The Pace of Plutonism PDF

The Pace of Plutonism Trango Towers, Gil...

The Pace of Plutonism Trango Towers, Gilgit-Baltis Drew S. Coleman1, Ryan D. Mills1, and Matthew J. Zimmerer2 1811-5209/16/0012-0097$2.50 DOI: 10.2113/gselements.12.2.97 B eneath volcanoes are magmas that never erupt but that become frozen liquid mixtures (clearly igneous), into feldspar- and quartz-rich rocks broadly called granite. Where the the evolution of plutonic rocks is not directly observable. Although crystallized magmas form bodies with distinctive textures, they are he ultimately fell on the losing side grouped into named units—plutons. The rate (pace) at which magmas accumu- of the debate, migmatist advocate late into plutons is fundamental to understanding both how room is made Prof. Herbert H. Read of Imperial College, London (UK) emphasized for the magmas and how unerupted and erupted magmas are connected. two important themes in his 1947 Dating plutonic rocks suggests that plutons accumulate slowly. Although the presentation. The fi rst theme was pace of magma accumulation does not preclude direct connections between the room problem: If granites intrude as magmas, how are the plutons and small volcanic eruptions, it appears to be far too slow to support wall rocks displaced in order to connections between most plutons and supereruptions. make space for the granite? The KEYWORDS : granite, pluton, U–Pb, Ar–Ar, geochronology second theme was the impor- tance of time in understanding the origin of granites. Read later INTRODUCTION devoted an entire paper to this second theme, his “A Contemplation of Time in Plutonism” As recently as the early nineteenth century, the contro- (Read 1949). In it, Read made the intriguing suggestion that versy surrounding the origin of plutons was staged as a metaphorical battle between the Roman gods. The Death, I agree, is an important event to a man, Neptunists (championed by German pioneering geologist but his life is much more important to posterity Abraham Werner [1749–1817]) believed that granites— – and so it is with plutonic rocks. Besides, just along with all rocks—were precipitated from the oceans. as all members of a family do not usually die on According to the Neptunists, the granitic layer of the Earth the same day, so genetically related granites may was the oldest layer, upon which younger fossil-bearing complete their courses at widely different ages… layers were deposited. The Plutonists (originally proposed by Italian abbot and naturalist Anton Moro [1687–1764], In many ways, our drive to understand the pace of pluto- but ultimately championed by the great Scottish geolo- nism follows from this quote. Much of the modern debate gist James Hutton [1726–1797]) posited that plutonic circles around the questions, “How do plutonic magmas rocks had their origin in fi re and were crystallized from accumulate?” and “In the life and death cycle of a pluton, magmas. Ultimately, it was the recognition of dikes (tabular what event(s) are dated? How are those ages interpreted?” bodies of rock that crosscut older rocks) that vanquished the Neptunists. Plutonists successfully argued that these DATING PLUTONIC ROCKS rocks could not have been deposited from the sea, and Techniques for determining the age of plutonic rocks have Pluto, Roman god of the underworld (FIG. 1), won the day. focused in the past 20 years on the decay of uranium to Fittingly, understanding the rates of magma movement and lead (U–Pb geochronology), and the decay of 40K to 40Ar. accumulation in dikes and other tabular plutonic bodies Dating with Ar includes the modification of measuring remains key to understanding the pace of plutonism. 40K through the proxy of the isotope 39Ar, which can be During a 1947 conference organized by the Geological produced artificially from 39K in a nuclear reactor: thus, it is Society of America (Gilluly 1948), the origin of plutons an 40Ar– 39Ar (or simply Ar–Ar hereafter) age. The details of had evolved to a debate between a melt-dominated mecha- modern U–Pb and Ar–Ar chronology are beyond the scope nism (crystallization from a magma—magmatism) or a of this contribution (see for example, “One Hundred Years metasomatic mechanism with minor melt (the dominantly of Isotope Geochronology” [Elements v9n1]). However, solid-state alteration of preexisting rock—migmatism). The several key facts about the methods bear mention here. conundrum of granite formation arose because, unlike For plutonic rocks, the most commonly dated mineral volcanic rocks that can be witnessed to erupt as crystal– by U–Pb is that particularly robust time capsule, zircon (ZrSiO4). It can retain its daughter Pb at temperatures in excess of magmatic crystallization temperatures (>900 ˚C). 1 Department of Geological Sciences University of North Carolina Consequently, a zircon age is generally thought to reflect Chapel Hill, NC 27599-3315, USA the integrated magma history between the time zircon E-mail: [email protected]; [email protected] began to crystallize (the zircon saturation temperature, 2 New Mexico Bureau of Geology and Mineral Resources typically >800 °C) through the time when the magma 801 Leroy Pl, Socorro, NM 87801, USA becomes completely solid. The mineral titanite (CaTiSiO5) [email protected]  The effort to resolve the pace of plutonism began in earnest with the effort to resolve the room problem. Understanding how the solid earth moves in order to accommodate space for an intruding magma is clearly linked to knowing the rate at which the solid earth has to move. Also linked to the room problem, however, were entrenched ideas about the nature of magma chambers and the connection between plutonic and volcanic rocks. Because texturally similar plutonic rocks were thought to be frozen magma chambers capable of wholesale convection and mixing, and because large silicic eruptions were thought to require extensive chemical evolution in shallow crustal magma chambers, rapid assembly of such bodies was required by basic petrology tenets. It is fair to say that there is no consensus in the granite community about many of the issues, but a look into the room problem and plutonic– volcanic connections provides a good view into why it is critical to understand the pace of plutonism. THE ROOM PROBLEM During the late 1980s to early 1990s, there was a revival of interest in the room problem for granite plutons (Hutton 1988; Tikoff and Teyssier 1992). Most approaches to under- standing the problem focused on the structural mecha- nisms for moving rocks out of the way of the rising magmas and on the tectonic settings that might best facilitate FIGURE 1 Pluton (French for “Pluto”) as sculpted by Henri Chapu and photographed by MeI22. LICENSED UNDER CREATIVE space-making. One of the most straightforward settings COMMONS ATTRIBUTION 3.0. where links between the pace of magmatism and plate tectonics can be made is at mid-ocean ridges. There, new is also commonly used in U–Pb geochronology; however, oceanic crust is made by intrusion and eruption of magmas unlike zircon, titanite leaks daughter Pb down to tempera- at a pace dominantly determined by the spreading rate tures of ~680 °C (the closure temperature). The relatively of the ocean plates on either side of the ridge. On the low closure temperature of titanite is useful in evaluating continents, however, most plutons are formed where one the pace of plutonism because comparison of U–Pb zircon plate is subducted beneath another. Whereas the room ages to titanite ages from a single sample offers insight problem for gabbro plutons (formed under the oceans) was into the cooling rate of the pluton in the interval between easily solved and the pace of gabbro plutonism was well- >800 °C and 680 °C. determined, making room for granite plutons (formed in continents) and the pace at which that might happen were Dating with Ar–Ar can be done using any potassium- more problematic. A link had to be made between magma bearing mineral in a rock. For plutonic rocks, this intrusion rates and reasonable rock deformation rates in commonly includes hornblende and biotite. Because the collisional settings. daughter isotope in this decay scheme is a noble gas, it tends to diffuse out of crystals at magmatic temperatures. Fortunately, at this same time, there was rapid advance- Consequently, like U–Pb in titanite, the age determined ment in U–Pb analytical techniques. Prior to the late using Ar–Ar is the time since the mineral dropped (and 1980s, the analysis of zircon required large amounts of stayed) below its closure temperature. The closure tempera- the mineral and, consequently, large amounts of rock. tures of hornblende and biotite vary somewhat depending Zircon is an accessory mineral commonly comprising less on the composition and size of each mineral, as well as the than 0.05% of a granitic rock and, depending on the age rate of cooling, but are nominally 550 °C for hornblende of the sample, obtaining enough zircon for dating could and 350 °C for biotite. Like the comparison of U–Pb ages mean collecting 50 kg of rock or more. Thus, sampling a for zircon and titanite, comparison of Ar–Ar ages between pluton was often restricted to one sample per map unit, and hornblende and biotite offer insights into the cooling usually from easily accessible locations. However, in order history of the rock. It is ideal to combine U–Pb and Ar–Ar to understand the rates of magma accumulation, multiple ages for a sample to get a nearly complete thermal history samples are necessary. The development of cleaner labora- from zircon saturation through biotite closure. tory techniques and instrumentation capable of measuring smaller masses of daughter isotopes of Pb allowed for THE SIGNIFICANCE OF THE PACE single-crystal analysis, with commensurate reduction in rock mass to less than a kilogram and the possibility to Until recently, the rate of magma accumulation to form better date plutons. This led to advances in attempts to plutons received little attention. Generally, it was thought solve the room problem (and other problems). that magmas rose through the Earth’s crust as diapirs (picture the ascending blobs in a lava lamp) and ultimately Surprises from the Tuolumne Intrusive Suite crystallized into plutons, requiring ascent and accumula- (California, USA) tion rates that were faster than crystallization. During the early days of geochronology, dating a single sample from a One of the fi rst studies to suggest age variations in mapped pluton was considered sufficient to know the entire pluton’s plutons was done by Kistler and Fleck (1994) of the US age. Discrepancies between zircon U–Pb and hornblende Geological Survey on the Tuolumne Intrusive Suite—best or biotite Ar–Ar ages were presumed to result from slow known as the bedrock for most of Yosemite National Park cooling histories deep in the crust, or subsequent reheating in California (FIG. 2). This work combined ages that were that resulted in resetting the Ar clock. determined by several techniques (K–Ar, Ar–Ar, Rb–Sr and U–Pb) on a host of rocks and mineral phases. Focusing on just hornblende Ar–Ar ages (for which there are the accumulate challenged most of what was thought about most samples for direct comparison), the plutons of the pluton assembly. Most relevant to solving the room problem suite preserve a 6 My history from approximately 89 Ma to is the question, “How do magmas ascend and accumulate 83 Ma. This study included only three zircon U–Pb ages, to form plutons if not as large diapirs?” which were clustered around 88–86 Ma and were tighter The key to understanding the ascent and assembly of than the spread in Ar–Ar ages. The data also include biotite plutons may go back to the same observations that sunk the Ar–Ar ages that are younger than hornblende Ar–Ar ages by Neptunists. It is widely recognized that intrusion through up to 4 My. Taken together, these suggested the possibility fractures as tabular bodies—dikes and sills—is an efficient that the dispersed Ar–Ar ages reflect cooling rather than mechanism for moving granitic magmas through the crust crystallization. The results called for more zircon ages. (Clemens and Mawer 1992; Petford et al. 1993). Few would It is no accident that the Tuolumne Intrusive Suite has argue against the idea that the sheeted dike complexes been the focus of understanding the dynamics of plutons that characterize the plutonic rocks of the oceanic crust and magma chambers for decades. Glaciers have carved a comprise a significant volume of the mafic crust formed spectacular 3-D exposure of the plutons, leaving almost in that extensional setting. Extrapolating that result to nothing to the imagination (FIG. 2). Thus, the suite provides granitic rocks has been significantly more controversial an ideal opportunity to unravel the pace of plutonism. because, for the most part, granitic plutons do not preserve The fi rst comprehensive zircon U–Pb investigation of the dike-like structures. The Main Donegal Granite of north- Tuolumne Intrusive Suite was completed by Coleman et west Ireland is perhaps the best known example of a pluton al. (2004), who published U–Pb zircon results for eleven thought to be assembled by the incremental intrusion of samples from across the suite. These authors concluded that tabular bodies (Pitcher and Berger 1972). There, preserva- the magmas had accumulated and solidified over a 10 My tion of numerous screens of the rocks into which the pluton interval between 95 Ma and 85 Ma. Additionally, they were intruded (wall rocks) defi nes a series of dike intrusions. able to document that an individual pluton, the Half Dome Bartley et al. (2008) proposed that such features were more Granodiorite, had assembled over at least 4 My. Subsequent common than appreciated and that most granitic plutons studies confi rmed these results (e.g. Memeti et al. 2010). could form through the amalgamation of many dikes and/or sills. However, they suggested that preservation of contacts between increments was most likely to be near the contacts between the earliest intruding granites and their nongranitic wall rocks. They noted that the earliest intrusions have the greatest contrast between wall-rock and magma tempera- ture, enhancing the likelihood that they would cool quickly and preserve contacts. Furthermore, with the recognition that in an incrementally assembled pluton, early intruding granites become the wall rocks for later intruding granites, the lack of nongranitic wall rocks screens in the youngest portions of granitic plutons is not surprising. Additionally, they argued that an incrementally assembled granite pluton spends significant intervals at or near magmatic temperatures, helping to erase contacts between increments. This is less probable in an oceanic setting where intrusive depths are generally shallower and the magmatic tempera- tures are significantly higher than those FIGURE 2 Granitic rocks of the Tuolumne Intrusive Suite, Yosemite National Park (California, USA). Essentially, for granites. That hypothesis can be tested with thermal everything in the photo that is not green or blue is a plutonic rock modeling and by taking advantage of the contrast in that accumulated by the incremental addition of magmas over closure temperatures of the U–Pb and Ar–Ar systems. millions of years. The double peak just left of center is Cathedral Peak. PHOTO BY A LLEN G LAZNER Cooling Rates Comparing temperature–time paths constructed from dating multiple minerals in individual samples with those predicted from thermal models can help in understanding How Do Plutons Form? the assembly of plutons. Incrementally assembled plutons Work in the Tuolumne Intrusive Suite revealed that some slowly add heat to the system over the duration of pluton individual map units—rocks once thought to represent assembly, but individual pulses initially cool quickly. single injections of magma, or multiple upwellings within a Thus, high-temperature chronometers may yield a tight single large injection of magma—accumulated and crystal- age cluster and low-temperature chronometers may yield a lized over millions years. Work on many plutons now spread in ages (FIG. 3). The Ar–Ar data from the Tuolumne demonstrates that the results from Yosemite are typical of Intrusive Suite suggest that at least some of those plutons plutons from a variety of times and tectonic settings, and spent millions of years at temperatures above ~350 °C. incremental assembly of plutons is now widely accepted. Subsequent studies have combined U–Pb zircon and titanite That plutons are amalgamated from multiple pulses of data, or Ar–Ar hornblende and biotite, to quantify the magma is an old idea (Pitcher 1979); however, the inferred temperature–time histories of plutonic rocks. small volume of pulses and long times during which they  At the high-temperature end of the spectrum, it is useful where secondary titanite is likely to form (FIG. 3). Davis to compare U–Pb zircon saturation ages with U–Pb titanite et al. (2012) were able to show that plutons in the Sierra closure ages. These can provide an estimate of the time Nevada batholith of California recorded up to a 5 My differ- between temperatures in excess of 800 °C and 680 °C. Work ence between hornblende and biotite ages. Moreover, they in the Adamello batholith of Italy reveals indistinguish- noted that biotite ages were more clustered than higher- able ages for the youngest zircon and magmatic titanite temperature (zircon) ages and suggested that this clustering from individual samples. This is consistent with rapid resulted from a relatively rapid decrease of temperature cooling of the magmas from zircon crystallization to below once pluton accumulation had stopped. 680 °C, respectively. Curiously, the Adamello batholith also includes a second population of titanite (nonmagmatic) The Crux is the Flux intergrown with chlorite that probably grew at lower Knowing the age range, potential intrusion and assembly temperatures (Schoene et al. 2012). These titanite crystals mechanisms, and cooling rate brings us closer to being are significantly younger than the magmatic crystals and able to estimate the flux of magma into plutons. The last suggest that the plutons remained (or were reheated) above remaining variable (i.e. question) requires some subjec- 300 °C to 400 °C for hundreds of thousands of years after tivity: “What is the volume of the pluton?” Whereas the magma intrusion. surficial area of a pluton can be known fairly accurately, The Ar–Ar system is, perhaps, better suited for exploring the evaluating what happens in the third dimension is more thermal history of plutons at, or just below, solidification problematic and two end-member approaches have been temperatures (Zimmerer and McIntosh 2012). The closure taken. The first recognizes that plutons are typically tabular temperature of hornblende is near that of granite solidifi- or wedge-shaped and that the thickness can be roughly cation, and the closure temperature of biotite is near that predicted by the area (Cruden and McCaffrey 2001). Thus, a pluton with a mean width of 20 km may be reasonably estimated to be approximately 3.5 km thick—a flat disk. 1000 instantaneous A The second recognizes that the contacts between wall rocks and plutons tend to be steep, and plutonic rocks of broadly intrusion the same composition may extend for tens of kilometers 800 below surface exposures (e.g. Lipman 2007). Thus, that zircon saturation (zs) same 20 km wide disk may be a 20 km tall cylinder. Neither of these approaches accounts for loss of volume to erosion. magmatic titanite closure (tc) Regardless, application of the fi rst approach yields magma 600 volume estimates that can be an order of magnitude less than those obtained by using the second approach. hornblende closure (hc) Another complicating factor for estimating magma flux biotite biotite temperature (˚C) into plutons is recognition that some of the magma might 400 breakdown closure have erupted. Given that exposures of demonstrably related secondary (bc) plutonic and volcanic rocks are sparse, knowing the volume titanite (st) of magma that a pluton lost to an eruption is difficult, 200 if not impossible. Systematic study of intrusive:extrusive 1000 ratios suggests a range of between 3:1 and 10:1 (White et B incremental al. 2006), once again adding a factor of three to the uncer- tainty in knowing the magma volume. Quantifying this 800 intrusion relationship is presently a hot topic of research. zs A fi nal complication in calculating magma flux is that the tc number of samples dated is fi nite. Consequently, the time interval of magma accumulation is always a minimum. 600 At some level, this mitigates problems introduced with hc knowing the volume. In our work, we have chosen to estimate the flux for the exposed volume. That is, if we have sampled a 20 km wide pluton over 1 km of relief, 400 duration of incremental we calculate the volume of a disk with a diameter of 20 st bc km and a height of 1 km. We recognize that this is likely emplacement to be an underestimate of the flux for all the reasons outlined above, but it avoids making the assumptions that 200 unseen rocks are present and are the same age as exposed 0 1 2 3 4 rocks. Despite these difficulties, however, and considering reasonable variations in the assumptions made, estimated time (My) magmatic fluxes are consistently between 10 −3 to 10 −4 km3 y−1 for plutons with volumes estimated to be from 101 to FIGURE 3 Examples of how chronometers can be used to decipher the pace of plutonism using HEAT 3-D 10 4 km3 (FIG. 4). models by Ken Wohletz. Both models show a pluton intruded at a depth of 5 km with a final diameter of 30 km and a final thickness The apparent flux of magmas into plutons is slow. It is of 10 km. Colored horizontal bars show the range of ages poten- orders of magnitude slower than fluxes required to sustain tially recorded by different chronometers. These are at the same a dominantly liquid magma chamber, and much more temperature in each panel. (A) Instantaneous emplacement of the entire volume. The pluton cools slowly and chronometers yield a in line with known tectonic strain rates. With geodetic range of ages. (B) Incremental assembly of the same volume as (A) imaging techniques, it is possible to detect inflation of over a 3 My period. Note that zircon, magmatic titanite and (poten- active volcanoes and estimate the flux of magma beneath tially) hornblende ages will overlap. However, many hornblende them. These estimates tend to be higher than long-term and biotite ages as well as titanite growth ages will be significantly younger. Many variables will play a role in determining a pluton’s estimates for plutonic systems; however, long-term pluton thermal history and different chronometers may yield different fluxes likely mute higher short-term, observable fluxes results, depending on those variables. under modern volcanoes (Pritchard and Gregg 2016 this  0.1 of 0.01 km3 y−1 needed to sustain a magma chamber that volcanic rocks leads to supereruptions. The comparison between small plutonic rocks eruptions and plutons shows the opposite—the paces of small volume volcanism and plutonism are comparable magma flux (km3/y) 0.01 (FIG. 4). An excellent example of the disparity between large plutons and supereruptions comes when comparing zircon ages for the Mount Givens Granodiorite (Sierra Nevada, USA) and 0.001 the Fish Canyon Tuff (Southern Rocky Mountain Volcanic Field, USA). These rocks are comparable in composition and volume (with all the caveats inherent in estimating volumes). However, the Mount Givens Granodiorite 0.0001 preserves a record of at least 7 My of zircon growth, in contrast to less than 0.5 My of zircon growth recorded by the Fish Canyon Tuff (FIG. 5). The contrast in ages preserved in this pluton and tuff 0.00001 suggests either dramatically different paces of magma 10 100 1000 10000 accumulation or significant differences in zircon crystalli- zation and/or preservation between the two environments. volume (km3) A commonly invoked explanation for the discrepancy is efficient zircon dissolution prior to tuff eruption (Miller FIGURE 4 Magma volume and flux estimates for granitic plutons et al. 2007). In this scenario, prior to eruption, the tuff (blue squares) and volcanic rocks (red circles). Deposits noted in the literature as supereruptions fall in the gray existed as a slowly accumulated (pluton-pace) shallow shaded area. One other high volume field with low flux includes magma body that, if completely crystallized, would have multiple units and is therefore not considered a supereruption. preserved millions of years of zircon growth. The reheating Supereruption magmas appear to accumulate orders of magnitude that triggered eruption is proposed to be sufficient to take faster than comparable volumes of plutonic rocks. M ODIFIED FROM M ILLS AND COLEMAN (2013) AND FRAZER ET AL. (2014). the magma body back above the zircon saturation tempera- ture for a long enough interval that all of the preexisting zircon dissolved, and the narrow time recorded by the zircon in the tuff represents new growth. The difficulties issue). Slow magma accumulation rates age range ~7 My were recognized by geophysical techniques 98 Pb/238U age (Ma) for Fish Canyon Tuff data (red) 12 samples, n = 54 for oceanic magma systems (Detrick et al. N Pb/238U age (Ma) for Mt. Givens data (blue) estimated 1990) around the same time that geologists volume ~4,500 km3 studying granites began earnestly searching 97 for a solution to the room problem. As in mid-ocean ridge settings, geophysical Mt. Givens a single sample analysis of continental magma systems has 96 Granodiorite is highlighted yet to reveal the presence of large volumes CA, USA in light blue 10 km of liquid magma beneath Earth’s largest active volcanoes. Even Yellowstone—with 95 29 a record of supereruptions, and often cited as having potential for future supererup- 94 tions—is sitting above partially molten 28 rock with only 2–15% melt present (Farrell et al. 2014; Huang et al. 2015). 93 N age range ~0.44 My La Garita 1 sample, n = 24 PLUTONIC–VOLCANIC ROCK 92 estimated volume ~5,000 km3 caldera CONNECTIONS CO, USA The apparently slow pace of magma accumulation in plutonic environments 91 10 km is at odds with the dominant paradigm 206 linking plutonic and volcanic rocks. Smith inset 100 µm 90 206 (1960) noted the likely “kinship” between large volcanic ash flows and shallow zircon ages in order from youngest to oldest plutonic rocks. This relationship was further developed by many authors, and FIGURE 5 Comparison of single grain U–Pb zircon analyses from Lipman (1984) championed a direct link between the two the Mount Givens Granodiorite (blue; Frazer et al. 2014) and the Fish Canyon Tuff (red; Wotzlaw et al. 2013). on the basis of field, petrologic, and chemical relationships. Horizontal axis sequences analyses from youngest to oldest for one The spatial connection between plutons and huge volcanic Fish Canyon Tuff sample and twelve Mount Givens samples distrib- ash-flows is undeniable, but the calculated magma fluxes uted throughout the intrusion. These two rock units are comparable for the two are quite different. A compilation of data for compositionally and likely comparable volumetrically. The scale for both samples is the same, but note the difference in absolute age plutons and ash-flows shows that most magmas that lead to for the Mount Givens Granodiorite (left axis) and Fish Canyon Tuff supereruptions (> 450 km3) are generated at rates between (right axis). Maps show outlines for the Mount Givens pluton and 1 and 2 orders of magnitude faster than similarly sized the La Garita caldera (the source of the Fish Canyon Tuff) at the plutons (FIG. 4). This tipping point for supereruptions seen same scale. INSET: A typical zircon from the Mount Givens Granodiorite. The Mount Givens Granodiorite records an age range from the geochronologic data is in accord with the thermal that is over an order of magnitude greater than that for the Fish models of Annen (2009), which predict a minimum flux Canyon Tuff. with this scenario are that zircon dissolution is inefficient ered. Whereas the slow pace of plutonism is not at odds and not likely to happen at the timescales for remobili- with a direct connection to geologically common small- zation and eruption (Frazer et al. 2014). Direct evidence volume volcanic eruptions, it is apparently far too slow against the efficacy of zircon dissolution in tuffs includes to permit a direct connection between most plutons and the common preservation of zircons inherited from slightly (fortunately!) rare massive supereruptions. older (