Volcanology General Contributions Posters
Presiding: O Bachmann, University of Geneva; M Jellinek, University of British Columbia
V13B-01 1330h
New Paradigms and Intractable Problems: A new Approach to Source Rock Characterization
Problems relating to the negative correlation between Sr and Nd isotopic ratios of recent basalts, initial isotopic ratio for the bulk Earth, excess radiogenic lead and a few others could be resolved if it is believed that all radiogenic isotopes are fractionated with respect to the non-radiogenic isotopes. First, a new approach to the source rock characterization is outlined here. It is based on the fact that there is a perfect correlation between Rb/Sr vs. ((Rb/Sr) / (87Rb/86Sr)) (termed WAR). The line formed by this relation is named here MUOX, whose significance is that for a rock of any age derived from an undepleted reservoir, its intercept at the Rb/Sr value of 0.031 remains constant, i.e. 0.34572097 (termed BEWRUB). In deriving the above relation, the following parameters were presumed: IBE 4.55 Ga ago: 0.69877; age of the Earth: 4.55 Ga; NBE: 0.09536; λ 87Rb: 1.408E-11/yr. The various symbols used in the text denote the following: I is the 87Sr/86Sr ratio; BE is the bulk Earth; N is the 87Rb/86Sr ratio 4.55 Ga ago; R is the source rock; λ is the decay constant; and ME is the time of magma extraction. The other relations that have been worked out are as follows: (1) The relation between wRb vs. dSr defines a slope (named RUH) of 29.697285 and an intercept of -0.0000000041; where wRb = BEWRUB - SOWRUB, and SOWRUB is the predicted value of the MUOX Intercept at the Rb/Sr value of 0.031; and dSr = (IR(ME) - IBE(ME)). (2) The relation between MUOX Intercept (MI) vs. IR(ME) defines a slope (termed PUCI) of -29.69176 and an intercept of 10.96976744. (3) The relation between the MUOX slope vs. the age of a rock (in Ma years) defines a slope (named Gargi) of -705465.433 and an intercept of 8.668. (4) The relation between the age of a rock (in Ma years) vs. 1/E defines a slope (termed Raman) of -0.0000001265 and an intercept of 0.000579232; where E = P / WRb; P is (NR / NBE - 1) * (100); and WRb = (wRb / BEWRUB) * (100). This relation can be used to find the NR of a source rock. The above relations when applied to the data from basaltic achondrites (Papanasstassiou, D.A. et al. 1969) yield the following results: wRb: -0.00000275; IR(ME) (PUCI): 0.69897705; IR(ME) (RUH): 0.69897659; Age (Gargi): 4.34 Ga (4.39 Ga if λ of 1.39E-11/yr is used); NR (Raman): 0.07097; Source Reservoir Depletion: 26%. The close match of these results with the actual data, strongly suggests that the presumed parameters are valid. The apparently depleted nature of the achondrites is not because of their depleted source reservoir, but because of the fractionation of 87Sr isotopes. It is likely that similar depletion in the early basic melt fractions may have resulted in the residual mantle to be enriched. The reason that the negative correlation between Sr and Nd isotopic ratios is not observed in older rocks, is because the production of new radiogenic isotopes obliterates the effect of fractionation. To prove fractionation, other examples from the literature will be discussed.
V13B-02 1330h
The Core as the Third Pivotal End Member of the Earth's Plate Tectonic Cycle: A New Theory
Existing data and use of a hypothetical model, post-Stishovite-Magnesiowustite-Iron, indicate that the Earth's core could be the the convertor end member of the Earth's Plate Tectonic Cycle (EPTC): a new theory. This third pivitol end member, the core, is the place where the cycle begins and ends, to begin again. The first pivotal end member of the EPTC, for a three end member system, is the global MORB end member where new oceanic crust and lithosphere are created. Sea-floor spreading connects it to the second end member, the subduction end member, where oceanic crust and lithosphere disappear to become cold lithospheric-crust complexes descending through the mantle toward the Earth's core. When complexes break into it they are slowed, turned and endothermally ingested. Partial melting frees lower mantle phases and iron while forming metallic liquid and a densified immiscible silicate liquid, of which 17 vol.% reduces the bulk density of a convecting outer core by 10 %. Freed crystalline phases form micro-phases of micrometer to millimeter in size that more or less fill mega-bodies of <83 vol.% metallic liquid and <17 vol.% immiscible silicate liquid, both of centimeters to kilometers in size. Excess core energy starting each cycle comes mainly from irreversible exothermal reactions at numerous unstable phase contacts by stable phases within and between mega-bodies to yield stable products of lower Gibbs free energy, only to make new contacts and react...and react again. Other sources of exothermal energy come from radioactive silicate liquid and friction at stable phase contacts during mega-body convection. Heat accumulated from these energy sources tends to expand the outer core as univariant boundary reactions of the core and the lower mantle reverse, with +5.0 cm3g-1comming from the inner core boundary reaction alone. The outer core's pervasive expansion against the passively resisting strength of the mantle results in explosive ejection of silicate liquid along a line of weakness or at a point of weakness of the CMB when it fails, as it must. Superheated actions of this liquid with lower mantle phases result in hybrid, hot and solid domains that ascent as hot basic plumes in a cooler and denser ultrabasic mantle. Decompression melting of the plume above 290 km contributes new basalt crust at mid-oceanic ridges, the MORB end member, from a line source of the CMB. New crust for oceanic islands basalts arrives there, after similar processes yield plumes from point sources of the CMB. Both of these basalt types are thought to be linked with silicate liquid ejected fron the Earth's outer core as expressed by the ubiquitous and unexplained C-component of Pb isotope ratios of Atlantic-, Indian- and Pacific mid-oceanic basalts, major segments of the MORB end member of the EPTC, and by its corresponding FOZO isotope component of oceanic island basalts. Oceanic crust and lithosphere disappear at the subduction end member before it is a twenty-third the age of the Earth, and leads to its ingestion and eventual energization in the Earth's outer core, the convertor, and third end member of the EPTC, to drive ensuing cycles. It seems unavoidable that egress features for silicate liquid on the CMB could serve as a template for the EPTC in a bottom-up dynamic, except for the actions of the subduction end member of the EPTC. There is no shortage of silicate liquid in the outer core as 17 vol.% is almost twice the volume of the entire crust of the Earth. Finally, it seems possible that some of the unexplained heat flux of the Earth, over and above the radioactivity of the mantle, may be attributed to this new paradigm that drives the Earth's Plate Tectonic Cycle.
http://www.EarthsPlateTectonicCycle.us
V13B-03 1330h
Nd and Hf Isotopic Compositions of Pan-African High-Pressure Mafic Granulites
High-pressure (HP) granulites and eclogites which define the Pan-African Dahomeyide suture zone in West Africa have major and trace element compositions which indicate geochemical imprints of their magmatic protoliths. Samples representing suites characterized by island arc tholeiitic (IAT) and MORB affinities have been analyzed for Nd and Hf isotopic compositions. The MORB-like sample has relatively high radiogenic Nd isotopic composition (143Nd/144Nd= .513167; epsilon Nd= 8.17) and Hf isotopic composition of (176Hf/177Hf=.283496; epsilon Hf=16.14) that yielded TDM= 0.77 Ga. Two samples with IAT-affinities gave lower radiogenic Hf isotopic composition (176Hf/177Hf ~ .282877, .282985; epsilon Hf= 9.43 and10.79) and lower epsilon Nd (5.3 and 6.6), that plot in the OIB field of the mantle Hf-Nd array and yield older TDM~0.96 Ga. The data suggest that magmatic protoliths of the HP granulites may have erupted ~200-300 my prior to suturing involving collisional orogeny and thus provide, for the first time, some constraint on the timing of the pre-Gondwana rifting of the WAC. The data also provide the opportunity to test models of the decoupling of Hf-Nd isotopic fractionation by garnet; these models predict that garnet production during partial melting could result in anomalous fractionation of crustal rocks out of the mantle Hf-Nd array.This is apparently not the case for the samples analyzed although the HP granulites which recrystallized ca 0.60 Ga (at 800 C, P >15 kbars) with as high as 20 percent modal garnet, require only ~450 my to evolve out of the juvenile crustal array (if partial melting is > 20%). Thus a plausible explanation for the preservation of the mantle array in the Pan-African HP mafic granulites array is that the degree of partial melting was significantly less than 10% and that this partial melt was largely retained as pervasive veins.
V13B-04 1330h
Geochemistry and Geochronology of Ngorongoro Crater, Tanzania: Implication for Magma Evolution, Duration of Volcanic Activity and Age of the Ngorongoro N-R Geomagnetic Polarity Transition
40Ar/39Ar dates on volcanic rocks from the Ngorongoro Crater (NC) in northern Tanzania indicate that NC activity was very short in duration lasting approximately 120 ka. Laser incremental heating experiments on lava from the bottom and top of the NC crater-wall section gave ages of 2.08 +/- 0.04 and 1.96 +/- 0.02 Ma respectively. Lavas from the same section show a change in magnetic polarity from normal (N) at the lower part to reverse (R) polarity at the upper part (Gromme et al. 1970). The new ages are about 400 ka younger than previously estimated by K-Ar technique. These new ages suggest correlation of the NC N-R polarity transition to the 2.1 Ma (N-R) Reunion-Matuyama boundary (Cande and Kent, 1995), instead of the Gauss-Matuyama boundary as proposed by Gromme et al. (1970). 87Sr/86Sr measurements on lavas from the NC section vary widely from 0.70801 in the trachydacite at the base to 0.70405 in the basaltic lava near the top. The lower part of the section is more radiogenic varying from 0.70592 to 0.70801 whereas the upper part is constrained to 0.70405 to 0.70450. The more radiogenic lower part is likely to have interacted with crustal rocks. Two possible contaminants are the Tanzanian Archean Craton to the west and the late Proterozoic Mozambican belt in the east. The crater-wall section is composed of trachydacite at the bottom that becomes trachyandesite in mid-section. The top section is mainly basaltic. Major and trace elements show an inverted geochemical signature that is typical of stratified magma chambers characterized by a silicic top and basaltic bottom. Olivine basalt at the upper part of the section has the highest Mg&35; (56.60) and in general the upper section is more mafic than the lower section as inferred from Mg&35;. The upper part of the section is high in TiO2, MgO, FeOT, and CaO wt% whereas SiO2 and K2O wt% are higher in lower part of the section. No significant variations are observed in N2O, Al2O3, P2O5 and MnO wt% up-section. Highly incompatible elements e.g. Zr, Nb and Hf concentrations are higher in the lower section compared to the upper part probably indicating the extent of magma evolution.
V13B-05 1330h
Kimberlite-Clan-Rocks in India: Significance of new VGP, T, and GP Observations.
Although India is acknowledged for the first description of diamond some 2000 BCE, it should also rightfully be credited for the 17th C recognition that diamond is a product of volcanism. With this extraordinary background, it is surprising that the host rocks remain controversial, being neither archetypical kimberlites, nor classic lamproites. Lacking affinities to micro-diamond-bearing UHPT metamorphic rocks and being unequivocally volcanic, the term Kimberlite-Clan-Rock (KCR) is applied. Over 200 KCR pipes and dikes, many of which are richly mineralized, have recently been discovered in the Diamond Corridor (1000 x 200 km) of the Eastern Dharwar Craton, and in adjoining cratons to the N and NE. From 32 absolute age determinations on KCRs, the remaining intrusions, in comparable stratigraphic settings, are assumed to be 1.1 Ga, equivalent to the Argyle lamproite (Australia, highest diamond grade), and the Premier kimberlite (RSA, largest known diamond). India has the largest number of known Proterozoic KCR intrusions that over the centuries have produced a wealth of famous diamonds. The primary source of these extraordinary stones, however, remains unknown, possibly because the unusual host rocks defy conventional exploration protocols. Six new observations make the setting even more unusual: (1) Coeval, large scale magmatism in the Kalahari (>2 m sq km) and Laurentian (>300 k cub km) Cratons at 1085-1112 Ma, during assembly of Rodinia, confirms the relation between and among KCR volcanism, LIPs, and supercontinents; Proterozoic Rodinia, into which the India KCRs, Argyle and Premier were intruded, was constructive, whereas the other, globally-wide, diamond-intrusive event, that occurred during the Mesozoic (80-120 Ma), was related to the breakup of Pangea and the dispersion of Gondwana; both events occurred during long period geomagnetic chrons implying a core relation, and superplume activity from the CMB. (2)The transcontinental Mumbai-Chennai gravity lineament that separates diamond (North) from barren (South) KCRs is interpreted as a sub-lithospheric, architectural discontinuity, with a shallow keel to the N and deeper penertration to the S. (3)An intrusive carbonatite stock into one KCR, and intensely carbonatized xenoliths in another, has rekindled the unsettled VGP debate of kimberlite-carbonatite relations; the new discoveries strongly support a kinship. (4) An eclogite xenolith from Kaliandurg has euhedral inclusions of re-equilibrated majorite in garnet; the assemblage has important implications for depths of origin (in or close to the TZ), and the controversial issue of Archean ocean crust recycling, recognizing that there are other ways to generate mantle eclogites. (5) Heterogeneities in the source regions of KCRs, and in the degrees of mantle melting, metasomatism and mixing, are well displayed in the Chigicherla KCR cluster; CC5 has unprecedented 2-10cm diameter autoliths (ol + cc + sp + perv), that are partially to totally assimilated in closely associated bodies, or are absent in others; (6)Groundmass olivine in KCRs from Naryanpet (non-diamond) and Wajrakur (diamond-bearing) have densely packed, crystal-oriented laths of rutile in association with blebs of metal + sulfide; these intergrowths bear on the unresolved solubility of Ti in olivine and its appeal to super high P-T conditions in diamond and coesite-bearing metamorphic terrains; the KCR settings, however, imply crystallization of olivine at low P, coupled with or followed by Ti metasomatism. With increased activities in diamond exploration many more VGP, T and GP mantle revelations are expected from this geologically classic and historically important part of the sub-continent.
V13B-06 1330h
Experiments of U Solubility in Earth's Core
The experimental work of Murrell et al (1984) on the solubility of U in FeS was carried out at low pressure (< or =1.5GPa). Their work showed that U may combine with CaS in the FeS phase. Later, McDonough (2003) argued that if U entered the Earth's core with Ca, then the current Ca/Ti and Ca/Al ratios of the mantle would be difficult to explain. Here we investigate the solubility of U in pure Fe using a mixture of uraninite, Fe and peridotite powder at pressures of 3-7.5GPa and at temperatures of 1900-2000 °C. Our results show that U is soluble in Fe melts. Recovered run products were analyzed by LA-ICP-MS (detection limit for U: 0.0122 ppm) and by Electron Microprobe (EM, detection limit for U: 149 to 167 ppm). The average concentration of U in pure Fe from LA-ICP-MS is: 3GPa: 7 ppm; 5GPa: 55 ppm; 6GPa:102 ppm and 7.5GPa:126 ppm and DU values, partition coefficient of U (concentration of U in Fe / concentration of U in silicate) at the above pressures are (×10-3): 0.219; 1.82; 2.69 and 4.59, respectively. The EM data support the trend defined by the LA-ICP-MS data but are too close to the detection limit to provide quantitative results. It can be seen that U in the Fe phase generally increases with pressure. We also found that Si in the Fe phase increases from 0.45 wt% to 2.13wt% from 3GPa to 7.5GPa resulting in a positive correlation between Si and U in the Fe phase. This also can be seen qualitatively from the LA-ICP-MS line scan figures. The concentration of CaO is small (<0.02 wt%) and no relation with U is found, which indicates that U may alloy with Fe directly. If Si concentration in the Fe phase can be used as an indicator of oxygen fugacity (Kilburn and Wood, 1997), then the increase in Si and U with pressure suggests a pressure dependent decrease in oxygen fugacity. This supports U (and possibly also Si) inclusion in the Earth's core at the time of core formation.
V13B-07 1330h
Ti and Al Pyroxene in the Egersund Dikes: Analogue for Martian Meteorites?
The Egersund dikes in SW Norway intruded at 650-600 Ma. Most formed by intrusion of olivine tholeiites that are geochemically similar to modern OIB. Olivines range in composition from Fo84 to Fo76 and show a variety of textural features including skeletal crystals, strained crystals, crystals intergrown with plagioclase, and resorbed crystals. Inclusions with high-Ca pyroxene, plagioclase, amphibole, Fe-Ti oxide, and very rare spinel and apatite appear to represent partly crystallized melt trapped in some olivines. The pyroxene is titanaugite with very high Al2O3 (up to 18 wt%) and TiO2 (up to 6.4 wt%), low SiO2 (40.6 ave wt%), and low Mg&35; (ave 0.58). Coexisting amphibole is kaersutite (Mg&35; 0.25-0.58) with up to 0.77 Ti afu and 2.8-3.5 Al afu. Si contents (5.74-5.34 afu) are the lowest recorded for this amphibole. Inclusions containing pyroxene and kaersutite occur in resorbed olivines. Some of the latter show reversed zoning (cores - Fo76, rims Fo82), and they are interpreted to be xenocrysts that formed in magma chambers at pressures of P=0.8-1.2 GPa. The pressures are constrained from phase equilibrium data, and indicate crystallization in chambers located at or near the crust-mantle boundary (ca 28 km). The inclusions possibly represent alkaline magmas that resided in the magma chamber prior to flushing by olivine tholeiitic magma. This is supported by the Ti-rich composition of the pyroxene by extensive substitution of Al into the pyroxene tetrahedral site, and by the occurrence of kaersutite, which does not form from tholeiitic melts based on experimental studies. Alkaline magmas (trachybasalts) occur in the Egersund dike swarm. The high Al2O3 and TiO2 character of pyroxene and the occurrence of kaersutite in the Egersund dikes are similar to that found in olivine hosted melt inclusions in Martian meteorites (Chassignites and Nakhlites). These similarities allow the Egersund dolerites to be used as a possible analogue for igneous processes on Mars. The results of this study suggest that melt compositions inferred from inclusions in olivine may not provide reliable estimates of the parent magma composition. Further study of these inclusions could provide a better understanding of the Al and Ti enrichment observed in Martian meteorite melt inclusions.
V13B-08 1330h
Mg-Fe2+ Exchange Between Olivine and Melt
Knowledge of exchange of Mg and Fe2+ (Kd) between olivine and melt is crucial for understanding the origin and evolution of magmas. It is essential to use accurate and precise values for Kd for quantitative studies of melting and crystallization. However, values of Kd derived from experimental studies are associated with relatively large uncertainties. Moreover, there is disagreement about the dependence of Kd on fO2, T, and melt composition. To obtain accurate and precise values of Kd as a function of fO2, T, and melt composition, electron microprobe data for equilibrium olivine-glass pairs were compiled from published experimental studies of natural samples at 0.1 MPa with controlled oxygen fugacity. The melt compositions in these experiments include tholeiitic and alkali-olivine basalts, picrites, andesites, and highly alkaline lavas (nephelinite, leucitites, ugandites). These data confirm that Kd is independent of fO2 and T as concluded by many previous workers. However, Kd is strongly dependent on melt composition, in particular on the molar alkalis and silica contents of the melt. These results are consistent with those obtained by Longhi et al (1978), and suggest that variations in Kd correlate with the degree of polymerization of the melts. Taking into account the non-ideal behavior of Mg-Fe2+ olivines does not affect this conclusion. An expression that describes lnKd in terms of molar alkalis and silica recovers experimental values with an accuracy of ±0.0001 and a precision of ±0.007. Comparable accuracy and precision is obtained using the expression given by Gee and Sack (1988) describing variations in Kd in terms of melt alkalinity. It is not presently clear that these expressions are applicable to Fe-rich melts (Mg&35; less than about 0.25) or MgO-rich melts such as komatiites. However, they provide a rigorous test for equilibrium between olivine and melt in the most abundant terrestrial magmas. In addition, accurate values of Kd allow melt redox state and hence fO2 to be calculated for melts with total Fe reported as FeO. Values of fO2 calculated for 189 experimental olivine-melt pairs agree with reported values to ~0.22 log bar units.
V13B-09 1330h
Geochemical Heterogeneity Within Lava Flows From 9°25'-9°51'N EPR
We evaluated the degree of geochemical heterogeneity and morphological variation within carefully mapped and sampled lava flow units collected by ALVIN from 9 25'-55'N along the EPR to better understand the distribution of lavas that comprise the neovolcanic zone of the East Pacific Rise (EPR) and how the upper oceanic crust is formed at fast-spreading mid-ocean ridges (MORs). The selection of sample sites was guided by microbathymetry and sidescan sonar images in areas where discrete flow units appear to emanate from within or near the axial summit trough (AST) and extend up to a few kilometers off-axis. Pillow mounds with possible off-axis origins were also mapped and sampled. While a range of lava morphologies including pillow, sheet, and lobate flows were recovered, preliminary observations indicate that a variety of morphologic facies are present in each individual lava flow unit and the morphologies can be related to flow dynamics. More than 25 individual flows were directly observed and/or sampled along the crestal plateau where ABE microbathymetry and DSL-120A side scan sonar images show acoustically reflective scalloped morphologies that are observed to be overlapping flow fronts 100's of meters in length. At many sites, lava samples were collected from the pillows comprising flow fronts as well as lobate flows that form the tops of eruptive units as well as underlying units. Non-reflective, dendritic, sinuous to linear features that are largely perpendicular to the AST represent distributary channels that serve to move lavas off-axis. Seven sites in five separate channel systems were examined and sampled in detail. While there is only slight chemical variability (usually within analytical uncertainty) in glasses from most of the individual lava flows, two complimentary dives on either side of the EPR axis near 9 50'N sampled lavas with significant major element differences between flows proximal to the AST and those farther away. In each case, the most primitive samples are from the first few flow units emanating from the AST, while those now located up to several kilometers away from the AST are significantly more evolved. The observed inter-flow or intra-flow major element variations cannot be explained by low-pressure fractional crystallization; which might be expected if cooling and crystallization proceeded as flows moved off-axis. Instead, the data suggest the near- and off-axis samples are not directly related and that other magmatic processes have affected interflow geochemical variation. Higher pressure crystallization (4 to 4.5 kbars) is required to explain the chemistry of some samples as well as mixing between more evolved and more primitive melts; both unlikely to happen during seafloor eruptive and transport processes. In addition, ~30% crystallization is required to explain the interflow variation for each dive while no greater than 10% crystals are present in any samples from this section of the EPR.
V13B-10 1330h
Self Potential and Gravity Study for Geothermal Exploration of Nevis
A geophysical team from the Massachusetts Institute of Technology and SP International, Inc. performed self potential and gravity surveys on the island of Nevis for geothermal exploration under contract from the Organization of American States, Geo-Caraibes Project. The preliminary geothermal model for the Charlestown area of interest based on SP current sources and gravity, places a large high gravity body under most of the area with geothermal fluids flowing on the edge of the body. We believe the Charlestown area is underlain by a heat source which may be capped by an impervious layer, therefore permeability for geothermal fluid upflow is mainly limited to faults and fractures and other high permeability conduits. The most intense current sources are located on or near the edge of the high gravity body.
V13B-11 1330h
Volcanic Chemostratigraphy on the Outcrop Using Field-Portable X-Ray Fluorescence: An Example from the Basaltic Flows of Hewitt's Cove, MA.
The area of Hewitt's Cove, Hingham, Massachusetts structurally represents the southern extent of the Boston Basin, and as such, provides the opportunity to identify and recognize basin-wide events. While the stratigraphy of the Boston Basin has been developed since the advent of U-Pb geochronology and formal stratigraphic facies descriptions (Socci and Smith, 2001), the stratigraphy of the Hewitt's Cove area has not been thoroughly addressed since the work of William O. Crosby in 1894 (Billings, 1976, Bailey and Bland, 2001). Hewitt's Cove consists of an andesitic basaltic flow, approximately 150 m thick, overlain by siltstone and conglomeratic sequences. Field-Portable X-Ray Fluorescence (FP-XRF) was used on fresh and weathered surfaces on the outcrop and in hand samples, and these analyses were compared with conventional laboratory XRF analyses. The in situ field-based analyses produced a reproducible chemostratigraphy that is consistent with subsequent laboratory-based analyses. These data suggest that the series of andesitic basalt flows at Hewitt's Cove are the result of compositionally different magmatic pulses. Additional analyses must be completed to determine the extent of these pulses and the extent of variation in this area before further conclusions can be made. This study particularly demonstrates the utility of using FP-XRF in igneous geologic applications.
V13B-12 1330h
The Effect of Composition on the Volatile Solubility of Mixed H2O-CO2 Fluids in Arc Basalts
The volatiles H2O and CO2 are abundant in melt inclusions from arc lavas and play an important role in magmatic processes associated with subduction zones. While there are a significant number of studies into the solubility behavior of these volatiles, they do not always cover the necessary range of geologic conditions needed to accurately interpret pre-eruptive conditions recorded by melt inclusions. Most studies have considered only a pure, single component fluid, and are at either very high (greater than 1.0 GPa) or low (less than 200 MPa) pressure. The few studies that have investigated mixed volatiles (H2O and CO2) in melts have only rarely measured the fluid composition in equilibrium with the melt. In addition, the overall effect of bulk melt composition is not well constrained. Therefore, because melt inclusions found in a single lava commonly have a wide bulk compositional variation, the intensive parameters (P-T) cannot be accurately modeled for the solubility surface using the bulk composition of the host lava. We present here the initial results of an experimental study aimed at defining the mixed volatile solubility surface in P-T-Xi space, for two associated melt compositions from Central America; one representing a "typical" lava composition and the other a Ca-rich melt observed in inclusions found in the same lava type. Superliquidus experiments (0.4-0.7 GPa; 1200° C) were conducted in a non end-loaded piston cylinder using a double capsule method. An outer Pt capsule containing an undersaturated hydrous rhyolite was used to prevent the brittle failure of an inner Au-Pd capsule containing basalt and added volatiles (variable H2O:CO2 ratios; XH2O(fluid) ~0.3-0.70). Oxygen fugacity was monitored in the inner capsule using a thin Pt strip and measuring the resulting Pt-Fe alloy. After equilibration and quenching, fluid compositions were measured by vacuum manometry, and the volatile content of the glasses was measured by high T vacuum manometry, secondary ion mass spectrometry, and infra-red spectrometry. Results at 0.4 and 0.6 GPa and XH2O fluid of ~0.6 show the dramatic influence of calcium on CO2 solubility. We observe an ~2-fold increase in CO2 content when the basalt CaO content is increased by ~20 percent (~6000 versus 2,500 ppm at 0.6 GPa), whereas H2O solubility is unchanged or even diminished. Our results indicate that current solubility models for do not accurately predict H2O and CO2 abundance in high Ca basalt at these pressures, and by extension, the intensive parameters predicted by these models for measured volatile abundances in high-Ca melt inclusions are inaccurate. The high solubility of CO2 in calcic basalts also has significant influence on arc geochemistry and volcanic behavior. For example, high-Ca basalts could act as carriers of volatiles and additional alkaline earths. Through underplating and mixing, high-calcic basalts could transport CO2 into magma chambers, thereby contributing to fluid saturation and other physical conditions leading to eruption.