V13D-2141
The Final Phase of Drilling of the Hawaii Scientific Drilling Project
The principal goal of the Hawaii Scientific Drilling Project (HSDP) was to core continuously deep into the flank of a Hawaiian volcano and to investigate the petrology, geochemisty, geochronology, magnetics, etc. of the recovered samples. Drilling in Hilo, on the island of Hawaii proceeded in three phases. A 1.06 km pilot hole was core-drilled in 1993; a second hole was core-drilled to 3,098 meters below sea level (mbsl) in 1999, then deepened in 2004-2007 to 3,509 mbsl. Although the final phase of drilling was at times technically challenging, core recovery was close to 100%. All rocks from the final phase of drilling were emplaced below sea level and are from the Mauna Kea volcano. On-site core logging identified 45 separate units (the 1999 phase of drilling yielded 345 units). Five lithologies were identified: pillows (~60%); pillow breccias (~10%); massive lavas (~12%); hyaloclastites (~17%); intrusives (~1%; these are mostly multiple thin (down to cm scale) fingers of magma with identical lithologies occurring over narrow depth intervals). The rocks are primarily tholeiitic, ranging from aphyric to highly olivine-phyric lavas (up to ~25% olivine phenocrysts), with considerable fresh glass and olivine; clays and zeolites are present throughout the core. Forty whole-rock samples were collected as a reference suite and sent to multiple investigators for study. Additionally, glass was collected at roughly 3 m intervals for electron microprobe analysis. Although continuous and consistent with the shallower rocks from the previous phase of coring, there are several noteworthy features of the deepest core: (1) Glasses from shallower portions of the core exhibited bimodal silica contents, a low SiO2 group (~48-50 wt.%) and a high SiO2 group (~50.5- 52 wt.%). Glasses from the last phase of drilling are essentially all in the high-silica group and are somewhat more evolved than the high-silica glasses from the shallower portion of the core (5.1-7.6 vs. 5.1-10.4 wt.% MgO). (2) The expected sequence of lithologies with depth in the core is subaerial lava flows, hyaloclastites (formed by debris flows carrying glass and lithic fragments from the shoreline down the submarine flanks of the volcano), and finally pillow lavas. This sequence was generally observed in the earlier phases of drilling, and it appeared that the deepest rocks from the 1999 phase of drilling were essentially all formed from pillow lavas (i.e., there were no more hyaloclastites). However, thick hyaloclastites reflecting long distance transport from the ancient shoreline reappear in the bottom ~100 m of the drill hole. Although it may be coincidence, pillow breccias occur in the shallower parts of the core from the final phase of drilling, but not in the deeper parts in which the hyaloclastites reappear. (3) Intrusive rocks make up a lower fraction (~1%) of samples from the final phase of coring than in the deeper parts of the section from the 1999 phase of drilling (3.8%). It had been suggested that intrusives might become more common the deeper the drilling, but this is not the case at depths down to 3.5 km. (4) There are three units classified as "massive" including one relatively thick (~40 m), featureless (no internal boundaries, no evidence of mixing or internal differentiation), moderately olivine-phyric basalt.
V13D-2142
Hawaii Scientific Drilling Project: Objectives, Successes, Surprises and Frustrations
The Hawaii Scientific Drilling Project (HSDP) is a long-running project undertaken with the objective of studying a mantle plume by drilling an extended sequence of lavas from a single Hawaiian volcano. The project originated with a proposal to NSF in late 1986 with the idea of drilling to the Moho under Hilo; the target depth was estimated at 12km, commensurate with the depth reached by the drilling program then being pursued by the USSR and that proposed in the U.S. for the southern Appalachians, and in line with the aspirations of the nascent DOSECC program. Subsequently, due to limitations in funding and reorganization of the drilling program into what later became the NSF Continental Dynamics Program, HSDP was re-scoped with the objective of drilling deeply enough (ca. 4.5km) to recover most of the eruptive history of a single volcano. The project first went to a pilot stage, which resulted in coring to a depth of 1.1km in late 1993. The pilot stage was relatively inexpensive ($1M including science) and productive. Funding was then obtained from NSF and ICDP in 1995 (ca. $12M) with the objective of drilling to 4.5km. Drilling was originally planned for a five-year period, in two campaigns. The first campaign, in 1999, resulted in efficient coring to a depth of 3.1km over a period of 6 months; it used about 40 percent of the funds and was also highly productive. Deepening the hole below 3.1km turned out to be both difficult and expensive, although for interesting reasons. To facilitate deeper drilling the hole needed to be reamed to a larger diameter; but when this was done the well unexpectedly started to flow. We now know that there are several deep pressurized aquifers, with varying salt content, but these hydrological phenomena were totally unanticipated. A key finding, also unanticipated, is that cold seawater circulates through the volcanic pile in volumes sufficient to refrigerate the entire section below 700m depth to temperatures about 25 degrees below a normal geothermal gradient. In early 1999 when the first drilling campaign was organized, the price of oil was 10 USD (rigs and drilling crews were available and reasonably priced); in early 2003 when hole opening was being arranged, the price of oil was 30 USD, and for the coring campaigns in 2005 and 2007 it was 50 to 70 USD. For these reasons, and because trip times were longer and deeply buried pillow basalts more difficult to drill, the remainder of the project funds (and then some) were needed to deepen the hole from 3.1 to 3.5km. Nevertheless, the project obtained a nearly continuous, and virtually unweathered, core consisting of lava flows, hyaloclastite, minor intrusives and sediment from a 3260m section of the Mauna Kea volcano, covering an age range from 200 to over 600 ka. It also recovered a 250m and a 280m section of the Mauna Loa volcano. A wealth of geological, volcanological, petrological, geochemical, geomagnetic, geodynamic, hydrological, and geobiological data have come from the core and the well, and more are coming in. The unprecedented geochemical-petrological data sets are a major success, as is the fact that geochemists can work together, but the hoped-for detailed geochronology for the core has proven difficult to obtain.
V13D-2143
Geochemical Stratigraphy of Core from the Hawaii Scientific Drilling Project: The Final Hurrah.
The Hawaii Scientific Drilling Project (HSDP) previously recovered about 3100 m of core, documenting up to 550 ka of Mauna Kea volcano's eruptive history. Here, we report on the geochemistry of an additional 408 m of core (3098 - 3506 mbsl), recovered during the final phase of the project, which extends the eruptive history of Mauna Kea by an additional 18 ka. Most of the analyzed samples (40) are of submarine pillow lavas, with minor hyaloclastites and pillow breccias. Two samples are intrusive. All the extrusive samples belong to the previously identified high-silica group (Rhodes and Vollinger, 2004; Stolper et al., 2005): the dominant magma type erupted throughout most of Mauna Kea's tholeiitic shield-building stage. The presence of this magma type is indicative of relatively uniform, and shallow, magma generation processes throughout most of the shield-building stage. In contrast with younger, high-silica lavas, these lowermost samples show more variation in trace element abundances and ratios (particularly Zr/Nb, Nb/Y etc). This indicates increased variability in source components during the last 18ka compared to the limited variability of the younger high-silica lavas that erupted over a much longer time span of around 270 ka. The two intrusive samples are low-silica lavas and were probably intruded during the relatively short period (490 - 530 ka) when both low- and high silica lavas were erupting. This implies that the eruption of low-silica lavas was a brief event and not reflective of an earlier period of prolonged low-silica, Loihi-like, magmatism prior to the eruption of the dominant high-silica lavas.
V13D-2144
HSDP: The Lost Volcano
We measured high-precision Hf, Nd, and Pb isotope compositions for 40 samples from the final leg of the Hawaii Scientific Drilling Project (HSDP) core using solution chemistry and MC-ICP-MS. This final leg extends the drill core from 3097.7 mbsl down to a depth of 3505.7 mbsl. The new isotope data are indistinguishable from those higher up in the Mauna Kea part of the core and vary from, respectively, +11.6 to +12.3, +6.4 to +7.0, and 18.2948 to 18.6819 for εHf, εNd, and 206Pb/204Pb. The Pb isotope data define three linear trends with a common intersection. Principal component analysis recognizes no more than three geochemical end-members (99.8% of the variance) among these trends. When all the Pb isotope data acquired so far ([1-3] and this work) are pooled, two contrasting groups stand out, mostly with respect to 208Pb/206Pb. (1) A first coherent, prevalent group is identified, which represents the Kea main eruptive sequence. This group combines the Kea-mid8 and the Kea-lo8 subgroups of [3]: although these subgroups define two separate trends in Pb-Pb isotope space, they smoothly follow each other in time. (2) A second group of 15 flows at depths below 2000 mbsl is characterized by distinctly higher source Th/U and, for the samples above the new core samples, also high 3He/4He. These samples belong to the K/L group of [2] and to the Kea-hi8 group of [3] and are about 0.4 (Hf) and 0.5 (Nd) ε units below the main sequence Kea values. The toggle between these two groups of flows is so sharp and repetitive and the gap between the values so conspicuous that the alternation must reflect eruptions from distinct volcanic centers, Mauna Kea and an unknown volcano (K/L) that bears some isotopic resemblance to Kilauea for Nd and Hf and to Loihi for He [2, 4]. This 'lost' volcano stopped erupting about 550 ka ago, i.e., 200 ka before the oldest ages known for Kilauea [5], and is presently buried under 2-3000 m of Mauna Kea flows. We left out the data on the lost volcano and combined the new Nd, Hf, and Pb isotope data with literature data to reevaluate the periodograms of [2] and [3] on an expanded database of about 150 samples. No particular wavelength has a power that unambiguously exceeds the 95% significance level. With the exception of a couple of minor wiggles in Hf at shallow depth, the isotopic signals show overall smooth up- core variations with little short-distance variability. This indicates that, at any given depth, the isotopic variability of the plume conduit is very small. The observed homogeneity is ascribed to radial mixing across the plume conduit. The wobble of rheological heterogeneities, which reflects the variable petrological nature of the upwelling mantle, in the plume flow is the likely cause of mixing. [1]Abouchami et al., 2000. Chem. Geol. 169, 187-209 [2] Blichert-Toft et al., 2003. Geochem. Geophys. Geosyst. 4, 8704, doi:10.1029/2002GC000340 [3] Eisele et al., 2003. Geochem. Geophys. Geosyst. 4, 8710, doi:10.1029/2002GC000339 [4]Abouchami et al., 2005. Nature 434, 851-856 [5] Quane et al., 2000. J. Volcano. Geotherm. Res. 102, 319-338.
V13D-2145
Olivine compositions from the Hawaii Scientific Drilling Project, Phase 2: Evidence for a peridotite mantle source region
To the extent that mantle plumes reflect whole mantle convection, Hawaii may provide the clearest window
into Earth's lower mantle. Samples from the Hawaii Scientific Drilling Project (HSDP) thus provide valuable
tests for models of mantle mineralogy and composition. In this vein, it has been argued recently that Hawaiian
olivines, especially those from the shield-building phase as sampled by HSDP, are so high in Ni (Sobolev et
al., 2005, 2007), and that Hawaiian whole rocks are so low in CaO (Herzberg, 2006) and high in SiO2
(Hauri, 1996) that a peridotite mantle source cannot generate such compositions. The Hawaiian plume, so
the argument goes, is thus supposedly rich in pyroxenite, and possibly olivine-free. However, comparisons of
HSDP olivines to lherzolites, and HSDP whole rocks to lherzolites and partial melting experiments belie these
premises. Testable predictions of the pyroxenite model also fail. New comparisons instead show that
Hawaiian lavas can be produced from a peridotite source.
First, it is unclear that the Hawaiian source is enriched in NiO. The NiO contents of olivines hosted by
lherzolites (GEOROC) have the same range as olivines from the HSDP; indeed, the maximum NiO for olivines
from lherzolites (0.6 wt. %) is as high as that reported for olivines from any oceanic volcano locality. There is
a compositional separation between lherzolite- and HSDP-hosted olivines. But HSDP olivines are not NiO
enriched so much as lherzolite olivines are higher in Fo at a given NiO. Lower Fo contents at Hawaii (at a
given NiO) ensue because olivine compositions there follow a liquid line of descent, where both Ni and Mg
decrease with differentiation. In contrast, subsolidus equilibria involving orthopyroxene enforce a higher and
less variable Fo content for lherzolite-derived olivines. Moreover, the pyroxenite mantle model predicts that
whole rocks with low CaO and high SiO2 should host olivines with high NiO. But in HSDP samples,
neither correlation is evident. Whole rock compositions also support the peridotite mantle view. Primitive
HSDP lavas (15%
V13D-2146
IRON-ENRICHMENT OR SECONDARY GARNET PYROXENITE: EVIDENCE FROM CaO- K2O-RICH HSDP2 MAUNA KEA GLASSES, HAWAII
Precise mass-spectrometric analyses of the Fe/Mn ratio in mantle-derived magmas reveal that Hawaiian shield magmas in general have higher Fe/Mn (>65) than MORB (52-57) and Icelandic magmas (<60). Under the assumption that the source of Hawaiian shield magmas is peridotite-dominated, Humayun et al. (2004) interpreted the high Fe/Mn in Hawaiian magmas as a result of excess iron in the source, attributed to core-mantle interaction. However, based on the high SiO2 content in Hawaiian magmas and the high Ni content in Hawaiian olivines, Sobolev et al. (2005; 2007) argued that an olivine-free, pyroxenitic source contributes these geochemical features of Hawaiian magmas. Additionally, the pyroxenitic source model might account for the elevated Fe/Mn of Hawaiian magmas because compared to Mn, Fe is more compatible in olivine and less compatible in pyroxenes and garnet, and consequently, partial melts of olivine-free garnet pyroxenite can have high Fe/Mn relative to the bulk source. These two rather different views of the origin of the sources of Hawaiian magmas (and their distinctive Fe/Mn ratios) have important geodynamic implications for the Hawaiian plume relating to either the depth of origin at the core-mantle boundary, or to the amount of recycled crust in the source. These contrasting models (i.e., an Fe-enriched peridotitic source vs. an olivine-deficient source produced via dacite interacting with a non-Fe-enriched peridotite) can be tested by comparing magmas produced by partial melting of peridotite with those produced by partial melting of pyroxenite. For example, a specific prediction of the dominantly pyroxenitic source model is that Hawaiian magmas generated by melting of peridotite should have Fe/Mn<60, similar to MORB and Icelandic magmas. In this study, we focus on a small group of Mauna Kea volcanic glasses, characterized by low SiO2 contents and high CaO and K2O contents, recovered from the Hawaiian Scientific Drilling Project Phase 2 (HSDP2) drill core (Stolper et al., 2004). Herzberg (2006) has argued that the high CaO contents of these glasses are strong evidence of a peridotitic source and while some other Hawaiian magmas (e.g., alkalic magmas from Loihi) are also generally accepted to have dominantly peridotitic source, a focus on the distinctive HSDP2 CaO- K2O-rich glasses provides a clean test of whether Hawaiian peridotitic source have elevated Fe/Mn ratios like other Hawaiian magmas, or more MORB-like ratios as predicted by Sobolev et al. (2005; 2007). We report precise mass-spectrometric Fe/Mn ratios and transition metal abundances determined in situ by laser ablation ICP-MS on a suite of seven HSDP2 CaO-K2O-rich volcanic glasses. Six of the seven glasses have Sc>38 ppm, higher than any previously published Hawaiian shield lavas (<34 ppm), and they plot well within the MORB field. Since Sc is compatible in garnet and clinopyroxene, the high Sc abundances in HSDP2 CaO-K2O-rich glasses are consistent with their peridotitic origin proposed by Herzberg (2006). These HSDP2 CaO-K2O-rich glasses have Fe/Mn>63, and plot at the low-MgO end of the olivine control trend defined by Hawaiian lavas. Therefore, provided the CaO-K2O-rich glasses are derived from a peridotitic source as concluded by Herzberg (2006), we conclude that the high Fe/Mn in Hawaiian lavas is not a reflection of partial melting of an olivine-deficient source but, rather, it reflects an iron enrichment in the lavas' source, the Hawaiian plume, which may originate from core-mantle interaction.
V13D-2147
U-Th/He Ages of HSDP-2 Submarine Samples
Hawaiian lavas have been used to investigate the life-cycles of hotspot-traversing volcanoes. The ~ 3500m core recovered by the Hawaii Scientific Drill Project, phase 2 (HSDP-2) has proven invaluable in refinement of models that link plume structures and melting rates to subaerial growth and geochemical evolution. Accurate age dating of lavas is critical in linking geochemical observations to plume characteristics; however, young ages and potassium-poor lavas limit the precision of argon methods. The U-Th/He method on olivine phenocrysts has been used to successfully date Hawaiian post-shield alkali basalts and flood basalts from the Snake River Plain. We are applying the method to olivine-rich lithologies in the HSDP-2 core in an attempt to gain further information about the growth rates of Hawaiian volcanoes. Preliminary results indicate that the method could help refine the flow chronology, but that modifications to the analysis procedure may be necessary to optimize the results. A subaerial Mauna Kea tholeiitic basalt from 528m depth yields a U-Th/He age of 485 +/- 100 ka (sample SR0222), slightly older than expected based on previous determinations on stratigraphically bounding flows by the argon isochron method (Sharp et al., Gcubed, 2005). A submarine hyaloclastite sample from 2931m depth (SR930) yields a preliminary age of 650+/-100 ka, which agrees well with previous Ar measurements. A pillow lava, SR0964, was also investigated, but it yielded a complicated He release pattern and no age can be obtained. U and Th concentrations in olivine separates from all three samples are low (2.6 - 5.2 ppb U; 4.5 - 8.0 ppb Th). The submarine samples appear to have a substantial amount of magmatic helium still remaining in the olivine after in vacuo crushing, as evidenced by high R/Ra values in gas released at high temperature. Residual gas left after crushing may be up to 85% magmatic, which makes the determination of radiogenic He less accurate. Sulfur contents of glass from the host submarine lava samples are high, indicating that the lavas were incompletely degassed. Helium isotopic ratios measured from the crushing step are within error of previously published values (R/Ra = 12.4 and 12.9). Extreme R/Ra values in the 530° C pre-fusion extraction (R/Ra = 756 for SR0964 and 68 for SR0930) suggest that high-T degassing may preferentially release 3He. We are continuing a helium isotopic exploration of the lowermost 450m of HSDP-2, which will provide important information about He isotopes in the core of the Hawaiian plume, and focusing further U- Th/He investigations on samples with low-S glass.
V13D-2148
Hydrothermal Alteration of Hyaloclastites Adjacent to Sill-Like Intrusives in the HSDP 3-km Core Hole.
Hyaloclastites at present depths below1880 mbsl on the submarine flanks of Mauna Kea volcano have been intruded by numerous, < 10 m-thick, sill-like bodies. The contact metamorphism of the hyaloclastites has resulted in up to 1 m-thick bleached zones, characterized by the presence of Na-and Mg-enriched alteration rinds on sideromelane clasts as well as the precipitation of drusy hydrothermal clinopyroxene (calcic augite to hedenbergite) and analcime within void spaces. The intrusive activity associated with contact metamorphism appears to have occurred early in the diagenetic history of the hylaoclastites, when they possessed porosities of 40-50%, because (1) early induration and pore-filling by hydrothermal minerals apparently strengthened them, preventing significant grain compaction during subsequent burial, (2) hydrothermal minerals have been coated or overgrown by smectite, zeolites, and palagonite during subsequent diagenesis and microbial innoculation, and (3) 87Sr/86Sr ratios of hydrothermal rinds on glass shards, averaging .7069 ± .0006, imply extensive interaction with seawater, whereas 87Sr/86Sr ratios of adjacent palagonitized glass, averaging .7042 ± .0002, imply interaction with comparatively less fluid, presumably after diagenetic pore-filling. Thermal modeling, which assumes (1) convective cooling, (2) that hydrothermal clinopyroxenes formed at minimum temperatures of 350°C, and (3) that hyaloclastite porosities approached 50% at the time of intrusion, implies that the observed contact aureoles must have been produced by mafic intrusions that maintained temperatures above the solidus rather than being rapidly cooled and frozen. This may have occurred because magma continued to flow in the intrusion conduit, consistent with the suggestion that these intrusions fed overlying pillow flows (Garcia et al., 2007). If this intrusive activity occurred at shallow depths within the edifice of Mauna Kea (Seaman et al. 2004), then hydrothermal clinopyroxene must have precipitated from superheated steam, as hydrostatic pressures would have been too low to allow precipitation from a liquid aqueous phase.
V13D-2149
Complex Isotope Mixing Arrays Arising From Upper Mantle Dynamics and Melting of a Veined Mantle Plume
Spatial patterns in geochemistry at hotspots follow mixing trajectories in compositional diagrams that offer clues regarding the mantle feeding the hotspot. At many hotspots, the observations can only be explained with three or more lithologic components in the mantle; the distribution of these components in the source is unknown but is important for understanding the scales and styles (layered versus non-layered) of heterogeneity in the deep source. For example, 208Pb/204Pb vs. 206Pb/204Pb data at Hawaii follow many mixing trajectories, which when interpreted as two-component binary mixing, suggest that a volcano samples many different components over its lifetime, and that the Kea and Loa sub-chains do not share at least one component. We use a 3D geodynamic model of a hot, veined mantle plume interacting with a lithospheric plate to explore the spatial patterns in magma geochemistry that result from upper mantle dynamics and melting of a mantle with small-scale heterogeneities. We assume that there are three compositional components in the mantle each with different mass fractions, solidi, and trace element and isotope compositions. The deeper melting components have wider melting zones, and thus can be sampled proportionally more heavily at the edges of the hotspot compared to near the center where the shallower melting components are prominently expressed. Also, plume-lithosphere interaction offsets the melting zones, which causes the predicted geochemical pattern at the surface to be asymmetric in the direction of plate motion. In general, the magma composition at the surface is very different from the initial average solid composition in the mantle. Models predict that in the lifetime of a volcano passing over the hotspot, magma composition follow pseudo-binary mixing trajectories that in some cases do not appear to share the same end-member compositions even though all components are present everywhere in the mantle. An interesting result is that predicted 208Pb/204Pb vs. 206Pb/204Pb mixing trajectories erupted by single volcano are not always linear or those that are linear have different slopes. These results are augmented if the mass fraction of one of the veined components increases with proximity to the center of the plume. The results illustrate that complicated, nonlinear mixing trajectories can result from variable melting of a relatively simple veined mantle with little or know regional scale zoning in compositions.
V13D-2150
Is the observed chemical heterogeneity of the Hawaiian mantle plume real or a result of melting and mixing processes?
The chemical and isotopic heterogeneity of lavas derived from Hawaiian volcanoes has resulted in a standard model of a compositionally zoned Hawaiian mantle plume. Particularly He isotope data from Loihi, Mauna Loa and Mauna Kea have been interpreted in terms of a concentrically zoned plume, with some of the Hawaiian volcanoes having never been above the plume center. Based on He, Ne and Ar fusion data from olivines derived from Mauna Kea (surface and Hawaii Scientific Drilling Project samples), Mauna Loa, Kilauea and Kohala (Big Island) as well as Haleakala (Maui) we show that the Hawaiian plume might not be zoned but that the heterogeneities observed in the lavas can be derived from melting and mixing processes. This is in accordance with other, recently developed models. The He isotopic ratios of the studied samples vary from 7 to 18 RA, whereas Ne isotopic ratios plot along the Loihi-Kilauea line in a Ne-three isotope plot. Thus the He isotopes range from ratios typical for the upper mantle to ratios more typical for a primitive mantle source. All Ne isotopic ratios are typical for a more primitive mantle. This indicates a decoupling of He from Ne as well as a homogeneous plume source. Combined He, Ne and Ar systematics show that this decoupling is caused by predegassing fractionation process, which leaves the He isotope ratio more susceptible for changes than Ne or Ar isotope ratios. Basically, this process caused a He deficit of melts generated by the plume, as shown by 3He/22Ne below current estimates of solar composition and 4He*/21Ne* and 4He*/40Ar* lower than the theoretical production ratios. These observations can best be explained by a model in which He is fractionated from Ne and Ar during formation of plume melts generated by low melting degrees and subsequent mixing with upper mantle melts generated by higher degrees of melting. This requires He to be more compatible during melt formation than Ne and Ar. Depending on the relative proportion of mixing the resulting lavas will cover the whole observed He isotope range, whereas Ne isotopes remain relatively homogeneously distributed.
V13D-2151
Tungsten Abundances in Hawaiian Picrites: Implications for the Mantle Sources of Hawaiian Volcanoes
Tungsten abundances have been measured in a suite of Hawaiian picrites (MgO >13 wt.%) from nine Hawaiian shield volcanoes (Mauna Kea, Mauna Loa, Hualalai, Loihi, Koolau, Kilauea, Kohala, Lanai and Molokai). Tungsten concentrations in the parental melts for these volcanoes have been estimated via the intersection of linear W-MgO trends with the putative MgO content of the parental melt (~16 wt.%). Tungsten behaves as a highly incompatible trace element in mafic to ultramafic systems; thus, given an independent assessment of the degree of partial melting for each volcanic center, the W abundances in their mantle sources can be determined. The mantle sources for Hualalai, Kilauea, Kohala and Loihi have non- uniform estimated W abundances of 11, 13, 16 and 27 ng/g, respectively, giving an average source abundance of 17±5 ng/g. This average source abundance is nearly six times more enriched than Depleted MORB Mantle (DMM: 3.0±2.3 ng/g) and slightly elevated relative to the Bulk Silicate Earth (BSE: 13±10 ng/g). The relatively high abundances of W in the Hawaiian sources relative to the DMM can potentially be explained as a consequence of crustal recycling. For example, incorporation of 30% oceanic crust (30 ng/g W), including 3% sediment (1500 ng/g W), into a DMM source could create the W enrichment observed in the Loihi source, consistent with estimates from earlier models based on other trace elements and isotope systems. The Hualalai source, however, has also been suggested to contain a substantial recycled component, as implied by similarly radiogenic 187Os/188Os, yet this source has the lowest estimated W abundance among the volcanic centers studied. The conflict between these results may: 1) reflect chemical differences among recycled components, 2) indicate a more complex history for Hualalai samples, e.g. involvement of a melt percolation component, or 3) implicate other sources of W.
V13D-2152
Mass intrusion beneath Kilauea Volcano, Hawaii, constraints from gravity and geodetic measurements (1975-2008)
Since January 3 1983, Kilauea Volcano, Hawaii, has erupted almost continuously from vents on the volcano's east rift zone. On March 19, 2008, an explosion at Halema'uma'u Crater, within the summit caldera of Kilauea, marked the opening of a second eruptive vent on the volcano. The east rift vent at Pu'u'O'o and the summit vent at Halema'uma'u continue to be active as of August 2008, marking the longest interval in Kilauea's recorded history of eruptive activity on the volcano. Four gravity surveys with a network covering Kilauea's summit area have been performed during 1975-2003. We reoccupied this 45-station network in January and July 2008 with three portable LaCoste-Romberg gravimeters (G209, G615 and EG026) using a double-looping procedure. These two most recent gravity surveys span the onset of summit eruptive activity. The micro-gravity data set, combined with existing geodetic data from leveling, GPS, EDM, and InSAR, allow us to investigate and model the shallow magma system under the summit caldera to roughly constrain its shape, position, volume change and density, and better understand its long and short term evolution. We corrected for the effect of vertical deformation on gravity data (the so-called free-air effect) using uplift measurements from annual surveys performed by the USGS Hawaiian Volcano Observatory. Preliminary analysis of this record, which covers more than 30 years, indicates a persistent positive residual gravity anomaly located at the southeast margin of Halema'uma'u Crater, very close to the location of the new summit eruptive vent. This anomaly suggests a long term mass accumulation beneath the summit caldera.
V13D-2153
Gravity Anomalies in the Northern Hawaiian Islands: Evidence for an Alternative Magma Chamber on Kauai and a Conjoined Niihau-Kauai Island
The shield stage evolution of the islands of Kauai and Niihau are poorly understood. Previous land-based gravity surveys provide only a coarse constraint on the observed gravitational field. Questions as to whether the island of Kauai was formed by a single or multiple shields and the developmental relationship between these neighboring islands are still debated. Our new land-based gravity survey of Kauai and ship-board gravity surveys around both islands identified large complete Bouguer gravitational anomalies under Kauai's Lihue Basin and offshore in the Kaulakahi Channel, a 30-km-long bathymetric ridge connecting the two islands. These gravitational highs are consistent in size and magnitude with those of other Hawaiian islands and imply local zones of high density crust, most likely attributed to magmatic intrusions; e.g. former magma chambers, or rift zones. The Lihue Basin anomaly observed is offset 20 km east from the geologically mapped caldera region. This offset implies either the unlikely case that the shield stage plumbing system connecting the magma chamber and caldera could have been inclined by up to 75 degrees from the vertical, or that the currently mapped caldera is a late feature, unrelated to shield volcanism. The location of the gravitational anomaly, in the Kaulakahi Channel, 20 km east of Niihau is consistent with geologic mapping, which indicates that Niihau is a remnant of an ancient shield volcano centered east of the island. The proximity of the Niihau gravitational anomaly 10 km from the western edge of Kauai supports the hypothesis that the two volcanoes were part of the same island.