Journal of Petrology Advance Access published online on January 7, 2009
Journal of Petrology, doi:10.1093/petrology/egn072
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subduction Influence on Oxygen Fugacity and Trace and Volatile Elements in Basalts Across the Cascade Volcanic Arc
Department of Geosciences, Oregon State University, Corvallis, or 97331, USA
Received July 10, 2008; Revised typescript accepted December 1, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Fluids or melts derived from a subducting plate are often cited as a mechanism for the oxidation of arc magmas. What remains unclear is the link between the fluid, oxygen fugacity, and other major and trace components, as well as the spatial distribution of the impact of those fluids. To test the potential effects of addition of a subduction-derived fluid or melt to the sub-arc mantle, olivine-hosted melt inclusions from primitive basaltic lavas sampled from across the central Oregon Cascades (43°–45°N) have been analyzed for major, trace and volatile elements and fO2. Oxygen fugacity was determined in melt inclusions from sulfur speciation determined by electron microprobe and from olivine–chromite oxygen geobarometry. The overall range in fO2 based on sulfur speciation measurements is from <–0·25 log units to + 1·9 log units (
FMQ, where FMQ is fayalite–magnetite–quartz buffer). Oxygen fugacity is positively correlated with fluid-mobile trace element and light rare earth element contents in basalts generated by relatively low-degree partial melting. Establishing a further correlation between fO2 and fluid-mobile trace element abundances with position along the arc requires the basalts to be subdivided into shoshonitic, calc-alkaline, low-K tholeiite and enriched intraplate basalt groups. Melt inclusions from enriched intraplate and shoshonitic lavas show increasing fO2 and trace element abundances closer to the trench, whereas calc-alkaline melt inclusions exhibit no significant across-arc variations. Low-K tholeiitic melt inclusions record an increase in incompatible trace elements closer to the trench; however, there is no correlated increase in fO2. The correlation observed in enriched intraplate and shoshonitic melt inclusions is interpreted to reflect a progressively greater proportion of a fluid-rich, oxidized subduction component in magmas generated nearer the subduction zone. Significantly, calc-alkaline melt inclusions with high ratios of large ion lithophile elements to high field strength elements, characteristic of ‘typical’ arc magmas, have oxidation states indistinguishable from low-K tholeiite and enriched intraplate basalt melt inclusions. The lack of across-arc geochemical variation in calc-alkaline melt inclusions may suggest that these basalts are not necessarily the most appropriate magmas for examining recent addition of a subduction component to the sub-arc mantle. Flux and batch melt model results produce a wide range of predicted amounts of melting and subduction component added to the mantle source; however, general trends characterized by increased melting and proportion of the subduction component from enriched intraplate, to low-K tholeiite, to calc-alkaline are robust. The model results do not require enriched intraplate, low-K tholeiite and calc-alkaline magmas to be produced from the same more fertile mantle source. However, enriched intraplate magmas, in contrast to calc-alkaline and low-K tholeiite magmas, cannot be generated from a depleted mantle source. Flux or batch melting of either the more fertile or depleted mantle sources used to generate the low-K tholeiite, calc-alkaline, and enriched intraplate magmas cannot reproduce shoshonitic compositions, which require a significantly depleted mantle source strongly metasomatized by a subduction component. The potential mantle source for shoshonitic basalts has a predicted fO2 (after oxidation) from + 0·3 to + 2·4 log units (FMQ) whereas the mantle source for low-K tholeiite, calc-alkaline, and enriched intraplate magmas may range from –1·1 to + 0·7 log units (FMQ). KEY WORDS: basalt; Cascades; melt inclusions; oxidation state; volatiles
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
The sub-arc mantle oxidation state in subduction zones can affect mineral phase equilibria, the transfer of multivalent trace elements (e.g. V, Eu, and Cr) and speciation of volatile phases in primitive magmas. However, direct measurement of mantle oxygen fugacity in volcanic provinces is often limited to study of peridotite xenoliths entrained within basalts (e.g. Brandon & Draper, 1996
; Parkinson & Arculus, 1999). Because of the rarity of such xenoliths in continental volcanic arcs, broad generalizations of sub-arc mantle fO2 conditions must often be made on the basis of relatively limited data (Brandon & Draper, 1996). As a consequence, measured, calculated, and inferred oxidation states of primitive basalts have been widely used in an attempt to estimate differences in mantle source oxidation states (e.g. Arculus, 1985; Christie et al., 1986; Ballhaus et al., 1990; Ballhaus, 1993; Conrey et al., 1997; Li & Lee, 2004; Lee et al., 2005).
Most studies of the impact of partial melting, degassing and crystallization on basaltic oxygen fugacity have focused on ocean-island and mid-ocean ridge basalts (OIB and MORB, respectively) where melting and crystallization processes are typically thought to be better constrained (Christie et al., 1986; Bezos & Humler, 2005). This investigation represents the first attempt to quantify the oxidation states of subduction-related basalts, taking into account the effects of addition of a subduction component and variable degrees of partial melting and thus putting mantle source fO2 variability into a spatial context across a subduction zone. The oxidation states of basalts may provide a reasonable measure of the relative differences in mantle oxidation state, although there is also the potential for modification resulting from processes such as partial melting, crystallization, crustal assimilation, degassing and diffusion (Sato & Wright, 1966; Mathez, 1984; Carmichael & Ghiorso, 1986, 1990; Bezos & Humler, 2005).
In subduction zones, as with other volcanic settings, basalt oxidation states may provide insights into the degree and type of compositional heterogeneity within the mantle source. Spatially associated, but compositionally distinct basaltic lavas are often observed in subduction zone magmatism (e.g. Bacon et al., 1997). This diversity has been attributed to variations in the degree of subduction-fluid induced flux melting (Stolper & Newman, 1994; Reiners et al., 2000; Walker et al., 2003) and/or mantle source heterogeneity (e.g. Bacon et al., 1997; Conrey et al., 1997; Hochstaedter et al., 2000; Churikova et al., 2001; Borg et al., 2002; Leeman et al., 2005). Addition of a subduction component (SC) to the subarc mantle has also been shown to exert significant control on the melting conditions and geochemical characteristics of basaltic magmas in subduction zones (e.g. Plank & Langmuir, 1993; Stolper & Newman, 1994; Hochstaedter et al., 2001). The subduction component is commonly interpreted as a fluid and/or hydrous silicate melt rich in incompatible, fluid-mobile elements (i.e. Cl, Ba, K, Na, La, Pb, Sr, U, Rb) derived from the dehydration or partial melting of the subducting oceanic plate and overlying sediments (e.g. Tatsumi et al., 1986; Schmidt & Poli, 1998; Stern, 2002).
The goal of this study is to evaluate the effects of the addition of a subduction component to basaltic magmas by comparing variations in oxidation state with major, trace and volatile (S, Cl) element variability from distinct basalt groups across the central Oregon segment of the Cascade volcanic arc. In an attempt to reduce the impact of crustal assimilation, degassing, and crystallization, which can alter the composition and oxidation state of magmas, we have focused on olivine-hosted (Fo77–90·5) melt inclusions from primitive lavas (whole-rock MgO >8 wt % for the fore-arc and arc and >6 wt % MgO for back-arc lavas). Basalt oxidation states are determined in situ for melt inclusions based on measurements of sulfur speciation (e.g. Carroll & Rutherford, 1988).
The central Oregon Cascade region has abundant olivine-phyric primitive basaltic and basaltic andesite flows and cinder cones, making this an exceptionally mafic segment of the Cascade volcanic arc (Fig. 1; Sherrod & Smith, 1990). In addition, this is one of the few localities (along with southern Washington and northern California) in the Cascades where there is abundant, young (<1 Ma) fore-arc and back-arc volcanism, providing over 100 km of semi-continuous, coeval samples of basaltic magmatism across the arc (Guffanti & Weaver, 1988). Taken together, these characteristics make the central Oregon Cascades a well-suited locality for a tightly constrained study of this type.
|
|
OREGON CASCADES BACKGROUND |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Petrology and geochemistry
Previous geochemical and petrological studies in the Oregon Cascades have identified up to four distinct basaltic compositions based largely on trace element concentrations (Bacon, 1990
; Leeman et al., 1990; Bacon et al., 1997; Conrey et al., 1997; Leeman et al., 2005; Schmidt et al., 2008). These endmembers are commonly referred to as low-K tholeiitic basalts (LKT; also referred to as MORB-like basalt and high-alumina olivine tholeiites or HAOT), enriched ocean island-like basalts (OIB-like) or high field strength element (HFSE)-enriched basalts, calc-alkaline basalts (CAB), and shoshonitic basalts [including both absarokites and high-K CABs of Conrey et al. (1997)]. Some ambiguity exists, however, regarding the distinction and classification of these compositional groups. For the purpose of this study, we will retain the LKT, CAB and shoshonitic nomenclature. However, because of the genetic implications of the term OIB we have chosen not to continue the usage of the term OIB-like. Similarly, HFSE-enriched is equally unsuitable in that these basalts are enriched in elements other than HFSE. We have therefore chosen a generic geochemical description of enriched intraplate basalt (EIB). Categorization of these basalts is based dominantly on whole-rock trace element characteristics; however, these groups also have distinctive phenocryst phases and groundmass textures. Calc-alkaline basalts are dominated by both olivine and plagioclase phenocrysts whereas EIB basalts have olivine phenocrysts with olivine + plagioclase in the groundmass. The phenocryst assemblage of shoshonitic basalts ranges from olivine and clinopyroxene in mafic compositions to olivine and plagioclase in more evolved compositions. Trace fluor-phlogopite and fine-grained apatite phenocrysts have also been identified in shoshonitic basalts. Low-K tholeiitic lavas are distinguished by a diktytaxitic texture with plagioclase + olivine ± clinopyroxene.
Distribution of basalts
The Oregon segment of the Cascade volcanic arc is divided into two geographical provinces, the Western Cascades (40–10 Ma) and the High Cascades (10–0 Ma). The high volume of mafic volcanism present in the central Oregon Cascades is largely the result of intra-arc rifting, estimated to have initiated 
5 Myr ago in the central Oregon High Cascades (Sherrod & Smith, 1990; Conrey et al., 2002). The northward propagation of the Cascade rift has been most often interpreted to be related to the encroachment of Basin and Range extension into the volcanic arc and the clockwise rotation of the Willamette microplate (Fig. 1; Wells, 1990; McCaffrey et al., 2000). Basaltic compositions of all of the previously discussed types are observed within the graben over a distance of less than 30 km across the arc (Conrey et al., 1997). The only young fore-arc basalt included in this study is a shoshonitic lava located within the Western Cascade province dated at 82·3 ± 3·1 ka (Rowe, 2006). Although a limited number of other relatively young calc-alkaline basalts are present in the fore-arc they have been dated at 3 Ma and are outside the range of the current study (Walker & Duncan, 1989).
Back-arc volcanism younger than 1 Ma along the Cascade volcanic arc is limited to three general locations; the Simcoe volcanic field in southern Washington, Newberry Volcano in central Oregon, and Medicine Lake in northern California (Fig. 1). Newberry is a shield volcano situated 50 km east of the axis of the High Cascades and encompasses an area of 1300 km2 (Fig. 1; Jensen, 2000). Despite its location distinctly behind the Cascade arc, the relation of Newberry to Cascadia subduction is difficult to interpret because of its complex tectonic setting and the convergence of multiple volcanic regimes. Newberry Volcano is situated at the intersection of the northernmost Basin and Range normal faults and the Brothers fault zone, a NW-trending fault zone intersecting the Cascades to the north of Newberry (Higgins, 1973; Jordan et al., 2002). Prior studies suggested that Newberry Volcano is related to High Lava Plains volcanism, extending east–west across Central Oregon along the northern boundary of the Basin and Range (Jordan, 2005). However, the presence of dominantly calc-alkaline volcanic rocks, compositionally similar to those found in the Cascades to the west, provides a basis for interpreting Newberry as a back-arc volcanic center, related to the Oregon Cascades in an analogous position to Medicine Lake volcano relative to Mount Shasta (e.g. Kinzler et al., 2000; Elkins Tanton et al., 2001). Additionally, helium isotope ratios for the Newberry lavas are consistent with those of Cascade arc magmas and substantially lower than High Lava Plains volcanic rocks (Graham et al., 2009).
As with the High Cascades, the primitive basaltic lavas at Newberry Volcano also exhibit a wide range of compositions. Although CABs predominate, both EIB and LKT lavas also occur, particularly on the flanks of the volcano. Only shoshonitic compositions appear to be absent in the Newberry Volcano region. Taken together, Newberry Volcano, the High Cascades and the young volcanism present in the fore-arc represent an opportunity to examine samples of primitive basalts corresponding to the distinct compositional endmembers from over 100 km across the Cascade volcanic arc (Fig. 1).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Sample collection
The region of study was limited to 43–45°N and basalts younger than
1 Ma. The goal in limiting the study spatially and temporally was to reduce the potential impact of along-axis variation in mantle composition and temperature, temporal variations in the subduction component, convergence rates, and thermal state of the subduction zone. Detailed mapping and whole-rock geochemical analysis of basaltic lavas within this region have identified young, compositionally diverse primitive lavas, from which we have chosen various compositions for the present melt inclusion study (Beyer, 1973; Higgins, 1973; MacLeod & Sherrod, 1988; Linneman, 1990; MacLeod et al., 1995; Conrey et al., 1997, 2002; Jordan, 2001; Sherrod et al., 2004; J. Donnelly-Nolan, unpublished data). High Cascades and fore-arc basalt samples in this study have >8 wt % MgO and back-arc basalts have >6 wt % MgO; all samples are olivine bearing (Table 1). Whole-rock major and trace element concentrations for basalts in this study were determined by X-ray fluorescence (XRF) and inductively coupled plasma mass spectrometry (ICP-MS), respectively, at Washington State University (Knaack et al., 1994; Johnson et al., 1999). Samples of lava and scoria were collected from vent or near-vent localities when possible. However, because of the rapid burial of lavas and erosion, resulting in the loss of vent material, more distal lavas had to be collected in some cases.
Sample preparation and re-homogenization
Olivine grains were handpicked from crushed lava and scoria samples. Melt inclusions in olivine grains were examined in ethanol under a petrographic microscope and the degree of post-entrapment crystallization was estimated. In most cases inclusions were found to be partially (<5%) to completely crystalline, requiring re-homogenization (excluding samples DB04-1 and QV04-3B). Melt inclusions were re-homogenized at 1 atm in a Deltech vertical furnace following the procedure described by Rowe et al. (2006). Inclusions were held at near-liquidus temperatures, calculated from bulk-rock compositions using COMAGMAT (Ariskin et al., 1993), for 10 min to ensure re-homogenization, and then rapidly quenched. Total heating time at temperatures above 1000°C was limited to 15 min to reduce the potential for hydrogen diffusion (Danyushevsky et al., 2002; Hauri, 2002). Oxygen fugacity within the furnace was maintained at the fayalite–magnetite–quartz (FMQ) oxygen buffer with a CO2–H2 mixture. Following re-homogenization, olivine grains were mounted in 25 mm diameter epoxy molds and polished to expose the melt inclusions. Reflected and transmitted light images of melt inclusions were used to document inclusion sizes and confirm that re-homogenization resulted in the melting of all daughter crystals.
Analytical techniques
Olivine grains, trapped spinel inclusions, and melt inclusions were analyzed by electron microprobe (Cameca SX-50 and SX-100) at Oregon State University for major and volatile (S, Cl) elements. The size of the electron beam used for the analysis depended on the size of the inclusion; a 4 µm beam was used for the smallest inclusions and a 10 µm beam for the largest. Olivine grains and spinel inclusions were analyzed with a focused 1 µm electron beam. Beam conditions were 15 keV accelerating voltage and 30 nA (for glass) and 50 nA (for mineral phases). The analytical procedure for glass was optimized for analysis of S (30 s peak count time) and Cl (100 s peak count time). BCR-2G (Kent et al., 2004, and references therein), a Loihi glass (LO-02-04ii; Kent et al., 1999), Makaopuhi Lava Lake glass (USNM 113498/1 VG-A99), chromite (USNM 117075), and Springwater meteorite olivine (USNM 2566) were analyzed under identical analytical conditions to monitor accuracy and precision.
Following analysis, melt inclusion compositions were recalculated based upon the assumed equilibrium composition relative to their host olivine, by incrementally adding or removing olivine to the melt composition until an olivine–melt KDFe–Mg of 0·3 was reached (Roedder & Emslie, 1970; Sobolev & Chaussidon, 1996). For the purpose of recalculations for CAB, LKT, and EIB melt inclusions, Fe3+/Fe2+ ratios were calculated using the formulation of Kress & Carmichael (1991), assuming an oxidation state at the FMQ oxygen buffer, following the method described by Kent et al. (1999). For the fore-arc shoshonitic basalt, melt Fe3+/Fe2+ ratios were calculated with the empirical calibration of Sack et al. (1980) using calculated fO2 from sulfur peak shift measurements (Rowe et al., 2007). This study used only those inclusion compositions requiring less than 10 wt % addition or subtraction of olivine to establish calculated re-equilibration. In addition, inclusion compositions were carefully examined with respect to evidence of post-entrapment re-equilibration or Fe loss (Danyushevsky et al., 2000, 2002). Only rehomogenized inclusions with sulfur concentrations at or above the Fe-sulfide saturation curve (Wallace & Carmichael, 1992) are considered in this study. Sulfur concentrations in basaltic melt are controlled dominantly by S speciation (Fig. 2) where S6+ is up to an order of magnitude more soluble than S2– (Jugo et al., 2005a). In reducing conditions (S dominantly as S2–) sulfur solubility is also a function of the FeO content of the glass (sulfide saturation), such that increasing FeO results in greater S solubility (Fig. 2; Wallace & Carmichael, 1992).
|
Sulfur concentrations below the Fe-sulfide saturation curve often occur in breached inclusions that have the greatest potential for contamination during surface weathering and rehomogenization and therefore may not maintain the composition of the original trapped melt (Nielsen et al., 1998
). Low Cl concentrations coinciding with S concentrations below the Fe-sulfide saturation curve provide additional evidence for degassing.
Trace element concentrations in melt inclusions were determined by laser ablation ICP-MS (LA-ICP-MS) analysis using a NewWave DUV 193 nm ArF Excimer laser and VG PQ ExCell Quadrupole ICP-MS system at OSU. Detailed analytical conditions and standard reproducibility have been reported by Kent et al. (2004) and Rowe et al. (2006). Melt inclusions with a diameter of less than 50 µm were ablated with a 30 µm laser spot, whereas a 50 µm laser spot was used for inclusions with a diameter greater than 50 µm. Trace element abundances were calculated relative to the USGS glass standard BCR-2G, with 43Ca as the normalizing isotope. Accuracy and precision of melt inclusion analyses were monitored by repeated analysis of USGS glass BHVO-2G over the course of each analytical session. Results of those tests indicate accuracy generally within 10% of accepted values (e.g. Kent et al., 2004).
Sulfur speciation
Sulfur speciation in melt inclusions was determined based on the relative shift of the S K
X-ray wavelength as determined by electron microprobe analysis after Carroll & Rutherford (1988) and Wallace & Carmichael (1994). Sulfur speciation measurements were made on a Cameca SX-50 at University of Oregon (UO) and a Cameca SX-100 at OSU using a 5 µm beam diameter operating at the same conditions as previously described for glass analysis. Wavelength scans of the S K X-ray were simultaneously made on three PET crystals for measurements at UO and on two PET and one LPET crystal for measurements at OSU. Each spectrometer was moved incrementally by 4·37 x 10–4 Å for 100 steps with a 5 s counting time for mineral standards (<9 min total time) and 30 s counting time for glasses (50 min total time). A Gaussian curve was then fitted to the wavelength scans to determine the sulfur peak position of the unknown and standards. Pyrrhotite (Fe1–xS; UO) and troilite (FeS; OSU) were analyzed before and after each unknown glass. Analysis of troilite at UO relative to the pyrrhotite standard did not indicate a significant shift between the two reduced sulfur standards allowing for a direct comparison between measurements made at UO and OSU. Additionally, pyrite (FeS2) and anhydrite (CaSO4) mineral grains were analyzed at the beginning and end of each analytical session to monitor precision and calculate S6+/Stotal, respectively (Rowe et al., 2007).
To avoid oxidation of the sulfur during the glass analysis, which can occur when beam exposure time is greater than 10 min, the microprobe stage was moved 1 µm/min during continuous wavelength scan analysis (Wallace & Carmichael, 1994; Rowe et al., 2007). As a result of the requirement for incremental stage movement, melt inclusions with a diameter of less than 30 µm were not suitable for analysis with this technique. Because of the short time of analysis required for high-S reference minerals a stationary stage for these is acceptable (Rowe et al., 2007). To determine reproducibility of sulfur speciation measurements we ran six repeat analyses of Galapagos Spreading Center (95·5°W) glass K 14-3 (Christie et al., 1986). The standard deviation (1
) of the six repeat analyses was ±3·5% S6+/Stotal, significantly greater than the standard deviation (1) of the three concurrent spectrometer measurements on each inclusion (Table 2).
|
|
Determination of fO2
One sample of primitive basalt from each of the shoshonitic, CAB, LKT, and EIB basalt groups was selected from the arc, back-arc, and for the shoshonitic basalts the fore-arc, to determine oxygen fugacity (total of eight samples). Selection criteria include major element and trace element composition, and the presence of large melt inclusions (required for the sulfur K
wavelength shift method).
The primary method of determining basaltic oxidation state in this study is based on the measurement of S K X-ray wavelength in melt inclusions to determine sulfur speciation and oxidation. This method is utilized because it provides an in situ oxidation state directly comparable with the melt inclusion compositions. Olivine–chromite oxygen geobarometry was utilized as a secondary technique to better evaluate the accuracy of the calculated oxygen fugacity from the measured sulfur speciation. A detailed discussion summarizing each of the techniques utilized to determine oxygen fugacity and the results obtained from comparing the two methods is presented in the Appendix.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Melt inclusion compositions and variability
Fore-arc, arc and back-arc basaltic melt inclusions are subdivided based on geochemistry as either shoshonitic, EIB, CAB, or LKT (see subsequent discussion; Conrey et al., 1997
; Leeman et al., 2005). Whereas trace element concentrations in melt inclusions are in some cases highly variable relative to whole-rock compositions, overall trends and ratios (e.g. Ce/Pb and K/Nb) are similar to those observed in the whole-rock samples (Table 1). CAB and shoshonitic melt inclusions generally have less than 50% variability (relative to whole-rock compositions) in trace element concentrations whereas EIB and LKT inclusions have trace element concentrations typically within 20% of the whole-rock values. All melt inclusion major and trace element data are given in the Electronic Appendix at http://petrology.oxfordjournals.org/.
Enriched intraplate basalts
Back-arc EIB melt inclusion compositions are generally more mafic, with an average of 8·4 wt % MgO (10·9–5·4 wt % MgO) compared with an average of 6·4 wt % MgO (7·8–4·8 wt % MgO) for arc melt inclusions (Table 1). Despite this distinction, K2O concentrations between the arc and back-arc are similar with an average of 0·54 wt % K2O for arc basalts and 0·50 wt % K2O for back-arc basalts (Table 1). On average, arc EIB inclusions are slightly enriched in TiO2 with 1·96 wt % TiO2 relative to back-arc EIB inclusions with an average of 1·71 wt % TiO2 (Table 1). Trace element concentrations are significantly different between the arc and back-arc lavas. Arc EIB melt inclusions are enriched in light to middle rare earth elements (LREE–MREE) but have heavy rare earth element (HREE) abundances equivalent to back-arc EIB lavas. Despite slightly lower Nb and Ta concentrations in arc EIB inclusions, Zr concentrations are, on average, 50 ppm greater than in back-arc inclusions (Fig. 3a). In addition, Ba, Pb, and Sr concentrations are enriched, based on average inclusion compositions, 60%, 100% and >200% respectively, in arc relative to back-arc EIB inclusions (Fig. 3a).
|
Sulfur concentrations are in the range of 0·24–0·74 wt % in arc and 0·07–0·24 wt % in back-arc EIB inclusions; however, the average S concentration is significantly greater in arc melt inclusions (0·13 wt % S) relative to back-arc inclusions (0·11 wt % S; Table 1). Chlorine concentrations are significantly higher in arc EIB inclusions, ranging from 0·05 to 0·17 wt % Cl (0·13 wt % average with one outlying inclusion at 0·38 wt % Cl) relative to back-arc inclusions with a range of 0·01–0·04 wt % Cl (0·03 wt % average; Table 1; Electronic Appendix). Chlorine concentrations are reported here normalized to Ti and Nb (HFSE) to reduce potential variations in Cl abundances resulting from variations in the degree of partial melting (Figs 4 and 5). Despite greater Nb and TiO2 concentrations in arc inclusions, Cl/Ti and Cl/Nb ratios, similar to Cl concentrations, are also enriched in arc EIB inclusions relative to back-arc inclusions (Fig. 5).
|
|
Low-K tholeiites
Melt inclusions from LKT back-arc lavas have major and trace element compositions similar to inclusions from arc lavas (Table 1). Arc LKT melt inclusions are commonly more primitive with an average of 7·8 wt % MgO vs 5·9 wt % MgO for back-arc inclusions. K2O concentrations are on average greater in LKT back-arc inclusions relative to LKT arc inclusions, 0·54 wt % vs 0·45 wt %, respectively (Table 1). TiO2 concentrations are similarly greater in back-arc LKT melt inclusions, with an average concentration of 1·99 wt % TiO2, relative to 1·75 wt % TiO2 in arc LKT inclusions (Table 1). Although generally higher in back-arc inclusions, Nb concentrations (9·3–12·3 ppm) overlap with those of arc inclusions (7·2–10 ppm). Arc LKT inclusions, relative to back-arc LKT inclusions, are enriched in Ba (233–389 ppm vs 142–186 ppm) and Sr (341–435 ppm vs 290–330 ppm) (Fig. 3b; Electronic Appendix).
Sulfur concentrations in LKT back-arc inclusions, ranging from 0·13 to 0·21 wt % S (0·14 wt % average), typically are slightly greater than concentrations in arc melt inclusions, which range from 0·08 to 0·16 wt % S (0·11 wt % average), contrary to across-arc trends in EIB inclusions (Table 1). Chlorine concentrations, however, are essentially identical between arc and back-arc (0·03 wt % average); LKT inclusions vary from 0·001 to 0·05 and from 0·02 to 0·04 wt % Cl, respectively (Table 1). As with Cl concentrations, Cl/Ti ratios overlap between arc and back-arc LKT inclusions; however, Cl/Ti ratios in arc LKT inclusions are more variable, in the range of 0–0·08 (0·03 average for both arc and back-arc inclusions; Fig. 4). Cl/Nb is greater in arc LKT inclusions, with an average of 39·3 (compared with 29 for back-arc LKT inclusions) and, as with Cl/Ti ratios, exhibits a significantly greater range (Fig. 5).
Calc-alkaline basalts
Calc-alkaline basalts are the most abundant and compositionally variable basalt group in both the arc and back-arc of the Oregon Cascades. Arc CAB melt inclusions are generally more primitive with an average of 8·1 wt % MgO compared with an average of 6·8 wt % MgO for back-arc CAB inclusions. K2O concentrations in arc and back-arc inclusions are similar, with an average of 0·60 wt % (Table 1). Despite the greater trace element variability in calc-alkaline basalts relative to other basalt groups, the melt inclusion compositions from the arc are similar to those of the back-arc and show no systematic variations across the arc. The overall variation in CAB trace element concentrations is greater in back-arc lavas and in some cases these melt inclusions are more enriched in the fluid-mobile elements Ba, K, and Pb, than comparable arc lavas (i.e. Ba is enriched up to 300% in back-arc relative to arc melt inclusions; Fig. 3c). Strontium concentrations from arc CAB inclusions vary from 554 to 413 ppm, within the range of, but on average greater than concentrations in back-arc CAB inclusions (532–259 ppm; Table 1). HFSE, in particular Nb, generally overlap in back-arc and arc melt inclusions, with Nb concentrations varying in the range of 4·4–10·1 and 5·9–10·7 ppm, respectively.
Despite greater variation in S concentrations in back-arc CAB inclusions, with a maximum concentration of 0·29 wt % vs 0·13 wt % in arc inclusions, average S concentrations are essentially identical in both arc and back-arc inclusions (average concentrations of 0·10 wt % S). Similarly, Cl concentrations overlap greatly between arc (0·05 wt % average) and back-arc (0·04 wt % average) melt inclusions; however, arc Cl concentrations vary significantly from sample to sample with the range 0·02–0·10 wt % Cl (Table 1; Electronic Appendix). Similar to LKT inclusions, Cl/Ti and Cl/Nb ratios overlap between arc and back-arc CAB inclusions (Figs 4, 5).
Shoshonites
Shoshonitic lavas are rare in the Oregon Cascades, with only two lavas identified within the study region; one in the fore-arc and one in the arc. Despite a more mafic whole-rock composition and being trapped within the most forsteritic olivines in this study (Fo86–90), fore-arc shoshonitic melt inclusions are generally less mafic (6·8 wt % MgO) than shoshonitic arc inclusions (7·2 wt % MgO; Table 1). Inclusions from the fore-arc shoshonitic lava have, on average, higher K2O (2·17 wt % vs 1·95 wt %), lower Na2O (2·49 vs 3·08 wt %) and significantly greater P2O5 (1·33 wt % vs 0·48 wt %). TiO2 concentrations are also greater in the fore-arc (1·68 wt % average) relative to arc (1·30 wt % average) inclusions (Table 1). Trace elements are almost universally enriched in fore-arc shoshonitic melt inclusions relative to arc inclusions, including a 2–3 times enrichment in LREE and MREE concentrations, with the notable exception of a stronger depletion in HREE, Y, and Yb in the fore-arc inclusions (Fig. 3d). Barium concentrations vary considerably in fore-arc inclusions, ranging from 2427 to 8330 ppm, vs 906-1917 ppm in arc inclusions (Fig. 4a; Table 1; Electronic Appendix). Strontium concentrations are equally enriched but generally less variable in the fore-arc inclusions, ranging from 3519 to 4503 ppm, whereas arc shoshonitic inclusions range from 1433 to 3488 ppm Sr (Table 1; Electronic Appendix). Nb concentrations also vary considerably between fore-arc and arc inclusions with an average of 9·0 ppm in fore-arc inclusions versus an average of 3·8 ppm in arc inclusions (Table 1).
Sulfur concentrations in the naturally quenched fore-arc shoshonitic lava range from 0·09 to 0·58 wt % whereas concentrations in the arc shoshonite have a lower overall range, varying from 0·01 to 0·49 wt % S (Table 1). Chlorine concentrations are significantly enriched in the fore-arc shoshonitic inclusions, in the range of 0·07–0·19 wt % compared with 0·01–0·12 wt % in the arc inclusions (Table 1; Electronic Appendix). Chlorine/Ti ratios, however, are comparable between the fore-arc and arc melt inclusions, in the range of 0·07–0·19 and 0·01–0·21, respectively (Fig. 4).
Basalt oxygen fugacity
Based on sulfur X-ray wavelength measurements, calculated oxygen fugacities for basalts from across the Cascade arc range from <–0·25 to +1·9 log units (FMQ; Fig. 4c; Table 2). EIB basalts, LKTs, and CABs have similar oxygen fugacities varying from <–0·25 to +0·65 log units (FMQ). However, oxidation states calculated for EIB arc melt inclusions are, on average 0·5 log units greater than EIB back-arc inclusions (Table 2). In addition, EIB arc inclusions exhibit significant variation in fO2, from –0·06 to 0·65 log units (FMQ). Low-K tholeiitic and CAB inclusions from arc lavas have oxidation states indistinguishable from the back-arc LKT and CAB melt inclusions, with average fO2 values ranging from less than –0·25 log units (FMQ) to FMQ.
In addition to the greatest across-arc variability (0·7 log units), shoshonitic inclusions are significantly more oxidized, with fO2 ranging from +1·1 to + 1·9 log units (FMQ), with the highest calculated oxygen fugacity from the fore-arc shoshonitic lava (Fig. 4c). The high oxidation states of shoshonitic basalts compared with LKT, CAB and EIB, in conjunction with the decreasing oxidation states from fore-arc to arc (shoshonites) and arc to back-arc (EIB) in the Oregon Cascades, probably indicate multiple variables influencing the oxidation state of the basalts, potentially including subduction modification, source fertility, and degree and style of melting.
Chromite compositions and chromite–olivine oxygen barometry results are summarized in Tables 3 and 4 and Figs 4c and 6. Average oxidation states determined from chromite–olivine pairs have a smaller overall range (+0·35 to +1·5 log units FMQ) than determined from sulfur speciation calculations on the same samples (Tables 3 and 4; Fig. 4c). Across the arc variation in fO2 in LKT, CAB and EIB is not systematic, with an average increase in LKT and CAB fO2 (although statistically indistinguishable) and a 0·5 log unit decrease in EIB fO2 from arc to back-arc. From chromite–olivine pairs, as with fO2 from sulfur speciation, shoshonitic basalts are up to 0·5 log units more oxidized than other basalt groups, with fO2 decreasing from fore-arc to arc (Fig. 4c).
|
|
Correlating fO2 between chromite–olivine pairs and sulfur speciation is problematic at low fO2 as a result of our inability to accurately determine oxidation state from sulfur speciation below
4% sulfate (–0·25 log units below FMQ), despite measured sulfur speciation in melt of essentially 0% sulfate. Chromite–olivine oxygen barometry is also problematic at low fO2 as a result of errors in determining the ferric iron content in spinel from microprobe analysis and greater imprecision at low fO2 (Ballhaus et al., 1991; Parkinson & Arculus, 1999). The standard deviation (1) of multiple chromite–olivine fO2 estimates is, however, significantly lower than the calculated precision, ranging from 0·1 to 0·46 log units (Table 4).
|
At high oxygen fugacities (>FMQ) calculated oxidation states from sulfur speciation and olivine–chromite oxygen geobarometry correlate reasonably well; however at low fO2 (<FMQ) oxygen fugacities calculated from chromite–olivine pairs are typically greater than predicted by sulfur speciation by as much as 1·25 log units (Fig. 6a). Although there is no evidence to suggest re-equilibration of chromite and olivine pairs during melt inclusion rehomogenization, re-equilibration between the olivine and spinel inclusions may have occurred naturally as a result of slower cooling (Scowen et al., 1991
; Barnes, 1998; Kamenetsky et al., 2001). Re-equilibration is most evident in a back-arc CAB sample (TB02-1) where olivine–spinel pairs record a range in temperatures (Fe–Mg exchange), calculated after Ballhaus et al. (1991), from temperatures slightly above estimated liquidus temperatures (1250°C) down to 840°C (Fig. 6b). Concurrent with the decrease in temperature is an increase in fO2 (FMQ) of over 0·5 log units, suggesting that the high calculated oxidation states in this sample may be the result of re-equilibration. Although re-equilibration is most easily observed in TB02-1, it is also likely to have occurred in the LKT (NEF03-1 and FLR03-1) lavas, which were collected further from estimated vent locations from relatively thick flows. As a result of the inherent problems associated with determination of low oxygen fugacities and because of the potential for olivine–spinel re-equilibration, we feel that estimates based on sulfur speciation measurements from inclusions without significant post-entrapment modification (e.g. H loss) provide a better approximation of the primitive basaltic fO2. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Classification of lavas and inclusions
As stated above, despite mineralogical and textural distinctions, division of basaltic lavas into four groups (CAB, LKT, EIB, and shoshonitic) is justified largely on whole-rock trace element abundances. Calc-alkaline basalts, the dominant basalt type in the central Oregon Cascades, are variably enriched in LILE and fluid-mobile elements, with K2O concentrations typically ranging from 0·5 to 1·0 wt % (Fig. 7a). LREE concentrations do not reflect the enrichment in more fluid-mobile elements and consequently fluid-mobile/LREE ratios are generally high (Fig. 7b). Additionally, CAB lavas are characteristically depleted in HFSE, which may reflect either a higher degree partial melt or alternatively a more depleted mantle source (e.g. Leeman et al., 2005
). Low-K tholeiites, relative to CAB compositions, generally have lower LILE and higher HFSE concentrations, for the most part consistent with the OIB–MORB array (after Leeman et al., 1990, 2005). Compositionally, LKT lavas may overlap with CAB lavas; however, the LKT magmas are texturally distinctive. EIB lavas are characterized by an enrichment in HFSE, especially evident for Nb and Ta, and LREE (La and Ce), and are generally thought to be derived from a more fertile mantle source (e.g. Conrey et al., 1997; Smith & Leeman, 2005). Shoshonitic lavas are extremely enriched in LILE and LREE, with K2O concentrations >1·5 wt %. Shoshonitic lavas, similar to CAB lavas, are depleted in HFSE; however, they are also depleted in HREE (Yb). Additionally, shoshonitic lavas have higher 87Sr/86Sr ratios relative to other basalt types in central Oregon (Schmidt et al., 2008). To make direct across-arc comparisons, in the following discussion we make the assumption that the mantle source for each basalt group (CAB, LKT, EIB, and shoshonite), barring modification from the addition of a subduction component, is essentially the same across the arc. This assumption is supported by similarities between more fluid-immobile trace element (HFSE, HREE) concentrations and ratios between melt inclusions within basalt groups from different regions across the volcanic arc. Despite greater variability, melt inclusion compositions reflect the same characteristic trace element abundances and ratios for the basalt groups as the whole-rock geochemistry. This observation is significant in that it suggests that melt inclusions are petrogenetically related to their host lavas, and that there is relatively limited interaction between basalt groups during melting or melt aggregation.
|
Correlation of fO2 and trace elements
Identification and interpretation of the effects of addition of a slab-derived melt or fluid (i.e. subduction component) on the generation of subduction zone magmas has been the focus of numerous studies (e.g. Stolper & Newman, 1994
; Churikova et al., 2001; Hochstaedter et al., 2001; Kent & Elliott, 2002; Grove et al., 2003; Eiler et al., 2005). Comparing trace and volatile element concentrations with sulfur speciation and oxygen fugacity for over 100 km across the Oregon Cascade volcanic arc allows us to directly examine the effects of a subduction component on the oxidation states of basaltic magmas as a function of distance to the trench and basalt geochemistry and petrogenesis (Figs 4 and 8).
|
For the purpose of this analysis, melt inclusion major and trace element compositions obtained in this study have been normalized to be in equilibrium with Fo90 olivine to better approximate mantle conditions, following the approach of Leeman et al. (2005
). Incompatible trace element concentrations are not significantly affected by this normalization, given the initial high Fo content of the olivine host crystals. Fluid-mobile trace elements (LILE and LREE in equilibrium with Fo90 olivine) and oxygen fugacity (from sulfur speciation) are well correlated in melt inclusions, with LILE and LREE concentrations increasing at higher fO2 (Fig. 8). As discussed above, concentrations of fluid-mobile trace elements such as Sr and Ba generally decrease from fore-arc to back-arc in the Oregon Cascades within each magma group (Fig. 4). Similarly, Ba/Nb (Fig. 5b) and Ba/Zr ratios decrease across the arc for LKT and EIB, although no apparent across-arc correlation is evident for CAB or shoshonitic melt inclusions. If the increase in fluid-mobile trace elements is the result of addition of a subduction component to the mantle source then the correlation between trace element concentrations and oxygen fugacity would suggest that the subduction component is the oxidizing agent. The decrease in Cl/Ti from the fore-arc to back-arc (Fig. 4b) and the broad positive correlation between Cl/Nb and fO2 (Fig. 8d) is also consistent with an oxidizing subduction component. However, the magnitude of the enrichment in both Ba/Nb and Cl/Nb is low relative to volatile and trace element data from other arcs (Fig. 5; e.g. Cervantes & Wallace, 2003; Wade et al., 2006; Sun et al., 2007).
Melt inclusions from calc-alkaline basalts do not exhibit the across-arc variation in fO2, fluid-mobile element concentrations and LILE/HFSE ratios apparent in melt inclusions in EIB, shoshonitic, and LKT basalts (Figs 4 and 5b). If the input from the currently subducting Juan de Fuca slab decreases from fore-arc to back-arc as suggested by other trace element and volatile concentrations and ratios from other basalt groups, this would imply that back-arc CAB melts are essentially ‘over enriched’ relative to comparable arc compositions. One interpretation of the greater enrichment and lack of across-arc variation is that the enriched trace element signature of back-arc CAB lavas may not be related to current subduction modification but may in fact result from melting of a previously enriched mantle source region. This has previously been suggested for the southern Washington and northern California segments of the Cascades (Borg et al., 2002; Leeman et al., 2004, 2005).
Whereas Cl/Nb concentrations broadly correlate with sulfur oxidation for arc and back-arc lavas, fore-arc shoshonitic melt inclusions appear to be significantly more oxidized despite a similar Cl enrichment (Fig. 8). This observation is in contrast to the good positive correlation between fluid-mobile trace elements (Ce, Ba, and Sr) and fO2 for fore-arc, arc, and back-arc basalts (Fig. 8a–c). This may suggest that Cl is behaving somewhat anomalously to other fluid-mobile elements or that the shoshonitic basalts, in particular the fore-arc basalt, are derived from a source that is generally more oxidized, regardless of subsequent Cl enrichment.
Mantle melting
Decompression melting and flux melting (melting induced by the addition of a hydrous component to the mantle) are considered here as two endmember styles of mantle melting in subduction zones. The relative significance of either melting scenario is strongly dependent on the sub-arc thermal structure and convection style, rate of subduction, and temperature of the subducting slab (Davies & Stevenson, 1992; Schmidt & Poli, 1998; Walker et al., 2003). In the Cascadia subduction zone close proximity of the Juan de Fuca spreading ridge to the convergent margin results in the subduction of a young (8 Ma), warm, thin oceanic plate. This in turn results in the dehydration of the subducting slab at relatively shallow depths (Leeman et al., 2004), as supported by the inferred serpentinization of the fore-arc mantle wedge based on S-wave tomography (Bostock et al., 2002).
One potential consequence of shallow dehydration is the implied early loss in available slab fluid that may be available to produce flux melting further from the trench (Harry & Green, 1999; Leeman et al., 2005). However, both the whole-rock and inclusion data suggest enrichment in fluid-mobile trace elements (Ba, Sr) across the Central Oregon segment of the arc, similar to enrichments observed by Walker et al. (2003) across the Central American arc. The degree of enrichment strongly implies a role for metasomatism in any melting scenario (decompression or flux melting).
For the purpose of this investigation, mantle melting is modeled to provide only general observations on the mantle source fertility, the degree of melting (F) and the amount of subduction component (SC) added to the mantle wedge. Two distinct mantle sources and subduction components, used in prior studies of the Cascadia subduction zone (Borg et al., 1997; Reiners et al., 2000), are utilized to demonstrate the dependence of modeling results on the initial mantle composition and the composition of the SC (Table 5). In an attempt to limit the number of variables, we have chosen to use a compositionally homogenous SC (‘SC’ of Reiners et al., 2000; and ‘Fluid’ of Borg et al., 1997); however, in detail this may not be realistic, given models for progressive dehydration of subducting slabs (Schmidt & Poli, 1998).
|
The mantle source of Reiners et al. (2000
) is a lherzolite with Ol:Opx:Cpx:Grt:Spl = 55:30:10:2·5:2·5. The modal mineralogy (Ol:Opx:Grt:Amph = 64:26:4:6) of the depleted mantle source is similar to depleted harzburgite xenoliths recovered from Simcoe Volcano, Washington (Brandon & Draper, 1996). In addition, 4% garnet and 6% amphibole have been added to the source composition, whereas clinopyroxene is absent. The presence of residual garnet, implying melting at greater pressure, is indicated by a strong depletion in HREE and Yb in the resultant melts. Addition of a hydrous phase (amphibole or phlogopite) to the depleted mantle is necessary to match the observed K, Rb, and Ba concentrations of the shoshonitic lavas (Fig. 5a). Such a source mineralogy may form during metasomatism of the mantle wedge by addition of a subduction component (e.g. Davies & Stevenson, 1992).
We have used an isenthalpic (heat balanced) flux melt model (Reiners et al., 2000) to constrain F and amount of SC added to the mantle. The coupled relationship between the amount of SC and F in a flux melt model provides reasonable approximations of both parameters while reducing the number of unconstrained variables present in a decompression melt model (e.g. Stolper & Newman, 1994; Reiners et al., 2000). As one would expect intuitively based on the known fluxing effects of fluids, the amount of melting increases with increasing SC. However, isenthalpic melting also results in a reduction in mantle temperature as the SC is added and generates melt. In detail, as melting progresses the efficiency with which each added increment of SC produces melt decreases (see fig. 6 of Reiners et al. 2000), such that a greater amount of subduction component must be added to generate the same increment of melt.
Compositionally, if one assumes the source compositions noted above, initial addition of a subduction component results in magmas with trace element concentrations similar to EIB magmas (Reiners et al., 2000). As F increases, melt trace element compositions resemble LKT magmas. Because relatively smaller increases in melt fraction will be produced at higher proportion of added SC in an isenthalpic model, eventually the SC will exert increasing dominance on the trace element composition of the magma, resulting in melts similar to CAB magmas.
Several general points become evident when comparing results from flux melt modeling using the various mantle source and subduction component compositions (Fig. 9). First and foremost, the flux melt model is extremely sensitive to changes in the composition of the subduction component and initial mantle source. As a result, we have chosen not to try to force a ‘best-fit’ subduction component to our data but rather have retained the compositions used in prior models (Borg et al., 1997; Reiners et al., 2000). This simplification results in a range of T, %SC, and %F dependent on the elements modeled; however, it does not affect the overall trends (Figs 9 and 10). Second, EIB compositions, both from the arc and back-arc, cannot be generated from a depleted mantle source (Borg et al., 1997). Despite the dependence of model results on starting compositions, this is a relatively robust conclusion and we suggest that the presence of EIB melts therefore requires the presence of a more fertile mantle source. This does not rule out a depleted mantle source for the generation of the LKT and CAB magmas. However, average compositions of chrome-spinel inclusions within olivine have Cr-numbers ranging from 45 to 27 in LKT, CAB, and EIB lavas (Table 3) and show no variation with indices of source fertility such as Yb (Clynne & Borg, 1997), consistent with the interpretation that a variably depleted mantle source is not necessary for the generation of the different basalt groups (Fig. 1). Thus we suggest that the mantle source required for EIB lavas could also be the source of CAB and LKT magmas. This is in contrast to the studies of Clynne & Borg (1997) and Smith & Leeman (2005), who interpreted LKT and EIB lavas to be derived from a more fertile mantle source than CAB and shoshonitic lavas in Cascade Range segments in northern California and southern Washington.
|
|
If a simple batch melting model (decoupled SC and F) is applied using the Reiners et al. (2000
) peridotite and SC, the overall %F that best fits observed compositions decreases only slightly in CAB, LKT, and EIB; however, the amount of SC required to generate the CAB, LKT and EIB magmas decreases substantially, based on the observed variation of Ba vs Nb (Fig. 10b). Thus, the general trend from EIB (low F, low SC) to CAB (high F, high SC) is preserved no matter which melt model of the enriched mantle source is applied. Using a depleted mantle source significantly reduces the %SC and %F of the LKT and CAB magmas but, as with the flux melting of a depleted mantle, EIB compositions cannot be reproduced (Figs 9 and 10b).
Another key observation from modeling is that, assuming the mantle source for a given basalt type remains constant across the arc, and regardless of the melting model used, EIB and LKT magmas exhibit an increase in %SC and %F from back-arc to arc, whereas CAB magmas show no clear across-arc variation. This result is consistent with earlier observations that CAB magmas demonstrated no systematic across-arc variation in either trace and volatile element concentrations or oxidation state (Fig. 4). Leeman et al. (2005) suggested that CAB magmas are derived from a shallower mantle source enriched by a ‘stored’ subduction component. The model of a ‘stored’ subduction component is consistent with our results, potentially explaining the lack of any across-arc variation in CAB fluid-mobile trace element concentrations. However, the mechanism by which the geochemical signature of a subduction component may be stored in the lithospheric mantle is not well understood, considering that the addition of a volatile-rich component to the hot Cascadia subduction zone mantle may be expected to induce melting. The systematic increase in fluid-mobile trace elements observed in EIB and LKT magmas derived from a deeper mantle source [in accordance with the Leeman et al. (2005) model] may represent the ‘continuing’ addition from the subducting slab. Alternatively, rather than changing the amount of subduction component added to the mantle source, across-arc variations could result from changing the composition of the subduction component as the depth to the subducting slab increases. Although it is beyond our model at the present to distinguish between these two scenarios, in terms of the ‘end result’ of less fluid-mobile elements derived from the subducting plate further from the trench, our general results remain unchanged.
Shoshonitic basalt compositions cannot be reproduced from either mantle source or subduction component using the flux melt or batch melt models; however, a low-degree batch melt (<10% F; arc shoshonite requires higher F than fore-arc) of a peridotite with 50% lower Nb than the model depleted mantle source, with 2–5% SC added (with the higher proportion added in the fore-arc relative to the arc), can approximate the shoshonitic compositions. Qualitatively, the higher La/Sm of fore-arc shoshonite inclusions (4·75 average) relative to arc shoshonite inclusions (3·75 average) is consistent with a lower degree of partial melting, assuming the same initial mantle composition (Table 1). High spinel Cr-numbers (71 ± 2) and a low (1·3 ppm) average Yb concentration in the Cascade fore-arc shoshonitic basalt are also consistent with shoshonitic melt derivation from a highly depleted source (Fig. 11; Tables 1 and 3; Clynne & Borg, 1997). Cr-rich spinel compositions are comparable with those observed in high-Mg andesites from the Mt. Shasta and Lassen region in northern California, suggesting that the fore-arc shoshonitic basalt may have been derived from a mantle source that had previously experienced 20–30% melting (Fig. 11; Baker et al. 1994; Clynne & Borg, 1997).
|
Assimilation of granitic crustal rocks could also produce the enriched trace element signatures characteristic of the shoshonitic lavas; indeed, partially melted granitic xenoliths have been recovered from the fore-arc shoshonitic lava (Conrey et al., 1997
). However, given the high forsterite content of the olivines (up to Fo90), high transition metal contents, and absence of systematic variation in potential indices of crustal contamination (e.g. K or Ba concentrations) with decreasing forsterite content or Mg-number, we feel that the shoshonitic compositions reflect enrichment of the source rather than crustal contamination.
Melt extraction and mantle oxidation state
As discussed in detail by Parkinson & Arculus (1999), the process of melt extraction from the mantle exerts a significant control on basalt oxidation state. For example, decompression melting may result in an increase of 0·6–0·8 log fO2 units for every 10 kbar of decompression, provided the melts stay in equilibrium with their mantle source (i.e. batch melting; Ballhaus & Frost, 1994). In contrast, an isobaric melt, closed to oxygen, will not undergo a significant change in fO2, provided there is a relatively rapid isochemical ascent (<0·5 log unit decrease for 30 kbar of decompression; Kress & Carmichael, 1991). In either scenario the basaltic melts appear to record the oxidation state of the mantle that they were last in equilibrium with, but not necessarily their parental mantle (Parkinson & Arculus, 1999).
Considering the above scenarios for melt extraction, the potential range of mantle source oxidation may be estimated if we make the following simplifying assumptions:
- melting is initiated at 90 km depth, consistent with melt segregation estimates from Leeman et al. (2005);
- oxidation occurs prior to melting (Parkinson & Arculus, 1999) rather than post-melting (Brandon & Draper, 1996);
- The LKT, CAB, and EIB originate from a common fertile mantle source whereas the shoshonites are derived from a different mantle source.
An average depth of 45 km is assumed for the base of the crust beneath the arc and back-arc (Catchings & Mooney, 1988; Stanley et al., 1990), above which the melts behave isochemically. Based on this simplified scenario, the potential range of mantle oxidation states may be estimated. The mantle source for shoshonitic basalts should be within the range of +0·3 to +2·4 log units (FMQ), whereas the mantle source for the CAB, LKT, and EIB ranges from –1·1 to +0·7 log units (FMQ), with the small overlap between the two sources decreasing as the initial depth of melting decreases.
The predicted source region fO2 for CAB, LKT, and EIB largely overlaps that of sub-oceanic mantle and falls within the range of ‘lightly metasomatized’ spinel peridotite xenoliths defined by Ballhaus (1993) [–3 to +1 log units (FMQ)]. The source region for shoshonitic basalts must be substantially more metasomatized, characteristic of the ‘strongly metasomatized’ mantle of Ballhaus (1993), consistent with trace and volatile element concentrations and melt modeling, which suggests that the shoshonitic basalts are derived from a depleted mantle source strongly modified by addition of a subduction component.
The predicted range of mantle fO2 for CAB, LKT, and EIB magmas is consistent with the range predicted based on V/Sc systematics for the Cascade Range (–1·25 to + 0·5 log units FMQ) and MORB source mantle (–1 to + 0·2 log units FMQ; Lee et al., 2005). This suggests that for the bulk of the Cascade magmas the mantle source is not significantly oxidized relative to the global range of basalts despite addition of a subduction component to the mantle wedge. From flux and batch melt modeling, a range in %SC from 1 to 4 wt % can be estimated for EIB arc melt inclusions relative to back-arc inclusions, making quantification of the impact of a subduction component on the mantle wedge difficult. However, the correlation between increasing fluid-mobile trace elements and fO2 for the EIB does suggest that oxidation of the mantle source from back-arc to arc (for EIB inclusions) results from increasing addition of a subduction component to the mantle wedge, consistent with prior models (Ballhaus, 1993; Parkinson & Arculus, 1999; Lee et al., 2005). Despite a similar overall estimated addition of a subduction component (2–5%) for shoshonitic inclusions, the depleted source region for these magmas is significantly more oxidized following addition of a subduction component. This may suggest that the oxidizing potential of the subduction component is largely dependent on the mineralogy and composition of the mantle peridotite prior to metasomatism, with a depleted mantle source essentially more sensitive to addition of an oxidizing component.
|
CONCLUSIONS AND BROADER IMPLICATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Olivine-hosted melt inclusions from primitive basalts across the Oregon Cascades provide important information on how oxidation states and major, trace, and volatile element variability are correlated with one another in subduction zone magmatism. These correlations provide constraints on our conceptual and numerical models of how partial melting and input of subduction-derived fluids across a subduction zone may influence magma oxidation states. Several key observations and model results are summarized below.
Of the four major basalt groups, only the relatively low-degree partial melt magmas (EIB and shoshonitic) record significant across-arc variations in trace and volatile element abundances. However, Cl enrichment in all basalt types is relatively low compared with basalts from other volcanic arcs (e.g. Mexican Volcanic Belt, Central American arc). The absence of a significant across-arc variation in the CAB melt inclusions, in conjunction with the elevated trace element abundances of back-arc CAB melt compositions, may indicate that the CAB trace element signatures represent an older mantle enrichment rather than, or in addition to, recording current slab dehydration, consistent with conclusions of Borg et al. (1997, 2002) and Leeman et al. (2005). Interpretation of the LKT inclusions is less straightforward, as LILE enrichment is evident in arc LKT inclusions relative to back-arc inclusions despite comparable fO2 and only slightly greater volatile abundances (Cl/Nb) in arc LKT inclusions.
The correlations of fluid-mobile trace element abundances and Cl enrichment with oxidation state for low-degree partial melts suggests that magma compositions are inherently linked to oxygen fugacity. The most straightforward interpretation of this overall correlation is that increased fluid-mobile trace elements, and fO2, result from addition of a subduction component to the sub-arc mantle wedge, although this chemical signal may vary with the degree of melting. This scenario is most likely for the shoshonitic basalts where fore-arc shoshonitic melt inclusions probably record lower degree melting than arc shoshonite inclusions. For these two shoshonitic lavas, despite the correlation between trace elements and fO2, no positive correlation is evident between Cl/Nb and Ba/Nb and fO2. This suggests that differences in trace element abundances between these two lavas may be a function of degree of melting rather than source variation. Despite this suggested correlation, melting alone cannot explain trends of observed fO2 in low-degree melts for both the EIB and shoshonitic magmas.
If the observed trace element and fO2 correlations are the result of addition of a subduction component to the mantle wedge, this implies that the subduction component was oxidized and that a greater proportion was added closer to the trench, resulting in more highly oxidized, trace element enriched lavas closer to the subduction zone. This model is consistent with the observations of Luhr & Aranda-Gomez (1997), who found that the oxidation state of mantle xenoliths across the Mexican Volcanic Belt increased closer to the trench. Although our results indicate that the subduction component was the source of the oxidation, the question as to what about the subduction component causes oxidation remains unsolved (Frost & McCammon, 2008). This study does not address whether the oxidation of the mantle results from addition of water, or addition of ferric iron or sulfate.
Melt modeling, in conjunction with chromite composition, indicates that the shoshonitic magmas require a strongly depleted mantle source, distinct from that of the other basalt types. The significant variation between the fO2 of the shoshonitic magmas and that of the other basalt types may imply that initial mantle composition is an important variable in controlling the fO2 of magmas. This raises the possibility that a strongly depleted mantle source is more susceptible to oxidation during metasomatism; potentially a result of the lower buffering capacity of depleted mantle, as indicated by Amundsen & Neumann (1992) and Parkinson & Arculus (1999). For example, a depletion in Fe by melting of spinel and clinopyroxene in the mantle source may result in a peridotite more susceptible to oxidation by addition of Fe3+ from an oxidizing melt.
Estimation of the potential range of mantle fO2 from the basalt compositions illustrates that the mantle source for most of the Cascade basalts (CAB, EIB, and LKT) ranges from –1·1 to + 0·7 log units (FMQ), consistent with sub-oceanic mantle, and hence not significantly oxidized. In contrast, shoshonitic basalts have an estimated source fO2 ranging from + 0·3 to + 2·4 log units (FMQ), which is significantly more oxidized than the sub-oceanic mantle.
Magma types proposed to be generated by relatively low-degree melting (EIB and shoshonites) exhibit across-arc geochemical variation whereas high-degree melts (CAB and LKT) do not. This lack of correlation in magmas thought to be generated by higher degree melting events is itself significant, although at present we do not have sufficient information to fully constrain the problem. Potential explanations are that (1) that metasomatism of the mantle wedge is relatively localized, and that the fO2 effects are preserved only during low-degree melting events, or (2) the fO2 of the basalt records the oxidation state of the mantle source it was last in equilibrium with, not necessarily its original source. Thus, if during ascent the CAB and LKT magmas were last in equilibrium with a more reduced mantle source, they may record a more reduced oxygen fugacity. This would also imply that the low-degree melts ascend through the mantle and crust with relatively little interaction and thus may more accurately record the initial fO2 of their respective mantle sources. If true, these conclusions suggest that although CAB are the dominant basalt type and typical ‘subduction-modified basalt’, they may not be the best choice for looking at the effects of addition of subduction-derived fluids across longer-lived volcanic arcs, whereas the low-degree melts may best preserve their original source fO2 and volatile budgets.
|
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Determination of fO2
Sulfur K
X-ray wavelength shiftBased on sulfur speciation (S6+/Stotal) measurements, oxidation state, reported relative to the fayalite–magnetite–quartz (FMQ) oxygen buffer, can be calculated empirically based on the relationship determined by Wallace & Carmichael (1994
; Fig. A1a). A similar empirical curve defined by Jugo et al. (2005b) is buffered at higher oxidation states by the presence of sulfite (S4+), resulting in a maximum X[S6+] equivalent of 0·86 (the term X[S6+] equivalent was used by Winther et al. (1998) and Jugo et al. (2005b) to indicate a relative change in speciation rather than a strict determination of S6+/Stotal as a result of the presence of sulfite). Although sulfite has been identified in highly oxidized melt inclusions the maximum reported proportion is only 16% with a relative error of 20%, and does not appear to vary systematically with either per cent sulfate or total sulfur concentration (Metrich et al., 2002). Because the significance and variability of sulfite in basaltic melts is still not well constrained and because at fO2 < +1·5 log units (FMQ) the Wallace & Carmichael (1994) and Jugo et al. (2005b) curves correspond reasonably well (within less than 0·5 log units), we have maintained the use of the Wallace & Carmichael (1994) curve to determine oxidation state from sulfur speciation (Fig. A1). The asymptotic relationship between sulfur speciation and fO2 at both very low and high per cent sulfate, coupled with an estimated reproducibility of ±3·5% sulfate, defines the lower and upper limits for determination of fO2 from sulfur speciation at –0·25 log units (FMQ) and +2·57 log units (FMQ), respectively (Fig. A1a). Therefore, when the proportion of sulfate is below 4% or above 96% an independent means of estimating fO2 is required.
|
Degassing of the melt during entrapment can significantly alter the sulfur speciation and oxidation state of the melt inclusion (Mathez, 1984
; Danyushevsky et al., 2002; Metrich et al., 2005; Rowe et al., 2007). Because SO2 is the dominant sulfur gas species during degassing of basaltic magmas, the sulfur speciation can be modified by either the reduction of S6+ or the oxidation of S2–, depending on the initial oxidation state and water content of the melt (Anderson & Wright, 1972; Carroll & Webster, 1994; Wallace & Carmichael, 1994).
Post-entrapment modification of the trapped melt by either hydrogen diffusion or Fe loss will result in the oxidation of the melt inclusion (Danyushevsky et al., 2002; Rowe et al., 2007). Fe loss during re-equilibration of a melt inclusion with its olivine host will result in increased Fe3+/Fe2+ of the melt. The resulting change in calculated oxidation state will vary depending on the initial Fe concentration and the Fe3+/Fe2+ ratio.
Diffusion of hydrogen out of melt inclusions is often observed in inclusions that have been slowly cooled or have been kept at high temperatures for an extended period of time (Hauri, 2002). As a result of hydrogen diffusion, excess oxygen may be retained within the melt inclusion, resulting in oxidation (Rowe et al., 2007). The loss of ferrous iron, either through oxidation from excess oxygen following hydrogen diffusion or from olivine–melt re-equilibration, may also result in the precipitation of sulfide globules, further increasing the S6+/Stotal, as the concentration of dissolved sulfur as sulfide is strongly dependent on the ferrous iron content of the melt (Metrich et al., 1999; Danyushevsky et al., 2002).
Melt inclusions from both arc and back-arc CAB lavas included in this study for which sulfur speciation was determined appear to have undergone hydrogen diffusion and/or Fe loss, resulting in a wide range of calculated oxidation states. If the melt is oxidized, through either re-equilibration or hydrogen diffusion, the calculated olivine–melt KDFe–Mg will become larger with increasing oxidation because of the overestimation of Fe3+ as a result of the higher fO2 (Rowe et al., 2007). We report calculated oxidation state data for these CAB lavas only for inclusions where olivine–melt equilibrium (KDFe–Mg of 0·3) is approached, based on measured S6+/Stotal. Multiple sulfur speciation measurements from a single sample are therefore required to distinguish inclusions that have been affected by degassing and post-entrapment modification.
Chromite–olivine oxygen geobarometry
Compositions of chromite inclusions within olivine phenocrysts provide a secondary means of estimating magmatic oxidation states based on the redox equilibrium reaction
|
| (1) |
Oxygen fugacity is calculated following the method of Ballhaus et al. (1990). Although this method requires the compositions of olivine, orthopyroxene and spinel equilibrium phases, the orthopyroxene component may be compensated through the reaction orthopyroxene = olivine + SiO2. In orthopyroxene-undersaturated magmas, therefore, a correction for the activity of silica (SiO2) must be applied, estimated from Ghiorso & Carmichael (1987), resulting in a reduction of the calculated fO2 by 0·23–0·3 log units (Ballhaus et al., 1991). The advantage of utilizing this technique is that oxidation state may be determined independently of the S peak shift technique and potentially within the same olivine grain.
Because chromite inclusions were not always present in olivine grains from which sulfur speciation in inclusions was measured, multiple chromite inclusions within olivines were analyzed from the same grain mounts to provide the most direct comparison. Further, because of the potential for re-equilibration between the host olivine and chromite inclusions (Ozawa, 1984; Clynne & Borg, 1997; Kamenetsky et al., 2001) at the high furnace temperatures required for melt rehomogenization, oxidation states were calculated from heated and unheated olivine–chromite pairs from three samples spanning the range of oxidation states measured in this study. Calculated oxygen fugacities from heated and unheated melt inclusions correlate within error (1), suggesting that little re-equilibration is occurring during the short rehomogenization time (Fig. A2). However, this does not indicate whether sub-solidus re-equilibration has occurred after entrapment of chromite grains within the olivine host prior to eruption or during cooling. Precision of this method was reported by Ballhaus et al. (1991) at ±0·41 log units at oxygen fugacities above FMQ and ±1·2–1·5 log units 2 log units below FMQ.
|
|
ACKNOWLEDGEMENTS |
|---|
The authors would like to thank J. Wickum, N. Dotson, C. Darr, S. Majors, E. Kohut, M. Schmidt, A. Jefferson (OSU geology students), J. Donnelly-Nolan (USGS—Menlo Park), and Carl Thornber (USGS—CVO) for their assistance in the field. R. Conrey (WSU) and D. Sherrod (USGS—CVO) provided invaluable compilations of published and unpublished whole-rock geochemical analyses for the Cascades and Newberry, respectively. P. Wallace, J. Donovan, P. Jugo, and C. Ballhaus provided early discussions of olivine–spinel oxygen barometry and sulfur speciation. D. Johnson and C. Knaack provided access to the Washington State University geoanalytical lab, including XRF and ICP-MS analysis of whole-rock samples. Critical reviews by P. Jugo, C. Mandeville, W. Leeman, and W. Bohrson significantly improved the manuscript. This project was funded by two Geological Society of America Student Research Grants and a Kleinman Grant through the USGS awarded to M.C.R. and National Science Foundation grant EAR 0440382.
*Corresponding author: Present address: Department of Geoscience, University of Iowa, Iowa City, IA 52242, USA. Telephone: (319) 384-3474. E-mail: michael-rowe{at}uiowa.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONOREGON CASCADES BACKGROUNDMETHODSRESULTSDISCUSSIONCONCLUSIONS AND BROADER...SUPPLEMENTARY DATAAPPENDIXREFERENCES |
|---|
Amundsen H. E. F., Neumann E. R. Redox control during mantle/melt interaction. Geochimica et Cosmochimica Acta (1992) 56:2405–2416.[CrossRef][Web of Science]
Anderson A. T., Wright T. L. Phenocrysts and glass inclusions and their bearing on oxidation and mixing of basaltic magmas, Kilauea Volcano, Hawaii. American Mineralogist (1972) 57:188–216.[Web of Science]
Arculus R. J. Oxidation status of the mantle: past and present. Annual Review of Earth and Planetary Sciences (1985) 13:75–95.[CrossRef][Web of Science]
Ariskin A. A., Frenkel M. Y., Barmina G. S., Nielsen R. L. COMAGMAT; a Fortran program to model magma differentiation processes. Computers and Geosciences (1993) 19:1155–1170.[CrossRef]
Bacon C. Calc-alkaline, shoshonitic, and primitive tholeiitic lavas from monogenetic volcanoes near Crater Lake, Oregon. Journal of Petrology (1990) 31:135–166.
Bacon C., Bruggman P., Christiansen R., Clynne M., Donnelly-Nolan J., Hildreth W. Primitive magmas at five Cascade volcanic fields: melts from hot, heterogeneous sub-arc mantle. Canadian Mineralogist (1997) 35:297–423.
Baker M. B., Grove T. L., Price R. Primitive basalts and andesites from the Mt. Shasta region, N. California: products of varying melt fraction and water content. Contributions to Mineralogy and Petrology (1994) 118:111–129.[CrossRef][Web of Science]
Ballhaus C. Redox states of lithospheric and asthenospheric upper mantle. Contributions to Mineralogy and Petrology (1993) 114:331–348.[CrossRef][Web of Science]
Ballhaus C., Frost B. R. The generation of oxidized CO2-beraing basaltic melts from reduced CH4-bearing upper mantle sources. Geochimica et Cosmochimica Acta (1994) 58:4931–4940.[CrossRef][Web of Science]
Ballhaus C., Berry R. F., Green D. H. Oxygen fugacity controls in the Earth's upper mantle. Nature (1990) 348:437–440.[CrossRef]
Ballhaus C., Berry R. F., Green D. H. High pressure experimental calibration of the olivine–orthopyroxene–spinel oxygen geobarometer: implications for the oxidation state of the upper mantle. Contributions to Mineralogy and Petrology (1991) 107:27–40.[CrossRef][Web of Science]
Barnes S. J. Chromites in komatiites, 1. magmatic controls on crystallization and composition. Journal of Petrology (1998) 39:1689–1720.[CrossRef][Web of Science]
Beyer R. Magma differentiation at Newberry Crater in central Oregon. (1973) Eugene: Ph.D. thesis, University of Oregon.
Bezos A., Humler E. The Fe3+/
Fe ratios of MORB glasses and their implications for mantle melting. Geochimica et Cosmochimica Acta (2005) 69:711–725.[CrossRef][Web of Science]
Borg L. E., Clynne M. A., Bullen T. D. The variable role of slab-derived fluids in the generation of a suite of primitive calc-alkaline lavas from the southernmost Cascades, California. Canadian Mineralogist (1997) 35:425–452.[Web of Science]
Borg L. E., Blichert-Toft J., Clynne M. A. Ancient and modern subduction zone contributions to the mantle sources of lavas from the Lassen region of California inferred from Lu–Hf isotopic systematics. Journal of Petrology (2002) 43:705–723.
Bostock M., Hyndman R., Rondenay S., Peacock S. An inverted continental Moho and serpentinization of the forearc mantle. Nature (2002) 417:536–538.[CrossRef]
Brandon A. D., Draper D. S. Constraints on the origin of the oxidation state of mantle overlying subduction zones: an example from Simcoe, Washington, USA. Geochimica et Cosmochimica Acta (1996) 60:1739–1749.[CrossRef][Web of Science]
Brenan J. M., Shaw H. F., Ryerson F. J., Phinney D. L. Experimental-determination of trace-element partitioning between pargasite and a synthetic hydrous andesitic melt. Earth and Planetary Science Letters (1995) 135:1–11.[CrossRef][Web of Science]
Carmichael I. S. E., Ghiorso M. S. Oxidation–reduction relations in basic magma: a case for homogeneous equilibria. Earth and Planetary Science Letters (1986) 78:200–210.[CrossRef][Web of Science]
Carmichael I. S. E., Ghiorso M. S. The effect of oxygen fugacity on the redox state of natural liquids and their crystallizing phases. In: Modern Methods of Igneous Petrology: Understanding Magmatic Processes. Mineralogical Society of America, Reviews in Mineralogy —Nicholls J., Russel J. K., eds. (1990) 24:191–212.
Carroll M., Rutherford M. Sulfur speciation in hydrous experimental glasses of varying oxidation state: Results from measured wavelength shifts of sulfur X-rays. American Mineralogist (1988) 73:845–849.[Abstract]
Carroll M. R., Webster J. D. Solubilities of sulfur, noble gases, nitrogen, chlorine, and fluorine in magmas. In: Volatiles in Magmas. Mineralogical Society of America, Reviews in Mineralogy —Carroll M. R., Holloway J. R., eds. (1994) 30:231–279.
Catchings R. D., Mooney W. E. Crustal structure of east central Oregon: relation between Newberry volcano and regional crustal structure. Journal of Geophysical Research (1988) 93:10081–10094.[CrossRef]
Cervantes P., Wallace P. J. Role of H2O in subduction-zone magmatism: New insights from melt inclusions in High-Mg basalts from central Mexico. Geology (2003) 31:235–238.
Christie D. M., Carmichael I. S. E., Langmuir C. H. Oxidation states of mid-ocean ridge basalt glasses. Earth and Planetary Science Letters (1986) 79:397–411.[CrossRef][Web of Science]
Churikova T., Dorendorf F., Worner G. Sources and fluids in the mantle wedge below Kamchatka, evidence from across-arc geochemical variation. Journal of Petrology (2001) 42:1567–1593.
Clynne M. A., Borg L. E. Olivine and chromian spinel in primitive calc-alkaline and tholeiitic lavas from the southernmost Cascade range, California: a reflection of relative fertility of the source. Canadian Mineralogist (1997) 35:453–472.[Web of Science]
Conrey R., Sherrod D., Hooper P., Swanson D. Diverse primitive magmas in the Cascade arc, Northern Oregon and Southern Washington. Canadian Mineralogist (1997) 35:367–396.[Web of Science]
Conrey R. M., Taylor E. M., Donnelly-Nolan J. J., Sherrod D. North–Central Oregon Cascades: exploring petrologic and tectonic intimacy in a propagating intra-arc rift. In: Field Guide to Geologic Processes in Cascadia. Oregon Department of Geology and Mineral Industries Special Paper —Moore G., ed. (2002) 36:47–90.
Danyushevsky L. V., Della-Pasqua F. N., Sokolov S. Re-equilibration of melt inclusions trapped by magnesian olivine phenocrysts from subduction-related magmas: petrological implications. Contributions to Mineralogy and Petrology (2000) 138:68–83.[CrossRef][Web of Science]
Danyushevsky L. V., McNeill A. W., Sobolev A. Experimental and petrological studies of melt inclusions in phenocrysts from mantle-derived magmas: an overview of techniques, advantages and complications. Chemical Geology (2002) 183:5–24.[CrossRef][Web of Science]
Davies J. H., Stevenson D. J. Physical model of source region of subduction zone volcanics. Journal of Geophysical Research (1992) 97:2037–2070.
Eiler J. M., Carr M. J., Reagan M., Stolper E. Oxygen isotope constraints on the sources of Central American arc lavas. Geochemistry, Geophysics, Geosystems (2005) 6. doi:10/1029/2004GC000804.
Elkins Tanton L. T., Grove T. L., Donnelly-Nolan J. Hot, shallow mantle melting under the Cascades volcanic arc. Geology (2001) 29:631–634.
Frost D. J., McCammon A. The redox state of the Earth' mantle. Anuual Review of Earth and Planetary Sciences (2008) 36:389–420.[CrossRef]
Ghiorso M. S., Carmichael I. S. E. Modeling magmatic systems: petrologic applications. In: Thermodynamic Modeling of Geological Minerals, Fluids and Melts. Mineralogical Society of America, Reviews in Mineralogy —Carmichael I. S. E., Eugster H. P., eds. (1987) 17:467–499.
Graham D. W., Reid M. R., Jordan B. T., Grunder A. L., Leeman W. P., Lupton J. E. Mantle source provinces beneath the northwestern USA delimited by helium isotopes in young basalts. Journal of Volcanology and Geothermal Research (2009) (in press).
Green T. H., Sie S. H., Ryan C. G., Cousens D. R. Proton microprobe-determined partitioning of Nb, Ta, Zr, Sr and Y between garnet, clinopyroxene and basaltic magma at high pressure and temperature. Chemical Geology (1989) 74:201–216.[CrossRef][Web of Science]
Grove T. L., Parman S. W., Bowring S. A., Price R. C., Baker M. B. The role of an H2O-rich fluid component in the generation of primitive basaltic andesites and andesites from the Mt. Shasta region, N. California. Contributions to Mineralogy and Petrology (2003) 142:375–396.[Web of Science]
Guffanti M., Weaver C. Distribution of late Cenozoic volcanic vents in the Cascade range: volcanic arc segmentation and regional tectonic considerations. Journal of Geophysical Research (1988) 93:6513–6529.
Harry D. L., Green N. L. Slab dehydration and basalt petrogenesis in subduction systems involving very young oceanic lithosphere. Chemical Geology (1999) 160:309–333.[CrossRef][Web of Science]
Hart S. R., Brooks C. Clinopyroxene–matrix partitioning of K, Rb, Cs, Sr, and Ba. Geochimica et Cosmochimica Acta (1974) 38:1799–1806.[CrossRef][Web of Science]
Hart S. R., Dunn T. Experimental cpx/melt partitioning of 24 trace elements. Contributions to Mineralogy and Petrology (1993) 113:1–8.[CrossRef][Web of Science]
Hauri E. H. SIMS analysis of volatiles in silicate glasses, 2: isotopes and abundances in Hawaiian melt inclusions. Chemical Geology (2002) 183:115–141.[CrossRef][Web of Science]
Higgins M. Petrology of Newberry Volcano, Central Oregon. Geological Society of America Bulletin (1973) 84:455–487.
Hochstaedter A. G., Gill J. B., Taylro B., Ishizuka O., Yuasa M., Morita S. Across-arc geochemical trends in the Izu–Bonin arc: constraints on source composition and mantle melting. Journal of Geophysical Research (2000) 105:495–512.[CrossRef]
Hochstaedter A., Gill J., Peters R., Broughton P., Holden P., Taylor B. Across-arc geochemical trends in the Izu–Bonin arc: Contributions from the subducting slab. Geochemistry, Geophysics, Geosystems (2001) 2. 2000GC000105.
Jensen RA. Roadside Guide to the Geology of Newberry Volcano (2000) 168. CenOreGeoPub, Bend, OR.
Johnson D. M., Hooper P. R., Conrey R. M. XRF analysis of rocks and minerals for major and trace elements on a single low dilution Li-tetraborate fused bead. Advances in X-ray Analysis (1999) 41:843–867.
Jordan BT. Basaltic volcanism and tectonics of the High Lava Plains, southeastern Oregon. In: Ph.D. thesis (2001) Corvallis: Oregon State University.
Jordan B. T. Age-progressive volcanism of the Oregon High Lava Plains; overview and evaluation of tectonic models. In: Plates, Plumes and Paradigms. Geological Society of America, Special Papers —Foulger G. R., Natland J. H., Presnal D. C., Anderson D. L., eds. (2005) 388:503–515.[CrossRef]
Jordan B. T., Streck M. J., Grunder A. L. Bimodal volcanism and tectonism of the High Lava Plains, Oregon. In: Field Guide to Geologic Processes in Cascadia. Oregon Department of Geology and Mineral Industries Special Paper —Moore G., ed. (2002) 36:23–46.
Jugo P. G., Luth R. W., Richards J. P. An experimental study of the sulfur content in basaltic melts saturated with immiscible sulfide ro sulfate liquids at 1300°C and 1·0 GPa. Journal of Petrology (2005a) 46:783–798.
Jugo P. J., Luth R. W., Richards J. P. Experimental data on the speciation of sulfur as a function of oxygen fugacity in basaltic melts. Geochimica et Cosmochimica Acta (2005b) 69:497–503.[CrossRef][Web of Science]
Kamenetsky V. S., Crawford A. J., Meffre S. Factors controlling chemistry of magmatic spinel: an empirical study of associated olivine, Cr-spinel and melt inclusions from primitive rocks. Journal of Petrology (2001) 42:655–671.
Kay R. W. Aleutian magnesian andesites; melts from subducted Pacific Ocean crust. Journal of Volcanology and Geothermal Research (1978) 4:117–132.[CrossRef][Web of Science]
Kent A. J. R., Elliott T. R. Melt inclusions from Mariana Arc lavas, implications for the composition and formation of island arc magmas. Chemical Geology (2002) 183:263–286.[CrossRef][Web of Science]
Kent A. J. R., Norman M. D., Hutcheon I. D., Stolper E. M. Assimilation of seawater-derived components in an oceanic volcano: evidence from matrix glasses and glass inclusions from Loihi seamount, Hawaii. Chemical Geology (1999) 156:299–319.[CrossRef][Web of Science]
Kent AJR, Stolper EM, Francis D, Woodhead J, Frei R, Eiler J. Mantle heterogeneity during the formation of the North Atlantic Igneous Province: constraints from trace element and Sr–Nd–Os–O isotope sytematics of Baffin Island picrites. Geochemistry, Geophysics, Geosystems (2004) 5. doi: 10.1029/2004GC000743.
Kinzler R. J., Donnelly-Nolan J. M., Grove T. L. Late Holocene hydrous mafic magmatism at the Paint Pot crater and Callahan flows, Medicine lake Volcano, N. California and the influence of H2O in the generation of silicic magmas. Contributions to Mineralogy and Petrology (2000) 138:1–16.[CrossRef][Web of Science]
Knaack C, Cornelius SB, Hooper PR. Trace element analyses of rocks and minerals by ICP-MS (1994) Technical Notes, GeoAnalytical Lab, Washington State University.
Kress C. V., Carmichael I. S. E. The compressibility of silicate liquids containing Fe2O3 and the effect of composition, temperature, oxygen fugacity and pressure on their redox states. Contributions to Mineralogy and Petrology (1991) 108:82–92.[CrossRef][Web of Science]
Lee C.-T. A., Leeman W. P., Canil D., Li Z.-X. A. Similar V/Sc systematics in MORB and arc basalts: implications for the oxygen fugacities of their mantle source regions. Journal of Petrology (2005) 46:2313–2336.
Leeman W. P., Smith D. R., Hildreth W., Palacz Z. A., Rogers N. W. Compositional diversity of late Cenozoic basalts in a transect across the southern Washington Cascades; implications for subduction zone magmatism. Journal of Geophysical Research (1990) 95:19561–19582.[CrossRef]
Leeman W. P., Tonarini S., Chan L. H., Borg L. E. Boron and lithium isotopic variations in a hot subduction zone; the southern Washington Cascades. Chemical Geology (2004) 212:101–124.[CrossRef][Web of Science]
Leeman W. P., Lewis J. F., Evarts R. C., Conrey R. M., Streck M. J. Petrologic constraints on the thermal structure of the Cascades arc. Journal of Volcanology and Geothermal Research (2005) 140:67–105.[CrossRef][Web of Science]
Li Z.-X. A., Lee C.-T. A. The constancy of upper mantle fO2 through time inferred from V/Sc ratios in basalts. Earth and Planetary Science Letters (2004) 228:483–493.[CrossRef][Web of Science]
Linneman SR. The petrologic evolution of the Holocene magmatic system of Newberry Volcano, central Oregon. In: Ph.D. thesis (1990) Laramie: University of Wyoming. 312.
Luhr J. F., Aranda-Gomez J. J. Mexican peridotite xenoliths and tectonic terranes: correlations among vent location, texture, temperature, pressure, and oxygen fugacity. Journal of Petrology (1997) 38:1075–1112.[CrossRef][Web of Science]
MacLeod N., Sherrod D. Geologic evidence for a magma chamber beneath Newberry Volcano, Oregon. Journal of Geophysical Research (1988) 93:10067–10079.[CrossRef]
MacLeod N., Sherrod D., Chitwood L., Jensen B. Geologic map of Newberry Volcano, Deschutes, Kalamath, and Lake Counties, Oregon. US Geological Survey Miscellaneous Map Investigations Series (1995) Map I-2455.
Mathez E. A. Influence of degassing on oxidation states of basaltic magmas. Nature (1984) 310:371–375.[CrossRef]
McCaffrey R., Long M. D., Goldfinger C., Zwick P. C., Nabelek J. L., Johnson C. K., Smith C. Rotation and plate locking at the southern Cascadia subduction zone. Geophysical Research Letters (2000) 27:3117–3120.[CrossRef][Web of Science]
McDonough W. F., Sun S. S. The composition of the Earth. Chemical Geology (1995) 120:223–253.[CrossRef][Web of Science]
McKenzie D., O’Nions R. K. Partial melt distributions from inversion of rare earth element concentrations. Journal of Petrology (1991) 32:1021–1091.
Metrich N., Schiano P., Clocchiatti R., Maury R. C. Transfer of sulfur in subduction settings: an example from Batan Island (Luzon volcanic arc, Philippines). Earth and Planetary Science Letters (1999) 167:1–14.[CrossRef][Web of Science]
Metrich N, Bonnin-Mosbah M, Susini J, Menez B, Galoisy L. Presence of sulfite (SIV) in arc magmas: Implications for volcanic sulfur emissions. Geophysical Research Letters (2002) 29. 10.1029/2001GL014607.
Metrich N, Berry A, O’Neill H, Susini J. A XANES study of sulfur speciation in synthetic glasses and melt inclusions (abstract). Geochimica et Cosmochimica Acta (2005) 69(10). Supplement 1, A51.
Nielsen R. L., Michael P., Sours-Page R. Chemical and physical indicators of compromised melt inclusions. Geochimica et Cosmochimica Acta (1998) 62:831–839.[CrossRef][Web of Science]
Ozawa K. Olivine–spinel geospeedometry: Analysis of diffusion-controlled Mg–Fe2+ exchange. Geochimica et Cosmochimica Acta (1984) 48:2597–2611.[CrossRef][Web of Science]
Parkinson I. J., Arculus R. J. The redox state of subduction zones: insights from arc peridotites. Chemical Geology (1999) 160:409–423.[CrossRef][Web of Science]
Plank T., Langmuir C. H. Tracing trace elements from sediment input to volcanic output at subduction zones. Nature (1993) 362:739–743.[CrossRef]
Reiners P. W., Hammond P. E., McKenna J. M., Duncan R. A. Young basalts of the central Washington Cascades, flux melting of the mantle and trace element signatures of primary arc magmas. Contributions to Mineralogy and Petrology (2000) 138:249–264.[CrossRef][Web of Science]
Roedder E., Emslie R. Olivine–liquid equilibrium. Contributions to Mineralogy and Petrology (1970) 29:275–289.[CrossRef][Web of Science]
Rowe MC. The role of subduction fluids in generating compositionally diverse basalts in the Cascadia subduction zone. In: Ph.D. thesis (2006) Corvallis: Oregon State University.
Rowe M. C., Nielsen R. L., Kent A. J. R. Anomalously high Fe contents in rehomogenized olivine hosted melt inclusions from oxidized magmas. American Mineralogist (2006) 91:82–91.
Rowe M. C., Kent A. J. R., Nielsen R. L. Determination of sulfur speciation and oxidation state of olivine hosted melt inclusions. Chemical Geology (2007) 236:303–322.[CrossRef][Web of Science]
Sack R., Carmichael I., Rivers M., Ghiroso M. Ferric–ferrous equilibria in natural silicate liquids at 1 bar. Contributions to Mineralogy and Petrology (1980) 75:369–376.[CrossRef][Web of Science]
Sato M., Wright T. L. Oxygen fugacities directly measured in magmatic gases. Science (1966) 153:1103–1105.
Schmidt M. E., Grunder A. L., Rowe M. C. Segmentation of the Cascade arc as indicated by Sr and Nd isotopic variation among diverse primitive basalts. Earth and Planetary Science Letters (2008) 266:166–181.[CrossRef][Web of Science]
Schmidt M. W., Poli S. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters (1998) 163:361–379.[CrossRef][Web of Science]
Scowen P. A. H., Roeder P. L., Helz R. T. Reequilibration of chromite within Kilauea Iki lava lake, Hawaii. Contributions to Mineralogy and Petrology (1991) 107:8–20.[CrossRef][Web of Science]
Sherrod D. R., Smith J. G. Quaternary extrusion rates of the Cascade range, northwestern United States and southern British Columbia. Journal of Geophysical Research (1990) 95:19465–19474.[CrossRef]
Sherrod DR, Taylor EM, Ferns ML, Scott WE, Conrey RM, Smith GA. Geologic map of the Bend 30- x 60-minute quadrangle, Central Oregon. US Geological Survey Geologic Investigations Series (2004) Map I-2683.
Smith D. R., Leeman W. P. Chromian spinel–olivine phase chemistry and the origin of primitive basalts of the southern Washington Cascades. Journal of Volcanology and Geothermal Research (2005) 140:49–66.[CrossRef][Web of Science]
Sobolev A. V., Chaussidon M. H2O concentrations in primary melts from suprasubduction zones and mid-ocean ridges: Implications for H2O storage and recycling in the mantle. Earth and Planetary Science Letters (1996) 137:45–55.[CrossRef][Web of Science]
Stanley W. D., Mooney W. D., Fuis G. S. Deep crustal structure of the Cascade range and surrounding regions from seismic refraction and magnetotelluric data. Journal of Geophysical Research (1990) 95:19419–19438.[CrossRef]
Stern RJ. Subduction zones. Reviews of Geophysics (2002) 40:1012. doi:10.1029/2001RG000108.[CrossRef]
Stolper E., Newman S. The role of water in the petrogenesis of Mariana trough magmas. Earth and Planetary Science Letters (1994) 121:293–325.[CrossRef][Web of Science]
Sun W. D., Binns R. A., Fan A. C., Kamenetsky V. S., Wysoczanski R., Wei G. J., Hu Y. H., Arculus R. J. Chlorine in submarine volcanic glasses from the eastern Manus basin. Geochimica et Cosmochimica Acta (2007) 71:1542–1552.[CrossRef][Web of Science]
Tatsumi Y., Hamilton D. L., Nesbitt R. W. Chemical characteristics of fluid phase released from a subducted lithosphere and origin of arc magmas: evidence from high-pressure experiments and natural rocks. Journal of Volcanology and Geothermal Research (1986) 29:293–309.[CrossRef][Web of Science]
Wade J. A., Plank T., Melson W. G., Soto G. J., Hauri E. H. The volatile content of magmas from Arenal Volcano, Costa Rica. Journal of Volcanology and Geothermal Research (2006) 157:94–120.[CrossRef][Web of Science]
Walker J. A., Roggensack K., Patino L. C., Cameron B. I., Matias O. The water and trace element contents of melt inclusions across an active subduction zone. Contributions to Mineralogy and Petrology (2003) 146:62–77.[CrossRef][Web of Science]
Walker GW, Duncan RA. Geologic map of the Salem 1° by 2° quadrangle, western Oregon. US Geological Survey Miscellaneous Investigations Series (1989) Map I-1893.
Wallace P. J., Carmichael I. S. E. Sulfur in basaltic magmas. Geochimica et Cosmochimica Acta (1992) 56:1863–1874.[CrossRef][Web of Science]
Wallace P. J., Carmichael I. S. E. S speciation in submarine basaltic glasses as determined by measurements of SK X-ray wavelength shifts. American Mineralogist (1994) 79:161–167.[Abstract]
Wells R. E. Paleomagnetic rotations and the Cenozoic tectonics of the Cascade arc, Washington, Oregon, and California. Journal of Geophysical Research (1990) 95:19409–19417.[CrossRef]
Winther K. T., Watson E. B., Korenowski G. M. Magmatic sulfur compounds and sulfur diffusion in albite melt at 1 GPa and 1300–1500°C. American Mineralogist (1998) 83:1141–1151.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||