- © 2015 by the Mineralogical Society of America
The magmatic history of a continental arc can be characterized as punctuated equilibrium, whereby long periods of low-level activity are interrupted periodically by short bursts of high-volume magmatism (“flare-ups”). Geochronological records, most notably from zircon, reveal episodicity in volcanism, pluton formation, and detrital sedimentation in, and associated with, arc segments and volcano-plutonic suites. Distinct tempos can be recognized at all resolvable spatial and temporal scales and are broadly fractal, with each scale reflecting the timescale of processes occurring at different levels in the arc crust. The tempos of continental arc magmatism thus reflect modulation of the mantle-power input as it is progressively filtered through the continental crust.
This thematic issue of Elements is centered on the recognition that magmatic processes in volcanic arcs vary in space and time. Much of the lifetime of an arc is characterized by the “steady state” magmatism that builds both the andesitic composite cones, which define the arc, and their intrusive equivalents, tonalitic plutons. This steady state magmatism is thought to be characterized by low mantle power (i.e. supply of thermal energy and volatiles to the base of the crust by basalt intrusion from the mantle), low magmatic addition rates (termed MARs in this thematic issue), and normal geothermal gradients. At the surface, these conditions produce small-volume, largely effusive eruptions, with only occasional explosive caldera-forming events (Hildreth 1981; de Silva 2008).
The “steady state” is periodically punctuated by short-lived, high-volume magmatic events termed “flare-ups” (Fig. 1). It is during flare-ups that primarily granodioritic Cordilleran batholiths and their eruptive equivalents, dacitic ignimbrites, form (Ducea et al. 2015 this issue). DeCelles et al. (2009) suggested that flare-ups are periods of increased rates of magma production, perhaps up to three or four times greater relative to the steady state rates of arc-magma production in the mantle wedge. Flare-ups are probably triggered by major changes in the geometry and geodynamics of subduction zones and lithospheric processes. Advection of mantle power through the crust sets up a series of feedbacks that lead to elevated geotherms and prodigious “upper-plate” (“crustal”) magmatism (Brown 2001; de Silva and Gregg 2014). Upper-plate magmatism may be imbued with an enriched chemical and isotopic character quite distinct from periods of steady state magmatism (e.g. Ducea and Barton 2007). Volcanic activity during arc flare-ups is dominated by the eruption of tens of thousands of cubic kilometers of ignimbrites (“ignimbrite flare-up”) and by the formation of multiple spatially and temporally related caldera complexes. These collectively define the ignimbrite plateaus that are interpreted as the surface expressions of incrementally constructed granodioritic to granitic batholiths (Lipman et al. 1972; Bachmann et al. 2007) and that are integral to the formation and stabilization of the continental crust. Closely associated with the magmatic history is a geomorphic evolution where erosion of the volcanic elements and exhumation of the plutonic roots lead to a sedimentary record that should, in principle, integrate the volcanic and plutonic records. Adequate integrated records are rare; however, zircon can act as a robust recorder of magmatic time-scales and tempos, which can be determined through high-resolution zircon geochronology. Such records that do exist reveal the episodic nature of flare-ups at a variety of space–time–volume scales. We characterize this episodic nature as “fractal” because the episodicity recurs at progressively smaller scales (see also de Saint Blanquat 2011), each with characteristic tempos (Fig. 1).
ZIRCON AS A RECORDER OF MAGMATIC TEMPOS IN CONTINENTAL ARCS
Zircon has long been recognized as an excellent geochronometer (U–Pb system) because of its high partition coefficients for uranium and thorium radionuclides, which are used in radioisotopic dating, and its low affinity for nonradiogenic lead. Additionally, zircon is a common, readily obtained, accessory phase in the silicic magmas that dominate during continental arc flare-ups, and zircon is thought to record a significant period of the magmatic history of the system. These characteristics, combined with the development of high-resolution isotopic dating techniques, have revolutionized the science of determining magmatic timescales (e.g. Schmitt 2011; Schmitz and Kuiper 2013). The data we discuss here rely on high-precision/low-spatial-resolution isotope-dilution thermal-ionization mass spectrometry (ID-TIMS; used for volcanic/plutonic records; error <1%) and the lower-analytical-precision, but much finer spatial resolution (tens of microns), in situ techniques of sensitive high-resolution ion microprobe (SHRIMP) mass spectrometry (used for volcanic/plutonic/sedimentary records; error 2–5%), and laser ablation inductively coupled mass spectrometry (LA-ICPMS; used for sedimentary records; error 2–5%). The magma systems that we discuss range in age from 200 Ma to 1 Ma. At the 1 Ma age range, age differences of 0.1 to 0.05 My can be resolved; but in 200 Ma old systems the errors may be 4 to 10 My. Events of ~0.1 My are unlikely to be recorded, and probability density functions, statistical representations of probability distributions of ages, derived from LA-ICPMS and SHRIMP detrital data tend to smooth out the error on individual grains. The variation in precision between systems may lead to offset in apparent derived ages.
One important caveat when using zircon ages is that they may not reflect the entire magmatic history. Zircon will rapidly crystallize once Zr-saturation conditions are met, but will dissolve slowly in a Zr-undersaturated melt. Furthermore, larger zircon phenocrysts preferentially survive over smaller ones. Thus, depending on magmatic conditions, some history may not be recorded. Alternatively, multiple histories may be recorded: in magmas with complex multistage histories, surviving zircon crystals provide nucleation sites for additional zircon growth. Thus, zircon can record multiple events or cyclicity in magmatic systems (e.g. Folkes et al. 2011; Lima et al. 2012).
INTEGRATING VOLCANIC, PLUTONIC, AND DETRITAL RECORDS TO REVEAL TEMPOS IN CONTINENTAL ARC MAGMATISM
Episodic behavior during flare-ups in continental arcs is often inferred from the volcanic record of ignimbrite plateaus. Eruptions occur sporadically, and by examining the space–time–volume record of volcanic activity, the episodic nature and variability of intensity of eruptions can be identified. Particularly in young arcs, or in environments where erosion is minimal and the volcanic record is well preserved, relatively robust extrusive fluxes (if the area through which a magma flows is known) can be delineated through mapping, stratigraphic correlation, and extensive geochronology (McIntosh et al. 1992; de Silva and Gosnold 2007; Lipman 2007; Best et al. 2013). Magmatic addition rates, or fluxes are determined by assuming the commonly quoted volcanic-to-plutonic ratios of 3:1, 5:1, and 10:1. However, it is important to understand that the eruptive record is not a complete record of the magmatic activity. Moreover, the largest silicic systems are often difficult to interpret because they are older, cyclic, incompletely preserved or exposed, and may be “resolution limited” because the ability to resolve absolute age differences decreases as age increases. Finally, most studies have focused on segments of arc or on individual volcanic complexes, so data for complete arcs are lacking. Sampling bias also creeps in: studies commonly focus on a few specific eruptions intrusions in preference to the entire region (Fig. 2).
The plutonic record that is exposed in broadly exhumed arcs is also examined to retrieve apparent intrusive magmatic fluxes. Systematic mapping and abundant geochronologic sampling allow us to illustrate the apparent flux as preserved on an inclined section through the upper crust. Such areal representation of flux, as estimated either by U–Pb zircon age or by area, is highly variable due to pulses and lulls in intrusive activity (Bateman 1992; Ducea 2001; DeCelles et al. 2009). Caution is warranted for three reasons when assessing for magmatic flux. First, geochronologic data are generally not collected in a spatially systematic manner. Second, inclined sections are incomplete exposures of less than fully exhumed plutonic arcs. Third, translation of such data into a volume typically relies on poorly constrained estimates of the thickness of intrusions. As in the volcanic record, the petrologist's effect is apparent in plutonic studies (Fig. 2).
The detrital-zircon record would be assumed to provide the most complete assessment of magmatic activity, owing to the very common occurrence of zircon in igneous rocks and the nature of sedimentation (e.g. Gehrels 2012). By extension, detrital zircon should record magmatic addition rates and provide a first approximation of the volume of zircon-bearing magmatic rock, with short-lived magmatic pulses and lulls also being apparent (e.g. Barth et al. 2012; Laskowski et al. 2013). Nonetheless, caution is warranted when using the zircon record as a proxy because seldom does the sedimentary record isolate a single part of a magmatic arc: sedimentary successions probably represent longer-term integrated “average” records.
MAGMATIC TEMPOS IN CONTINENTAL ARCS—THE RECORD IN FLARE-UPS
The Mesozoic Sierra Nevada–Salinia–Mojave arc of the western USA and Mexico (Dickinson 1981; Saleeby 2003) provides an example of how plutonic and volcanic zircon records, together with the derivative sedimentary record, show the episodicity apparent on a scale of several tens of millions of years (Fig. 3). Remnants of the Jurassic parts of the Salinia–Mojave arc cover a region about 200 km across strike by 300 km along strike, comparable in scale to the aforementioned ignimbrite plateaus. An extensive plutonic age dataset shows a distinct primary pulse lasting 40–60 My with a peak at 166 Ma, reflecting the assembly of a typical Sierran granodiorite-to-granite intrusive suite during a flare-up (Fig. 3B). The more defined/pronounced peaks in the volcanic record may implicate shorter windows when conditions for eruption are met, while the plutonic record presents a more continuous history of magma accumulation and batholith construction. The detrital data from the fore-arc and back-arc sections of this Mesozoic arc complex show a broad correlation depending on whether the sedimentary basin records a volcanic or plutonic terrain uniquely or dominantly. The back-arc signature closely matches the plutonic record, whereas the fore-arc signature shows a younger dominant peak that matches the 150 Ma volcanic peak (Fig. 3C). The sample base in both the fore-arc and back-arc regions is hundreds of zircon grains from several units spanning several hundred kilometers parallel to the arc; this contrasts with the Salinia–Mojave portion as represented by plutonic and volcanic zircons. The double peak in the volcanic record together with the correlative peak in the detrital fore-arc data, when contrasted with the irregular plutonic curve, also illustrates the longitudinal asymmetry of the arc. Over a longer timescale, as much as 100 My, the distal back-arc record possibly mimics the greater flux curve for the Sierra Nevada batholith in Figure 3B (Laskowski et al. 2013).
FRACTAL TEMPOS IN CONTINENTAL ARC MAGMATISM
Similar, but not necessarily identical, patterns of arc magmatic activity that recur at progressively smaller scales define fractal tempos in continental arc magmatism. For instance, secondary pulses can be seen within the primary 40–60 My p2 pulse (Fig. 3C). Further insight into these pulses requires data of higher resolution that could be readily available in the space–time–volume patterns of ignimbrite plateaus. Among the best documented are those of the Altiplano–Puna Volcanic Complex (APVC) of the Central Andes (de Silva et al. 2006), and the various middle Cenozoic volcanic fields of the North American Cordillera (e.g. McIntosh et al. 1992; Bryan et al. 2007; Lipman 2007; Best et al. 2013).
In the APVC, over 15,000 km3 of largely high-K, crystal-rich dacitic magma was erupted from five distinct caldera systems and smaller ignimbrite shields during an ~10 My ignimbrite flare-up. The spatiotemporal pattern of volcanism, its concordance with a negative residual gravity anomaly, and a seismic low-velocity zone are inferred to map out the incremental construction of an upper-crustal granodioritic batholith, which reflects the remnant plutonic roots of the erupted magmas. A clear pattern that emerges from the space–time–volume data of the APVC is that magmatic activity is episodic: activity clusters at approximately 2 My intervals (Fig. 4), and these define tertiary (third-order) pulses. Furthermore, the volume–time pattern of eruptions is defined by an initial waxing stage of relatively small eruptions clustered around 10 ± 0.5 Ma, followed by a climactic stage ~4 My long during which clusters of supereruptions define third-order pulses of activity at 8 ± 0.5, 6 ± 0.5, and 4 ± 0.5 Ma, with a final waning stage, less than 3 My in duration, of successively smaller eruptions (de Silva et al. 2006).
Volume-normalized SIMS U–Pb age data for the APVC eruptions show a similar pattern to the eruptive record, with distinct pulses of zircon ages coincident with the pulses of eruptive activity (Fig. 4B). The broader peaks in zircon crystallization reflect two things. First, because zircon ages record a period of crystallization prior to eruption, the zircon ages from a single ignimbrite may define a relatively broad age-population extending from hundreds of thousands to a million years before eruption right up to the eruption age. Secondly, these data show that each tertiary pulse of ~2 My consists of three or more quaternary (fourth-order) pulses, defined by distinct magmatic histories of several hundred thousand years that culminated in eruptions. In summary, the secondary pulse of the 10 My ignimbrite flare-up of the APVC consists of four tertiary pulses of intrusion/eruption with a periodicity of ~2 My. The tertiary pulses, in turn, consist of three or more distinct quaternary pulses each of <1 My duration.
While plutons are not exposed in the APVC, a plutonic record of similar temporal scale is that of the Tuolumne Intrusive Suite (TIS) and John Muir Intrusive Suite, which are part of the Mesozoic Sierra Nevada arc in California (Fig. 3; Bateman 1992). Here, ~140 high-precision zircon ages are available from multiple samples. If most zircon growth records solidification at the level of emplacement (e.g. Coleman and Glazner 1997), then the accumulated zircon analyses could be used as a proxy for emplacement rates of these intrusive suites, which occupy areas of ~103 km2 and depths of up to 5 km. These nested intrusive suites might be the plutonic equivalent of a composite volcanic field (analogous to the APVC or a large nested caldera system), but it is not clear how these mappable compositional units may record discrete “plutons.” The available zircon data set for the TIS, if taken as collectively defining a secondary pulse, clearly indicates zircon growth over an integrated period of ~9 My (Figs. 2, 5). The data also then suggest three or more tertiary pulses of zircon growth of perhaps 1 to 2 My duration each with comparable lulls. If so, the pulses of growth roughly correspond to the nested suites as mapped and these would be equivalent to the tertiary pulses for the APVC. Resolution of quaternary pulses, evident in the volcanic record, is not possible with the current data for the TIS (Figs. 4C-F).
Correlation between ages derived from the TIS plutonic rocks and the detrital signature (Fig. 5) highlights that, at the scale of 5–10 My (i.e. the secondary pulses), the plutonic and detrital records are not as well correlated as they are at the broader scale of the primary pulses (cf. Fig. 3). Probability curves from the fore-arc and back-arc have broadly similar shapes. At first glance these are different from the curve for the TIS, suggesting that the 10+ My scale is best for understanding how detrital zircons reflect the tempo of magmatism. At this scale, however, the errors assigned to grains can become important. The probability curve smooths out error on individual grains, which is 1–2%, but the error on CA-TIMS data is <1%. Thus, some of the offset of plutonic and detrital data apparent in Figure 4 may be due to errors on grains. Other discrepancies, for example the pronounced trough in TIS data at 92 Ma, which is not reflected in either fore-arc or back-arc data, point out the potential for detrital data to reflect multiple sources, arc migration, or other factors that would blur distinctions.
MAGMATIC TEMPOS AND HOW CONTINENTAL ARCS WORK
The hierarchy of pulses in continental arcs can be considered as successively coupled space/time scales related to specific scales of continental-arc magmatic processes. The primary pulse, which is the high-volume event or flare-up, reflects the largest space/time scale of an excursion in mantle-power input from steady state magmatism, possibly coupled to deformation patterns in the upper plate (DeCelles et al. 2009). This primary 40–60 My mantle signal may occur over a large area: examples include the entire Mesozoic Californian arc, in which flare-ups were driven by delamination events (Ducea 2001); or the middle Cenozoic of the western United States, where the flare-up appears to have been triggered by a regional transition from low-angle plate convergence to an increasingly extensional regime and where peak volcanism largely preceded the bulk of the extension (Lipman 1972; Best et al. 2013).
Secondary pulses are most readily defined on a smaller spatial scale than the primary pulses, reflecting the localization of magmatism into regional “nodes.” This suggests that the primary pulse can be modulated by processes of melting, segregation, and transfer through the crust, along with local upper-plate tectonics to produce secondary pulses in smaller regions. For instance, each of the various volcanic fields of the middle Cenozoic flare-up of the western United States, or the plutonic and volcano-plutonic suites of the Mesozoic California arc, define the individual components of much larger flare-ups. The APVC and the TIS are the examples used here. These secondary pulses, ~10–20 My in duration, must reflect the timescale of melt production and delivery from the melting and assimilation (MASH) zone at or near the mantle/lower crust boundary and its interaction with upper-plate tectonics over the area of each component. The magmatic histories of many individual volcano-plutonic systems tend to be characterized by an ~10 My timescale (Grunder et al. 2008), which suggests that this is a fundamental signal in continental magmatism.
The tertiary pulses in the APVC, and perhaps the TIS, indicate processes that not only divide the secondary pulse into shorter pulses (~2 My magmatic events in the APVC that culminate in eruptions) but produce the characteristic space–time–volume pattern of waxing, climax, and waning. The tempo of these tertiary pulses may ultimately reflect the influence of a combination of mantle, crustal, and upper-plate tectonic modulation (or other nonmagmatic upper-crustal process) (Fig. 6). The waxing–climax–waning pattern seen in these tertiary pulses may reflect the evolution of crustal melt production with time, thereby mimicking the mantle pulse topology. Alternatively (or additionally), the pattern may reflect a progressive evolution of crustal rheology resulting from the thermal signal from the mantle progressing through the crust by intrusion and advection (Fig. 1). This progression may result in the formation of a mid- to upper-crustal MASH zone and magma “staging” area and an elevation of the brittle–ductile transition to shallow levels in the crust. Over time, these processes conspire to allow successively larger magma bodies to be built in the uppermost crust before they erupt (e.g. de Silva and Gosnold 2007).
Quaternary pulses are defined by individual eruptions from the magma bodies that were produced during the tertiary pulses. Quaternary pulses have timescales of ~1 My, based on zircon crystallization timescales for large calderas in the APVC. This tempo is controlled by the thermomechanical evolution of the upper-crustal magma reservoirs. As magma accumulates in the upper crust, feedbacks between temperature, host-rock mechanics, and chamber pressurization results in ductile host-rock rheologies, which promotes storage and growth over eruption (de Silva and Gregg 2014). Eventual eruption is due to crossing a mechanical threshold. Putting aside any tectonic overprint, the eruptive tempo of a flare-up may be dominantly modulated by the thermomechanical properties of upper-crustal magma reservoirs.
SUMMARY AND RECOMMENDATIONS FOR THE FUTURE
The time–volume evolution of continental arcs is described as punctuated equilibrium where high-volume events (“flare-ups” of 40–60 My duration) interrupt steady state arc activity or lulls. Combined volcanic, plutonic, and detrital U–Pb zircon records from flare-ups reveal a fractal tempo to the magmatism. From the examples in this paper, a hierarchy of pulses is identified, which define four successively faster tempos. These represent how the magmatic front, driven by the primary mantle-power input, is transmitted through the crust. This modulation of mantle power is reflected by sequentially smaller spatial and temporal scales that progress from arc-segment (first-order), to volcano-plutonic suites (second-order), to individual plutons or magma reservoirs (third-order) to caldera-forming eruptions (fourth-order).
Future research should aim at better defining the hierarchy of pulses and the resulting tempos, their apparent fractal character, and the link between these temporal patterns and magmatic and tectonic processes. Particularly important will be efforts to understand magmatic tempo in the context of rates of mantle-power input and the intrinsic and structural properties of the crust and of upper-plate tectonics. Linking volcanic, plutonic, and detrital records are the way forward, but challenges remain in understanding how multiple sources and arc migration might affect the mutual correlations of these records.
The authors thank the guest editors for their invitation to contribute this article and for their deft handling of the manuscript, which benefited from thoughtful reviews by Olivier Bachmann and Keith Putirka for Elements. Discussions over the years with A. Schmitt, R. Economos, A. Grunder, J. Miller, D. Coleman, C. Jacobson and J. Wooden have been invaluable. Our research on continental arc magmatism in the Andes and California is funded by National Science Foundation grants EAR-0711119, EAR-0838536, and EAR-0908324.