Eruption and emplacement of a basaltic welded ignimbrite during caldera formation on Gran Canaria

The 14.1 Ma old composite ignimbrite cooling unit P1 (45 km³) on Gran Canaria comprises a lower mixed rhyolite-trachyte tuff, a central rhyolite-basalt mixed tuff, and a slightly rhyolite-contaminated basaltic tuff at the top. The basaltic tuff is compositionally zoned with (a) an upward change in basalt composition to higher MgO content (4.3–5.2 wt.%), (b) variably admixed rhyolite or trachyte (commonly <5 wt.%), and (c) an upward increasing abundance of basaltic and plutonic lithic fragments and cognate cumulate fragments. The basaltic tuff is divided into three structural units: (I) the welded basaltic ignimbrite, which forms the thickest part (c. 95 vol.%) and is the main subject of the present paper; (II) poorly consolidated massive, bomb- and block-rich beds interpreted as phreatomagmatic pyroclastic flow deposits; and (III) various facies of reworked basaltic tuff. Tuff unit I is a basaltic ignimbrite rather than a lava flow because of the absence of top and bottom breccias, radial sheet-like distribution around the central Tejeda caldera, thickening in valleys but also covering higher ground, and local erosion of the underlying P1 ash. A gradual transition from dense rock in the interior to ash at the top of the basaltic ignimbrite reflects a decrease in welding; the shape of the welding profile is typical for emplacement temperatures well above the minimum welding temperature. A similar transition occurs at the base where the ignimbrite was emplaced on cold ground in distal sections. In proximal sections the base is dense where it was emplaced on hot felsic P1 tuff. The intensity of welding, especially at the base, and the presence of spherical particles and of mantled and composite particles formed by accretion and coalescence in a viscous state imply that the flow was a suspension of hot magma droplets. The flow most likely had to be density stratified and highly turbulent to prevent massive coalescence and collapse. Model calculations suggest eruption through low pyroclastic fountains (<1000 m high) with limited cooling during eruption and turbulent flow from an initial temperature of 1160°C. The large volume of 26 km³ of erupted basalt compared with only 16 km³ of the evolved P1 magmas, and the extremely high discharge rates inferred from model calculations are unusual for a basaltic eruption. It is suggested that the basaltic magma was erupted and emplaced in a fashion commonly only attributed to felsic magmas because it utilized the felsic P1 magma chamber and its ring-fissure conduits. Evolution of the entire P1 eruption was controlled by withdrawal dynamics involving magmas differing in viscosity by more than four orders of magnitude. The basaltic eruption phase was initially driven by buoyancy of the basaltic magma at chamber depth and continued degassing of felsic magma, but most of the large volume of basalt magma was driven out of the reservoir by subsidence of a c. 10 km diameter roof block, which followed a decrease in magma chamber pressure during low viscosity basaltic outflow.     

Ignimbrites of the Roque Nublo group, Gran Canaria, Canary Islands

 Non-welded, lithic-rich ignimbrites, hereintermed the Roque Nublo ignimbrites, are the most distinctive deposits of the Pliocene Roque Nublo group, which forms the products of second magmatic cycle on Gran Canaria. They are very heterogeneous, with 35–55% volume lithic fragments, 15-30% mildly vesiculated pumice, 5–7% crystals and 20–30% ash matrix. The vitric components (pumice fragments and ash matrix) are largely altered and transformed into zeolites and subordinate smectites. The Roque Nublo ignimbrites originated from hydrovolcanic eruptions that caused rapid and significant erosion of vents thus incorporating a high proportion of lithic clasts into the eruption columns. These columns rapidly became too dense to be sustained as vertical eruption columns and were transformed into tephra fountains which fed high-density pyroclastic flows. The deposits from these flows were mainly confined to palaeovalleys and topographic depressions. In distal areas close to the coast line, where these palaeovalleys widened, most of the pyroclastic flows expanded laterally and formed numerous thin flow units. The combined effect of the magma–water interaction and the high content of lithic fragments is sufficient to explain the characteristic low emplacement temperature of the Roque Nublo ignimbrites. This fact also explains the transition from pyroclastic flows into lahar deposits observed in distal facies of the Roque Nublo ignimbrites. The existence of hydrovolcanic eruptions generating high-density pyroclastic flows, unable to efficiently separate the water vapour from the vitric components during transport, also accounts for the intense zeolitic alteration in these deposits. Received: 5 November 1996 / Accepted: 3 March 1997     

Large-scale rheomorphic shear deformation in Miocene peralkaline ignimbrite E,Gran Canaria

the single ignimbrite cooling unit E (average thickness, 28 m; volume, ca. 30 km₃) forms the uppermost member of the Miocene Upper Mogán Formation on Gran Canaria. It is strongly chemically zoned from basal, first-erupted comendite (peralkaline rhyolite) to late-erupted trachyte, and, apart from an upper trachytic zone, it is densely welded. E was emplaced onto a surface inclined ca. 2–5° from the source caldera. Detailed mapping of key sections, up to 300 m long, exposed in barranco walls, ca. 10 km from the caldera margin, reveals structures that are interpreted to have been produced by rheomorphic deformation of the ignimbrite along shear zones. The shear zones formed within the lower-viscosity comenditic tuff. Extensional structures include mega-boudinage and decapitated sequences and compression resulted in sequence repitition by overthrusting. Mechanisms traditionally thought to be important during rheomorphic deformation of welded tuffs (compaction, lateral creep, folding, vertical density-driven diapirism) cannot account for these features, which reflect lateral (post-compactional) rheomorphic movement locally in excess of 800 m. We suggest the following sequence of events: emplacement of the several flow units; compaction, with little lateral movement; rheomorphic deformation. During and after compaction, layers of secondary porosity developed within the comenditic tuff, possibly where upward escape of gas was prevented by overlying, relatively impermeable layers of densely compacted ignimbrite. These structurally weak layers of high porosity subsequently acted as shear zones.     

Genesis of roque nublo formation: A special kind of ignimbritic eruptions in Gran Canaria

A series of comagmatic volcanic materials originated by subaerial eruptions that happened during the Neogene in Gran Canaria (Canary Islands) are studied in this work. The sequence consists of flows (basalts, basanites, tephrites) underlying volcanic agglomerate sheets. The whole unit, which reaches a maximum thickness greater than 700 m, is cut by a number of phonolite plugs. Flows, agglomerate sheets and plugs are genetically related, forming a differentiation series whose evolution has been rather complex: crystal fractionation, amphibole resorption, changes in oxygen pressure and gaseous transfer have played a role in the genesis of these volcanics. From an evolutionary point of view, the Roque Nublo Formation can be described as an alkaline series with two different undersaturated zones (tephritic flows and phonolitic domes) separated by a maximum of saturation (the agglomerate matrix is chemically a trachyte). In this respect, the Roque Nublo Formation is similar to the alkaline Cantal series, although in Gran Canaria the relations between trachytes and phonolites seem to be better defined. As for the agglomerates, their uncommon characteristics (heterogeneous and very poorly sorted boulders predominating over a vitroclastic welded matrix) lead one to think that they were produced by an unusual kind of cruption; certainly not by the « nuée ardente » types which have been repeatedly postulated before. The author’s suggestion is that these agglomerates (« Roque Nublo type ») were formed in ignimbritic-style eruptions of higy-viscosity magmas contained in very high-pressure chambers.     

Low-temperature hydrothermal alteration of intra-caldera tuffs,Miocene Tejeda caldera,Gran Canaria,Canary Islands

The Miocene Tejeda caldera on Gran Canaria erupted ~ 20 rhyolite–trachyte ignimbrites (Mogán Group 14–13.3 Ma), followed by ~ 20 phonolitic lava flows and ignimbrites (Fataga Group 13–8.5 Ma). Upper-Mogán tuffs have been severely altered immediately within the caldera margin, whereas extra-caldera Mogán ignimbrites, and overlying Fataga units, are apparently unaltered. The altered intra-caldera samples contain minerals characteristic of secondary fluid–rock interaction (clays, zeolites, adularia), and relics of the primary mineral assemblage identified in unaltered ignimbrites (K-feldspar, plagioclase, pyroxene, amphibole, and groundmass quartz). Major and trace-element data indicate that Si, Na, K, Pb, Sr, and Rb, were strongly mobilized during fluid–rock interaction, whereas Ti, Zr, and Nb behaved in a more refractory manner, experiencing only minor mobilization. The δ¹⁸O values of the altered intra-caldera tuffs are significantly higher than in unaltered extra-caldera ignimbrites, consistent with an overall low-temperature alteration environment. Unaltered extra-caldera ignimbrites have δD values between − 110‰ and − 173‰, which may reflect Rayleigh-type magma degassing and/or post-depositional vapour release. The δD values of the altered intra-caldera tuffs range from − 52‰ to − 131‰, with ambient meteoric water at the alteration site estimated at ca. − 15‰. Interaction and equilibration of the intra-caldera tuffs with ambient meteoric water at low temperature can only account for whole-rock δD values of around − 45‰, given that ?D_clay–water is ca. − 30‰ at 100 °C, and decreases in magnitude at higher temperatures. All altered tuff samples have δD values that are substantially lower than − 45‰, indicating interaction with a meteoric water source with a δD value more negative than − 15‰, which may have been produced in low-temperature steam fumaroles. Supported by numerical modeling, our Gran Canaria data reflect the near-surface, epithermal part of a larger, fault-controlled hydrothermal system associated with the emplacement of the high-level Fataga magma chamber system. In this near-surface environment, fluid temperatures probably did not exceed 200–250 °C.     

Geochronology of Gran Canaria,Canary Islands: Age of shield building volcanism and other magmatic phases

Forty-six new K-Ar age determinations are presented on whole rock samples and mineral separates from volcanic and subvolcanic rocks of Gran Canaria. The main subaerial shield building basaltic volcanism with estimated volume of about 1000 km³ was confined to the interval about 13.7 m.y. to 13.5 m.y. ago in the middle Miocene. Substantial volume (～100 km³) of silicic volcanics (trachyte and peralkaline rhyolite) were erupted with no detectable time break following the basaltic volcanism, essentially contemporaneous with formation of a large collapse caldera at 13.4±0.3 m.y. ago. Trachytic to phonolitic volcanism continued intermittently in the waning states of activity until about 9 m.y. ago. Following a long hiatus there was resurgence of volcanism with eruption of about 100 km³ of basanitic to hauyne phonolitic rocks of the Roque Nublo Group between about 4.4 m.y. and 3.4 m.y. ago in the Pliocene. After a hiatus of less than 1.0 m.y., olivine nephelinite magmas were erupted and this activity continued intermittently until relatively recent times, the younger eruptives being mainly basanitic in composition. The volume of volcanic products in this phase probably does not exceed 10 km³. Thus the volume of all the resurgent volcanism comprises less than 10 percent of the subaerially exposed part of Gran Canaria. The results show that the subaerial main shield building phase of volcanism in Gran Canaria, consisting of mildly alkali to transitional basalts, occurred over a time interval that was less than 0.5 m.y. Magmatic evolution on Gran Canaria appears to be similar to that found on other basaltic volcanoes in oceanic regions. Thus volcanoes in the Hawaiian, Marquesas and Society Islands all were built by basaltic lavas in similar short-lived episodes of volcanism. In some Hawaiian volcanoes, a resurgent phase of volcanism of strongly undersaturated basalts of small volume is recognized following a long hiatus, again similar to that found on Gran Canaria. The relatively large volume of silicic lavas erupted in Gran Canaria immediately following the main basaltic shield building phase is, however, not matched in the Pacific volcanoes mentioned.     

Probabilistic risk assessment of common booster biocides in surface waters of the harbours of Gran Canaria (Spain)

The presence of booster biocides in the aquatic environment has been associated with a risk to non-target species due to their proven toxicity. The aim of the present study was to determine the spatial and temporal distribution of common booster biocides in different harbours of the island of Gran Canaria (Spain) and evaluate, by means of a probabilistic risk assessment (PRA), the ecological risk posed by these compounds. With these objectives, a monitoring campaign was conducted between January 2008 and May 2009, collecting a total of 182 seawater samples. Four common booster biocides (TCMTB, diuron, Irgarol 1051 and dichlofluanid) were monitored. Diuron levels ranged between 2.3 and 203 ng/L and Irgarol 1051 between 2.4 and 146.5 ng/L. The ecological risk associated with these levels was always low, however, with probabilities of exceeding the 10th percentile of autotroph toxicity below 3.5%.     

Ignimbrite volcanism in North Wales

Many episodes of ignimbrite volcanism have occurred in North Wales. Ignimbrites can be recognised amongst the Pre-Cambrian rocks and from almost every stage of the Ordovician succession. The best known and most instructively exposed ignimbrite volcanies are those of Caradocian age in Snowdonia. In this deeply incised mountainous area it is possible to demonstrate the major characteristics of ignimbrite volcanism and to examine the relationship between numerous intrusive rhyolite masses and the extrusive rocks. Both acid and basic magmas were available in North Wales during Ordovician times, and at several of the volcanic centres the rocks show a differentiation sequence from pyroxeneandesite to alkali-rhyolite. The emergence and growth of volcanic islands upon which the ignimbrites were deposited is revealed in the stratigraphical record. An intimate relationship exists between magmatism and crustal unrest, and it is possible to discuss certain problems regarding the petrogenesis of the rocks, the location and character of the volcanic vents and the palaeogeography of North Wales during the Ordovician. Definition of the terms employed and criteria used in the identification of Welsh ignimbrites is given, and the field relationships and petrology of a number of areas are described in detail.     

Origin and diagenesis of the Roque Nublo breccia,Gran Canaria (Canary Islands) — Petrology of roque nublo volcanics,II

The Pliocene Roque Nublo Series, the second of three major magmatic series on Gran Canaria (Canary Islands), consists of a lower sequence (200 m) of alkalic lavas (basanite to phonolite) and a thicker upper section (600 m) of interlayered lava and widespread breccia sheets: encompassing pyroclastic flow deposits, lahars and reworked epiclastic rocks. Components in the poorly sorted block — and ash — flow deposits are (unwelded) pumice, rock fragments, crystals, glass shards and, locally, bread-crust bombs. Some flow units are graded with fine-grained basal zones and lithic-rich lower and pumice-rich upper parts. Some have strongly grooved the underlying rocks, directions of these striations being independent of preexisting topography and are constant in direction for more than 5 km. The flows are thought to have been emplaced below minimum welding temperatures by collapse of eruption columns. They are similar in many respects to coarse-grained pyroclastic flow deposits found in andesite volcanoes. Glass of tephritic to phonolitic composition of clasts of the breccias is generally altered to «palagonite» and is partly replaced by clay minerals and zeolites (mainly chabazite and phillipsite). Palagonitization was a low temperature diagenetic process, resulting in the hydration of glass accompanied and followed by precipitation of zeolites and clay minerals. Electron-microprobe data suggest the following decreasing order of mobility of selected elements during palagonitization: Na, K, Al, Si, Ca, Mg, and Fe; Ti was assumed to be inert.     

L’Ignimbrite et la Caldera de Batur (Bali,Indonesie)

Bali island may be considered as consisting of two distinct units: the western part, essentially made by a ridge of volcanic breccias, and the eastern part, made by a series of active volcanoes,viz. the strato-volcano Agung and two large calderas, Tjatur and Batur. The latter results from the collapse of a strato-volcano following the outpouring of an ignimbritic unit (ash flow) covering the northern and the southern flanks of Batur. This event occurred some 22,000 years ago. The whole island is tilted northwards around its long axis; this tilt is made obvious 1°) by the different slopes of the ignimbritic flow on the northern and southern sides of the volcano, 2°) by the island morphology and 3°) by the altitude — several hundred meters above sea level — where are found presently on the southern foot of Agung volcano recent submarine basalts (with pillow lavas and hyaloclastites). The tilt of Bali could be due to the presence of a major normal fault (or a system of faults), which itself explains the generation of the ignimbritic magma and, accordingly, the collapse of the caldera. The outflow of the ignimbrite indeed followed a long period of andesitic activity; it has been preceded — and followed — by flows of bandaite, a leucocratic lava with highly basic plagioclase (about 80 to 90% An); according to the authors, this kind of lava was generated, at shallow dephts by the assimilation of aluminous strata by a basaltic magma. Such an assimilation, as well as the anatexy generating the ignimbritic magma, is made possible thanks to important amounts of heat; most probably this heat has been supplied by basaltic magma rising in the crust through large normal faults, the magnitude and dip of which are shown by the northward tilting of the island.     

Anisotropy of magnetic susceptibility studies of depositional processes in the Campanian Ignimbrite,Italy

The late Pleistocene trachytic Campanian Ignimbrite underlies much of the Campanian Plain near Naples, Italy, and occurs in valleys in the mountainous area surrounding the plain out to about 80 km from its source, the Campi Flegrei caldera. At sites within 15 km of the Campi Flegrei, anisotropy of magnetic susceptibility (AMS) principal directions indicate that, in the absence of significant topography, deposition came from a flow moving in a roughly radial direction. AMS studies of the more distal ignimbrite reveal downhill and/or downvalley flow directions prior to deposition, even where these directions are at high angles to a generally radial transport direction from the vent. On the flanks of Roccamonfina Volcano, flow was directly downhill, as if the source of the ignimbrite was the summit of the volcano. In most localities, the ignimbrite consists of a single massive deposit. In a few localities in the Apennine Mountains, however, the confluence of multiple drainage systems off mountains resulted in multiple local flow units that cannot be correlated between valleys. A detailed study of the ignimbrite in the flat Titerno River valley near Massa shows that the AMS fabrics are not due to late-stage creeping during deposition or compaction. Well-defined, but non-parallel AMS fabrics from vertical and lateral sections in the Massa area are best explained by the merging of gravity currents flowing down the valley and steep valley sides to form a single aggradational deposit. Clast compositions and AMS axes at Mondragone indicate that the pyroclastic flow encountered the Monte Massico massif and was partially blocked, so that flow during deposition was toward the Campi Flegrei. Similar AMS data from sites along the edge of the Campanian Plain indicate back-flow off the first ridge of the Apennine Mountains reached at least 5 km from their base. The Campanian Ignimbrite was deposited from a density-stratified pyroclastic flow. The depositional system consisted of the lower, denser portion of the current, and was controlled by topography. The grouping of the AMS axes is interpreted to indicate that deposition occurred under laminar flow conditions.     

Timing of magma extraction during the Campanian Ignimbrite eruption (Campi Flegrei Caldera)

A core drilled within the northern part of the city of Napoli has offered the unique opportunity to observe in one single sequence the superposition of the four pyroclastic flow units emplaced during the Campanian Ignimbrite (CI) eruption. Such a stratigraphic succession has never been encountered before in natural or in man made exposures. Therefore the CI sequence was reconstructed only on the basis of stratigraphic correlations and compositional data (in literature). The occurrence of four superposed CI flows, together with all the data available (in literature) allowed us to better constrain the chemical stratigraphy of the deposit and the compositional structure of the CI magma chamber. The CI magma chamber includes two cogenetic magma layers, separated by a compositional gap. The upper magma layer was contaminated by interaction with radiogenic fluids. The two magma layers were extruded either individually or simultaneously during the course of the eruption. In the latter case they produced a hybrid magma. But no evidence of input of new geochemically and isotopically distinct magma batches just prior or during the eruption has been found. Comparison with the exposed CI deposits has permitted reconstruction of variable eruption phases and related magma withdrawal and caldera collapse episodes. The eruption was likely to have began with phreatomagmatic explosions followed by the formation of a sustained plinian eruption column fed by the simultaneous extraction from both magma layers. Towards the end of this phase the upward migration of the fragmentation surface and the decrease in magma eruption rate and/or activation of fractures formed an unstable pulsating column that was fed only by the most-evolved magma layer. This plinian phase was followed by the collapse of the eruption column and the beginning of caldera formation. At this stage expanded pyroclastic flows fed by the upper magma layer in the chamber generated. During the following major caldera collapse episode, the maximum mass discharge rate was reached and both magma layers were tapped, generating expanded pyroclastic flows. Towards the end of the eruption, only the deeper and less differentiated magma layer was tapped producing more concentrated pyroclastic flows that traveled short distances.     

Ignimbrite morphology and the effects of erosion: a New Zealand case study

Ignimbrite morphology, previously generalised using aspect ratios, is here quantified as the relationships between the various thicknesses of material forming an ignimbrite and the areas and volumes represented by those thicknesses. The morphology can be measured for the deposit in its present-day, eroded condition, or reconstructed for the original deposit. The reconstructed morphology of the 22 500 year BP, ca. 11 500 km², ca. 300 km³ Oruanui ignimbrite in New Zealand is documented to illustrate the latter approach. The Oruanui ignimbrite is an intermediate aspect ratio deposit and shows broadly linear relationships between (1) In thickness and the cumulative area occupied by that thickness or less of material and (2) thickness and the volume represented by that thickness or less of material. Two theoretical morphologies, one where thicknesses exponentially decay with distance from a maximum and the other of uniform thickness (slab), are compared with the Oruanui data. Limited comparative data suggest that low aspect ratio (violently emplaced) ignimbrites will show upward-concave curves (at one extreme following the exponential decay model) and high aspect ratio (gently emplaced) ignimbrites downward-concave curves (with the slab model as an extreme) when plotted on diagrams where the Oruanui data show linear trends. The effects of erosion on Oruanui and model ignimbrite morphologies are modelled using two theoretical erosion scenarios: (1) material is evenly removed from the land surface, and (2) thinner, non-welded material is preferentially removed. For the Oruanui ignimbrite data, area is lost much more rapidly in the first instance than volume; for example, 5 m of erosion is sufficient to remove 50 area %, whereas 40 m (scenario 1) or 120 m (scenario 2) of erosion is required to remove 50 volume %. In old ignimbrites, volume estimates may be reasonably accurate even after strong erosion, provided the original thicknesses of ponded/landscape-forming material can be inferred, but estimates of original area and aspect ratio will be inaccurate. An envelope enclosing all known outcrops of an ignimbrite will give a better estimate of original area and aspect ratio than simply summing the areas of known outcrops.     

Testing the ABOx-SC method: Dating known-age charcoals associated with the Campanian Ignimbrite

Over the past decade several studies have shown the improvements to radiocarbon chronologies that arise when Acid Base Oxidation-Stepped Combustion (ABOx-SC, Bird et al., 1999) pretreatment methods are applied to the dating of charcoal thought to be >30 ka BP. However, few studies have examined whether the use of ABOx-SC produces dates that are not only older, but accurate on known-age charcoal samples that could not be decontaminated using the routine Acid–Base–Acid (ABA) pretreatment protocol. In this study we date 9 charcoal fragments found below the Campanian Ignimbrite (CI) tephra layer, dated by ⁴⁰Ar/³⁹Ar to 39,230 ± 45 years (De Vivo et al., 2001, Rolandi et al., 2003), from three Palaeolithic sites. When treated with the ABOx-SC pretreatment protocol, the radiocarbon dates provide an accurate terminus post quem for the CI. In contrast, the ABA protocol consistently underestimates the age of the tephra. These results serve as a warning against the use of consistency as an indicator for reliability, demonstrate that the routine ABA method is not sufficient to decontaminate charcoal samples from sites of Palaeolithic age, and show that ABOx-SC produces not only older, but accurate age estimates.     

Dunes of the Gran Desierto Sand-Sea,Sonora, Mexico

The Gran Desierto Sand-Sea contains dunes of crescentic and star form in simple, compound, and complex varieties. The dunes have developed in bimodal to complex wind regimes of intermediate energy. Transitions from simple through compound to complex crescentic dunes are associated with regional changes in wind regimes. Growth of large star dunes takes place by merging of smaller crescentic and reversing dunes from southerly directions and reworking of sand by northerly and westerly winds. Although wind regimes appear to be the major control of dune morphology in this sand-sea, there is a close correlation between the spacing of simple crescentic dunes and the grain size of the coarse 20th percentile.     

Interaction between caldera collapse and eruptive dynamics during the Campanian Ignimbrite eruption,Phlegraean Fields,Italy

The Campanian Ignimbrite (36000 years B.P.) was produced by the explosive eruption of at least 80 km³ DRE of trachytic ash and pumice which covered most of the southern Italian peninsula and the eastern Mediterranean region. The eruption has been related to the 12-x15-km-diameter caldera located in the Phlegraean Fields, west of Naples. Proximal deposits on the periphery of the Phlegraean Fields comprise the following pyroclastic sequence from base to top: densely welded ignimbrite and lithic-rich breccias (unit A); sintered ignimbrite, low-grade ignimbrite and lithic-rich breccia (unit B); lithic-rich breccia and spatter agglutinate (unit C); and low-grade ignimbrite (unit D). Stratigraphic and componentry data, as well as distribution of accidental lithic types and the composition of pumice clasts of different units, indicate that coarse, lithic-rich breccias were emplaced at different stages during the eruption. Lower breccias are associated with fines-rich ignimbrites and are interpreted as co-ignimbrite lag breccia deposits. The main breccia unit (C) does not grade into a fines-rich ignimbrite, and therefore is interpreted as formed from a distinct lithic-rich flow. Units A and B exhibit a similar pattern of accidental lithic types, indicating that they were erupted from the same area, probably in the E of the caldera. Units C and D display a distinct pattern of lithics indicating expulsion from vent(s) that cut different areas. We suggest that unit C was ejected from several vents during the main stage of caldera collapse. Field relationships between spatter agglutinate and the breccia support the possibility that these deposits were erupted contemporaneously from vents with different eruptive style. The breccia may have resulted from a combination of magmatic and hydrothermal explosive activity that accompanied extensive fracturing and subsidence of the magma-chamber roof. The spatter rags probably derived from sustained and vigorous pyroclastic fountains. We propose that the association lithic-rich breccia and spatter agglutinate records the occurrence of catastrophic piecemeal collapse.