V52B-01 10:20h
Not so fast: Contrasting timescales of crystallization and magma storage beneath the Aleutian Island arc
U-Th isotope data from Pleistocene-Recent basaltic to rhyolitic lavas and their phenocrysts from Seguam Island, Aleutian Island arc, reveal a monotonic evolution consistent with radiogenic ingrowth of $^{230}$Th in a long-lived magma reservoir between 142 and 9 ka. Internal U-Th mineral isochrons from six lavas and tephras are indistinguishable from their eruption ages as constrained by $^{40}$Ar/$^{39}$Ar dating, which implies a short period of crystallization. These results can be reconciled if small batches of magma are repeatedly extracted from a deep, thermally buffered, basaltic reservoir and rapidly cool and differentiate in shallow, ephemeral chambers or conduits immediately prior to eruption. Cone collapse and caldera formation at 9 ka correspond to influx of new basaltic magma into the deep reservoir. Our interpretation, that integrates petrologic, geophysical, and now U-series observations from Seguam, is at odds with the conclusion that magma residence times in the crust are short ($<$ 10$^{3}$ years) and that most variation in the transit time between partial melting and eruption originates in the mantle wedge, a widely held view that is based mainly on U-Th-Ra series isotope data from whole rocks and Sr and Mg diffusion profiles of plagioclase in but a few arc magmas. If correct, our model implies that bodies of basaltic magma can reside undisturbed within the lower crust of an island arc for $>$ 10$^{5}$ years. Suppression of crystallization for 10$^{5}$ years requires minimal heat loss; suggesting either storage in unusually hot wall rocks, or heating from still deeper magma ponded below this reservoir. In any case, the U-series data from Seguam vividly illustrate the importance of a suite of well-dated lavas spanning a significant period of the eruptive history of a single volcano when attempting to constrain rates of magma ascent, crystallization, and differentiation.
V52B-02 10:35h
Timescales and Mechanisms of Batholith Construction, Coastal Cordillera, Northern Chile, From Precise U-Pb Zircon Ages and Regional Geochronological Data
Analysis of $>$200 K-Ar and Ar-Ar ages on intrusive rocks combined with new single zircon U-Pb analyses on individual composition layers within three plutonic complexes of the Coastal Cordillera batholith of N Chile define distinctive patterns that provide insight into the upper crustal growth of a Mesozoic batholith in space and time. Triassic-Jurassic plutons, which dominate the western part of the arc display a diffuse age distribution pattern with no temporal zonation, indicating a period of widespread extension and magmatism. Cretaceous plutons young systematically west to east and define an average magmatic migration rate of about 500 m/Myr. In detail 20-30 km wide Cretaceous plutonic complexes remain fixed in space for ca. 3 Myr, then the locus of magmatism jumps 2-10 km east during a 1-3 Myr time gap. The 2-3 Myr emplacement times for individual plutons is confirmed by precise U-Pb zircon ages on subhorizontal, 100-500 m thick layers that make up the intrusions. The Carrizal Bajo pluton grew in ca. 2 Myr by initial injection of a diorite sheet dated at 207.7+/-1.3 and 207.5+/-0.5 Ma that was split by a 206.2+/-0.3 Ma granodiorite sheet and fed by a sheeted dyke complex with units dated at 208.5+/-0.5 Ma and 208.2+/-0.2. The Infiernillo pluton yielded ages from top to bottom of 131.4+/-0.2, 128.7+/-0.2, and 129.0+/-0.2 Ma and layers in La Borracha pluton were dated at 116.3+/-0.2, 114.5+/-0.2, and 113.0+/-0.1 Ma. Both plutons grew from the top down by floor depression over ca. 3 Myr. The two ages from the lower unit of Infiernillo indicate incremental filling from bottom to top during progressive, asymmetric fault-assisted floor subsidence over $<$600 Kyr. Both the vertical growth of plutonic complexes and their west-east localization within the batholith are attributed to reactivation of strands of arc-parallel faults during the Cretaceous. As magmatism migrated eastward across the arc, melt transport and pluton growth was controlled and stabilized for ca. 3 Myr intervals by the nearest crustal-scale fault. Further migration eventually led to the capture of the magma transport system by the adjacent fault strand, and the observed across-arc jumps in intrusive ages. West-to-east migration of magmatism and concurrent regional transtension are attributed to the interaction between westward slab roll-back and the dynamics of upper plate extension. The spatial-temporal distribution of upper crustal plutons in the batholith was controlled by the migration rate and direction of lower crustal melting and the horizontal spacing of crustal-scale faults.
V52B-03 10:50h
U-series Isotopic Evidence for Remelting of Kilauea Volcano's Mantle Source Region During the Puu Oo Eruption
Partial melting of the mantle to produce ocean-island and mid-ocean ridge basalts is generally assumed to be a steady-state process in which the continuous flux of upwelling solid into the melting region is balanced, after melt generation and extraction, by the permanent removal of the residue from the system. Here we present high-precision measurements of the U-series isotope abundances of lavas from the current Puu Oo eruption of Kilauea Volcano to investigate the dynamics of partial melting within the actively upwelling Hawaiian mantle plume. Our results show that Puu Oo lavas display small, but significant, temporal decreases in their 230Th-238U and 226Ra-230Th disequilibria from 2.5 to 1.4% excess 230Th and 14 to 12% excess 226Ra (relative to our +/-2 sigma analytical uncertainty of 0.3%). These changes in the (230Th)/(238U) and (226Ra)/(230Th) activity ratios of the lavas correlate systematically with larger decreases in the abundance ratios of highly versus moderately incompatible trace elements from a maximum of 23% for Ba/Yb to a minimum of 4% for Nd/Sm. These geochemical signatures cannot be explained by an increase in the degree of partial melting during the eruption. Instead, this volcano has tapped a mantle source component that was depleted of incompatible elements by "recent" prior melting. The systematic correlation between the 230Th-238U and 226Ra-230Th disequilibria and these trace element ratios shows that this depletion must have occurred within several half-lives of 226Ra ($<$8 kyr ago) because 226Ra will return to secular equilibrium with 230Th on this time scale. Furthermore, modeling indicates that the depletion of the volcano's source region may have occurred during the Puu Oo eruption itself. Thus, we propose that relatively fertile mantle within the Hawaiian plume may have been progressively exhausted during the eruption, allowing the increasingly depleted, and thus, more refractory, residue to continue melting. Our results suggest that the dynamics of melt extraction from the mantle may have a dramatic effect on the geochemical signatures of basaltic lavas.
V52B-04 11:05h
Laser Ablation Pb Isotopes as Tracers of Rhyolite Evolution and Crystal Residence Times, Long Valley, CA
At Long Valley (LV) the Sr isotopic composition of phenocrysts from the caldera-related Bishop Tuff (BT) imply that some crystals are $>$1 Myr older than eruption. In our studies of LV rhyolites, we find that zircon U-Pb ages are generally only hundreds of ky older than eruption, resulting in weak to strong disparities with the Rb/Sr mineral-melt model ages. Since Rb/Sr model ages depend on host glass compositions, these differences could result from crystal-liquid exchange between magmas or, given the relatively low Sr contents ($<$2 ppm) of the glasses, post-crystallization magma modification. To test the isotopic affinities of the various phenocryst populations to both their host melts and to secular changes in the Pb isotopic composition of the LV magma system, we performed {\it in situ} Pb isotopic analyses of feldspars and their host glasses. Lead contents are relatively high (10's ppm) in rhyolitic glass and feldspar and Pb isotope compositions are not expected to vary appreciably due to intrinsic contamination or radiogenic ingrowth ($^{207}$Pb/$^{206}$Pb decreases $\sim$0.00006 per Myr), but may vary if the contributions from different sources change with time. Lead isotopes were obtained for 8 host glasses and 113 feldspar phenocrysts by UV (213 nm) laser ablation MC-ICPMS. We use a sample-standard comparison approach for mass fractionation correction (rather than Tl-doping) during analyses. Measurements are age-corrected to initial ($^{207}$Pb/$^{206}$Pb, $^{208}$Pb/$^{206}$Pb, $^{206}$Pb/$^{204}$Pb) values. With this approach we obtain accurate results with an external precision of \pm0.00016 (2 s.d.) for the $^{207}$Pb/$^{206}$Pb ratio of the NIST SRM 612 glass, comparable to TIMS double-spike methods. The Pb isotope data for feldspars and their host glasses lie along a mixing line between regional mantle ($^{207}$Pb/$^{206}Pb$=0.81366) and crust ($^{207}$Pb/$^{206}$Pb=0.82698) and span $\sim$10 % of that isotopic difference. Differences between the Pb isotope compositions of older ($\sim$1.7-2.2 Ma) precaldera Glass Mountain (GM) rhyolites (e.g., $^{207}$Pb/$^{206}$Pb$_{o}$=0.81856) and younger LV rhyolites, including the BT ($^{207}$Pb/$^{206}$Pb$_{o}$=0.81745), are found. Considering also existing Nd and O isotope data for the same rhyolites, the Pb isotope data record a progressively greater mantle contribution to the LV rhyolites over time. Most feldspars have Pb isotope compositions that are broadly similar to their host glasses. Notably, the strong contrast between the isotopic composition of the BT and the older GM magmas is mirrored in the feldspars they contain. Thus, there is no indication that BT feldspars grew from older GM-like magmas. A number of grains from several GM rhyolites do, however, have distinct Pb isotopic compositions (e.g., $^{207}$Pb/$^{206}$Pb $>$+0.0005) that indicate contamination from a granitic country rock or possibly LV-related intrusions. Moreover, the Pb isotope heterogeneity within individual GM rhyolites suggests that some magmas experienced open system processes that may lead to inaccurate model ages. For this reason, the absolute age constraints we have obtained by U-Pb dating are better guides to interpretations about the timescales of magmagenesis than Rb/Sr model ages. Collectively, the distinct Pb isotope signatures for individual rhyolites and their feldspars imply that LV volcanism did not erupt from a single long-lived magma chamber but rather tapped a number of transient and distinct magma bodies.
V52B-05 11:20h
Thermal Modeling of Large Composite Plutons
Field and geochronologic evidence indicate that large plutons commonly amalgamate from many small intrusive increments [e.g., {\it Glazner et al., 2004, GSA Today}; {\it Coleman et al., 2004, Geology}]. To investigate the thermal consequences of this process, we model the growth of composite intrusions using the program HEAT. HEAT uses a finite-difference scheme to track transport and storage of heat, and resulting temperature variations, in a 2-D or 3-D intrusion and its wall rocks. Heat advected by wall-rock displacement is conserved kinematically by distributing the heat of intruded rock into adjacent mesh locations following a system determined by the aspect ratio of each intrusive increment. Modeling to date focuses on laccolithic plutons formed by stacking of sills. In a typical model, an 1100°C sill of intermediate-composition magma with a solidus of 750°C, 100-m-thick and 10-km-wide, is emplaced every 10 ka (10 mm/yr vertical inflation rate) for 600 ka to form a tabular pluton 6 km thick. Model runs to date have yielded at least three intriguing results. 1) Stacking of sills from bottom to top produces higher sustained temperatures than stacking from top to bottom. Over the first half of a 600 ka emplacement time, a bottom-up intrusion completely solidifies between increments but, during the latter half, a partial-melt zone becomes a steady-state feature. Dimensions of the partial melt zone vary through a 10 ka intrusive cycle, but the thickness reaches nearly 2 km by the time the last sill is emplaced. An otherwise identical top-down intrusion solidifies completely after each intrusive increment until the last 35 ka, when a small zone of partial melt persists between increments. The large composite complexes of the Sierra Nevada such as the Tuolumne and Whitney intrusive suites appear to have grown from their tops down, but bottom-up plutons also have been widely reported. 2) The partial-melt zone in a bottom-up intrusion mainly forms below each new intrusive increment. Leucocratic partial melt formed under a newly emplaced intermediate-composition sill thus may ascend buoyantly through the sill, causing in situ magma mixing and mingling. In contrast, partial melting of wall rock in top-down intrusions occurs above newly emplaced sills, making in situ mixing and commingling less likely. 3) Although intrusive increments generally solidify in 1 - 10 ka, the resulting rocks remain at temperatures of 400-600°C on time scales of 10$^{2}$ - 10$^{3}$ ka. This result is consistent with measured cooling histories of the Tuolumne Intrusive Series [{\it Coleman et al., this volume}]. Widespread subsolidus textural annealing thus appears likely and this may be responsible for gradational or otherwise cryptic contacts commonly observed in the field between intrusions that yield measurably different U/Pb zircon ages.
http://www.ees1.lanl.gov/Wohletz/Heat.htm
V52B-06 11:35h
Timescale for Incremental Construction of the Silurian Vinalhaven Intrusive Complex, Coastal Maine, USA
Subvolcanic plutons, such as the Vinalhaven intrusive complex, preserve field evidence for incremental growth from ephemeral chambers due to crystallization, replenishment, rejuvenation and accumulation. In the Vinalhaven intrusion, field relations indicate that the lower portion of the intrusion, which is characterized by a layered section of gabbro-diorite and cg granite, is stratigraphically older than the upper portion of the intrusion which is characterized by homogeneous, cg granite with a wide variety of schlieren structures (Wiebe and Hawkins, this volume). To evaluate the distribution of time within this stratigraphic framework, we determined high-precision U-Pb ages (ID-TIMS) on single crystals and single crystal fragments of zircon from four samples in the intrusion. Two samples of cg grained, one collected from the layered section near the base of the intrusion and the other along the western margin of the intrusion along strike from the layered section, yield concordant U-Pb crystallization ages of 421.5 $\pm$ 0.4 Ma and 421.3 $\pm$ 0.4 Ma, respectively. A third sample of cg granite collected near the top of the intrusion yields a preliminary concordant U-Pb crystallization age of 419.8 $\pm$ 0.7Ma. A fourth sample, collected from the fg granite in the core of the intrusion, a body we interpret as a silicic replenishment emplaced during the middle portion of the growth history, yields a concordant crystallization age of 420.3 $\pm$ 0.4 Ma. These four concordant U-Pb crystallization ages are consistent with the field relationships and indicate that, despite its moderate size (about 80 sq km of exposure) and shallow level of emplacement, the Vinalhaven intrusion was constructed over a nominal interval of 1.7 m.y. Inheritance patterns in the zircon populations of these samples may provide additional insights into the construction of the intrusion. Inherited zircon is difficult to avoid in the sample of fg granite; numerous grains and grain fragments yield discordant U-Pb dates with Pb-Pb dates ranging from 1465 to 460 Ma. These ages extend well beyond the ages of the exposed country rocks, and the inherited zircon crystals were probably derived from the crust beneath the intrusion. In contrast, inherited zircon is less common in the three samples of cg granite and the dates they yield are no older than 427 Ma, well within the age range of the exposed country rocks. Moreover, two zircon crystals from the youngest granite overlap in age (weighted mean 421.3 Ma) with the samples from the lower portion of intrusion. Although these inheritance patterns may reflect sampling bias, we think it is more likely that they reflect magmatic processes during pluton growth. If so, we suggest two implications. First, we see field evidence that fg silicic dikes (silicic replenishments) disrupt crystal mush as they intrude. Thus, a recently replenished chamber contains zircon crystals remobilized from older granite mush within the intrusion. If such grains erupted, they would yield unrealistically long residence times. Second, if fresh inputs of magma that replenish a chamber carry inherited zircon, then granite mush accumulating below that chamber could preserve a `layer' of cg granite that is relatively enriched in inherited zircon. Such layers could be used as stratigraphic markers of successive silicic replenishment events in the solidified pluton.
V52B-07 11:50h
Short Timescales of High-silica Rhyolite Generation in the Mono-Inyo Craters Indicated by U-Th Isotopic Disequilibrium
Uranium and Th isotope measurements for some of the youngest basaltic- to rhyolitic lava flows associated with the Long Valley caldera and Mono-Inyo Craters indicate that the lavas have pronounced U and Th isotopic disequilibrium. The (238U/232Th) activity ratios range from 0.69 to 1.23, and the (230Th/232Th) activity ratios range from 0.80 to 1.03. The majority of the lavas plot to the right of the equiline, which corresponds with a subduction zone influence on the U and Th isotope compositions. A basaltic inclusion from the Mono Inyo dacite (age = 18 Ka, epsilon-Nd = +2) is one of the samples farthest to the right of the equiline, which indicates that the basaltic magmatism associated with the formation of the silicic magma system is subduction-like. In contrast, an older (ca. 100 Ka) trachybasalt lava from the Long Valley north moat (epsilon-Nd= -3) plots to the left of the equiline. We interpret these differences as reflecting lithospheric sources for the small-volume alkalic lavas and asthenospheric (subduction-affected) sources for the larger volume silicic system. The preservation of strong U-Th isotopic disequilibrium in high-silica lavas clearly indicates that the rhyolites do not have long magma chamber residence times as has been inferred for other silicic systems in western North America. This observation is in accord with the young age of the Mono Craters (less than about 18 Ka) and suggests that this system has not yet stored up a large amount of silicic magma that could feed an eruption of Bishop Tuff size.
V52B-08 12:05h
Establishment and Evolution of a new Silicic Magma System North of Yellowstone Caldera: Geochronology, Geochemistry and Petrographic Relationships of Extracaldera Basalts and Rhyolites in the Norris-Mammoth Corridor
Basalts and rhyolites erupted north of the Yellowstone Caldera following collapse at 640 ka. The geochronology, petrographic relationships and geochemistry of these rocks record the development of a new rhyolitic magma system in the Yellowstone Plateau Volcanic Field. Basaltic lavas were erupted from at least seven shield volcanoes and a cinder cone between 358 ka to 209 ka (40Ar/39Ar eruption ages). Lavas for each center are associated with separate magma batches derived by melting of asthenospheric mantle and incorporating differing amounts of lithospheric mantle. For example, the Swan Lake Flat basalt erupted from four centers and displays decreasing 87Sr/86Sr and increasing $\epsilon$Nd with decreasing age (87Sr/86Sr from 0.7062 to 0.7053, $\epsilon$Nd from -2.5 to +0.18). Trace-element concentrations decrease (e.g., Ba decreases from OIB levels to 0.1 OIB). This change is modeled by a larger MORB and a smaller lithospheric component in the younger lavas. These patterns are similar to those seen prior to previous caldera-forming eruptions at Yellowstone. For example, just before the eruption of the Huckleberry Ridge tuff at 2.2 Ma, the Hepburn Mesa basalt evolved from $\epsilon$Nd -6 to -1 signifying an increased input of asthenospheric mantle derived basalt just prior to caldera formation. Aside from two older rhyolite lavas erupted at 358 ka and 526 ka which are spatially and chemically distinct, all extracaldera rhyolites have characteristics which suggest they are derived from an evolving silicic magma system of substantial longevity ($>$326 ka to 80 ka). Eruptions occurred in an area ~5-7 km wide by 22 km long (north-south). Mingled lavas consisting of andesite enclaves in rhyolite characterize the early magma system from 316 ka to 263 ka, contemporaneous with Swan Lake Flat basaltic volcanism. Andesite enclave chemistry can be modeled as mixing of Swan Lake Flat basalt with rhyolite from mingled lavas. 206Pb/238U zircon ages from these early rhyolites indicate mean magma residence ages of 30-40 ka, with some zircons exhibiting simple magmatic oscillatory zoning (CL images) having ages 80-140 ka older than 40Ar/39Ar eruptive ages. Subsequently, high-silica rhyolites exhibit Sr, Nd and Pb isotope systematics consistent with derivation from a common source having subequal mantle and crustal contributions. 226 ka to 134 ka porphyritic rhyolites record progressive changes in chemistry consistent with simple FXL of modal phenocryst phases. From 118 ka to 80 ka rhyolites are sparsely porphyritic to aphyric and exhibit a shift to significantly less evolved trace element compositions followed by evolution consistent with renewed FXL. 206Pb/238U zircon ages from younger rhyolites indicate magma residence times of up to ~100 ka prior to eruption. Identification of one mingled lava in this younger group suggests renewed/enhanced basaltic input into the crustal magma system. Eruptions of associated basalts and rhyolites between 358 ka and 80 ka north of Yellowstone caldera record establishment of a new rhyolitic magma system at the leading edge of the melting anomaly, and may represent the initial stages of a 4th caldera cycle. Early input of basaltic magmas into the crust established this silicic magma system prior to 326 ka, and undoubtedly played a fundamental role in driving it for the subsequent ~250 ka.