout dispersed via lower level winds, before fallout fromhigher levels with differing atmospheric wind conditions occurred.This is consistent with recent work suggesting decoupled falloutpaths and independent dispersal of coarse and fine particles inPlinian eruption clouds (Hildreth and Drake, 1992; Fierstein andHildreth, 1992; Adams et al., 2001). Pyroclastic flow activityoccurred after the main Plinian event, generating a poorly-sortedflow deposit (unit D).
Although still a significant eruptive event, the 1362 eruption isvolumetrically smaller than previously thought. A total of 2.3 km3
(∼1.2 km3 DRE) was erupted during the energetic Plinian eruption,accompanied byminor pyroclastic flow activity yielding small volumedeposits (∼0.1 km3).
Low-SiO2 rhyolite pumice dominates the 1362 deposit; however aminor amount of intermediate composition material is also foundintermingled within unit C. These clasts consist of two distincttypes— grey pumice (high-SiO2 andesite— 63 wt.% SiO2) and bandedpumice that consists of white, rhyolite pumice with a distinct darkband (dacite bands — 66 wt.% SiO2). Our sample suite limits furtherconclusions regarding the origin of these mafic/intermediatecomponents. However, we conclude that a mafic to intermediatemagma was involved in the 1362 eruption. Future studies will needto consider the origin of these mixed melts, the degree of magmamixing, and their possible role in triggering rhyolite volcanism atÖræfajökull.
Only a relatively small amount of SO2 (∼1 Mt) was releasedinto the stratosphere during the 1362 event; thus a large amount ofstratospheric sulphate aerosol was not produced. Glass analysesalso indicate that fluorine and chlorine did not degas significantly.Thus, despite the violent nature and relatively high intensity ofthe eruption, the environmental and atmospheric impact of the1362 activity was limited to the injection of fine ash and dust tostratospheric levels and its subsequent widespread, fallout.
737K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
Acknowledgements
This research was funded by an Open University studentshipto KS. The authors wish to thank Andy Tindle for much assistancewith microprobe work, Andrea Di Muro (IPGP, Paris) for ionchromatography analysis, and Armann Hoskuldsson, ChristianLacasse, Dave McGarvie, Greg Zielinski, Rebecca Carey, and JohnStevenson for useful discussions and data. We also wish to thankJames White and an anonymous reviewer for constructive reviews ofthis manuscript.
Appendix A. Analytical setup and geochemical dataset
Analytical techniques
Pumice samples were crushed by hand and sieved into smallersize fractions from which matrix glass shards and crystals werehand-picked for electron microprobe analysis. To analyse the maficmaterial from the banded pumice, the band was trimmed andcrushed separately and small glass fragments were picked foranalysis. Major element and S, Cl and F analyses of glasses wereconducted at The Open University on a Cameca SX-100 electronmicroprobe, using a 20 kV accelerating voltage, 20 nA beam currentand a 10–20 μm diameter beam. These operating conditions wereselected to minimize Na and K loss during analysis of pumice glassfragments. 1σ analytical precision based upon replicate analyses
Table A1Major and trace element whole-rock geochemistry for selected Öræfajökull 1362 samples
a “Mafic” pumice samples found within unit C, Ö2-06B is the banded pumice sample (i.e. apumice from location 4.
b Indicates sample analysed by ICP-MS (for trace element data).
of the glass standard VG-568 (Yellowstone obsidian, USNM 72854)is b1% for major elements and b5% for minor elements (e.g., Mn,K, P). All reported analyses are the average of 3–10 spots. For S, Cl,and F precision, replicate analyses of the glass standard VG-2 (Juande Fuca Ridge glass, USNM 111240/52) gave 1480±50 (2σ) ppm S(n=80), and replicate analyses of the glass standard VG-568 gave2000±70 (2σ) ppm Cl (n=100) and 1200±100 (2σ) ppm F (n=100),in good agreement with values reported in previous studies(e.g., Sharma et al., 2004; Davis et al., 2003; Thordarson et al., 1996).
H2O and CO2 contents of the 1362 rhyolite magma were deter-mined from doubly polished wafers of glass inclusions and matrixglasses using Fourier Transform Infrared (FTIR) spectroscopy.Spectra were collected with a Thermo Nicolet Nexus FTIR spectro-meter coupled with a Thermo Continuμm IR microscope. Forall spectra, standard EverGlo mid-infra-red source optics, a Ge-on-KBr beamsplitter, and MCT-A⁎ detector (11,700–750 cm− 1)were used. The full quantitative procedure is described in Ohlhorstet al. (2001).
Weight percent H2O and CO2 concentrations were ascertainedusing Beer's law:
c ¼ MAρdeð Þ
� �� 100
where c is the species concentration, M is the molecular weight(18.02 for total H2O and 44 for CO3); A is the absorbance, ρ is the
and 1066 l mol−1 cm−1 respectively (Tamic et al., 2001). In rhyolitemelt, dissolved H2O occurs as two different species: molecular H2O
Table A3Glass inclusion geochemistry data for selected Öræfajökull 1362 samples
Shaded analyses indicate glass inclusions selected for petrologic method calculation (refer t
and OH− (Wallace and Anderson, 2000). H2O concentration wasobtained by measuring the height of the total water peak (i.e.molecular H2O+OH−) at 3550 cm−1, and CO2 concentration wasacquired by measuring peak height of CO2 at 2346 cm−1. Traceelement chemistry for the banded pumice sample (Ö2-03) wasobtained using ICP-MS analysis as the sample of dark material thatwe derived by picking and cutting was too small for standard XRFanalysis. A pumice sample from unit C (Ö4-06) was also analysed inthis way to test and compare the results of both techniques. Both XRFand ICP-MS analyses are shown in Table A1 Geochemical Datasets.
o Section 9 in text for more details).
739K. Sharma et al. / Journal of Volcanology and Geothermal Research 178 (2008) 719–739
References
Adams, N.K., de Silva, S.L., Self, S., Salas, G., Schubring, S., Permenter, J.L., Arbesman, K.,2001. The physical volcanology of the 1600 eruption of Huaynaputina, southernPeru. Bulletin of Volcanology 62, 493–518.
Anderson, D.J., Lindsley, D.H., Davidson, P.M., 1993. QUILF: a Pascal program to assessequilibria among Fe–Mg–Mn–Ti oxides, pyroxenes, olivine and quartz. ComputingGeosciences 19, 1333–1350.
Bacon, C.R., Hirschmann, M.M., 1988. Mg/Mn partitioning as a test for equilibriumbetween co-existing Fe–Ti oxides. American Mineralogist 73, 57–61.
Bonadonna, C., Ernst, G.G.J., Sparks, R.S.J., 1998. Thickness variations and volumeestimates of tephra fall deposits: the importance of particle Reynolds number.Journal of Volcanology and Geothermal Research 81, 173–187.
Carey, S., Sigurdsson, H., 1989. The intensity of Plinian eruptions. Bulletin of Volcanology51, 28–40.
Carmichael, I.S.E., 1967. The iron–titanium oxides of salic volcanic rocks and theirassociated ferromagnesian silicates. Contributions to Mineralogy and Petrology 14,36–64.
Davis, M.G., Garcia, M.O., Wallace, P.J., 2003. Volatiles in glasses from Mauna LoaVolcano, Hawaii: Implications for magma degassing and contamination, andgrowth of Hawaiian volcanoes. Contributions to Mineralogy and Petrology 144,570–591.
Devine, J.D., Gardner, J.E., Brack, H.P., Layne, G.D., Rutherford, M., 1995. Comparison ofmicroanalytical methods for estimating H2O contents of silicic volcanic glasses.American Mineralogist 80, 319–328.
Fierstein, J., Nathenson, M., 1992. Another look at the calculation of fallout tephravolumes. Bulletin of Volcanology 54, 156–167.
Fierstein, J., Hildreth, W., 1992. The Plinian eruptions of 1912 at Novarupta, KatmaiNational Park, Alaska. Bulletin of Volcanology 54, 646–684.
Gardner, J.E., Rutherford, M., Carey, S., Sigurdsson, H., 1995. Experimental constraints onpre-eruptive water contents and changing magma storage prior to explosiveeruptions of Mount St Helens volcano. Bulletin of Volcanology 57, 1–17.
Ghiorso, M.S., Sack, O., 1991. Fe–Ti oxide geothermometry: thermodynamic formulationand the estimation of intensive variables in silicic magmas. Contributions toMineralogy and Petrology 108, 485–510.
Hall, V.A., Pilcher, J.R., 2002. Late-Quaternary Icelandic tephras in Ireland and GreatBritain: detection, characterization and usefulness. The Holocene 12 (2), 223–230.
Herzog, M., Graf, H., 2007. Volcanic plumes: what is the realistic neutral buoyancyheight? EOS Transactions 88 (52) AGU Fall Meeting Supplement, abstract V31A-0291.
Hildreth, W., Drake, R.E., 1992. Volcan Quizapu, Chilean Andes. Bulletin of Volcanology54, 93–125.
Hoblitt, R.P., Wolfe, E.W., Scott, W.E., Couchman, M.R., Pallister, J.S., Javier, D., 1996. Thepreclimactic eruptions of Mount Pinatubo, June 1991. In: Newhall, C.G., Punong-bayan, R.S. (Eds.), Fire and Mud: Eruptions and Lahars of Mount Pinatubo,Philippines. University Washington Press, Quezon City, pp. 457–512.
Houghton, B.F., Wilson, C.J.N., Pyle, D.M., 2000. Pyroclastic fall deposits. In: Sigurdsson,H., Houghton, B.F. (Eds.), Encyclopaedia of Volcanoes. Academic Press, San Diego,pp. 555–570.
Houghton, B.F., Wilson, C.J.N., Fierstein, J., Hildreth, W., 2004. Complex proximaldeposition during the Plinian eruptions of 1912 at Novarupta, Alaska. Bulletin ofVolcanology 66, 95–133.
Koyaguchi, T., 1996. Volume estimation of tephra-fall deposits from the June 15, 1991,eruption of Mount Pinatubo by theoretical and geological methods. In: Newhall, C.G.,Punongbayan, R.S. (Eds.), Fire and Mud: Eruptions and Lahars of Mount Pinatubo,Philippines. University of Washington Press, Quezon City, pp. 513–543.
Lacasse, C., 2001. Influence of climate variability on the atmospheric transport of Icelandictephra in the subpolar North Atlantic. Global and Planetary Change 29, 31–55.
Larsen, G., Gudmundsson, M.T., Bjornsson, H., 1998. Eight centuries of periodicvolcanism at the center of the Iceland hotspot revealed by glacier tephrostrati-graphy. Geology 26, 943–946.
Larsen, G., Dugmore, A.J., Newton, A.J., 1999. Geochemistry of historical age silicictephras in Iceland. The Holocene 9, 463–471.
Luhr, J.F., 2000. The geology and petrology of Volcan San Juan (Nayarit, Mexico) and thecompositionally zoned Tepic Pumice. Journal of Volcanology and GeothermalResearch 95, 109–156.
McCormick, M.P., Thomason, L.W., Trepte, C.R., 1995. Atmospheric effects of the MountPinatubo eruption. Nature 373, 399–404.
McPhie, J., Doyle, M., Allen, R., 1993. Volcanic Textures: a Guide to the Interpretation ofTextures in Volcanic Rocks. Tasmanian Government Printing Office, Hobart.
Neri, A., Esposito, O.T., Macedonio, G., Gidaspow, D., 2003. Multiparticle simulation ofcollapsing volcanic columns and pyroclastic flows. Journal of Geophysical Research108 (B4), 2202.
Ohlhorst, S., Behrens, H., Holtz, F., 2001. Compositional dependence of molarabsorptivities of near-infrared OH− and H2O bands in rhyolitic to basaltic glasses.Chemical Geology 174, 5–20.
Palais, J.M., Sigurdsson, H., 1989. Petrologic evidence of volatile emissions frommajor historic and pre-historic volcanic eruptions. In: Berger, A., Dickinson, R.E.,
Prestvik, T., 1979. Petrology of hybrid intermediate and silicic rocks from Öræfajökull,southeast Iceland. GFF 101, 299–308.
Prestvik, T., 1982. Petrography, chemical characteristics, and nomenclature ofÖræfajökull rocks. Jokull 32, 69–76.
Pyle, D.M., 1989. The thickness, volume and grain size of tephra fall deposits. Bulletin ofVolcanology 51, 1–15.
Pyle, D.M., 1995. Assessment of the minimum volume of tephra fall deposits. Journal ofVolcanology and Geothermal Research 69, 379–382.
Pyle, D.M., 2000. Sizes of volcanic eruptions. In: Sigurdsson, H., Houghton, B.F., et al.(Eds.), Encyclopaedia of Volcanoes. Academic Press, San Diego, pp. 263–269.
Scaillet, B., Clemente, B., Evans, B., Pichavant, M., 1998. Redox control of sulfur degassingin silicic magmas. Journal of Geophysical Research 103, 23937–23949.
Scasso, R.A., Corbella, H., Tiberi, P., 1994. Sedimentological analysis of the tephra fromthe 12–15 August 1991 eruption of Hudson volcano. Bulletin of Volcanology 56,121–132.
Selbekk, R.S., Trønnes, R.G., 2006. The 1362 AD Öræfajökull eruption, Iceland: petrologyand geochemistry of large-volume homogeneous rhyolite. Journal of Volcanologyand Geothermal Research 160, 42–58.
Self, S., Gertisser, R., Thordarson, Th., Rampino, M.R., Wolff, J.A., 2004. Magma volume,volatile emissions, and stratospheric aerosols from the 1815 eruption of Tambora.Geophysical Research Letters 21 (L20608).
Sharma, K., 2006. The Eruptions of Öræfajökull 1362 (Iceland) and Lanzarote 1730-36(Canary Islands): Sulphur Emissions and Volcanology. The Open University, MiltonKeynes. 292 pp.
Sharma, K., Blake, S., Self, S., Krueger, A.J., 2004. SO2 emissions from basaltic eruptionsand the excess sulfur issue. Geophysical Research Letters 31, L13612.
Sigurdsson, H., Carey, S., 1989. Plinian and co-ignimbrite fall from the 1815 eruption ofTambora volcano. Bulletin of Volcanology 51, 243–270.
Sparks, R.S.J., 1986. The dimensions and dynamics of volcanic eruption columns. Bulletinof Volcanology 48, 3–15.
Sparks, R.S.J., Wilson, L., Sigurdsson, H., 1981. The pyroclastic deposits of the 1875eruption of Askja, Iceland. Philosophical Transactions Royal Society of London 229(A), 241–273.
Stevenson, J.A., McGarvie, D.W., Smellie, J.L., Gilbert, J.S., 2006. Subglacial and ice-contact volcanism at the Öræfajökull stratovolcano, Iceland. Bulletin of Volcanology68, 737–752.
Suzuki, Y.J., Koyaguchi, T., Ogawa, M., Hachisu, I., 2005. A numerical study of turbulentmixing in eruption clouds using a three-dimensional fluid dynamics model. Journalof Geophysical Research 110 (B08), 201.
Tamic, N., Behrens, H., Holtz, F., 2001. The solubility of H2O and CO2 in rhyolitic melts inequilibrium with a mixed CO2–H2O fluid phase. Chemical Geology 174, 333–347.
Thorarinsson, S., 1958. The Öræfajökull eruption of 1362. Acta Naturalia Islandica II (2).Thordarson, T., Self, S., Oskarsson, N., Hulsebosch, T., 1996. Sulfur, chlorine and fluorine
degassing and atmospheric loading by the 1783–1784 AD Laki (Skaftar Fires)eruption in Iceland. Bulletin of Volcanology 58, 205–225.
Walker, G.P.L., 1972. Crystal concentration in ignimbrites. Contributions to Mineralogyand Petrology 36, 135–146.
Walker, G.P.L., 1981. Plinian eruptions and their products. Bulletin of Volcanology 44 (2),223–224.
Wallace, P.J., 2001. Volcanic SO2 emissions and the abundance and distribution ofexsolved gas in magma bodies. Journal of Volcanology and Geothermal Research108, 85–106.
Wallace, P.J., 2004. Frommantle to atmosphere: magma degassing, explosive eruptions,and volcanic volatile budgets. In: De Vivo, B., Bodnar, R.J. (Eds.), Melt Inclusionsin Volcanic Systems: Methods, Applications and Problems. Elsevier, Amsterdam,pp. 105–127.
Wallace, P.J., Anderson, A.T., 2000. Volatiles in magmas. In: Sigurdsson, H., Houghton, B.F.(Eds.), Encyclopedia of Volcanoes. Academic Press, San Diego, pp. 495–511.
Wilson, C.J.N., Hildreth, W., 1997. The Bishop Tuff: new insights from eruptive stra-tigraphy. Journal of Geology 105, 407–439.
Wilson, L., Sparks, R.S.J., Walker, G.P.L., 1980. Explosive volcanic eruptions — IV: thecontrol of magma properties and conduit geometry on eruption column behaviour.Geophysical Journal International 63, 117–148.
Zielinski, G.A., 1995. Stratospheric loading and optical depth estimates of explosivevolcanism over the last 2100 years derived from the Greenland Ice Sheet Project 2ice core. Journal of Geophysical Research 100 (D10), 20937–20955.
![Electronics(S.E.) Syllabus New[1]](https://static.documents.pub/doc/80x56/55cf96ae550346d0338d1805/electronicsse-syllabus-new1.jpg)
