Volcanism and Global Environmental Change, eds. Anja Schmidt, Kirsten E. Fristad and Linda T. Elkins-Tanton. Published by Cambridge University Press. © Cambridge University Press 2015.
Large igneous provinces are recognized from the Precambrian at 3.79 Ga (Ernst, 2013), and extend through well-preserved examples from the Mesozoic and Cenozoic (Ross et al., Reference Ross, Ukstins Peate and McClintock2005; Bryan and Ferrari, Reference Bryan and Ferrari2013, and references therein). While originally inferred to consist of a layer-cake sequence of massive and laterally continuous effusive basaltic lava flows, detailed volcanostratigraphy studies have generated a more nuanced view of province architecture, highlighting that many provinces include a significant component of clastic material derived from volcanic and sedimentary formation mechanisms. Conversely, some of the volumetrically largest basaltic volcaniclastic deposits appear to be associated with large igneous provinces (Ross et al., Reference Ross, Ukstins Peate and McClintock2005).
The importance of volcaniclastic deposits – and the implications for paleoenvironmental reconstructions, eruption dynamics, and climate impact – is one of the key concepts to emerge from scientific studies of large igneous provinces over the last 25 years. Ross et al. (Reference Ross, Ukstins Peate and McClintock2005) recognized, and highlighted, the near-ubiquitous occurrence of mafic volcaniclastic deposits as an integral component in large igneous provinces. These deposits contain information – some unique – on primary fragmentation mechanisms, eruptive processes, and depositional environments. Mafic volcaniclastic deposits provide a record of what we now recognize as complex temporal and spatial volcanic heterogeneity in large igneous provinces, and allow us to reconstruct their tectonic and physical evolution as an equally significant and complementary story to that of the geochemical evolution of magmatism. We provide a brief overview of mafic volcaniclastic deposits and formation mechanisms, and spotlight recent work highlighting their utility for interpreting large-scale tectonic evolution and climate impact issues related to large igneous province emplacement.
1.2 Mafic volcanic-derived clastic deposits
Clastic deposits composed of mafic volcanic particles – in any proportion from partly to entirely – can be generated by a wide variety of mechanisms spanning the full range of volcanic to sedimentary processes, and the resultant textures and morphologies are likewise highly variable. Three genetic categories of mafic volcanic-derived clastic deposits, based on formation mechanisms, are primary and reworked volcaniclastic deposits, and epiclastic deposits (White and Houghton, Reference White and Houghton2006).
1.2.1 Primary volcaniclastic deposits
Primary volcaniclastic deposits are formed from fragmental material deposited as a direct result of explosive or effusive eruptions. There are four end-member types of deposits: autoclastic, pyroclastic, hyaloclastic, and peperitic (White and Houghton, Reference White and Houghton2006; White et al., Reference White, Bryan, Ross, Self and Thordarson2009). The main factors controlling formation of these deposits are magma eruption rates, concentration of magmatic volatiles, and presence and relative abundance of external water, either as freestanding bodies or in saturated sediments. Mobilization of magma-generated particles is unique compared to other sedimentation processes that depend exclusively on gravity, because primary volcanic particles may acquire transport energy from their source – e.g. explosive expansion, lava flow velocity – and may initially be independent of slope or depositional base level (Fisher and Smith, Reference Fisher and Smith1991).
Autoclastic deposits are products of auto brecciation, and are generated as effusive lavas that exceed the viscosity-strain rate threshold and fragment (Peterson and Tilling, Reference Peterson and Tilling1980); rapid cooling promotes groundmass crystallization and crust disruption (Cashman et al., Reference Cashman, Thornber and Kauahikaua1999). A’a lava flow morphology is characterized by brecciated upper and lower surfaces; slabby or rubbly pahoehoe have broken or brecciated upper crusts and are transitional between pahoehoe and a’a (e.g. Guilbaud et al., Reference Guilbaud, Self, Thordarson and Blake2005).
Pyroclastic deposits form from explosive volcanic eruption plumes and jets or pyroclastic density currents and can be generated by magmatic or phreatomagmatic fragmentation mechanisms, or a complicated interplay of both (e.g. Graettinger et al., Reference Graettinger, Skilling, McGarvie and Hoskuldsson2013). Magmatic fragmentation represents a minor mechanism for generating mafic volcaniclastic deposits in large igneous provinces, and documented examples are rare. The Columbia River Basalt vent sites in the Roza Member have densely agglutinated and welded spatter and highly vesicular scoria fall deposits (Brown et al., 2014).
Involvement of aquifer or surface water in mafic eruptions leads to large-scale phreatomagmatic volcanism, and can generate deposits with volumes of up to 102 to 105 km3 (Ross et al., Reference Ross, Ukstins Peate and McClintock2005). Phreatomagmatic pyroclastic and hyaloclastic deposits, along with peperite, represent the full spectrum of products from magma–water interaction (Wohletz, Reference Wohletz2002). External water is integral for formation, but magmatic volatiles are not precluded, and in the case of phreatomagmatic eruptions, may also play a role in fragmentation. Clast-forming processes during hydromagmatism include four primary mechanisms: magmatic explosivity, steam explosivity, cooling-contraction granulation, and dynamic stressing; all are dependent on the magma to water ratio (Wohletz, Reference Wohletz1983; Kokelaar, Reference Kokelaar1986). Phreatomagmatic pyroclastic deposits are produced by the optimal fuel (magma) to coolant (water or sediment-laden water) mixture to generate explosivity (magma to pure water mass ratio of ~ 0.33: White, Reference White1996), whereas hyaloclastic deposits are volumetrically dominated by water and peperite dominated by wet sediment (wet sediment to magma mass ratios > 1: Wohletz, Reference Wohletz2002).
Hyaloclastic deposits are solely generated by quench fragmentation during magma–water interaction, and result from effusive magma contacting abundant water, in either marine or continental settings. Pillow lavas, pillow–palagonite breccias, and hyaloclastites are the most typical products of mafic magma quenching and spalling in a subaqueous environment.
Peperitedeposits result from magma interaction with unconsolidated, water-bearing clastic deposits in shallow intrusions, subaqueous or surface environments (White et al., Reference White, McPhie and Skilling2000; Skilling et al., Reference Skilling, White and McPhie2002). Experimental and theoretical studies suggest that mechanisms of magma–water interaction and magma–sediment–water interaction may be similar (Kokelaar, Reference Kokelaar1986), and peperites rely on fluidization and vigorous injection and mixing of water-saturated sediments and lava (Kokelaar, Reference Kokelaar1982). Recognition of peperite indicates contemporaneity of sedimentation and volcanism (Busby-Spera and White, Reference Busby-Spera and White1987). Given the ubiquity of environments that could generate peperite, there are relatively few documented examples in large igneous provinces. However, this may be a function of identification rather than absence. Even in the predominantly arid desert environment that the Paraná–Etendeka Large Igneous Province was emplaced into, peperites formed where pahoehoe lava flows interacted with wet lacustrine silts and clays, which collected in low-lying topography of lava flow surfaces (Waichel et al., Reference Waichel, de Lima, Sommer and Lubachesky2007).
1.2.2 Reworked volcaniclastic and epiclastic deposits
Reworked volcaniclastic deposits are composed of particles sourced from primary volcaniclastic deposits that have been redeposited by surface processes (wind, water, ice, gravity) either concurrent with eruption or after a period of immobility. In reworked volcaniclastic deposits, the volcanic processes that create the particles are not the same as those that transport the particles to their final depositional site. Epiclastic (or volcanogenic) sediments are formed from weathering and erosion of volcanic rocks, including previously lithified volcaniclastic rocks. Lithification can occur as part of volcanic emplacement (e.g. welding) or as a secondary process of cementation or compaction.
1.3 Spatial and temporal occurrence of mafic volcaniclastic deposits
One of the strengths of mafic volcaniclastic deposits is in the record they preserve of tectono-volcanic facies and province architecture evolution over time. The recognition that pre-volcanic kilometer-scale doming is not an unequivocal feature of large igneous provinces, coupled with recent numerical modeling indicating that large igneous province emplacement can generate substantial and complex patterns of pre- and syn-volcanic subsidence and/or uplift (+/– hundreds to thousands of meters: e.g. Czamanske et al., Reference Czamanske, Gurevitch, Fedorenko and Simonov1998; Ukstins Peate and Elkins-Tanton, Reference Ukstins Peate and Elkins-Tanton2009; Elkins-Tanton and Ukstins Peate, Reference Elkins-Tanton and Ukstins-Peate2010; Sobolev et al., Reference Sobolev, Sobolev and Kuzmin2011), suggests that tectonic evolution may be a significant factor controlling the broad-scale distribution of these deposits. Provinces that contain significant volcaniclastics include the middle-Jurassic Kirkpatrick section of the Ferrar flood basalts in Antarctica, with tuff-breccias up to 400 m thick (Elliot and Flemming, Reference Elliot and Flemming2008); the Kachchh region in the northwest of the Deccan flood basalts, with lapilli and lithic blocks (Kshirsagar et al., Reference Kshirsagar, Sheth and Shaikh2011); and the Karoo (McClintock et al., Reference McClintock, Marsh and White2008). Here, we briefly describe three additional significant examples: the North Atlantic, the Emeishan and the Siberian.
1.3.1 East Greenland, North Atlantic Igneous Province
Detailed volcanostratigraphic studies in East Greenland illustrate a cyclicity of phreatomagmatism and subsidence during the initial stages of province emplacement, recording three phases of subaqueous to subaerial volcanism, with progressively less hydrovolcanic influence (and inferred downdropping) in each cycle (Figure 1.1; Ukstins Peate et al., Reference Ukstins Peate, Larsen and Lesher2003). Initiation of volcanism is represented by subaerially deposited phreatomagmatic lapilli-tuffs with accretionary lapilli and abundant quartz and feldspar grains (~ 50%) sourced from underlying upper shoreface sandstones and mid-Paleocene fluvial clastic deposits (Larsen et al., Reference Larsen, Fitton and Pedersen2003). Overlying these are a series of hyaloclastites and pillow lavas, some forming foreset-bedded units > 300 m thick (Nielsen et al., Reference Nielsen, Soper and Brooks1981), suggesting that water depth increased dramatically with the initiation of basaltic volcanism. Hydromagmatic deposits transition to 500 m of compound lava flows, and the entire volcanic succession forms a shield-like structure with a diameter of ~ 40 km (Ukstins Peate et al., Reference Ukstins Peate, Larsen and Lesher2003). This is, in turn, overlain by a sequence of mafic volcaniclastic deposits that preserve a lateral facies change from primary units – including vent sites – in the northeast to reworked volcaniclastic and epiclastic deposits to the southwest.
Primary deposits consist of c. 300 m of: fallout tuffs; surge deposits with abundant accretionary and armored lapilli; bomb beds; and scoria deposits with three-dimensional cone morphology (Ukstins Peate et al., Reference Ukstins Peate, Larsen and Lesher2003). These transition to 1000 m of reworked and epiclastic deposits of siltstone and sandstone containing up to 80% volcanic material: altered basaltic glass (tachylite, palagonite), clinopyroxene crystals, basaltic lava clasts, and pyroclastic lithic fragments, with minor intercalated tuffs. Correlation of reworked and epiclastic deposits highlight the development of regional syn-volcanic basins with cumulative thicknesses > 3000 m (Larsen et al., Reference Larsen, Fitton and Pedersen2003; Passey and Bell, Reference Passey and Bell2007).
Overlying this is the main phase of flood basalt lavas, with a few thin magmatic tuffs containing Pele’s tears and glass shards, concentrated in the lowermost part of the sequence (Ukstins Peate et al., Reference Ukstins Peate, Larsen and Lesher2003). A final transition from effusive flood lava to highly explosive basaltic phreato-Plinian eruptions occurs in the uppermost sequence, when active lithospheric rifting and subsidence resulted in flooding of the nascent North Atlantic Rift proto-ocean basin (Larsen et al., Reference Larsen, Fitton and Pedersen2003; Jolley and Widdowson, Reference Jolley and Widdowson2005).
1.3.2 Emeishan large igneous province
Research on the Emeishan large igneous province highlights the utility of mafic volcaniclastic deposits in addressing questions of large-scale tectonic evolution during flood volcanism. A thick and laterally extensive wedge of clastic deposits (170 m thick, 30 to 80 km wide, 400 km long), emplaced near the base of the Emeishan lavas, was initially interpreted as an alluvial fan conglomerate, and was attributed to pre-volcanic, kilometer-scale domal uplift and erosion of underlying carbonate (He et al., Reference He, Xu, Chung and Wang2003). The ubiquity of dense to poorly vesicular blocky sideromelane, pyroclastic textures such as accretionary lapilli, volcanic bombs with bomb sags, and ductile deformation of mafic clasts unequivocally identifies these rocks as phreatomagmatic lapilli-tuffs and tuff-breccias, and likely represent near-vent deposits (Figure 1.1; Ukstins Peate and Bryan, Reference Ukstins Peate and Bryan2008, Reference Ukstins Peate and Bryan2009). The abundance of marine limestone lithic fragments – some containing mafic clasts themselves – and the presence of unbound shelly fossil material, strongly suggests that active carbonate deposition was contemporaneous with volcanism, and that these units were emplaced near sea level (Ukstins Peate and Bryan, Reference Ukstins Peate and Bryan2008, Reference Ukstins Peate and Bryan2009).
Continuing work, focusing on the zone of inferred maximum uplift, has identified a protracted and extensive record of hyaloclastic and phreatomagmatic volcanism (Figure 1.1). Microfossil studies show nascent carbonate platform collapse immediately prior to initiation of volcanism (> 200 m: Sun et al., Reference Sun, Lai and Wignall2010). The first phase of volcanism is laterally heterogeneous but dominated by phreatomagmatic and subaqueous volcanism. Eruptions through shallow-water carbonates generated thin subaqueous hyaloclastites and subaerial tuff deposits near Daiquo (Ukstins Peate and Bryan, Reference Ukstins Peate and Bryan2008), whereas in the Dali area (the core of inferred maximum uplift), volcanism initiated with a succession (c. 750 m) of pillow lavas and hyaloclastites with intercalated marine limestones and submarine tuffs (Zhu et al., Reference Zhu, Guo, Liu and Du2014). Eruptions transitioned to phreatomagmatic lapilli-tuffs and tuff-breccias intercalated with basaltic lava sheet flows displaying peperitic basal zones and carbonates (c. 2000 m: Zhu et al., Reference Zhu, Guo, Liu and Du2014), suggesting a very shallow subaqueous to subaerial depositional environment. This is overlain by > 2500 m of thick a’a and pahoehoe basalts and rhyolite lavas, intercalated with minor, thin (~ 1 m), oxidized basaltic tuffs dominated by glassy vesicular ash shards (Zhu et al., Reference Zhu, Guo, Liu and Du2014), likely derived from subaerial phreatomagmatic to magmatic pyroclastic eruptions.
1.3.3 Siberian flood basalts
The Siberian flood basalts contain intercalated volcaniclastics to varying extent throughout the most studied sections of that province, for example in Noril’sk (e.g. Fedorenko et al., 1996). A vast literature on the Siberian province exists, but here we focus on the understudied volcaniclastics and present some new results. The most significant volcaniclastics in the Siberian province are the thick, primarily phreatomagmatic deposits underlying the lavas. In the northeast and northwest sections the majority of the basal volcaniclastics are less than 30 m in thickness, and are sometimes absent (Figure 1.2). In the central, eastern and southern regions, however, the volcaniclastics are voluminously present in largely massive, featureless outcrops.
Along almost 200 km of the Angara River north of Ust Ilim’sk, all the river cliffs consist of volcaniclastics, and visible outcrops are as much as 250 m thick, with erosional upper and unexposed basal contacts (Naumov and Ankudimova, Reference Naumov and Ankudimova1995). Volcaniclastic units are massive, unbedded and sediment-rich, though near the Kata River there is local bedding and accretionary lapilli. Some outcrops have lithic blocks of underlying sedimentary strata; peperites and sediment dikes indicate an active aquifer and driving force for eruption from depth. Notably absent are pillow basalts and hyaloclastites.
Similar deposits occur along 200 km of the Nizhnaya Tunguska River, stretching east–west past the middle Siberian town of Tura. In Tura, drill cores indicate at least 500 m of tuffs transitioning to overlying effusive lavas (Drenov, Reference Drenov1985). These drill cores demonstrate that voluminous phreatomagmatism immediately preceded the main stage of effusive lava emplacement.
1.4 Evidence for volatile loads, temperatures and plume heights
Chemicals released by volcanism will have the greatest effect on global climate, both in terms of destructive chemical reactions and longevity, if they reach the stratosphere. Material is rapidly washed from the troposphere by rain. Basaltic magmas are generally less gas-rich (with the possible exception of sulfur) and less viscous than more silicic eruptions, and are generally less explosive without interactions with external volatiles. However, basaltic Hawaiian-style fire fountains are capable of injecting material into the stratosphere (Stothers et al., Reference Stothers, Wolff, Self and Rampino1986; Woods, Reference Woods1993). This is corroborated by the Laki eruption (Iceland 1783–1784, Thordarson et al., Reference Thordarson, Self, Óskarsson and Hulsebosch1996). Laki was largely effusive, but it had significant fire-fountaining episodes. Witness accounts describe an eruption column more than 9 km high (Thordarson and Self, Reference Thordarson and Self2003), and the detection of Laki chemicals in Greenland ice cores confirms that material reached the stratosphere (Fiacco et al., Reference Fiacco, Thordarson and Germani1994; Wei et al., Reference Wei, Mosley-Thompson, Gabrielli, Thompson and Barbante2008).
Basaltic eruptions that produce volcaniclastic deposits, however, may have more capacity to implant material into the stratosphere, particularly at northern latitudes with a low-altitude tropopause (~ 9 km near the poles compared to 17 km at the equator, but sensitive to a wide variety of external parameters (Thuburn and Craig, Reference Thordarson, Self, Óskarsson and Hulsebosch1997; see also Ross et al., Reference Ross, McClintock and White2008)). Basaltic volcaniclastic deposits are commonly phreatomagmatic, and incorporation of water into an eruptive plume can have a range of effects. If the water vaporizes, the density of the plume decreases with addition of steam, but the temperature also decreases through the consumption of latent heat. The decrease of density means a higher plume may develop with a lower eruption velocity.
Conversely, if added water cannot vaporize, the eruptive fountain will collapse and flow laterally (Koyaguchi and Woods, Reference Koyaguchi and Woods1996). As phreatomagmatic basaltic eruptions proceed, they may transition between these states. Koyaguchi and Woods (Reference Koyaguchi and Woods1996) suggest that accretionary lapilli can form in both the fountains and wet lateral flows, and find that for an initial eruptive velocity of 100 m/s, a temperature of 1000 K, and a volatile content of 3 wt% water, a plume may reach as high as 35 km, at which point its temperature would be below 300 K. In agreement, Walker et al. (Reference Walker, Self and Wilson1984) estimate a vent exit velocity of 250–350 m/s for the basaltic plinian Tarawera eruption of 1886, which generated a column of ash that ascended ~ 30 km.
While individual eruptions may not always reach these conditions, the broad thermal perturbation of a flood basalt province may produce its own weather pattern of thermals in the atmosphere. This concept was first investigated by Emanuel et al. (Reference Emanuel, Speer, Rotunno, Srivastava and Molina1995), in which they posited that both very large bolide strikes and large-scale volcanic eruptions could produce exceptionally violent storms termed hypercanes, capable of injecting large amounts of mass into the stratosphere. More recent work by Kaminski et al. (Reference Kaminski, Chenet, Jaupart and Courtillot2011) described penetrative convection above large lava flows, where broad temperature perturbations drive large upwellings past the tropopause.
1.5 Summary: potential for climate change
Mafic volcaniclastics make up a significant fraction of large igneous province eruptive volume, including in the Siberian, Emeishan, North Atlantic, Karoo, Ferrar, and Columbia River flood basalts. The type and distribution of volcaniclastics can determine the relative impacts of volcanism and tectonism on a region, by filling pre-effusive topography in some cases, and by tracking surface or ground water in others.
Globally, volcaniclastic eruptions are now recognized to have potential for climate-changing atmospheric effects. Their eruptions can inject material into the stratosphere, either from their eruptive plume or with the help of regional weather effects produced by the large igneous province itself.
A significant next step for the scientific community will be estimating the volatile load of these eruptions. Some efforts have been made in Siberia, and demonstrate that Siberian volcaniclastics carried significant sulfur, chlorine and fluorine from the fluids in their eruption-driving aquifer (Black et al., Reference Black, Elkins-Tanton, Rowe and Ukstins Peate2012). Further, these eruptions naturally produced halocarbons sufficient to destroy ozone (Black et al., Reference Black, Lamarque, Shields, Elkins-Tanton and Kiehl2013) through reaction in the bedrock and the eruptive plume (Svensen et al., Reference Svensen, Planke and Polozov2009; Black et al., 2014, this volume).
The volcaniclastics in flood basalts, therefore, may be the major missing link between flood basalts and extinctions. The underlying cause may be climate-changing volatiles sourced from continental crustal rocks that chamber the flood basalt magmas, which are missing in ocean basins.