Hot clasts and cold blasts: thermal heterogeneity in...

Hot clasts and cold blasts: thermal heterogeneity in boiling-over

pyroclastic density currents

ERIKA RADER1*, DENNIS GEIST1, JOHN GEISSMAN2,3, JOE DUFEK4 & KAREN HARPP5

1Department of Geological Sciences, University of Idaho, 825 West 7th Street, 322 Mines

Building, PO Box 443022, Moscow, ID 83844-3022, USA2Department of Earth and Planetary Sciences, MSC 03 2040, 1 University of New

Mexico, Albuquerque, NM 87131-0001, USA3Present address: Department of Geosciences, ROC 21, The University of Texas

at Dallas, 800 West Campbell Road, Richardson, TX 75080, USA4School of Earth and Atmospheric Sciences, Georgia Institute of Technology,

311 Ferst Drive Atlanta, GA 30332, USA5Department of Geology, Colgate University, 13 Oak Drive, Hamilton, NY 13346, USA

*Corresponding author (e-mail: [email protected])

Abstract: Partial thermal remanent magnetization data from clasts in pyroclastic density current(PDC) deposits provide information on the emplacement temperatures of both lithic and juvenilemagmatic clasts contained in the deposits. We collected palaeomagnetic data from clasts in PDCdeposits emplaced during historical eruptions of two volcanoes in Ecuador, the 2006 eruption atTungurahua and the 1877 eruption at Cotopaxi. These eruptions were characterized by emplace-ment of PDCs mainly related to boiling-over activity. The deposits of these eruptions aresimilar and are characterized by cauliflower-textured juvenile scoria clasts up to 1 m in diameterand a diverse assemblage of lithic clasts surrounded by an unwelded ashy matrix. On the basis ofprogressive thermal demagnetization experiments, we infer that emplacement temperatures formost of the lithic clasts in PDC deposits are below 90 8C. In contrast, palaeomagnetic data fromjuvenile clasts from the same deposits provide emplacement temperatures higher than 540 8C.These data indicate the PDC were thermally heterogeneous over short length scales (decimetres)also after deposition. We hypothesize that PDCs emplaced by the boiling-over mechanism coolquickly owing to atmosphere entrainment, causing the juvenile clasts to form a rind that retainsheat and that also prevents lithic clasts from appreciable heating. Several deposits on Cotopaxi,despite being morphologically similar to the PDC deposits, contain both cold lithic and juvenileclasts, which we interpret to be lahar deposits formed by PDCs travelling across glacial ice andsnow. Rare deposits containing both hot lithic and hot juvenile clasts are classified as well-mixed, hot PDCs, and were erupted during a more energetic phase at Tungurahua.

Eruptions in 2006 at Tungurahua volcano and in1877 at Cotopaxi volcano, northern Ecuador, pro-duced similar deposits, which resulted from similareruption styles that have been described as ‘boilingover’ (Wolf 1878; Hall & Mothes 2008; Kelfounet al. 2009; Pistolesi et al. 2011). ‘Boiling over’occurs when strongly vesiculated magma fountainsor froths over a crater rim without forming a con-vective plume (or the plume is a small fraction ofthe eruptive output). The process is a poorly under-stood phenomenon, despite being widely cited toexplain thermally differing eruption products (e.g.Carrasco-Nunez & Rose 1995; Freundt 1999; Fulop2002; Soler et al. 2007).

The 1877 Cotopaxi and 2006 Tungurahuaboiling-over eruptions produced a characteristic

type of pyroclastic flow deposit, which is unusu-ally rich in giant scoria bombs. These deposits areparadoxical because charred wood and fluidal-textured, extremely fragile juvenile clasts suggesthot emplacement temperatures, whereas the pres-ence of uncharred wood and unmelted plastic indi-cates low temperatures. In this study we haveestimated the deposition temperatures of lithic andjuvenile clasts within these pyroclastic densitycurrent (PDC) deposits using the common palaeo-magnetic method of progressive thermal demagneti-zation (Paterson et al. 2010 and references therein)to examine the possibility that a pre-existing rema-nence in the lithic and juvenile clasts has been par-tially to completely thermally unblocked after thedeposits came to rest at elevated temperatures, and

From: Ort, M. H., Porreca, M. & Geissman, J. W. (eds) The Use of Palaeomagnetism and RockMagnetism to Understand Volcanic Processes. Geological Society, London, Special Publications, 396,http://dx.doi.org/10.1144/SP396.16# The Geological Society of London 2015. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics

by guest on January 9, 2015http://sp.lyellcollection.org/Downloaded from

http://sp.lyellcollection.org/

a subsequent, partial or complete thermorema-nent magnetization has been blocked in cooling.Because the PDCs have strongly impacted the com-munities around Tungurahua and Cotopaxi, a betterunderstanding of the eruption styles and emplace-ment mechanisms will advance efforts to minimizedamage in future eruptions.

Background

Location and eruptive history

Cotopaxi and Tungurahua are stratocones locatedin the Andes of north-central Ecuador (Fig. 1). Coto-paxi is 50 km south of the capital city of Quito, andTungurahua is 90 km further south. Both Cotopaxiand Tungurahua have erupted PDCs generated bya boiling-over style. The deposits emplaced by this

eruptive activity are characterized by prevalenceof scoria-bomb-rich PDC deposits at both volca-noes (Wolf 1878; Hall & Mothes 2008; Kelfounet al. 2009). Products from both volcanoes rangefrom mafic andesite to dacite and rhyolite (Hall &Mothes 2008; Samaniego et al. 2011) but the mostrecent eruptions are almost entirely andesitic.

Between 1600 and 2000 CE, Tungurahua experi-enced 15 eruptions. Eruptions were moderatelyexplosive, typically VEI (Volcanic ExplosivityIndex) 2 to 4, and produced lava flows, tephra andPDCs. The flanks of the volcano are populated withfarms and villages, and the town of Banos (currentpopulation 9500: Fig. 2a) is approximately 8 kmaway from the summit vent. Several larger erup-tions damaged property and caused fatalities,including the eruptions of 1886, 1916 and therecent cycle, which began in 1999. From 1999 to

Fig. 1. Location map of Tungurahua and Cotopaxi in Ecuador. The capital city of Quito is indicated by a star.Volcanoes represented by triangles from north to south are: 1, Cayambe; 2, Reventador; 3, Guagua Pichincha;4, Chacana; 5, Antisana; 6, Sumaco; 7, Cotopaxi; 8, Quilotoa; 9, Chimborazo; 10, Tungurahua; 11, Sangay.

E. RADER ET AL.



Fig. 2. Map showing the extent of pyroclastic flows and sample locations on (a) Tungurahua and the lahar extent on(b) Cotopaxi. Matching symbols represent samples collected from a single flow. The city of Banos is indicated inpink in (a).

HOT CLASTS, COLD BLASTS



2006, the activity at Tungurahua was mostlyStrombolian but, in 2006, it became more explosive(Steffke et al. 2010). Steffke et al. (2010) usedthermal infrared cameras to demonstrate that largeamounts of air were entrained into the PDCs andthe plume during this last eruption. The peak ofactivity occurred in July and August 2006, whichproduced gas emissions, explosions, lahars, lavaflows, tephra and PDCs (Samaniego et al. 2011;Douillet et al. 2013a, b). We focused on the 2006PDC deposits in this study. Tungurahua’s activitycontinues to the pres-ent day.

Twenty-four eruptions have occurred at Coto-paxi since the early 1700s, the most noteworthybeing in 1768 (Wolf 1878; Pistolesi et al. 2011),which produced a lahar that travelled over 320 kmin 18 h (Mothes et al. 1998). The eruption in 1877produced PDCs that melted much of the glacialicecap and produced large-volume lahars (Fig. 2b).Tephra fell for several hours and the final gaspof the eruption resulted in several thick andesitelava flows, which only travelled a few hundredmetres from the summit (Pistolesi et al. 2011). Wefocused on the 1877 deposits at Cotopaxi.

Pyroclastic density currents from

boiling-over eruptions

Pyroclastic density currents (PDCs) are producedby lateral blasts, dome collapse, column collapseand pyroclastic fountaining (Branney & Kokelaar2002). Most commonly, strong pyroclastic erup-tions generate columns that lose their initial buoy-ancy and collapse to form PDCs. Low fountainslack convective columns and begin to flow asdensity currents as soon as they breach the crater

rim (Fig. 3). These are referred to as ‘boiling-over’eruptions, a term coined from the 1877 eruption atCotopaxi (Wolf 1878).

Boiling-over eruptions have been witnessedand described at several volcanoes (e.g. Wolf 1878;Taylor 1958; Hoblitt 1986; Clarke et al. 2002; Sheaet al. 2011; Hall et al. 2013), and inferred by inspec-tion of deposits from unwitnessed eruptions (e.g.Carrasco-Nunez & Rose 1995; Fulop 2002; Soleret al. 2007). Numerical models have simulatedPDCs sourced by boiling-over fountains (Freundt1999; Clarke et al. 2002; Dufek & Bergantz 2007).Conditions in the models that result in boiling-overeruptions are high volume flux coupled with lowexit velocity (Dufek & Bergantz 2007) and a lowvolatile content (Clarke et al. 2002). As pre-eruptivewater content in melt inclusions from 2006 Tun-gurahua juvenile clasts indicate volatile concentra-tions of up to 4% (Myers 2012), the low viscosityof mafic andesite may play a greater role than vola-tile contents in producing a boiling-over fountain asopposed to a buoyant column.

Pyroclastic density currents produced by boiling-over eruptions have short run-out distances due totheir low initial velocity, and they tend to producethick, poorly sorted deposits (Dufek & Bergantz2007). Several studies interpret PDC deposits thatlack contemporaneous large-volume fallout depos-its as resulting from a boiling-over mechanism,exemplified by Chitelaltepetl volcano, Mexico(Carrasco-Nunez & Rose 1995; Branney & Koke-laar 2002). Some workers proposed that boiling-over PDCs entrain less of the ambient atmospherethan other PDC types, resulting in little cooling ofthe flow prior to deposition (Freundt 1999; Cioniet al. 2004). ‘Boiling over’ has also been invoked

Fig. 3. Tungurahua emitting a boiling-over pyroclastic density current on 4 December 2010, as indicated by the arrow.PDCS travelled down the north and west drainages that day. Image retrieved from http://i2.cdn.turner.com/cnn/2011/images/04/26/t1larg.tungurahua.volcano.afp.gi.jpg accessed on 3 March 2014.

E. RADER ET AL.


http://i2.cdn.turner.com/cnn/2011/images/04/26/t1larg.tungurahua.volcano.afp.gi.jpg




for silicic eruptions on the basis of interpretations ofrelatively high emplacement temperatures (Lestiet al. 2011), which are attributed to the lack ofatmospheric entrainment during flow. These depos-its include welded units in the Gutai Mountains ofRomania, Vilama in the Central Andes, Grey’sLanding in the USA, the Garth tuff in Wales, the79 AD eruption at Vesuvius in Italy and the CerroGalan ignimbrite of NW Argentina (McArthuret al. 1998; Fulop 2002; Cioni et al. 2004; Soleret al. 2007; Andrews & Branney 2011; Cas et al.2011; Lesti et al. 2011). We question the inferencethat ‘boiling over’ results in lower entrainment ofair and less cooling during transport and deposi-tion of PDCs, interpretations that are inconsistentwith abundant evidence for substantial entrainmentin collapsing fountains and jets (Suzuki et al. 2005;Hall et al. 2013 and references therein), as well asour observations of the physical characteristics ofthe Cotopaxi and Tungurahua deposits.

Estimation of temperature of volcanic

deposits using palaeomagnetic methods

The extent of carbonization of wood (Sawada et al.2000; Scott & Glasspool 2005), observations madeduring eruption (Fujinawa et al. 2008), thermalimaging and palaeomagnetic data in the form of pro-gressive thermal demagnetization results from lithicclasts embedded within the PDC deposits (e.g.Sawada et al. 2000; Cioni et al. 2004), have beenused to estimate emplacement temperatures. Temp-eratures of emplacement for PDC deposits rangefrom around 48 8C (Sparks et al. 2002) to approxi-mately 700 8C (Nogami et al. 2001), and estimatesof temperature can vary by 200–300 8C within dif-ferent parts of a single deposit (e.g. Zlotnicki et al.1984; Bardot 2000; Paterson et al. 2010). In prin-ciple, progressive thermal demagnetization dataprovide a robust method to determine the emplace-ment temperature of individual clasts, allowingfor a detailed perspective of the thermal conditionsduring deposition of a PDC.

Heating of (accidental) lithic clasts incorpo-rated within a PDC thermally resets the magnetiza-tion by partial to complete unblocking of a previousremanence, usually a thermoremanent magneti-zation (TRM) if the clast is of volcanic origin.Heating takes place from the outside of the lithicfragment to its core, whereas the cooling of a juven-ile (magmatic) clast blocks the TRM and the frag-ment cools from the rim inwards. As a clast cools,magnetization blocking will result in a total orpartial TRM aligned with the Earth’s magnetic fieldat the time of cooling. The total TRM is acquired asa single magnetization component if the fragmentremains stationary during complete magnetiza-tion blocking from elevated temperatures. If the

fragment changes its orientation during magnetiza-tion blocking, then the total TRM acquired is a com-posite of magnetizations of different directions. Inthis case, the orientation of the last componentof magnetization acquired will be parallel to theEarth’s magnetic field at the moment of deposition.Each component of magnetization acquired at a dis-crete temperature interval will be unblocked ifheated above about that same temperature, assum-ing that heating and cooling are both rapid (McClel-land & Druitt 1989; Paterson et al. 2010). It ispredicted that if a clast was deposited above theCurie point (maximum blocking temperature) orwas never reheated in the deposit, it will have asingle component of magnetization, as revealed byunidirectional decay of the remanence. In the caseof no reheating, the direction of the magnetizationshould be random among different clasts with thesame thermal history. If a reheated clast attains atemperature below the Curie point after it has beendeposited, or if it is transported as it cools, then aclast can subsequently block multiple componentsof magnetization, as partial thermoremanent mag-netizations (pTRMs). Progressive thermal demag-netization, in principle, can isolate componentsacquired as a result of thermal resetting and thusprovide estimates of the emplacement temperatureof the clasts within a deposit.

Typically, the emplacement temperature of aclast is interpreted as the maximum laboratoryunblocking temperature (Tub) of the remanencethat is well grouped and is aligned with the geo-magnetic field at the time of cooling, and acquired(blocked) after the deposit came to rest (e.g.Clement et al. 1993; Bardot 2000; Paterson et al.2010). Alternatively, another group of researchersrecommends excluding extreme values of highand low laboratory unblocking temperatures (Tub)to determine the predominant temperature of thedeposit (Cioni et al. 2004; Zanella et al. 2007,2008; Di Vito et al. 2009). For this study, ‘low’ or‘cold’ emplacement temperatures are defined asthose that range from 50 to about 210 8C, while‘hot’ emplacement temperatures are greater than500 8C.

Methods

Sample collection and preparation

Clasts were collected from two PDC depositsemplaced during the August 2006 Tungurahua erup-tion and from three valley-filling PDC deposits fromthe 1877 eruption at Cotopaxi (Fig. 2). We collected18 orientated samples from nine different locationsat Tungurahua (one lithic clast and one juvenileclast at each site). An additional 18 orientatedclasts were collected (one lithic clast and one




juvenile clast) from each of nine locations at Coto-paxi. Between two and three sections from rimto core were cut from a single sample, making atotal of 84 specimens from these clast samplesthat were subjected to progressive thermal demag-netization. The location of the investigated PDCdeposits, sample type and number of sections canbe found in Table 1 and the locations can be foundin Figure 2.

Juvenile clasts from these scoria-rich PDCdeposits have a highly vesicular, cauliflower-like

morphology, and are fragile and polygonally jointed.The clasts have a breadcrust texture and manydroop over irregularities in the surface on whichthey were deposited (Fig. 4). Most of the lithicclasts are dense and have a weathered or alteredexterior. We cut sections from rim to core of eachjuvenile clast and prepared each section into6.25 cm3 cubes, while maintaining their orientation,using a non-magnetic diamond saw blade. Lithicclasts were cored using a 2.5 cm-diameter, water-cooled non-magnetic diamond bit with a drill press.

Table 1. Estimated minimum emplacement temperatures

Sample No. Clast type No. ofsectionsanalysed

Emplacementtemperature

(8C)

Flow type

Tungurahua

Tu-10-01 Lithic 2 210 Warm PDCJuvenile 3 500

Tu-10-02 Lithic 2 540 Hot PDCJuvenile 3 530

Tu-10-03 Lithic 2 ,90 Cold PDCJuvenile 4 575




Tu-10-07 Lithic 2 ,90 Cold PDCJuvenile 2 Drifts

Tu-10-08 Lithic 2 ,90 Cold PDCJuvenile 3 .590


Cotopaxi

Co-10-01 Lithic 2 ,90 Cold PDCJuvenile 3 590


Co-10-03 Lithic 3 ,90 LaharJuvenile 2 ,90





Co-10-08 Lithic 2 ,90 LaharJuvenile 2 ,90


A single sample number consists of one lithic clast and one juvenile clast. Each clast was sectioned fromrim to core into two–four cubes that were progressively demagnetized. The emplacement temperatureswere estimated by inspection of progressive demagnetization diagrams and assessment of the directionalstability of magnetization components during progressive heating steps.

E. RADER ET AL.



Magnetization measurements

The monitoring of bulk susceptibility as a func-tion of heating and cooling to temperatures aboveabout 640 8C was conducted on an Advanced Geo-sciences Instrument Company (AGICO) MFK1-FAsusceptibility instrument interfaced with an AGICOCS4 furnace apparatus at the University of Texas atDallas. In order to identify the main magnetic car-riers and detect alteration in magnetic minerals,heating and cooling curves were obtained whilethe specimen, either crushed whole rock or a mag-netic separate, was immersed in argon.

Rim and core specimens from the 36 orientatedclast samples were progressively thermally demag-netized in the Paleomagnetism Laboratory at theUniversity of New Mexico. Demagnetizations werecarried out in air with an average of 11 steps in amagnetically shielded furnace (ASC TD 48) up totemperatures of about 590 8C. The remanence wasmeasured for each sample after every heating stepusing standard procedures on a 2G EnterprisesModel 760R three-axis DC SQUID-based super-conducting rock magnetometer in a magneticallyshielded environment constructed by Lodestar Mag-netics. Demagnetization data were inspected onorthogonal demagnetization diagrams and equal-area projections, and evaluated using PaleoMag v3.1b2 (Jones 2002). The remanence unblocked atthe lowest temperature steps (25 and 90 8C) was

excluded from analysis to avoid possible inclusionof viscous remanent magnetization (VRM: Bardot& McClelland 2000). Because of the relativelylow variation in Tub values estimated in this study,we do not exclude any measured values based ondeviation from the mean. For reasons discussedbelow, we prefer not to assign an emplacement tem-perature to each entire deposit, and only refer toemplacement temperature estimates with regardsto individual clasts.

Results

Description of deposits

Most of the PDC deposits emplaced during the 2006eruption at Tungurahua volcano and the 1877 erup-tion at Cotopaxi volcano were deposited in largefans at slope breaks at the mouths of deep valleys(Fig. 5). The PDC deposits at both volcanoesexhibit lateral facies variations. The PDC depositsare 1–10 m thick and exhibit reverse grading ofthe coarse particles, with large cauliflower scoriabombs sitting on top of an ash-rich matrix. Detaileddescriptions of the deposits can be found in Douilletet al. (2013a, b) and Hall et al. (2013).

Granulometric analyses performed at 16 sites invalley-filling deposits at Tungurahua reveal a fines-skewed deposit with a median grain size of 2.5 mm.

Fig. 4. Field photographs showing juvenile bombs at both Cotopaxi and Tungurahua sagging under their ownweight (a), drooping over other objects (b) and a cauliflower texture (c) that forms from the expansion of the hotinner part of the bomb while the cool outer crust cracks. These features all indicate a hot emplacement temperaturefor the bombs.




A second facies comprises fine ash, typically,1 mm diameter, which was deposited as a veneerhigh on the walls of the canyons. The smaller grainsin the deposit are dominantly juvenile ash withlithic clasts, typically having a grain size .8 mm.At both volcanoes, lithic clasts vary widely in shapeand composition, with some having edges that aremore angular and others being well rounded. Lithicclasts range from fragments of old, altered andesiticlava to dense fresh andesite that was probablyeroded from the conduit.

Cotopaxi PDC deposits from the 1877 eruptionare similar to Tungurahua eruption deposits, withmetre-scale cauliflower scoria bombs resting onbeds of finer-grained lapilli and ash. They are chan-nelized and contain lithic clasts that consist of denseandesitic blocks. In addition to PDC deposits, Coto-paxi also has lahar deposits. The largest lahar,Chillos Valley Lahar, which occurred 4.5 kyr BP,flowed north and west of the summit, and left depos-its 35–40 m thick (Mothes et al. 1998). Lahardeposits from the 1877 eruption flow to the northon top of the Chillos Valley Lahar and are up to1 m thick (Pistolesi et al. 2013). The lahar depositsextend 6–12 km from the vent, and, similar to thePDC deposits, contain scoria bombs, lithic clastsand leveed channels of ash. The 1877 depositsat Cotopaxi are described in detail in Motheset al. (1998), Garrison et al. (2011) and Pistolesiet al. (2011).

Rock magnetism

Low-field magnetic susceptibility v. temperaturecurves were obtained from a total of 20 juvenileand lithic clasts. The curves have a reversible behav-iour in most of the cases, suggesting a negligiblealteration of the magnetic phases during heatingand cooling experiments. The inferred Curie temp-eratures range from 250 and 630 8C (Figs 6 & 7).Curie temperatures at Cotopaxi are around 200 8Chigher than at Tungurahua, indicating they have adifferent mineral assemblage. Lithic clasts at Coto-paxi have systematically higher Curie temperaturesthan juvenile clasts, which drop off steeply at about360 8C (Fig. 6). Lithic clasts collected from Tungur-ahua have the lowest and the highest Curie tempera-tures (up to 550 8C), while juvenile clasts haveCurie temperatures of 300–350 8C (Fig. 7). Thesetemperatures are well below the Neel temperatureof hematite (about 680 8C), suggesting that nogrowth of this phase occurred.

In some cases (e.g. TU-10-04 Juvenile (core andrim): Fig. 7) there is a clear mixture of phases withdifferent Curie temperatures. In most cases, theheating and cooling curves are close to identical, ifnot virtually reversible (e.g. CO-10-08 Juvenile(core); TU-10-05 Juvenile (core and rim)). Rimand core samples from the same clast have similarCurie temperatures except for the lithic clast fromCo-10-09, where the rim had a lower Curie

Fig. 5. Photograph of a PDC deposit at Cotopaxi. Pyroclastic flows are largely channelized in steep-sided quebradas(gullies) at both Tungurahua and Cotopaxi. Many of the preserved deposits are found at the base of the quebradasor where the PDCs jumped the steep walls. Cauliflower bombs can be seen dispersed along the bottom of thedrainage below Cotopaxi.

E. RADER ET AL.



temperature by about 100 8C (Fig. 6). Many of thesamples show a strong Hopkinson effect, in bothheating and cooling.

Palaeomagnetism of juvenile clasts

Most of the juvenile clasts from Cotopaxi and Tun-gurahua yield a well-defined response to thermaldemagnetization, with NRM intensity ranging from0.04 to 10 A m21. The magnetization of each sam-ple was fully unblocked after heating to between500 and 590 8C. Demagnetization trajectories aretypically well defined until magnetization unblock-ing of more than 99% of NRM has been reached,at which point essentially random directions aremeasured at each subsequent demagnetization step(Fig. 8). On the basis of demagnetization behav-iour, the likely principal magnetization carrier isa low-Ti titanomagnetite. There is no evidence ofan appreciable magnetization carried in secondaryhematite, which could result in a CRM that wouldgive rise to an overestimation of the emplacementtemperature, as found in some juvenile deposits(e.g. McClelland & Druitt 1989; Airoldi et al. 2012).

Analysis of the directional components in thejuvenile clast samples reveals two components. Thelow temperature component is defined as Tub , 2108C and typically has an intensity of around 1–4 A m21. The higher unblocking temperature com-ponent is typically unblocked over the temperaturerange from 210 to about 500 8C and is directionallyuniform in the juvenile clasts. The remanence typi-cally drops to very low values, ,1% of the NRM,at temperatures above 500 8C, close to completeunblocking. Progressive thermal demagnetizationdata indicate that the multiple specimens measuredfrom each clast have slight scattering in the low-temperature component, and a well-grouped, singledirectional component between 210 and 500 8C.Sample Tu-10-06 is an example of this behaviour(Fig. 8). It yields a consistent direction that is essen-tially parallel to the magnetic field at the time ofdeposition up to, in this case, about 590 8C.

The three exceptional juvenile clasts (Tu-10-07,Co-10-06 and Co-10-03: Table 1, Fig. 2) reveal asystematic change in direction with progressivedemagnetization (Fig. 9). The lower Tub com-ponents that are initially unblocked, however, do

Fig. 6. Plots of bulk, low field susceptibility (x) as a function of temperature for powders of specimens from bothjuvenile and lithic clasts from the Cotopaxi deposits. Heating curves (grey line) and cooling curves (black line) obtainedwith the specimen in an inert (Ar gas) atmosphere.




not have directions that are consistent with the geo-magnetic field at the time of emplacement and aredifficult to interpret. Instead, they yield magnetiza-tions that are .308 from the geomagnetic field at

the time of emplacement. For specimens from thecore of sample Tu-10-07, the first removed and thefirst magnetization component, which persists upto a temperature of about 500 8C, yields a direction

Fig. 7. Plots of bulk, low field susceptibility (x) as a function of temperature for powders of specimens from bothjuvenile and lithic clasts from the Tungurahua deposits. Heating curves (grey line) and cooling curves (black line)obtained with the specimen in an inert (Ar gas) atmosphere.

Fig. 8. Intensity decay, equi-areal projection and orthogonal demagnetization diagram (Zijderveld 1967) showing theresponse to progressive thermal demagnetization by a typical juvenile clast. In the orthogonal diagram, solid (open)symbols refer to projections on the horizontal (vertical) plane. Laboratory unblocking temperatures are distributed overthe interval 90–610 8C. Typical example of single component clast. At T ¼ 560 8C, the magnetization has almosttotally removed (less than 5% remains). The data after this temperature are scattered and not reliable. Grey squares arelower-hemisphere projections, and open squares are upper-hemisphere projections. The ellipses are the a95 confidenceinterval about the estimated mean direction.

E. RADER ET AL.



that differs from the geomagnetic field at the time ofemplacement by at least 308 (Fig. 10). At progress-ively higher Tub, the inclination remains constantbut the declination changes to about 3208. Speci-mens from the rim of this sample yield magnetiza-tions that exhibit a progressive shift in declinationfrom about 3208 at low unblocking temperaturesto approximately 1808 at more elevated tempera-tures (Fig. 10).

Like the other juvenile clasts, specimens fromclast samples Co-10-06 and Co-10-03 yield a sin-gle direction of magnetization over the entire Tub

spectra. Because the directions of the magnetizationresolved are consistent throughout the entire rangeof laboratory unblocking temperatures and similarto one another, the two samples probably cooledin a single location; this observation is consistentwith the interpretation that the clasts are juvenilematerial. However, unlike other juvenile clasts, the

directions resolved from these clasts differ by morethan 1008 from the geomagnetic field (Fig. 9). Onlytransportation of the clasts after cooling below thesetemperatures can result in this observation.

We compared the directions from clasts withsingle magnetization components with the present-day Earth’s magnetic field for Tungurahua in 2006.Declination of 21.58 and inclination of 20.38 arecalculated for this locality using the IGRF-11model (IGRF reference: http://wdc.kugi.kyoto-u.ac.jp/igrf/point/index.html). The magnetic field forCotopaxi in 1900 had a declination of 7.38 and aninclination of 16.08 (Figs 11 & 12). Maximum Tub

for specimens of juvenile clasts is about 590 8C,and the distribution of magnetizations isolated fromalmost all the samples from these clasts overlaps theambient magnetic field at the time of deposition(Fig. 11). The number of specimens accepted to pro-vide an estimated average declination includes all

Fig. 9. Equal-area projections showing orthogonal demagnetization diagrams showing two juvenile samples fromCotopaxi that do not align with NRM, indicating they were deposited after cooling in what is interpreted as a laharrunout from a PDC deposit. Symbols are the same as in Figure 8.

Tu-10-07 Juvenile Clast

0.7 A/m0.7 A/m

Rim

NRM

NRM

Core

Rim

590oC

590oC90oC

280oC N

W

S

E 610 o C

Core

Fig. 10. Equal-area projection and orthogonal diagrams showing the response by a juvenile sample from Tu-10-07at Tungurahua. The demagnetization data define a trajectory that suggests the single clast rotated in a single axisperpendicular to the slope as it cooled. This could be a manifestation of slow creep or slump in the deposit as the clastcooled. Symbols are the same as in Figure 8.



http://wdc.kugi.kyoto-u.ac.jp/igrf/point/index.html





the sections from each sample, including rim andcore, as the directions from each did not appreciablydiffer. Most of the single components are orientatedclose to the present-day (1900) geomagnetic fieldfor Cotopaxi (Fig. 11a) and present-day (2006) fieldfor Tungurahua (Fig. 11b). These data are interpre-ted to indicate that these magnetization componentswere acquired during cooling and after emplace-ment of deposits at temperature higher than 590 8C(i.e. the maximum blocking temperature: Table 1).

Palaeomagnetism of lithic clasts

Most of the lithic clasts exhibit progressive ther-mal demagnetization behaviour that is similar to

juvenile clasts. NRM intensities of the juvenileclasts range from 0.01 to 25 A m21. Most of thespecimens are completely thermally unblocked by575–600 8C. (Table 1). The single component ofmagnetization that is identified for most juvenileclasts is also observed in lithic samples. This com-ponent is well defined and typically unblocks overa temperature range from about 210 to 590 8C.The directions of this component are dispersedabout the ambient geomagnetic field directions forthe deposits. We interpret this component as thecharacteristic remanent magnetization (ChRM).

The directions of the ChRMs in lithic clastsare compared to the expected geomagnetic fieldsfor the Cotopaxi and Tunguarahua volcanoes. The

Fig. 12. Equal area projections of magnetizations of lithic clasts isolated in progressive thermal demagnetization of(a) Cotopaxi, and (b) Tunguraua. Lithic clasts show greater scatter than juvenile clasts and do not align with the moderngeomagnetic field. Most of these clasts are therefore interpreted to have been emplaced cold. Symbols are the sameas Figure 8.

Fig. 11. Equal-area projections showing that juvenile clasts have palaeomagnetic directions which closely match NRMeven at high temperatures in (a) Cotopaxi and (b) Tungurahua deposits. Symbols on the equal-area projection arethe same as in Figure 8.

E. RADER ET AL.



clasts from Cotopaxi yield relatively scattered direc-tions with an estimated mean direction of declina-tion (D) ¼ 14.48 and incliation (I) ¼ +33.88(n ¼ 20, a95 ¼ 16.48), with the ambient field direc-tion included in the 95% confidence limit (Fig. 12a).Magnetization directions from clasts from Tun-gurahua are more dispersed than those from Coto-paxi, as shown in Figure 12b. The fact that mostof the magnetization directions lie within the NEquadrant implies a non-random population, yet thegreater dispersion of these directions, in contrastto magnetizations isolated from juvenile clasts atboth high and low Tub intervals (Fig. 12), is nota-ble. Two lithic clasts from Tungurahua have highTub magnetizations that are similar in direction tothat of the ambient geomagnetic field. Specimensfrom sample Tu-10-01 yields single-componentmagnetizations of low Tub that cluster around thegeomagnetic field up to about 210 8C (Fig. 13).

Specimens from sample Tu-10-02, however,exhibit a single component of magnetization, witha broad Tub spectrum ranging from about 90 to575 8C, and the magnetizations isolated in thesespecimens cluster around the ambient geomagneticfield, similar to those from the juvenile samplesfrom this deposit (Fig. 10). Tu-10-02 is the onlyexample of a lithic clast that appears to have beensufficiently heated in situ to completely unblockits initial remanence.

Discussion

Emplacement temperatures of boiling-over

Tungurahua and Cotopaxi PDC deposits

Locally uncharred wood (Pollock et al. 2010),unmelted plastic (Hall et al. 2013) and undamagedvegetation were all found in parts of the TungurahuaPDC deposits after the 2006 eruption, indicating thatnot all parts of these PDCs were hot. At otherlocations, charred wood surrounding and withinthe 2006 PDC deposits indicates local temperaturesas high as 300 8C (Pollock et al. 2010).

Because the remanence preserved in an assem-blage of magnetic phases in volcanic rocks isblocked over a range of temperatures during cool-ing, palaeomagnetic data in the form of progres-sive thermal demagnetization measurements canbe used to estimate the emplacement temperatureof volcanic deposits (e.g. Aramaki & Akimoto1957; Mandeville et al. 1994; Bardot 2000). Theemplacement temperature can be accurately esti-mated in the case in which the clasts have a well-defined, first-removed component of magnetizationisolated over a range of relatively low laboratoryunblocking temperatures that is similar in direc-tion to the ambient expected geomagnetic fieldattending emplacement of the deposit. When theclast is characterized by single magnetic compo-nent, stable up to its maximum laboratory unblock-ing temperature, then this temperature can be usedas the minimum emplacement temperature forthat clast.

Other palaeomagnetic studies conducted onjuvenile materials with high laboratory unblockingtemperature magnetizations have concluded thata chemical remanent magnetization (CRM) canmask the thermoremanent magnetization (TRM),which typically is apparent by curvatures in demag-netization trajectories in progressive demagnetiza-tion diagrams (Zlotnicki et al. 1984; McClelland& Druitt 1989; Bardot et al. 1996; Bardot & McClel-land 2000) as a result of overlapping unblockingtemperature spectra of different components. Littlecurvature in behaviour is observed in our samplesand the data available do not support the presenceof appreciable hematite in these rocks. Conse-quently, we believe that the emplacement tempera-tures estimated from the scoria bombs are notaffected by CRM contamination. One possibleexplanation for the importance of CRM in sometemperature emplacement studies is that they havebeen mostly carried out on on silicic clasts, whichmight be more susceptible to secondary crystalliza-tion, whereas Cotopaxi and Tungurahua eruptedmafic andesite.

Fig. 13. Intensity decay, equal-area and orthogonal projections of a lithic clast from Tungurahua with a singlecomponent of magnetization stable up to 210 8C. This clast was heated at temperatures equal or higher than 210 8C.Symbols are the same as in Figure 8.




All of the specimens prepared from juvenileclasts at Tungurahua exhibit a single component ofmagnetization that is isolated up to 500–590 8C.The estimated mean direction derived from theserocks is similar in direction to the expected geo-magnetic field direction at the time of emplacement.These palaeomagnetic results are consistent withthose obtained by Roperch et al. (2014) from thesame PDC deposits. The authors, who reportedpalaeointensity data from these deposits, also con-cluded there was no CRM augmenting the emplace-ment temperature as the magnetic grains in thesamples were predominantly single-domain titano-magnetite with no hematite overgrowths.

For most of the lithic clasts at Tungurahua, thesingle magnetization component is considerablydispersed and thus not well aligned with the geo-magnetic field at the time of eruption. Thus, weinterpret these data to indicate that these clastsremained below about 90 8C as they were incorpor-ated, transported and deposited by the PDC. Thispalaeomagnetic signal is consistent for both thecores and the rims of the lithic clasts. Our interpret-ation of a high emplacement temperature for twoexceptional clasts assumes that they were heatedat source (e.g. possibly in the conduit) and retainedheat during transport (Hall et al. 2013). If correct,our interpretations imply large variations of ther-mal heterogeneity (.500 8C difference) betweenjuvenile and incorporated lithic clasts. This thermalheterogeneity is observed at the scale of a singleoutcrop in which hot juvenile scoria and coldlithic clasts coexist without efficient heat transfer.

It is important to note that the deposit was not hotenough to heat even the outermost part of the lithicclasts. We found that specimens prepared from theouter 2–3 cm rim of the clasts record the samepalaeomagnetic orientations as those specimensobtained from the core. At Cotopaxi, seven of thenine sampling sites include juvenile clasts thatwere emplaced above 590 8C, whereas all of thelithic clasts appear to have been emplaced below90 8C. This indicates that the thermal heterogeneityseen at Tungurahua was also present in the depositsat Cotopaxi.

Similar evidence for highly variable emplace-ment temperature of PDC deposits has been recog-nized in other studies. Previous studies of twodifferent PDC deposits have estimated emplace-ment temperatures that range from about 48 8C,measured by thermocouple (Sparks et al. 2002), toaround 700 8C, estimated by the mobility of chlor-ine and sulphur between the gas and ash of a PDC(Nogami et al. 2001). The highest emplacementtemperature estimated from progressive thermaldemagnetization data is approximately 680 8C(Lesti et al. 2011). Cooler temperatures are alsocommonly measured.

The lack of juvenile material accompanyingsteam eruptions and pyroclastic surges at theAdatara, Bandai and Zao volcanoes, in Japan, ledto the hypothesis that the deposits were all emplacedat temperatures of less than 100 8C (Fujinawa et al.2008). Other relatively low emplacement tempera-ture estimates for PDC deposits were measuredwith thermocouples, including deposits from MountEtna at about 200 8C (Belousov et al. 2011), Mont-serrat at 48 8C (Sparks et al. 2002), and Mount St.Helens between 50 and 250 8C (Banks & Hoblitt1996).

Palaeomagnetic studies of PDC deposits fromsilicic volcanic systems have also yielded relativelylow emplacement temperature estimates: ColliAlbani, .160 8C (Porreca et al. 2008); Ruapehu,,100 8C (Smith et al. 1999); Vesuvius, .180 8C(Cioni et al. 2004); Santorini, .160 8C (Bardot2000); and Taupo, .150 8C (McClelland et al.2004). In many of these examples, the presenceof uncharred organic matter is cited as evidenceindicating moderate emplacement temperaturesbecause temperatures must have been above 250–300 8C to char the wood (Sawada et al. 2000;Scott & Glasspool 2005; Scott et al. 2008). Physicalcharacteristics of low-temperature PDC depositsinclude unwelded, poorly sorted deposits (Fujinawaet al. 2008), uncharred vegetation (e.g. Sparks et al.2002; McClelland et al. 2004) and, in some, accre-tionary lapilli (e.g. Cioni et al. 2004; Porreca et al.2008; Belousov et al. 2011). In summary, a widerange of emplacement temperatures for PDC depos-its has been documented in volcanoes around theworld, and emplacement temperatures for clastsbetween 50 and 300 8C in these deposits are fre-quent. As far as we know, however, a broad rangeof emplacement temperatures has not been deter-mined for both lithic and juvenile clasts in a singledeposit, as is suggested by the results reportedhere. The major outcome of this study is that theCotopaxi and Tungurahua clasts show evidence ofboth hot and cold regions within the same PDC, atthe scale of the outcrop.

We propose that the matrix insulated the lithicclasts from the juvenile clasts, and the matrix musthave cooled during transport by air entrainment.We hypothesize that air entrainment and coolingduring transportation is especially important inboiling-over PDCs, which is consistent with obser-vations of relatively low flow velocities at Tungura-hua (Hall et al. 2013). This contradicts previousspeculations on boiling-over eruptions, whichassume that low fountaining produces little entrain-ment and promotes low cooling rates during flow(Freundt 1999). Limited entrainment of ambientair during the fountain phase may initially inhibitcooling of the PDC, but simulations by Clarkeet al. (2002) and Dufek & Bergantz (2007) show a

E. RADER ET AL.



greater degree of air entrainment during flow thanPDCs originating from a buoyant eruptive column.

The deposits at Tungurahua and Cotopaxi do notcontain significant amounts of pumice or densejuvenile clasts, and, therefore, cannot be classifiedas either an ignimbrite or a block-and-ash deposit(Branney & Kokelaar 2002). Instead, they are con-stituted of scoria bombs and large lithic clastsembedded in a coarse-ash and lapilli matrix, andare similar to deposits observed at arc volcanoessuch as Aso and Arenal (e.g. Mothes et al. 1998;Alvarado & Soto 2002; Cole et al. 2005; Miyabuchiet al. 2006). We interpret the palaeomagnetic datapresented here to indicate that the cauliflower-textured scoria bombs are associated with thermallyheterogenous gravity flows. The scoria bombs havea texture that is consistent with substantial airentrainment, namely a quenched outer rind, but anexpanded interior. We suggest scoria-rich PDCsare a common characteristic of thermally heteroge-nous boiling-over eruptions.

Model of thermal heterogeneity within

the deposit

The extent to which a lithic clast is heated afterbeing incorporated into a hot flow of pyroclasticmaterial consisting of gas, ash and clasts can bereadily estimated. We use the temperatures deter-mined from the progressive thermal demagn-etization of lithic clasts to model the minimumtemperature of the surrounding material at thetime of deposition. A numerical model of a spheri-cal clast with a radius of 10 cm being heated by

conduction from rim to core was devised. Themodel tests the time needed to heat a layer withinthe clast at a particular depth to a given temperature(Fig. 14).

The temperature of a spherical lithic clast that isbeing heated conductively by its surroundings isgoverned by a heat conductivity equation (Carslaw& Jaeger 1959):

T(r,t) =T0 − Te

2V(1,t) + Te

where

V(1,t) ; erf1+ 1

2p 1/2− erf

1− 1

2p 1/2

− 2t 1/2

1p 1/2e−

(1− 1)24t − e−

(1 + 1)24t

( )

where r is the distance from the centre of the clast, tis time, T0 is clast temperature at t ¼ 0, Te is thetemperature of the medium, 1 is the dimensionlessdistance from the centre of the clast ; (r/a), t isthe dimensionless Fourier number ; (kt/a2), a isthe radius of the clast and k is the thermal diffusiv-ity. We assume that the matrix of the deposit has auniform temperature (i.e. a constant temperatureboundary condition for the lithic clast’s surround-ings). This provides an extreme condition for heat-ing of the clasts (i.e. cooling of the matrix andremoval of heat by convection of gas would resultin cooler temperatures everywhere in space and

0 5 15140

180

200

240

Time (hours)

Tem

p (

oC

)

8 cm5 cm1 cmCore

10

Rim

Run at 230oC

220

160

Fig. 14. Example of a thermal model to constrain the temperature at different depths in a lithic clast deposited in amatrix with an initial temperature of 230 8C. The three depths in this simulation are illustrated by different colouredlines on the graph. The graph illustrates that a lithic clast must be in contact with an ashy matrix (at least .20 8C abovethe final temperature) for several hours to be reset.




time). This model has been used to calculate coolingof a PDC deposit by entrainment of cold lithic clasts(Marti et al. 1991), as well as to demonstrate that thelithic clasts at Santorini were already at elevatedtemperatures before incorporation into the PDC,rather than being reheated by the PDC (Bardot2000).

A clast achieves significant heating through thecore when t ¼ 1, which indicates a timescale ofbetween 7 and 11 h for 90% heat exchange in a10 cm-diameter clast with porosities of between10 and 60%, and k values of 4.0 × 1027–2.5 ×1027 m2 s21 (Riehle et al. 1995). As the clast temp-erature adjusts to the volumetrically dominantmatrix, the bulk deposit simultaneously coolsthrough heat loss to the ground and atmosphere.Although cooling of the matrix will result in adiminished extent of heating of the lithic clasts,the temperature of the matrix will cool less than1 8C in 7–11 h (Fig. 14). This can be demonstratedwith the equation describing cooling of a sheet:

x = 0.5 erfz + 1

2��t

√ − erfz − 1

2��t

√( )

− 0.5 erfw + 1

2��t

√ − erfw − 1

2��t

√( )

where z ¼ 1/a, w ¼ 2–z and a is the thickness ofthe sheet (Jaeger 1968).

For the few lithic clasts (Tu-10-01 and Tu-10-02)that reached temperatures sufficient to reset theprevious remanence (higher than 210 8C for Tu-10-01 and higher than 550 8C for Tu-10-02), ourmodel indicates that the matrix temperature mustbe slightly higher (10–20 8C) than the maximumtemperature of a clast to heat a 10 cm lithic clastuniformly before the entire deposit cools. For themajority of the lithic clasts, which were not reheatedabove 90 8C, this indicates the temperature of em-placement more the surrounding ashy matrix couldbe no higher than 100 8C where in direct contactwith lithic clasts.

Importance of sampling lithic and juvenile

clasts

Because temperatures can range by over 500 8Cin Cotopaxi and Tungurahua PDC deposits over thescale of centimetres to metres, it would be mislead-ing to assign a single, average emplacement temp-erature to a scoria-rich PDC or the deposits thatit forms. The results presented here support thehypothesis that PDC deposits are exceedingly ther-mally heterogeneous at the time of deposition (e.g.Marti et al. 1991). Furthermore, the temperaturedifferences between lithic and juvenile clasts high-light the importance of air entrainment and cooling

during flow. The palaeomagnetic data reported hereindicate that the Cotopaxi and Tungurahua PDCsare a ‘cold-matrix, hot-bomb’ type of PDC, support-ing the hypothesis that the current cooled by inges-tion of large amounts of air during flow. In fact, thefine-grained material cooled to nearly ambienttemperature before deposition, and few lithic frag-ments were heated to appreciable temperatures,even within their outermost centimetre. At the sametime, juvenile magmatic clasts barely cooled andremained hot enough to deform after deposition.The proximity of cold lithic and hot juvenileclasts, and the progressive thermal demagnetizationresults, indicate local (,1 m scale) temperaturevariations of .500 8C within the deposit. Emplace-ment temperatures estimated only from the palaeo-magnetic data from juvenile scoria bombs wouldhave led to the interpretation of a ‘hot PDC’ (depo-sition .500 8C), whereas measurements of onlylithic clasts would have indicated a ‘cold PDC’ oreven a lahar. Consideration of both types of clastspermits more thorough documentation of thethermal history of PDCs and characterization ofboiling-over pyroclastic flows.

Out of 13 studies that report progressive thermaldemagnetization data to estimate deposition temp-eratures, most have examined only lithic clasts(Bardot 2000; Cioni et al. 2004; McClelland et al.2004; Porreca et al. 2008; Gernon et al. 2009; Pater-son et al. 2010; Lesti et al. 2011). Two focus onjuvenile material (Saito et al. 2003; Paquereau-Lebti et al. 2008), and one study does not specifythe nature of the materials analysed (Aramaki &Akimoto 1957). Only two studies of which we areaware since the work of Aramaki & Akimoto(1957) examined both juvenile material and lithicclasts (Mandeville et al. 1994; Sawada et al. 2000).Most clasts (pumice, obsidian and lithic) from the1883 Krakatau ignimbrite yielded emplacementtemperatures of 475–550 8C, indicating that allclasts were heated during flow or after deposition(Mandeville et al. 1994). These PDC deposits arelarge-volume siliceous ignimbrites. The wide rangein inferred emplacement temperatures in the Taihei-zan, Japan, block-and-ash deposit (100–500 8C forlithic clasts and 250–560 8C for juvenile clasts) isnot discussed and interpreted in detail (Sawadaet al. 2000). The reported range in emplacementtemperatures at Taiheizan may be an indicationof thermal heterogeneity attributable to differentextents of air entrainment over short length scalesin block-and-ash flows.

Implications of boiling-over PDCs for

hazard assessment

This study demonstrates that boiling-over eruptionsgenerated by fountaining or low column collapse

E. RADER ET AL.



from andesitic volcanoes produce density currentswith low matrix temperatures and hot (magmatictemperature) juvenile clasts. By definition, boiling-over eruptions do not produce large-scale buoyantplumes via buoyancy reversals, and the flows maybegin with relatively high densities because theydo not entrain air during plume ascent or plungingfrom high altitude. They must engulf large volumesof air as they descend, however, which can aid trans-port of clasts (Dufek et al. 2009).

Hazard mitigation requires accurate knowledgeof the temperature of pyroclastic flows (e.g. Zanellaet al. 2007). The presence of scoria bombs andunheated lithic clasts signifies that thermal damagemay not be widespread, but damage and injuryfrom abrasion, force of collision and suffocationremain dangerous. Likewise, the scoria clasts, whichconstitute a significant fraction of the deposits, areclose to magmatic temperatures and present a ther-mal hazard.

Conclusions

Palaeomagnetic estimates of emplacement temp-eratures of PDC deposits from the 1877 eruptionat Cotopaxi and the 2006 eruption at Tungurahuareveal a thermally heterogeneous mixture of lithicand juvenile clasts at the time of deposition. Coldlithic clasts were emplaced less than 1 m awayfrom hot juvenile clasts, and are interpreted toimply a cool and insulating ashy matrix due to theextent of air entrainment during flow. Most of thelithic clasts from Tungurahua were never heatedabove about 90 8C, even in their outer rims, indicat-ing they were deposited from currents that werebelow magmatic temperatures at the time of empla-cement. Two sample sites have lithic clasts thatrecord elevated emplacement temperatures (.210and .540 8C), indicating deposition of hot materialfor a limited time or in isolated locations. Most ofthe Cotopaxi juvenile clasts have emplacementtemperatures that certainly are higher than 560–590 8C, but lithic clasts were deposited cold, sim-ilar to most of the Tungurahua lithic clasts. TwoCotopaxi deposits contain juvenile and lithic claststhat both record cold (,90 8C) emplacement temp-eratures and are interpreted as originating from hotPDC deposits that were then remobilized aftercooling.

Future palaeomagnetic studies of pyroclasticdeposits should obtain data from both juvenile andlithic clasts in order to distinguish between coldand hot PDCs. Cauliflower-like scoria bombs areassociated with thermally heterogenous deposits atCotopaxi and Tungurahua, suggesting that scoria-rich PDC deposits are good indicators of a boiling-over eruptive style that rapidly cools during flow.

This work was funded by NSF grant EAR-0838153 toD. Geist and EAR-0838171 to K. Harpp. Numerouspeople helped collect field observations and sampleswhile in Ecuador including M. Myers, M. Benage,N. Pollock and J. Meyers. Additional financial supportduring the development of this study was provided bythe University of Idaho. The authors also thank L.Gurioli, C. Mac Niocaill and M. Porreca for helpfulreviews, and M. Ort for editorial handling. We wouldlike to thank the Patty Mothes and the rest of the InstitutoGeofisico for introducing us to Tungurahua and helping uswith this project, and OVT for securing our safety duringfield work.

References

Airoldi, G., Muirhead, J. D., Zanella, E. & White,J. D. 2012. Emplacement process of Ferrar Doleritesheets at Allan Hills (South Victoria Land, Antarctica)inferred from magnetic fabric. Geophysical JournalInternational, 188, 1046–1060.

Alvarado, G. & Soto, G. 2002. Pyroclastic flow gener-ated by crater-wall collapse and outpouring of thelava pool of Arenal Volcano, Costa Rica. Bulletin ofVolcanology, 63, 557–568.

Andrews, G. D. & Branney, M. J. 2011. Emplace-ment and rheomorphic deformation of a large, lava-like rhyolitic ignimbrite: Grey’s Landing, southernIdaho. Geological Society of America Bulletin, 123,725–743.

Aramaki, S. & Akimoto, S. 1957. Temperature esti-mation of pyroclastic deposits by natural remanentmagnetism. American Journal of Science, 255,619–627.

Banks, N. G. & Hoblitt, R. P. 1996. Direct Tempera-ture Measurements of Deposits, Mount St. Helens,Washington, 1980–1981. United States GovernmentPrinting Office, Washington, DC.

Bardot, L. 2000. Emplacement temperature determi-nations of proximal pyroclastic deposits on Santorini,Greece, and their implications. Bulletin of Volcanol-ogy, 61, 450–467.

Bardot, L. & McClelland, E. 2000. The reliabilityof emplacement temperature estimates using palaeo-magnetic methods: a case study from Santorini,Greece. Geophysical Journal International, 143,39–51.

Bardot, L., Thomas, R. & McClelland, E. 1996.Emplacement temperatures of pyroclastic depositson Santorini deduced from palaeomagnetic measure-ments: constraints on eruption mechanisms. In:Morris, A. & Tarling, D. H. (eds) Palaeomagnetismand Tectonics of the Mediterranean Region. Geo-logical Society, London, Special Publications, 105,345–357.

Belousov, A., Behncke, B. & Belousova, M. 2011.Generation of pyroclastic flows by explosive inter-action of lava flows with ice/water-saturated substrate.Journal of Volcanology and Geothermal Research,202, 60–72.

Branney, M. J. & Kokelaar, P. 2002. PyroclasticDensity Currents and the Sedimentation of Ignimbrites.Geological Society, London, Memoirs, 27.




Carrasco-Nunez, G. & Rose, W. I. 1995. Eruption of amajor Holocene pyroclastic flow at Citlaltepetlvolcano (Pico de Orizaba), Mexico, 8.5–9.0 ka.Journal of Volcanology and Geothermal Research,69, 197–215.

Carslaw, H. S. & Jaeger, J. C. 1959. Conduction of Heatin Solids, 2nd edn. Oxford University Press, Oxford.

Cas, R. A., Wright, H. M., Folkes, C. B., Lesti, C.,Porreca, M., Giordano, G. & Viramonte, J. G.2011. The flow dynamics of an extremely largevolume pyroclastic flow, the 2.08-Ma Cerro GalanIgnimbrite, NW Argentina, and comparison withother flow types. Bulletin of Volcanology, 73,1583–1609.

Cioni, R., Gurioli, L., Lanza, R. & Zanella, E. 2004.Temperatures of the A.D. 79 pyroclastic densitycurrent deposits (Vesuvius, Italy). Journal of Geophy-sical Research, 109, B02207, http://dx.doi.org/10.1029/2002JB002251.

Clarke, A. B., Voight, B., Neri, A. & Macedonio, G.2002. Transient dynamics of vulcanian explosionsand column collapse. Nature, 415, 897–901.

Clement, B. M., Connor, C. B. & Graper, G. 1993.Paleomagnetic estimate of the emplacement tempera-ture of the long-runout Nevado de Colima volcanicdebris avalanche deposit, Mexico. Earth and PlanetaryScience Letters, 120, 499–510.

Cole, P., Fernandez, E., Duarte, E. & Duncan, A.2005. Explosive activity and generation mechanismsof pyroclastic flows at Arenal volcano, Costa Ricabetween 1987 and 2001. Bulletin of Volcanology, 67,695–716.

Di Vito, M. A., Zanella, E., Gurioli, L., Lanza, R.,Sulpizio, R., Bishop, J. & Laforgia, E. 2009.The Afragola settlement near Vesuvius, Italy: Thedestruction and abandonment of a Bronze Age villagerevealed by archaeology, volcanology and rock-magnetism. Earth and Planetary Science Letters, 277,408–421.

Douillet, G. A., Pacheco, D. A. et al. 2013a. Dune bed-forms produced by dilute pyroclastic density currentsfrom the August 2006 eruption of Tungurahuavolcano, Ecuador. Bulletin of Volcanology, 75,1–20.

Douillet, G. A., Tsang-Hin-Sun, E. et al. 2013b. Sedi-mentology and geomorphology of the deposits fromthe August 2006 pyroclastic density currents at Tun-gurahua volcano, Ecuador. Bulletin of Volcanology,75, 1–21.

Dufek, J. & Bergantz, G. W. 2007. Dynamics anddeposits generated by the Kos Plateau Tuff erup-tion; controls of basal particle loss on pyroclasticflow transport. Geochemistry, Geophysics, Geosys-tems, 8, Q12007.

Dufek, J., Wexler, J. & Manga, M. 2009. Transportcapacity of pyroclastic density currents: experimentsand models of substrate-flow interaction. Journal ofGeophysical Research, 114, B11203.

Freundt, A. 1999. Formation of high-grade ignimbrites;II, A pyroclastic suspension current model with impli-cations also for low-grade ignimbrites. Bulletin of Vol-canology, 60, 545–567.

Fujinawa, A., Ban, M., Ohba, T., Kontani, K. &Miura, K. 2008. Characterization of low-temperature

pyroclastic surges that occurred in the northeasternJapan arc during the late 19th century. Journal of Vol-canology and Geothermal Research, 178, 113–130.

Fulop, A. 2002. Transport and emplacement of ignim-brites and resedimented volcaniclastics from GutaiMts., Eastern Carpathians, Romania. Studia Universi-tatis Babes-Bolyai, Geologia, 49, 65–73.

Garrison, J. M., Davidson, J. P., Hall, M. & Mothes, P.2011. Geochemistry and petrology of the most recentdeposits from Cotopaxi Volcano, Northern Vol-canic Zone, Ecuador. Journal of Petrology, 52,1641–1678.

Gernon, T. M., Field, M. & Sparks, R. S. J. 2009.Depositional processes in a kimberlite crater: theupper cretaceous Orapa South Pipe (Botswana).Sedimentology, 56, 623–643.

Hall, M. & Mothes, P. 2008. The rhyolitic–andesiticeruptive history of Cotopaxi volcano, Ecuador. Bulle-tin of Volcanology, 70, 675–702.

Hall, L. M., Steele, A. L., Mothes, P. A. & Ruiz, M. C.2013. Pyroclastic density currents (PDC) of the 16–17 august 2006 eruptions of Tungurahua volcano,ecuador: geophysical registry and characteristics.Journal of Volcanology and Geothermal Research,265, 78–93.

Hoblitt, R. P. 1986. Observations of the Eruptions of July22 and August 7, 1980, at Mount St. Helens, Washing-ton. United States Geological Survey, ProfessionalPapers, 1335.

Jaeger, J. C. 1968. Cooling and Solidification of IgenousRocks. Interscience, New York.

Jones, C. H. 2002. User-driven Integrated SoftwareLives: ‘PaleoMag’ Paleomagnetics Analysis onthe Macintosh. Computers and Geosciences, 28,1145–1151.

Kelfoun, K., Samaniego, P., Palacios, P. & Barba, D.2009. Testing the suitability of frictional behaviourfor pyroclastic flow simulation by comparison witha well-constrained eruption at Tungurahua volcano(Ecuador). Bulletin of Volcanology, 71, 1057–1075.

Lesti, C., Porreca, M. et al. 2011. High-temperatureemplacement of the Cerro Galan and ToconquisGroup ignimbrites (Puna plateau, NW Argentina)determined by TRM analyses. Bulletin of volcanology,73, 1535–1565.

Mandeville, C. W., Carey, S., Sigurdsson, H. & King,J. 1994. Paleomagnetic evidence for high-temperatureemplacement of the 1883 subaqueous pyroclastic flowsfrom Krakatau Volcano, Indonesia. Journal of Geo-physical Research, 99, 9487–9504.

Marti, J., Diez-Gil, J. L. & Ortiz, R. 1991. Conductionmodel for the thermal influence of lithic clasts in mix-tures of hot gases and ejecta. Journal of GeophysicalResearch, 96, 21 879–21 885.

McArthur, A. N., Cas, R. A. F. & Orton, G. J. 1998. Dis-tribution and significance of crystalline, perlitic andvesicular textures in the Ordovician Garth Tuff(Wales). Bulletin of Volcanology, 60, 260–285.

McClelland, E. A. & Druitt, T. H. 1989. Palaeomag-netic estimates of emplacement temperatures of pyro-clastic deposits on Santorini, Greece. Bulletin ofVolcanology, 51, 16–27.

McClelland, E. A., Wilson, C. J. N. & Bardot, L.2004. Palaeotemperature determinations for the

E. RADER ET AL.


http://dx.doi.org/10.1029/2002JB002251

http://dx.doi.org/10.1029/2002JB002251

http://dx.doi.org/10.1029/2002JB002251


1.8-ka Taupo ignimbrite, New Zealand, and impli-cations for the emplacement history of a high-velocity pyroclastic flow. Bulletin of Volcanology,66, 492–513.

Miyabuchi, Y., Watanabe, K. & Egawa, Y. 2006.Bomb-rich basaltic pyroclastic flow deposit fromNakadake, Aso Volcano, southwestern Japan. Journalof Volcanology and Geothermal Research, 155,90–103.

Mothes, P. A., Minard, L. H. & Janda, R. J. 1998. Theenormous Chillos Valley Lahar: an ash-flow-generateddebris flow from Cotopaxi Volcano, Ecuador. Bulletinof Volcanology, 59, 233–244.

Myers, M. 2012. Melt inclusions of Tungurahua. Mastersthesis, University of Idaho.

Nogami, K., Hirabayashi, J. et al. 2001. Temporal vari-ations in the constituents of volcanic ash and adherentwater-soluble components in the Unzen Fugendakeeruption during 1990–1991. Earth Planets and Space,53, 723–730.

Paquereau-Lebti, P., Fornari, M., Roperch, P.,Thouret, J. C. & Macedo, O. 2008. Paleomagnetism,magnetic fabric, and 40 Ar/39 Ar dating of Plioceneand Quaternary ignimbrites in the Arequipa area,southern Peru. Bulletin of Volcanology, 70, 977–997.

Paterson, G. A., Roberts, A. P. et al. 2010. Paleomag-netic determination of emplacement temperatures ofpyroclastic deposits: an under-utilized tool. Bulletinof Volcanology, 72, 309–330.

Pistolesi, M., Rosi, M., Cioni, R., Cashman, K. V., Ros-

sotti, A. & Aguilera, E. 2011. Physical volcanologyof the post-twelfth-century activity at Cotopaxi vol-cano, Ecuador: behavior of an andesitic central vol-cano. Geological Society of America Bulletin, 123,1193–1215.

Pistolesi, M., Cioni, R., Rosi, M., Cashman, K. V.,Rossotti, A. & Aguilera, E. 2013. Evidence forlahar-triggering mechanisms in complex stratigraphicsequences: the post-twelfth century eruptive activityof Cotopaxi Volcano, Ecuador. Bulletin of volcanol-ogy, 75, 1–18.

Pollock, N., Harpp, K. S., Geist, D., Dufek, J. &Mothes, P. A. 2010. Vegetation damage as a proxyfor physical characteristics of PDCs. In: AmericanGeophysical Union, Fall Meeting 2010, 13–17 Decem-ber, San Francisco, California. American GeophysicalUnion (AGU), Washington, DC, abstract V13A-2337.

Porreca, M., Mattei, M., MacNiocaill, C., Giordano,G., McClelland, E. & Funiciello, R. 2008. Paleo-magnetic evidence for low-temperature emplacementof the phreatomagmatic Peperino Albano ignimbrite(Colli Albani volcano, Central Italy). Bulletin of Volca-nology, 70, 877–893.

Riehle, J. R., Miller, T. F. & Bailey, R. A. 1995.Cooling, degassing and compaction of rhyolitic ashflow tuffs: a computational model. Bulletin of Volca-nology, 57, 319–336.

Roperch, P., Chauvin, A., Le Pennec, J. L. & Lara,L. E. 2014. Paleomagnetic study of juvenile basaltic–andesite clasts from Andean pyroclastic densitycurrent deposits. Physics of the Earth and PlanetaryInteriors, 227, 20–29.

Saito, T., Ishikawa, N. & Kamata, H. 2003. Iden-tification of magnetic minerals carrying NRM in

pyroclastic-flow deposits. Journal of Volcanologyand Geothermal Research, 126, 127–142.

Samaniego, P., Le Pennec, J., Robin, C. & Hidalgo, S.2011. Petrological analysis of the pre-eruptive mag-matic process prior to the 2006 explosive eruptions atTungurahua Volcano (Ecuador). Journal of Volcanol-ogy and Geothermal Research, 199, 69–84.

Sawada, Y., Sampei, Y., Hyodo, M., Yagami, T. &Fukue, M. 2000. Estimation of emplacement tempera-tures of pyroclastic flows using H/C ratios of carbo-nized wood. Journal of Volcanology and GeothermalResearch, 104, 1–20.

Scott, A. C. & Glasspool, I. J. 2005. Charcoal reflec-tance as a proxy for the emplacement temperature ofpyroclastic flow deposits. Geology, 33, 589–592.

Scott, A. C., Sparks, R. S. J., Bull, I. D., Knicker, H. &Evershed, R. P. 2008. Temperature proxy data andtheir significance for the understanding of pyroclasticdensity currents. Geology, 36, 143–146.

Shea, T., Gurioli, L., Houghton, B. F., Cioni, R. &Cashman, K. V. 2011. Column collapse and genera-tion of pyroclastic density currents during the AD 79eruption of Vesuvius: the role of pyroclast density.Geology, 39, 695–698, http://dx.doi.org/10.1130/G32092.1.

Smith, G. A., Grubensky, M. J. & Geissman, J. W. 1999.Nature and origin of cone-forming volcanic breccias inthe Te Herenga Formation, Ruapehu, New Zealand.Bulletin of Volcanology, 61, 64–82.

Soler, M. M., Caffe, P. J., Coira, B. L., Onoe, A. T. &Kay, S. M. 2007. Geology of the Vilama caldera: a newinterpretation of a large-scale explosive event in theCentral Andean plateau during the Upper Miocene.Journal of Volcanology and Geothermal Research,164, 27–53.

Sparks, R. S. J., Barclay, J. et al. 2002. Generation of adebris avalanche and violent pyroclastic densitycurrent on 26 December (Boxing Day) 1997 at Sou-friere Hills Volcano, Montserrat. In: Druitt, T. H. &Kokelaar, B. P. (eds) The Eruption of SoufriereHills Volcano, Montserrat, from 1995 to 1999. Geo-logical Society, London, Memoirs, 21, 409–434.

Steffke, A. M., Fee, D., Garces, M. & Harris, A. 2010.Eruption chronologies, plume heights and eruptionstyles at Tungurahua Volcano: integrating remotesensing techniques and infrasound. Journal of Volca-nology and Geothermal Research, 193, 143–160.

Suzuki, Y. J., Koyaguchi, T., Ogawa, M. & Hachisu,I. 2005. A numerical study of turbulent mixingin eruption clouds using a three-dimensional fluiddynamics model. Journal of Geophysical Research,110, B08201, http://dx.doi.org/10.1029/2004JB003460.

Taylor, G. A. 1958. The 1951 Eruption of Mount Laming-ton, Papua. 2nd edn. Bureau of Mineral Resources,Geology and Geophysics, Bulletin, 38.

Wolf, T. 1878. Memoria Sobre el Cotopaxi y su UltimaErupcion Acaecida el 26 de Junio de 1877. Imprentade El Comercio, Guayaquil.

Zanella, E., Gurioli, L., Pareschi, M. T. & Lanza, R.2007. Influences of urban fabric on pyroclastic densitycurrents at Pompeii (Italy): 2. Temperature of thedeposits and hazard implications. Journal of Geophysi-cal Research, 112, B05214.



http://dx.doi.org/10.1130/G32092.1

http://dx.doi.org/10.1130/G32092.1

http://dx.doi.org/10.1130/G32092.1

http://dx.doi.org/10.1029/2004JB003460

http://dx.doi.org/10.1029/2004JB003460

http://dx.doi.org/10.1029/2004JB003460


Zanella, E., Gurioli, L., Lanza, R., Sulpizio, R. &Bontempi, M. 2008. Deposition temperature of theAD 472 Pollena pyroclastic density current deposits,Somma-Vesuvius, Italy. Bulletin of Volcanology, 70,1237–1248.

Zijderveld, J. D. A. 1967. AC demagnetization ofrocks: analysis of results. In: Collinson, D. W.,

Creer, K. M. & Runcorn, S. K. (eds) Methods inPaleomagnetism. Elsevier, Amsterdam, 254–286.

Zlotnicki, J., Pozzi, J. P., Boudon, G. & Moreau, M. G.1984. A new method for the determination of thesetting temperature of pyroclastic deposits (exampleof guadeloupe: French west indies). Journal of Volca-nology and Geothermal Research, 21, 297–312.

E. RADER ET AL.



Hot clasts and cold blasts: thermal heterogeneity in...

Documents

Transcript of Hot clasts and cold blasts: thermal heterogeneity in...