Michael F. Roden

Geochemical characteristics of Koolau Volcano: Implications of intershield geochemical differences among Hawaiian volcanoes

The voluminous shields of Hawaiian volcanoes are dominantly composed of tholeiitic basalts, but there are important intershield geochemical differences. The subaerial lavas forming the ~2–3 Ma Koolau shield have several extreme characteristics: relatively high abundances of SiO2, low abundances of total iron and CaO, and high ratios of LaNb and SrNb. In addition, they range to near bulk-earth strontium, neodymium, and lead isotopic ratios. Although postmagmatic alteration has significantly affected the compositions of some Koolau lavas (decreases in SiO2, K2O, and rubidium contents, increases in total iron and in unusual cases, increases in yttrium and REE abundances), the geochemical charac teristics of unaltered Koolau lavas reflect a distinctive primary magma composition. Within a stratigraphic sequence of lavas, Koolau lavas vary significantly in incompatible element abundance and isotopic ratios, but these variations are not systematic with eruption age, and they are smaller than the differences between Hawaiian shields. Intershield differences in some incompatible element abundance ratios, LaNb and SrNb, are correlated with intershield differences in isotopic ratios, thereby indicating that each shield formed from a compositionally distinct source. However, other intershield compositional differences are not correlated with differences in radiogenic isotope ratios. Some of these compositional differences probably reflect variations in the melting process; e.g., inverse correlations between SiO2 and total iron contents may reflect differences in the pressure of melt segregation, differences in abundances of incompatible elements may reflect variations in mean degree of melting, and variations in ratios like SmNd may reflect the presence of residual garnet. Each shield appears to reflect a unique combination of source components and variables, such as extent of melting and pressure of melt segregation. Consequently, the intershield geochemical differences have important implications for plume structure. Either a relatively large plume has a spatially systematic distribution of geochemical heterogeneities which are sampled by the overlying shields, or each shield is derived from a small radius (<20 km) conduit composed of geochemically distinct diapirs or solitary waves.

New He, Nd, Pb, and Sr isotopic constraints on the constitution of the Hawaiian plume: Results from Koolau Volcano, Oahu, Hawaii, USA

Abstract Most analyzed tholeiitic basalts from Koolau Volcano, Oahu, USA, have strontium ( 87 Sr 86 Sr = 0.7040–0.7043 ), neodymium ( 143 Nd 144 Nd = 0.51270–0.51276 ), and lead ( 206 Pb 204 Pb = 17.8–17.9 ) isotopic compositions near that of the bulk silicate earth, and 3 He 4 He isotopic ratios of 11–14 times the atmospheric ratio. These helium ratios are higher than MORB, but lower than those of lavas from Loihi seamount. Moreover, the source for the Koolau tholeiites is inferred to have non-bulk earth abundance ratios of highly incompatible elements. Consequently, the source of the Koolau lavas is not primitive, undegassed mantle. The abundance ratios La Nb , Zr Nb , and Sr Nb correlate with 87 Sr 86 Sr and 143 Nd 144 Nd ratios in Hawaiian tholeiites. The enriched (Koolau) source component has relatively high La Nb , Zr Nb , and Sr Nb ratios; in fact, Koolau tholeiites have higher Zr Nb and La Nb , and lower Th/Nb than most other OIB. These combined trace element and isotopic signatures of the enriched component are not consistent with derivation from primitive mantle, recycled crustal material, or a carbonatite metasomatized source. A simple explanation is that the enriched component is residual material, formed recently when a small amount of melt was extracted from primitive mantle, perhaps during the incorporation of the Koolau component into the plume.

Ion microprobe analyses bearing on the composition of the upper mantle beneath the Basin and Range and Colorado Plateau Provinces

Ion microprobe analyses of clinopyroxene from predominantly spinel-bearing peridotite xenoliths found in volcanic rocks of the Colorado Plateau (Buell Park, Green Knobs localities) and the Basin and Range (Black Canyon, Lunar Crater localities) provinces are highly variable in light and heavy rare earth element (LREE, HREE), Sr, Zr and Ti contents. Abundances of these elements vary by 2–3 orders of magnitude. The LREE are fractionated from the HREE such that relatively LREE-enriched and LREE-depleted inclusions occur at all localities. However, LREE-depleted clinopyroxenes predominate at the Colorado Plateau localities, whereas LREE-enriched clinopyroxenes or those with chondritic REE patterns predominate at the Basin and Range localities. In addition to fractionation of the LREE from the HREE, the high field strength elements Zr and Ti are generally depleted relative to REE of similar compatibility. Although some of this depletion reflects subsolidus partitioning between orthopyroxene and clinopyroxene, the extremely low Ti abundances (50–500 ppm) of some clinopyroxenes must reflect very low bulk rock contents of Ti. This depletion does not correlate with LREE/HREE ratios. The bulk of the peridotites from the Colorado Plateau are interpreted to be residues from partial melting; extremely low Ce/Sm]cn ratios (less than 0.1) are similar to these ratios in clinopyroxenes from abyssal peridotites and require that the melting process in some cases approached ideal fractional melting; the very low Sr content of these clinopyroxenes (less than 10 ppm) suggest that clinopyroxene was not a residual phase. The association of these peridotites with LREE-depleted websterites which may be crystal segregates from a magma similar to modern mid-ocean ridge basalts reinforces the link between modern oceanic lithosphere and the upper mantle beneath the Colorado Plateau. In contrast, the clinopyroxenes from the Basin and Range tend to have relatively high Ti contents (greater than 2000 ppm) and only moderate fractionation of the LREE from the HREE or to be LREE-enriched and have low Ti contents (as low as 100 ppm) but high Sr contents (to 220 ppm). The former clinopyroxenes are derived from fertile peridotites which could yield typical basaltic magmas upon melting or may have equilibrated with such magmas in the upper mantle. The latter clinopyroxenes are from metasomatic peridotites which likely reacted with a carbonatitic melt These features may be a consequence of widespread Cenozoic magmatism of the Basin and Range Province.

Trace element abundances in mantle-derived minerals which bear on compositional complexities in the lithosphere of the Colorado Plateau

Abstract Proterozoic crust of the Colorado Plateau, southwestern USA, was intruded in the mid-Tertiary by hypabyssal minettes and ultramafic breccias which contain mantle-derived xenoliths and xenocrysts. Pyrope-rich garnets and Cr-rich diopsides from these fragments collected at Red Mesa, Garnet Ridge and The Thumb diatremes were analyzed by secondary ion mass spectrometry for the rare earth elements (REE), Sr, Ti, Zr, Y, V, and Cr. Typical garnet xenocrysts derived from shallow, low temperature peridotites are light REE (LREE) depleted with convex-upward REE patterns, and have relatively high Y abundances and low Zr and Ti abundances. These xenocrysts are characterized by relatively low Zr/Y and high Al/Cr ratios consistent with a lherzolitic source. Uncommon garnet xenocrysts have “sinusoidal” REE patterns and low Y abundances and are Cr-rich but not subcalcic. Sinusoidal REE patterns also characterize some garnets from deeper, high-temperature garnet peridotites; a zoned garnet from one of these xenoliths preserves a core with a sinusoidal pattern overgrown by a rim with a convex-upward REE pattern. The rim of this garnet is also relatively rich in Ti, Zr and Y. Garnet xenocrysts with sinusoidal REE patterns have relatively low Al/Cr, Ti and Y abundances and some contain chlorite inclusions and probably have relatively refractory, and in some instances, hydrous source rocks. Our data further document the occurrence of these garnets outside Archean cratons and although their petrogenesis remains controversial, the garnets examined here could plausibly have formed by subsolidus processes in refractory peridotite. Garnets from more fertile, high temperature lherzolites have convex upward REE patterns and relatively high abundances of Ti, Y and Zr. Compared to the garnet xenocrysts, garnets from the higher temperature garnet peridotites tend to have lower Al/Cr and higher Zr/Y ratios indicating that the xenocrysts are derived from relatively fertile source rocks. Hence the mantle in this region may be stratified with more refractory rocks at greater depths. The mantle beneath the Colorado Plateau lacks pronounced, widespread silica enrichment and subcalcic garnets and is only rarely enriched in incompatible elements at shallow levels. At the time of minette magmatism, the Colorado Plateau was characterized by a cool root analogous to those beneath Archean cratons but compositionally more akin to modern abyssal peridotites.

Weathering of ilmenite from granite and chlorite schist in the Georgia Piedmont

Abstract Ilmenite grains from weathering profiles developed on granite and ultramafic chlorite schist in the Georgia Piedmont were studied for evidence of morphological and chemical alteration. Ilmeniterich concentrates from the fine sand (90-150 μm) component were studied to test the assumption that there is no difference between ilmenite in the parent rock and that in colluvium delivered to primary drainage systems. Ilmenite grains in the granite profile are rounded to subhedral, and commonly contain hematite exsolution blebs. Dissolution pits are observed along the boundaries of the exsolution blebs, with goethite occurring as an alteration product. Ilmenite grains in the schist profile occur as fractured anhedral grains with uncommon lamellae of rutile. Grain fractures are filled with goethite and hematite, particularly in the B-horizon. Ilmenite from the granite profile is Mn rich (7-15 mol% MnTiO3), whereas ilmenite from the schist profile contains only 1-2 mol% MnTiO3 and up to 8 mol% MgTiO3. Two populations of grains develop in both profiles. Grains with abundant exsolution blebs and fractures alter through a proposed two-step reaction mechanism. It is proposed that ilmenite first undergoes a solid-state transformation to pseudorutile via an anodic oxidation mechanism. Oxidized Fe and Mn diffuse from the structure and precipitate as goethite and MnO2. Pseudorutile is ephemeral and undergoes incongruent dissolution to form anatase, hematite, and goethite. The second population of grains experienced only slight oxidation and dissolution on grain surfaces, and they persist through the weathering profile. The Fe2+ content of competent ilmenite grains is somewhat lower in the C-horizon, compared with grains in the host rock. In horizons above the C-horizon, the Fe2+ contents of the ilmenite are similar to those in the host rock. This study shows that using ilmenite minor-element chemistry as a tracer for sediment provenance is a valid technique, however, textural features of ilmenite in colluvium may be distinct from those in the parent rock. Also, the production of secondary phases, such as anatase, goethite, and hematite, in soil profiles results in part from the alteration of ilmenite.

Upper Eocene impact horizon in east-central Georgia

We report the discovery of shocked quartz grains in upper Eocene sediments from the Coastal Plain of east-central Georgia. The grains exhibit low refractive indices and contain one or more sets of planar deformation features formed during a hypervelocity impact. These grains were collected from a sand layer near the base of the Twiggs Clay Member of the Dry Branch Formation and most likely are ejecta from the Chesapeake Bay impact, which occurred between 35.7 and 36.0 Ma. In this layer, ∼1 in 250 quartz grains from the fine-sand size fraction and fewer medium-sand– sized grains show some evidence of shock metamorphism. This horizon could represent the source stratum for the Georgia tektites and potentially contains important evidence revealing the dynamics and environmental effects of the late Eocene cataclysm.

Petrology and geochemistry of the Huerto Andesite, San Juan volcanic field, Colorado

The Huerto Andesite is the largest of several andesite sequences interlayered with the large-volume ash-flow tuffs of the San Juan volcanic field, Colorado. Stratigraphically this andesite is between the regions largest tuff (the 27.8 Ma, 3,000 km3 Fish Canyon Tuff) and the evolved product of the Fish Canyon Tuff (the 27.4 Ma, 1,000 km3 Carpenter Ridge Tuff), and eruption was from vents located approximately 20–30 km southwest and southeast of calderas associated with these ashflow tuffs. Olivine phenocrysts are present in the more mafic, SiO2-poor samples of andesite, hence the parent magma was most likely a mantle-derived basaltic magma. The bulk compositions of the olivine-bearing andesites compared to those containing orthopyroxene phenocrysts suggest the phenocryst assemblage equilibrated at 2–5 kbar. Two-pyroxene geothermometry yields equilibrium temperatures consistent with near-peritectic magmas at 2–5 kbar. Fractionation of phenocryst phases (olivine or orthopyroxene + clinopyroxene + plagioclase + Ti-magnetite + apatite) can explain most major and trace element variations of the andesites, although assimilation of some crustal material may explain abundances of some highly incompatible trace elements (Rb, Ba, Nb, Ta, Zr, Hf) in the most evolved lavas. Despite the great distance of the San Juan volcanic field from the inferred Oligocene destructive margin, the Huerto Andesite is similar to typical plate-margin andesites: both have relatively low abundances of Nb and Ta and similar values for trace-element ratios such as La/Yb and La/Nb.Deriving the Fish Canyon and Carpenter Ridge Tuffs by crystal fractionation from the Huerto Andesite cannot be dismissed by major-element models, although limited trace-element data indicate the tuffs may not have been derived by such direct evolution. Alternatively, heat of crystallization released as basaltic magmas evolved to andesitic compositions may have caused melting of crust to produce the felsic-ash flows. Mafic magmas may have been gravitationally trapped below lighter felsic magmas; mafic magmas which ascended to the surface probably migrated upwards around the margins of silicic chambers, as suggested by the present-day outcrops of andesitic units around the margins of recognized ash-flow calderas.