V31H-01 INVITED
Ascent, Degassing, and Crystallization of Water-Rich Mafic Magmas in the Trans-Mexican Volcanic Belt: A Melt Inclusion Perspective
Mafic cinder cones are common in subduction-related volcanic arcs. Such cones exhibit a wide range of eruptive styles and are often violently explosive, but the mechanisms driving explosive cinder cone eruptions are still poorly understood, as is the nature of their underlying plumbing systems. To investigate magmatic processes beneath cinder cones, we have analyzed major elements and volatiles in olivine-hosted melt inclusions from 14 cones in the central and western Trans-Mexican Volcanic Belt (TMVB). Our samples include the historic eruptions of Paricutin and Jorullo and the highly potassic minette and basanite cinder cones in the Colima Graben. Melt inclusions from each cone have variable H2O (typically <1 wt% to maximum values of 4 to 5.5 wt%) and CO2 (<50 to 5000 ppm), corresponding to crystallization pressures of <100 bars to ~5 kb. This indicates that olivine crystallized from variably degassed melts over a range of depths extending from the mid- crust to near-surface, probably within a network of dikes (±sills). The highest CO2 contents (≤5000 ppm) are found in the potassic minettes and basanites from the western TMVB, whereas values ≤1800 ppm are typical of calc-alkaline basalts and basaltic andesites. Modeling of major element variations and comparison to phase equilibria demonstrate that polybaric olivine crystallization beneath these cones was driven by H2O loss from melts during ascent. A puzzling feature of the inclusion data is that CO2 does not decrease as rapidly with degassing as predicted by degassing models. A likely explanation involves fluxing of magma in the conduit system with CO2-rich gas released by degassing magma at greater depths. Decreases in both olivine crystallization depths and the vigor of gas fluxing over time in longer lived cinder cone eruptions reflect evolution of the conduit system and deeper storage reservoir.
V31H-02
Using H2O and trace element ratios to produce a spatial map of magmatic H2O contents throughout the Trans-Mexican Volcanic Belt
Along with fluids, trace elements are released during dehydration of subducting sediment and altered oceanic crust. Large ion lithophile elements are typically fluid mobile, and thus may be used as tracers for fluid fluxing. We used melt inclusion H2O and trace element data from nine cinder cones across the subduction-related Michoacan-Guanajuato Volcanic Field (MGVF) of central Mexico to assess the fluid mobility of trace element species. We found correlations between H2O and Sr/La, Ba/Nb, Ba/Y, Pb/Y, Sr/Ti, suggesting that Sr, Ba, and Pb are present in fluids released from the downgoing slab. Additionally, we used regression lines for these correlations to estimate magmatic H2O for cinder cones across Mexico. We have applied the Sr/La and Sr/Ti relationships to the extensive dataset of cinder cone lava and scoria analyses from the MGVF by Hasenaka and Carmichael (1985). In order to see 2-D spatial patterns in H2O across the MGVF, we plotted the localities and the calculated H2O contents on a digital elevation model of Mexico. Initial results from this modeling show that, like our melt inclusion data, magmatic H2O contents are generally high (>3 wt%) across a broad region from the volcanic front to ~100 km behind the front. High H2O concentrations (4-6 wt%) are most abundant along the volcanic front, whereas much lower values (1-2 wt%) occur in an extensional region far behind the front. The relationship between H2O and trace element ratios can also be extended to other regions of Mexico, as the correlation between H2O/La and Sr/La is consistent to the east in the Chichinautzin Volcanic Field (Cervantes and Wallace, 2003) and to the west in the Colima Graben. Using analyses from these and other regions, we have created a spatial map of H2O contents across the Trans- Mexican Volcanic Belt, enabling us to see trends both along and across the arc. We can then use these spatial maps to relate patterns in H2O content to subduction processes such as arc migration over time and slab devolatilization. Using 2-D geodynamic models we can correlate the breakdown of hydrous minerals in the slab and mantle wedge to spatial variations in magmatic H2O.
V31H-03
Modeling the 1913 eruption of Colima volcano, Mexico, based on data collected by Jim Luhr and colleagues
Jim Luhr and colleagues spent more than a decade characterizing the explosive eruptions of Colima volcano, particularly the January 20, 1913 Plinian eruption that sent a tephra cloud to the NNE of the volcano, by some reports depositing tephra up to 725 km from the volcano. Their data are modeled using TEPHRA2, a computer model that calculates the expected accumulation of tephra at specific geographical locations as a result of a volcanic eruption with specific input parameters using the advection diffusion equation. TEPHRA2 has numerous input parameters so it is literally impossible to find a best-fit solution using brute force iteration. Instead, we use nonlinear inversion techniques to explore best-fit solutions. Here we use a downhill simplex inversion algorithm. No parameter correlations (for example between eruption column height and eruption mass) are assumed a priori in the inversion. Overall, it appears from inversion results that acceptable solutions for total eruption mass lie between 0.8 x 1011 kg and 1.3 x 1011 kg and acceptable solutions for eruption column height lie between about 20 and 38 km above mean sea-level. In order to better understand the solution space, we ran the inversion numerous times, each time limiting the ranges of eruption column height and erupted mass. All other eruption parameters are allowed to vary over wide ranges to identify best-fit solutions. These results show that best-fit solutions for total erupted mass are constrained between approximately 0.6 x 1011 kg and 1.6 x 1011 kg. Best-fit solutions of essentially equal quality are identified for a wide range of eruption column heights (20-40 km). The plot of best-fit solutions suggests that slightly better results are obtained by the model in the region of 30-38 km and 1.4 x 1010 to 1.8 x 1011 kg, with all other parameters allowed to vary over their entire ranges. We note that eruption physics places some additional constraints on the maximum column height. For an instantaneous explosion, and 1.3 x 1011 kg mass in the plume, and the maximum column height above sea level should be approximately 36 km. Given that the entire plume was not likely released instantaneously, we regard this to be an approximate maximum possible plume height.
V31H-04
Water in anhydrous minerals of the upper mantle beneath continental margins
Although nominally anhydrous, olivine and pyroxenes can accommodate small amounts of water, as protons inserted in mineral defects. These apparently trivial water contents actually have a disproportionate influence on many physical and chemical properties of peridotite, and therefore on that of the entire upper-mantle. Caution is required in interpreting measured water contents in olivines because they can loose a significant portion of their water during transport of the xenolith to the surface, in particular when the xenolith is brought up by alkali basalts, as is typical at continental margins. Pyroxenes, on the other hand, appear mostly immune to this H loss. Review of our and literature data on natural samples (FTIR and SIMS analyses) suggests that water content ranges are similar in the lithospheric mantle beneath continental margins and beneath cratons, namely 20-160 ppm water in bulk-rock peridotite. The main controls on hydrogen incorporation in mantle anhydrous minerals appear to be pressure, oxygen fugacity and mineral composition. Mantle wedge environments, represented by mantle xenoliths with un-ambiguous subduction-zone fluid or melt influence, will be a special focus of this study to assess if their water content is any different than that of mantle lithospheres with no subduction influence.
V31H-05
Comparative light element â€" Li isotope systematics of arc volcanic rocks from the TMVB to Guatemala: Roles for the slab, crust, and mantle in magma genesis
Active arc volcanism occurs under dramatically variable conditions of slab age, dip, temperature and subduction rate; and crustal thickness and composition along the southern margin of Mexico (the TMVB) into the Central American Arc in Guatemala (Ipala Graben region). This study has engaged both undergraduate and graduate researchers over four years in an examination of B-Be-Li and δ7Li systematics of young lavas from the TMVB and the Ipala Graben. Dr. Jim Luhr was generous in his support of this effort, hosting undergraduate researchers at NMNH to examine and select samples for analysis, and providing numerous insights into the volcanic petrology and history of the Trans-Mexican Belt. Lavas of the TMVB differ from those in Guatemala in overall lower B enrichments and B/Be ratios (avg. B/Be <15 in the TMVB vs. >20 in Guatemala), though as one moves eastward along the TMVB arc front B/Be increases toward values observed at the Guatemalan volcanic front. Behind-the-front arc lavas show lower B/Be in both arcs, though in the Ipala Graben regular cross-arc declines are not observed, and young lavas from the far eastern end of the graben may have elevated B contents and B/Be. Li/Yb ratios of the Ipala Graben suite are typical for arcs (2-6, at Dy/Yb of ~2), while TMVB lavas range to Li/Yb =13 and Dy/Yb = 4, indicative of mineralogically anomalous mantle sources. However, δ7Li signatures of the TMVB and Ipala Graben samples examined span nearly identical ranges of +2‰ to +5.5‰. In both the Ipala Graben and TMVB suites, δ7Li values indicate a role for assimilated, isotopically distinct crustal rocks (TMVB: +1.5‰ to +3.4‰; Ipala: +0.9‰ to -2.6‰). Markedly higher B/Be in the Ipala suite points to a greater role for low-T slab-derived fluid inputs. Elevated LILE abundances in some western TMVB suites have been interpreted as the result of earlier, integrated slab additions (Hochstaedter et al, 1996), but the similar and MORB-like δ7Li in both systems point to the importance of an upper mantle reservoir uniform in its Li isotopic signature in the genesis of arc lavas along these margins.
V31H-06 INVITED
Is new Plinian Eruption Imminent at Colima Volcano in Mexico? Insights From Mineral Chemistry, Melt Inclusion Volatiles and Bulk Rock B-Li-Sr-Nd-Hf-Pb Isotopic Systematics
Volcan Colima is among the largest and most active volcanoes on the planet. The extensive and detailed work of James F. Luhr has made Volcan Colima a "living petrological laboratory" on the compositional evolution of an active volcanic complex through time. The possibility that the current eruptive cycle (1961-2007) will culminate in a major Plinian (VEI 4) explosive event similar to those of 1818 and 1913 is an important focus of ongoing work. Lavas and ash erupted since 1961 show increasing compositional similarity to the mafic 1913 scoria, with trends towards lower SiO2 (from ~ 61.5 to 59 wt% in whole-rock; 72 to 66 wt% in groundmass), lower Ba (500 to 400 ppm), and higher MgO (2.7 to 4.5 wt% in whole-rock). Phenocryst rims mimic these trends with opx and cpx increasing in Mg# from 71 to 73 and 73 to 76 (respectively) and plagioclase increasing in An content (47 to 54 %). Unlike the 1913 scoria, all recent eruptive products lack the micro-vesiculated groundmass textures (signifying expansion of in-situ magmatic gas) that characterized the 1913 tephra. Samples from 1961-2007 are poor in hornblende, have melt H2O contents that decline with time (from 4 to 1 wt %.) and lower volatile contents in olivine hosted melt inclusions (H2O=2.1-2.3 wt%;CO2=17 to 122 ppm;F=1147 to 1354 ppm;Cl=1657-1771 ppm,S=539-675 ppm) as compared to the 1913 scoria. The current magma batch is relatively poorer in water than its predecessor and it appears that volatile-poor mafic mantle melts are currently reaching the base of the Colima conduit system. Although these melts are similar in major and trace element composition to the 1913 scoria they fortunately enter the upper volcano conduit system in a significantly degassed state. Thus the observed Vulcanian (dome blasting) summit explosions could be phreatomagmatic, rather than a product of magmatic degassing. Fluid and crustal assimilation-sensitive isotopic tracers such as B, Li, Sr, Pb, Nd, and Hf isotopes can be used to examine the source of volatiles and differences between the 1913 scoria and more recent magmatism. All of the Colima samples examined show low 87Sr/86Sr (<0.704), MORB-like Li isotope ratios,εNd (+3 to +6) and εHf (+7 to +10) i.e. representing mantle melts with little or no sign of crustal contamination. LA-MC-ICP-MS B isotope measurements show that the 1961-2007 Colima andesites possess typical arc-like values (δ11B = + 2 to +3‰) similar to samples from the basaltic cinder cone Jorullo (+ 2.1‰). However, these lavas are much lighter than those of nearby alkaline mantle-derived lavas erupting from older small volume monogenetic cinder cones within the Colima graben (up to +10 ‰). Their heavy B isotope ratios may represent fluid-enriched mantle source. The spikes in volatile- rich mafic melts at Colima seem to occur episodically (every ~ 100 years). These variations in fluid sensitive B isotope ratios and melt volatile content reveal episodicity of slab fluid releases below the Colima graben. Coincidence of mafic magmatism and slab fluid release may be necessary for a new Plinian eruption at Colima.
V31H-07
Glass Inclusions and Volatile Contents of Lamprophyres, Basanites, and Basalts from the Colima Rift and Mascota Volcanic Field (Western Mexican Volcanic Belt)
We present geochemical data for coeval basaltic to lamprophyric tephra erupted during Quaternary time in the Colima and Mascota volcanic fields at the western end of the Mexican Volcanic Belt. Focusing on glass inclusions (124 analyzed) trapped within olivine, our goal was to evaluate a vein/wall-rock melting relationship between the lamprophyres (vein-dominated melts) and the mafic calc-alkaline rocks (diluted by partial melting of peridotite wall rock after exhaustion of phlogopite and other vein minerals). We measured major-element compositions of glass inclusions and host olivines by electron microprobe. Volatile contents in glass inclusions were measured by electron microprobe (S, %S6+, Cl), Fourier-transform infrared spectroscopy (H and C species), and secondary-ion mass spectrometry (H, C, F, S, Cl). Estimates of magmatic fO2 from measured bulk-rock Fe3+/Fe2+ and glass-inclusion %S6+ reveal that all studied magmas were highly oxidized, up to several log units above the Ni-NiO buffer. The highest concentrations of water and most other volatile elements (7% H2O, 1,460 ppm CO2, ~2% SO3Total, 2,400 ppm Cl, and ~1% F) were recorded for a glass inclusion from a Colima minette (48.2 wt.% SiO2, 6.0% K2O, and 1.2% P2O5, normalized anhydrous), and are consistent with a depth of entrapment of 24.1 km (calculated pressure of 6,660 bars). We consider these chemical properties of the melt to be little changed since it formed at a depth of 90-100 km in the mantle wedge from nearly pure partial melting of phlogopite-garnet-pyroxenite veins. The mafic calc-alkaline bulk-rock and glass-inclusion compositions are consistent with formation through approximately 7-fold dilution of the lamprophyre-type vein melts with partial melts from the surrounding depleted peridotite wall-rock. Among the calc-alkaline glass inclusions, the Mascota basaltic andesite has the highest concentrations of water and most other volatile elements with 49.6 wt.% SiO2, 1.0% K2O, 0.3% P2O5 (normalized anhydrous), 2.8% H2O, 296 ppm CO2 (1,425 bars pressure and 5.2 km depth of entrapment), 0.8% SO3Total, 870 ppm Cl, and 720 ppm F. Such mafic calc-alkaline melts are envisioned as parental to the volumetrically dominant andesites of western Mexico.
V31H-08
The Complicated Life of a Little Monogenetic Volcano in Mexico: Volcan El Jorullo
Volcan El Jorullo was produced from 1759 to 1774 and represents one of two historical eruptions within the Michoacan-Guanajuato Volcanic field in the western Trans Mexican Volcanic Belt. Inspired by the results of an early detailed geological, petrological and geochemical study by Luhr and Carmichael (CMP 90, 1985), and the dearth of subsequent research, our group began a detailed assessment of spatial and temporal compositional variations within volcanic products of this eruption in the early 1990s. Collectively the studies have uncovered a complex (and non-linear) sequence of melting, ascent, differentiation, assimilation and eruption that betray a much larger spatial and temporal variability in magma storage and migration conditions than one might anticipate for a one-time volcanic event in a volcanic field fed sufficiently with magma to have produced nearly 1000 similar volcanic cones in the Holocene and Pleistocene. These complexities at Jorullo offer a rare opportunity to study mafic magma generation at a near arc volcano, a prospect that excited Jim Luhr and made him an engaged collaborator until his untimely passing. This presentation will summarize our data (mapping, major and trace element, Sr-Nd-Pb and U-series isotopic data) and synthesize it with those of the seminal Luhr and Carmichael study. We show that high to moderate MgO melts supplied to Jorullo were a mixture of those generated from an arc-fluid fluxed mantle and the lower crust and that magmas carry a signature of two distinct assimilation events (one deep, one probably shallow) that are distributed differently within the Jorullo compound lava flow field. Further, conditions of magma accumulation, differentiation, storage and transport were sufficiently heterogeneous to produce other spatial and temporal variations in lava compositions erupted in at least 8 distinct effusive events during this monogenetic activity. We suggest that there are direct feedbacks between what, where and how various magma compositions were erupted at Jorullo and contrast these conditions to those at Paricutin, at which an outwardly similar but ultimately simpler sequence of events prevailed (e.g., McBirney et al CMP 95, 1987).