Oxygen and strontium isotopic studies of basaltic lavas from the Snake River plain, Idaho

Abstract

We present oxygen and strontium isotopic data for olivine tholeiites, (that is, differentiated and (or) contaminated) lavas, rhyolites, and crustalderived xenoliths from the Snake River Plain. These data show that the olivine tholeiites are fairly uniform in 6 180 (5.1 to 6.2) and 87Sr/ 86 Sr (0.7056 to 0.7076) and reveal no correlation between these ratios. The tholeiites are considered representative of mantle-derived magmas that have not interacted significantly with crustal material or meteoric water. The lavas display a wider range in 6 180 (5.6 to 7.6) and 87Sr/86Sr (0.708 to 0.717) with positive correlations between these ratios in some suites but not in others. Crustal xenoliths have high and variable 87Sr/ 86 Sr (0.715 to 0.830) and 6 18 0 values that vary widely (6.7 to 9.2) and are a few permil greater than 6 18 0 values of the Snake River basalts. Thus, isotopic data for the lavas are permissive of small degrees of contamination by crustal rocks similar to the most 18 Q-depleted xenoliths. The 6 180 enrichments in some lavas also are consistent with crystal fractionation processes and do not necessarily recaiire bulk interaction with crustal rocks. Enrichment in 18 0 but not in 8 'Sr/ 86Sr in one suite of lavas suggests that crustal contamination may not be essential to the petrogenesis of those lavas. Other suites of lavas display large variations in 87Sr/86Sr that reflect at least some selective contamination with 87Sr. Bulk solid/liquid oxygen-isotope fractionation factors (a's) calculated for the lavas from Craters of the Moon National Monument are comparatively large. These a's are dependent upon the nature and proportions of phases removed by crystal fractionation; basaltic lava a's differ from latitic lava a's in accordance with different phenocryst assemblages in these rocks. Snake River Plain rhyolites are isotopically distinct from both the analyzed crustal xenoliths and olivine tholeiites. Their origin remains poorly understood, but crustal or sub-crustal sources may be viable. In the first case, they must be derived by anatexis of material distinct from the analyzed crustal xenoliths. In the second case, they must be derived from material unlike the source for tholeiites. No cogenetic relation with the tholeiites seems likely on the basis of available data. INTRODUCTION The Snake River Plain (SRP) is a bimodal basalt-rhyolite volcanic province that extends for 500 km across southern Idaho from the Columbia River Plateau eastward to Yellowstone National Park. Structurally, the western SRP is considered to be a graben bounded by high-angle faults (Malde, 1959), whereas the eastern SRP is essentially a downwarp controlled in part by margin faults (Kirkham, 1931; Trimble and Carr, 1976). Stratigraphic relationships within the SRP (for example, Malde and Powers, 1962; Cam and Trimble, 1963; Stearns and others, 1938) and absolute ages of the volcanic rocks (Armstrong and others, 1975) indicate that volcanic activity became progressively younger toward the east. In any locality within the SRP, volcanism began with voluminous eruptions of high-silica rhyolitic flows and ash-flow tuffs, and was followed by dominantly basaltic eruptions. Initial rhyolitic volcanism ranges in age from about 15 m.y. old to about 2 m.y. old from west to east whereas the oldest basalts range in age from about 13 m.y. old in the west to less than 1. m.y. old in the east. The basaltic rocks are dominantly olivine tholeiites characterized by comparatively high iron enrichment, high TiO£ (average 2.7 weight percent) and P2 05 (0.6 percent), low Si0 2 (46.9 percent) and total alkalies (3.1 percent) (Powers, 1960a; Stone, 1967; Leeman, 1974). The basalts are relatively uniform in composition regardless of age or location. In addition, minor volumes of evolved lavas occur situated mainly on the margins of the SRP. These rocks are typified by lavas at Craters of the Moon (COM) National Monument, which range widely in composition (from ferrobasalt to tholeiitic latite) as the result of low pressure crystal fractionation (Leeman and others, 1976; Stone, 1970). The COM-type lavas may be related to olivine tholeiite parental magmas via high-pressure differentiation (Thompson, 1975; Leeman and others, 1976). However, the presence of crustal-derived xenoliths in some of the COM-type lavas, and anomalies in their strontium and lead isotopic compositions (Leeman and Manton, 1971; Leeman, 1974 and in preparation) suggest that they may also be contaminated by crustal material. The eastward younging of SRP volcanism, as well as geologic and geophysical information for this province, have led numerous authors to suggest that volcanism there is related to movement of the North American plate over a mantle hot-spot presently located beneath the Yellowstone volcanic complex (for example, Smith and Sbar, 1974; Suppe and others, 1975; Eaton and others, 1975; Thompson, 1977; Brott and others, 1978). To evaluate this model it is desirable to determine the source(s) of the SRP magmas. Here we report oxygen isotopic analyses of representative basalts and crustal xenoliths from the SRP for which strontium isotopic analyses are available. Analyzed samples include: (a) 19 olivine tholeiites covering a broad temporal and spatial distribution, (b) 26 COM-type lavas from five localities, (c) 8 crustal xenoliths from three localities, and (d) 6 rhyolite flows and ash-flow tuffs. These data augment analyses by Muehlenbachs and Stone (1973) for nine other SRP samples and, in combination with strontium isotopic and major-element data, provide important constraints on possible source materials for SRP magmas and on processes by which they evolved. ANALYTICAL METHODS Identical splits of sample powders (about 200 mesh) were used in this study for both strontium and oxygen isotopic analyses. Many of the strontium isotopic analyses are taken from Leeman and Manton (1971), and the remaining samples were analyzed on a 6-inch-radius NBS design mass spectrometer with triple filament ionization located at the U.S. Geological Survey, Denver, Colorado. All isotopic ratios are normalized to a 87Sr/ 86Sr ratio of 0.1194, and 87 Sr/86Sr ratios are adjusted to a value of 0.7080 for the Eimer and Amend SrCOs standard. Withinand between-run precision for this ratio is about +0.0002 (two-sigma) or better. After drying the sample powders for one hour at 200°C, oxygen was extracted by reaction with BrF 5 at 550°C in nickel vessels (Clayton and Mayeda, 1963). Oxygen was converted to COz by combustion with a resistance-heated carbon rod (Taylor and Epstein, 1962). Isotopic analyses of CO were performed on a Nuclide 60 RMS mass spectrometer with modified source, collector, and emission regulator. The isotopic data are expressed as 6, the per mil deviation relative to SMOW (Craig, 1961). Twenty-two analyses of NBS-28 reference standard performed during this study yielded a mean 6 180 value of 9.53 + 0.36 (two-sigma), and three analyses of St. Peter sandstone yielded a mean value of 11.10 + 0.16 (two-sigma). The unusually large 2-sigma of NBS-28 reflects some severe Oz-extraction problems during the final stages of the study. The average 6 180 difference between replicate analyses of 22 individual samples (tables 1, 2, and 3) is 0.1 permil. RESULTS 01ivine tholeiites SRP olivine tholeiites were selected for analysis to provide representative spatial and temporal coverage. Our samples are divided into two age groups in table 1; lavas of the Snake River Group (that is, less than about 1 m.y. old), and those equivalent to the Glenns Ferry and Banbury formations of Malde and Powers (1962). The latter rocks range in age from about 3 to 10 m.y. old and, despite an attempt to sample fresh material, some of these rocks display incipient iddingsitic alteration of olivines. Both groups include samples from the western and eastern parts of the SRP. The Snake River Group lavas display a wide range in ICO content (0.14 to 0.83 weight percent), but narrow ranges in 6 180 (5.12 to 5.99) and 87Sr/ 86 Sr (0.7056 to 0.7068), and show little or no correlation between these compositional features. The older lavas also display a wide range in KzO (0.10 to 0.64) and similar ranges in 6 180 (5.06 to 6.21) and 87Sr/ 86Sr (0.7065 to 0.7076). Average 6 18 0 and 87 Sr/86Sr ratios (table 1) are slightly higher for the older group, but the differences are statistically insignificant. Other compositional features of these respective groups are similar except that the older lavas have higher average potassium/rubidium (580) than the younger group (470) (Leeman, 1974). The average <5 18 0 (5.59 + 0.30) and 87Sr/86 Sr (0.7064 + 0.0005) for all analyzed SRP tholeiites are considered representative of these lavas (Leeman and Manton, 1971). m«%a SRP OL THOL COM KING HILL IS^g^^ggg^SS^^Jj MISC COM-TYPE XENOLITHS msmm SRP RHYOLITES E;^i%8&:g^ YNP RHYOLITES (GL) (SAN) a b B NEO-VOLC ZONE REYKJANES M ALKALI BASALTS ICELAND MISC OCEAN IS H MORB r-18 s°o Figure 1A. Range in 6 18 0 values for Snake River Plain (SRP) and Yellowstone National Park (YNP) volcanic rocks and crustal xenoltihs. Data for basalts from this report. Data for rhyolites: SRP (this report), YNP non-hydrated rhyolite glasses (Friedman, personal commun., 1975), YNP sanidines (Friedman and others, 1974) (a Mesa Falls Tuff and Plateau Rhyolite; b Huckleberry Ridge Tuff). (IB) Range in 6 18 0 values for selected oceanic island and mid-ocean-ridge basalts. References are given in text. Table 1. Oxygen and initial strontium isotopic compositions and K 0 content of Snake River olivine tholeiites