Journal of Petrology Advance Access published online on July 19, 2007
Journal of Petrology, doi:10.1093/petrology/egm034
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Oxygen Isotope Evidence for Chemical Interaction of K
lauea Historical Magmas with Basement Rocks
1Department of Geology & Geophysics, University of Hawai’i, Honolulu, HI 96822, USA
2Department of Geology & Geophysics, University of Minnesota, Minneapolis, MN 55455, USA
3Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USA
Received January 30, 2007; Revised typescript accepted June 13, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
K
lauea historical summit lavas have a wide range in matrix 
18OVSMOW values (4·9–5·6
) with lower values in rocks erupted following a major summit collapse or eruptive hiatus. In contrast, 18O values for olivines in most of these lavas are nearly constant (5·1 ± 0·1). The disequilibrium between matrix and olivine 18O values in many samples indicates that the lower matrix values were acquired by the magma after olivine growth, probably just before or during eruption. Both Mauna Loa and Klauea basement rocks are the likely sources of the contamination, based on O, Pb and Sr isotope data. However, the extent of crustal contamination of Klauea historical magmas is probably minor (< 12%, depending on the assumed contaminant) and it is superimposed on a longer-term, cyclic geochemical variation that reflects source heterogeneity. Klauea's heterogeneous source, which is well represented by the historical summit lavas, probably has magma 18O values within the normal mid-ocean ridge basalt mantle range (5·4–5·8) based on the new olivine 18O values.
KEY WORDS: Hawaii; Klauea; basalt; oxygen isotopes; crustal contamination
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
K
lauea volcano has undergone dramatic changes over the last 
320 years varying from continuous effusion to explosive eruptions with major summit collapses (>100 m) (Macdonald et al., 1983
; Holcomb 1987; Garcia et al., 2003). The 1924 summit explosions and collapse were accompanied by a reversal in the temporal trend in lava composition (including incompatible trace element abundances and Pb, Sr and Nd isotope ratios) that was interpreted to reflect changes in melting processes and compositional heterogeneity in the source (Pietruszka & Garcia, 1999a; Garcia et al., 2003). The additional effects of crustal contamination were superimposed on the compositions of lavas erupted just after the collapse (Pietruszka & Garcia, 1999a). Klauea's historical period has been punctuated by numerous collapses (1823, 1832, 1840, 1868, 1924; Table 1). Here we evaluate the role of crustal contamination during the entire historical period using O isotopes, which have been shown to be a sensitive indicator of crustal contamination (e.g. Taylor, 1974).
|
Crustal contamination has been documented at various oceanic islands (e.g. Pitcairn; Eiler et al. 1995
; Canary, Thirlwall et al., 1997; Hawai’i, Gaffney et al., 2005) including the current eruption of Klauea, where magmas were stored for many years in the rift zone prior to eruption (Garcia et al., 1998). It should also be noted that 18O variations could also reflect source heterogeneity (e.g. Eiler 2001). To unravel the potentially offsetting influences of source heterogeneity vs crustal contamination, we analyzed O isotopes in coexisting olivine and matrix material from well-characterized historical lavas from Klauea volcano (20 eruptions during 400 years; Pietruszka & Garcia, 1999a; Garcia et al., 2003). These results provide a detailed time series of O isotopes, which are compared with those for another active hotspot volcano, Grimsvötn in Iceland, where 800 years of volcanic activity (including the Laki flood basalt) have been investigated (Bindeman et al., 2006). After evaluating the effects of crustal contamination on Klauea historical lavas, we examine the issue of 18O values for the Hawaiian plume source feeding this volcano. |
KLAUEA ERUPTIVE HISTORY SUMMARY
|
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
K
lauea's historical activity (1790 to present) includes prolonged periods of quiescent lava effusion, interrupted by major collapses (
100 m) of the caldera floor, rare violent explosions, and brief to prolonged rift zone eruptions (Table 1). The most dramatic event occurred in 1790, when violent explosions killed 80 Hawaiians (McPhie et al., 1990). This event was accompanied by a major lower east rift zone eruption (Macdonald et al., 1983). It was preceded by several explosive eruptions dating back to AD 1670 (Swanson et al., 2006) and followed, sometime between 1820 and 1823, by high lava fountaining within the caldera (Sharp et al., 1987). In 1823, a ‘great’ crack extending 22 km down the SW rift produced a large lava flow causing the summit lava lake to drain and a summit collapse of >100 m (Brigham, 1909; Macdonald et al., 1983). Eruptive activity resumed until 1832, when cracks formed on the NE rim of the caldera producing a small lava flow, which was followed by the summit lava lake draining and a >100 m collapse of the caldera floor (Brigham, 1909). Lava-lake activity restarted a few months later and persisted until 1840, when the caldera floor sank 100 m during a major east rift zone eruption (Dana, 1891). 
Crater, in the SW corner of the caldera (Fig. 1), was the focus of subsequent lava lake activity until a M 7·9 earthquake in 1868 caused the caldera floor to sink 90 m to 180 m (Brigham, 1909; Wyss & Koyanagi, 1992). Vigorous activity resumed immediately in and continued until 1894, when the lava lake drained. For the next 13 years, eruptive activity was episodic with a 5·5 year hiatus (Brigham, 1909). Lava-lake activity returned to in 1907, with major overflows and fissure eruptions from 1918 to 1921, covering large portions of the caldera floor.
|
In 1924, the
lava lake drained, its crater floor collapsed 300 m, and numerous violent phreatic explosions ensued (Jaggar, 1924). This event marked a reversal in Klauea's lava compositional trend that has persisted for >80 years (Pietruszka & Garcia, 1999a; Garcia et al., 2003) and a major change in its eruptive style. Summit eruptions became short and infrequent (seven brief eruptions during the next 10 years; Jaggar, 1924; Finch, 1940). From 1934 to 1952, the volcano remained quiet. The next 30 years (1952–1982) witnessed mostly short-lived eruptions from fissures in and near crater (Fig. 1) and the more frequent east and SW rift zone eruptions, following a 115 year hiatus in rift zone eruptive activity (Macdonald et al., 1983). There have been no summit eruptions since the start of the current east rift zone, 
eruption in 1983 (e.g. Garcia et al., 2000). |
SAMPLES |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Well-dated and chemically characterized summit flow and tephra samples were utilized for this study [see Pietruszka & Garcia (1999a
) and Garcia et al. (2003) for locations, petrography and compositional information (all for XRF major and trace elements, and many for Pb, Sr and Nd isotopes)]. This sample suite spans 310 years of Klauea's eruptive history (1670 to September 1982) and includes two young prehistoric samples [dates of 1790 were given by Pietruszka & Garcia (1999a) based on the work of McPhie et al. (1990); however, subsequent 14C dating shows these samples are 120 years older (Swanson et al., 2006) than previously thought]. Unfortunately, only one sample (1832) is available for the period 1832–1865, during which two major collapse events occurred (1832 and 1840). Klauea erupted just prior to the 1832 collapse. Subsequent major collapses (1868 and 1924) and eruptive gaps (1897–1902 and 1934–1954) are well bracketed by the sample suite (Table 2). A previous study determined 18O values on whole-rocks for five historical summit samples, including samples from three of the same eruptions examined in our study (1885, 1954 and 1971), obtaining 18O values (all values in this paper are reported relative to Vienna Standard Mean Ocean Water, VSMOW), of 5·5–5·6 (Kyser et al., 1982). Another study of O isotopes of olivine in 10 SOH-4 drill-core lavas from Klauea's east rift zone yielded 18O values of 4·9–5·2 , averaging 5·1 ± 0·1 (Eiler et al., 1996), consistent with the Kyser et al. (1982) summit results. In contrast, olivines from the current eruption are more variable, ranging in 18O values from 4·4 to 4·9 for 16 samples, with low values during the first 2 years of episodic eruptions (4·8 ± 0·1), followed by even lower values for the continuous effusion phase (4·64 ± 0·2; Garcia et al., 1998). The matrix for these and other lavas (24 total) also ranges widely in 18O values (4·56–5·25), averaging 4·8 ± 0·2 for the early lavas and 5·1 ± 0·1 for later lavas (Garcia et al., 1998).
|
Most of the summit samples are simple in mineralogy, consisting of rare to common olivine crystals (0·1–4·7 vol.%) in a glass matrix. Insufficient olivine was present in samples erupted between 1929 and 1934 to allow for O isotope analyses of olivine in these samples. Glass is present in most samples; in those samples where it is absent, matrix material was used for
18O determinations. Olivine elemental compositions for many of the summit samples were reported by Garcia et al. (2003). Most are in equilibrium with the whole-rock composition, except a few crystals from prehistoric samples (1790-1) and the early 19th century samples (1820 and 1832). The 1790 and 1820 samples contain a few lower forsterite crystals (Fo 84–85) that are weakly reversely zoned and the 1832 sample has rare higher forsterite olivines (88%; Garcia et al., 2003). Glass MgO contents of the summit samples range widely (from 9·3 wt% in the 1820 samples to 6·2 wt% in the September 1982 sample, 1982S-12; Table 3). Whole-rock MgO contents (labeled ‘R’ in Table 3 for samples without glass) range from 12·3 to 6·6 wt%. These values are considered reasonable indicators of matrix MgO content in the olivine-poor samples (< 2·5 vol.%). Sample 1790-1, with 4·7 vol.% olivine, is the exception.
|
To evaluate possible crustal contaminants for K
lauea magmas, new samples were collected from the accidental debris from the 1924 explosions (denoted by the prefix Z followed by a number and either ‘b’ for basalt or ‘g’ for gabbro) and from well cuttings collected every 3 m from the Puna Geothermal Venture well KS-3 in the lower east rift zone of Klauea (labeled PGV followed by the depth of sample collection in meters). This well, which reached temperatures of 320°C (D. Thomas, personal communication, 2006), provides deep samples (up to 2237 m) from the interior of Klauea volcano. The cuttings are 2–4 mm across and most have a gray matrix with sparse to common olivine. Within each sample interval studied, the chips are petrographically similar.
The accidental blocks were derived from the shallow section (< 300 m) overlying the summit magma reservoir (Jagger, 1924; Macdonald, 1944) and are thought to be possible contaminants for the summit reservoir, especially the 1924 collapse (Pietruszka & Garcia, 1999a). The blocks range in diameter from 25 to 150 cm, contain abundant olivine (30 vol.%) and have a grayish green coating. Sample Z19b is distinctive for its black olivine, which has been interpreted to result from high-temperature metamorphism (>700°C; Haggerty & Baker, 1967). Olivines in sample Z19b range from forsterite 94 to 98, the highest ever reported for a Hawaiian rock. These high forsterite contents result from exsolution of magnetite within olivine. Jaggar (1924) measured accidental block temperatures to 700°C, a minimum pre-eruption value. Mineral compositions (olivine, magnetite and orthopyroxene) within this block yielded temperatures of 970 ± 170°C using the Quilf95 program (D. Lindsley, personal communication, 1999). Sample Z6g has mildly oxidized olivine (metallic gray color), whereas olivines in sample Z2b are unoxidized. Thus, these blocks show a broad range of metamorphic histories from mild (Z2b) to intense baking (Z19b).
|
METHODS AND RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Whole-rock XRF
The new accidental block and well cutting samples were analyzed by X-ray fluorescence (XRF) at the University of Massachusetts for whole-rock major and trace (Nb, Rb, Zr, Sr, Y, V, Zn, Ni, Cr, Ba, La, and Ce) elements using methods described by Rhodes & Vollinger (2004
). The analytical uncertainties for these analyses have been given by Rhodes & Vollinger (2004). Samples were washed in deionized water until the solution was clear. Clean chips were hand picked before crushing in a tungsten carbide mill for less than 2 min.
The KS-3 geothermal well cuttings have typical Hawaiian tholeiite major element compositions with MgO concentrations of 6·3–9·0 wt%. Their compositions resemble Klauea lavas in having relatively high TiO2, moderate SiO2, and Zr/Nb of 9·6–11 (Table 2). These samples have experienced at least moderate levels of alteration based on their loss on ignition (LOI) of 1· 4–2·6 wt%, their somewhat low K2O/P2O5 compared with magmatic values (< 1·5 vs >1·5; Wright, 1971), and their low Rb contents (< 5 ppm, except the deepest sample). There is no correlation of these alteration indicators with the depth at which the cuttings were collected (Table 2). Except for these indicators, the major and trace element compositions of the five chip samples are remarkably similar to those of typical Klauea lavas (e.g. Nb/Zr of 10–11; Rhodes et al., 1989; Garcia et al., 2000, 2003).
The accidental blocks are high-MgO tholeiites (18·9–21· 6 wt%) with correspondingly high Cr and Ni concentrations (Table 2). The blocks have a wide range in K2O, TiO2 and incompatible trace elements (Table 2), magmatic K2O/P2O5 (1· 6–2·1) and low LOI values (<0·25 wt%). The blocks have Klauea-like Zr/Nb ratios (12–13; Rhodes et al., 1989). They are petrographically and geochemically similar to the suite C picritic basalts from Klauea's caldera cliffs (see Casadevall & Dzurisin, 1987).
Matrix oxygen isotopes
Oxygen isotope analyses of matrix material were made by conventional methods (Clayton & Mayeda, 1963) at the University of Minnesota using ClF3 (Borthwick & Harmon, 1982). CO2 gas from each extraction carried out during 1993 and 1994 was analyzed on a Finnigan MAT delta E mass spectrometer. CO2 gas extracted during 1997 and 1998 was analyzed with a Finnigan MAT 252 mass spectrometer. Each sample was split into 2–5 aliquots depending on the total sample weight, and at least two aliquots were analyzed to check for analytical reproducibility (see Table 3 for the number of replicates for each). For all but one sample (accident block Z19b, with the black olivines), analytical reproducibility is 0·1 or better. For 1993 and 1994 extractions, one laboratory standard (Nain plagioclase) was analyzed for every four matrix samples. The overall reproducibility for Nain plagioclase for the duration of this work was ±0·20 (1
) for 125 analyses. A mid-ocean ridge basalt (MORB) glass from East Pacific Rise (EPRD 10RB), which had previously been analyzed at the Geophysical Laboratory (Ito et al., 1987), was analyzed as a laboratory standard during 1997 and 1998. The overall reproducibility for EPRD 10RB for the two years was ±0·17 (1 ) for 56 analyses. The results reported here are relative to 18OVSMOW , normalized against SLAP (Standard Light Antarctic Precipitation, which is defined as –55 relative to VSMOW; Coplen, 1988).
The matrix 18O values for the summit suite of samples vary from 4·90 to 5·62 (Table 2) with no systematic variation with the date of eruption (Fig. 2). PVG K3-3 cuttings show a wide variation 18O from 1·75 to 5·31, generally decreasing with depth (Table 3). The three accidental blocks are remarkably similar in their O isotope values (5·50 ± 0·04) given their wide range in degree of metamorphism.
|
) and 
) lavas. Gray vertical bars show times of major collapses (dates are listed at top of figure); stippled zones show major gaps in eruptive activity (see 
, Ito et al. (1987
, Eiler (2001Olivine oxygen isotopes
Olivines were separated from historical K
lauea lavas and tephra by hand picking from coarsely crushed, but otherwise untreated, samples. Oxygen isotope compositions of c. 2 mg aliquots of these separates were determined by laser fluorination, using a 50W CO2 laser and BrF5 as reagent (Valley et al., 1995). Product O2 was converted to CO2 by reaction with hot graphite; CO2 was then analyzed for its isotopic composition by dual-inlet gas source mass spectrometry on a Thermo Finnegan 253 system at California Institute of Technology. Data are reported in units of per mil versus the VSMOW standard. Analyses were standardized by comparison with measurements of Gore Mountain garnet standard (Valley et al., 1995) and a San Carlos olivine (Eiler et al., 1995, 1996; Bindeman et al., 2006). Both standards were analyzed between two and five times on each day of analysis, and data for unknowns analyzed on that day were corrected by the average difference between measured and accepted values for these standards. The external precision of repeat measurements of separate splits of unknown samples averaged ±0·05 (1). This is comparable with the typical external precision for replicate measurements of silicate standards for this laboratory and technique (e.g. Eiler et al., 1995, 1996; Bindeman et al., 2006), suggesting that the olivine separates analyzed in this study are homogeneous in 18O at the scale of c. 2 mg aliquots. Garcia et al. (1998) evaluated the interlaboratory differences between Caltech and the University of Minnesota by comparing the results for a matrix glass from the eruption. The sample was run in duplicate in both laboratories. The results were 4·68 by conventional methods vs 4·79 by laser, which are within the analytical error of the two methods.
Ten of 13 olivine separates yield average 18O values that group tightly around a value of 5·11 ± 0·07 (Fig. 3). Members of this population are indistinguishable from one another, within analytical precision, and equal in 18O to olivine in equilibrium with typical MORB (Eiler et al., 2000; Cooper et al., 2004). This average value is also similar to the 18O values of olivines from Mauna Loa, L
’ihi, shield-building Haleakal
(i.e. Honomanu series), submarine Mauna Kea lavas, and olivines recovered from the Klauea SOH4 drill core (all of which have suite-average 18O values from 5·05 to 5·22; Eiler et al., 1996; Wang et al., 2003), and within the range typical of olivines from most ocean-island basalts and mantle peridotites (18O values of 5·0–5·4; Eiler, 2001).
|
Three historical K
lauea olivine samples are distinguishable from this tightly grouped set of 10 samples. Olivine from sample 1894-2 has a 18O of 4·91, just below the range typical of mantle-derived olivines and similar to previously observed averages for the Kula series of Haleakal. Samples 1971S and 1982-A-20 have the lowest values for summit lavas, 4·70 and 4·60, respectively, overlapping the range of values for olivines from the continuous phase of Klauea's eruption (Fig. 3) and within the range typical of subaerial Mauna Kea lavas (Eiler et al., 1996; Wang et al., 2003). Relatively low 18O values in Hawaiian olivines have been previously interpreted to result from lithospheric contamination (Eiler et al., 1996), a low-18O plume component (Lassiter & Hauri, 1998) or contamination within the volcanic edifice (Garcia et al., 1998; Wang et al., 2003). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
The wide range in matrix
18O values (4·9–5·6) during the 310 year period for Klauea summit lavas is in striking contrast to the nearly constant matrix 18O values (3·1 ± 0·1) for 800 years at Iceland's Grimsvötn volcano, including the 15 km3 Laki eruption (Bindeman et al., 2006). The variation observed at Klauea is a magmatic signature rather than a secondary feature because all of the summit samples are fresh, rapidly quenched volcanic rocks (Garcia et al., 2003). Could the Klauea variation be related to mantle source variations, as suggested for oceanic island basalts (e.g. Harmon & Hoefs, 1995)? The rapid variations in Klauea summit 18O values are difficult to reconcile with components in the Hawaiian mantle, especially given the lack of correlation of 18O values with incompatible trace element ratios (Nb/Y; Fig. 4) that are considered a good source indicator (Rhodes & Hart, 1995). Furthermore, 18O values of olivines from summit lavas are relatively constant from prehistoric times to 1982 (5·1 ± 0·1 ; Fig. 3), whereas Sr, Pb and Nd isotope ratios varied considerably during this period (e.g. 206Pb/204Pb 18·39–18·66; Pietruszka & Garcia, 1999a). These results are also distinct from those for subaerial HSDP2 core olivine, which showed lower average values (4·79 ± 0·13 ; Wang et al., 2003). Variable amounts of peridotite partial melting or crystallization of olivine and minor spinel are likely to have opposite but minimal impacts on the matrix 18O value (+0·1 for 20% crystallization vs –0·1 for up to 30% partial melting; Eiler 2001; Wang et al., 2003). Also, there is no correlation in the extent of fractionation (MgO content; Fig. 4) with matrix 18O value (e.g. at 7· 0 MgO, 18O values span the entire range of values; Table 3). Below we examine the temporal variation and discuss the possible causes for the 18O variation in historical Klauea lavas.
|
, rock) and whole-rock Nb/Y in historical KTemporal variation in oxygen isotopes
Two features emerge when the data are viewed in the context of K
lauea summit activities (Table 1): (1) matrix 18O values are lower following major summit collapses (100+ m; 1868 and 1924) or significant gaps in eruptive activity (>5 years; 1898 and 1934; Fig. 2); (2) they increase during periods of nearly continuous eruptive activity between these events (1877–1894 and 1912–1921), except for the period of short, infrequent eruptions from 1952 to 1982 (when very little lava was erupted; Dvorak & Dzurisin, 1993) and during the continuing eruption (Fig. 2). The drop in matrix 18O values following summit collapses and during gaps in eruptive activity or reduced levels of effusion may represent times when the summit reservoir magma is contaminated. The second feature, increasing matrix 18O values, may indicate that contaminated magma was progressively diluted with new, uncontaminated magma. New magma was more or less continuously added to the summit reservoir during historical times (e.g. Finch, 1940; Macdonald et al., 1983; Dvorak & Dzurisin, 1993). For the eruption, there was a 0·2–0·3 increase in matrix 18O values when the eruptive style switched from episodic to continuous effusion, although there was no change in other geochemical parameters (Garcia et al., 1998). These fluctuations in matrix 18O values were interpreted to represent variable amounts of crustal contamination within Klauea's east rift zone, with less contamination during periods of higher effusion (Garcia et al., 1998).
The olivine 18O values for summit historical lavas record a different history. The values are all relatively high, comparable with those for MORB olivine, and nearly constant except for the 1961 and 1971 lavas (Fig. 3). The lavas also have low olivine values. These results are in contrast to the wide variation in olivine isotope values for the Laki eruption (2·3–5·2), which were attributed to the incomplete equilibration of olivine from influxes of mantle-derived magma into a large chamber with 18O-depleted magma (Bindeman et al., 2006). The Klauea summit olivines apparently grew in magmas with 18O values of 5·45–5·75, assuming equilibrium growth and an olivine–melt isotopic fractionation of 0·55 [average of 0·7 (Ito & Stern, 1986) and 0·4 (Wang et al., 2003)]. The summit olivine data do not provide independent evidence for contamination except for the 1961 and 1971 summit and lavas (Fig. 3).
If the matrix and olivine 18O values are compared (Fig. 4), a record of disequilibrium in Klauea summit magmas is revealed. The extent of disequilibrium was greatest following the late 17th century major explosions and progressively decreased through the end of the 19th century when equilibrium values were achieved (Fig. 4). The 1924 major summit collapse and explosions, and the 1934–1954 eruptive hiatus marked another major matrix–olivine disequilibrium event, with the magnitude of the disequilibrium decreasing from 1954 to 1971 (Fig. 5). However, the April 1982 lava and many of the lavas (those erupted during periods of lower eruption rate) show disequilibrium values. No event was recorded that might explain contamination in the 1982 lavas. The presence of normal 18O values in olivine (5·0–5·2) from the September 1982 eruption (Fig. 3) suggests that these magmas were contaminated after the olivines grew, probably during eruption. All of the post-1950 lavas have low matrix 18O isotope values indicating that there was a major disruption in the magmatic plumbing system, which may have allowed for more magma contamination. As discussed above, the lavas with disequilibrium matrix–olivine values were erupted during periods of lower eruption rates. In summary, the two large summit events in the late 17th and early 20th centuries had a profound impact on the 18O values of Klauea magmas. If contamination caused the drop in O isotope values, where did it occur?
|
|
) lavas. The isotopes show a broad correlation, which may result from contamination of KImplications of O, Sr and Pb isotope data for contamination source and location
There is a general correlation of
18O with 87Sr/86Sr ratios, to a lesser extent with 206Pb/204Pb (Fig. 6) but none with 
Nd. The Klauea summit samples with the lowest 18O values tend to have higher Sr and somewhat lower Pb isotope ratios, although there is a broad range at a given 18O value (Fig. 6). These results help to identify potential sources of the contamination that created the lower 18O values in historical Klauea summit lavas. Shallow-level, low-temperature processes should produce higher 18O values (Gregory & Taylor, 1981). For example, values up to 8·7 have been reported for rocks from the shallow levels of Klauea geothermal wells (Smith & Thomas, 1990). Deeper in these wells where temperatures are higher (300°C), the 18O values are lower (3·2; Smith & Thomas, 1990). This trend is confirmed with our new data for well KS-3, which generally decrease with depth from 5·32 at intermediate depths (1353 m) to 1· 86 at 2237 m (Table 3), the lowest value reported for Hawaiian lavas. The Klauea high-temperature metamorphosed accidental blocks from the 1924 eruption, which were derived from depths <300 m (Jaggar, 1924), lack low 18O values (Table 3). Therefore, they are unsuitable contaminants for generating the lower 18O values observed in some summit lavas (Fig. 6).
The Pacific oceanic crust section, which was invoked by Gaffney et al. (2005) to explain low 18O values in West Maui lavas, has 206Pb/204Pb ratios too high (18·5–19·1; King et al, 1993; Fekiacova et al., 2007) to explain the range or correlation in Klauea matrix O and Pb isotope values (Fig. 6). The only other potential contaminant for Klauea magmas is Mauna Loa lavas (Fig. 7), which have lower 206Pb/204Pb and higher 87Sr/86Sr ratios than Klauea lavas (Fig. 6). Glasses from fresh Mauna Loa submarine lavas have a wide range in 18O values (4·7–5·5 ; Garcia et al., 1989). The lavas with lower values would be suitable contaminants for Klauea magmas to generate the positive O–Sr and O–Pb isotope correlation (Fig. 6). An alternative contaminant would be moderately hydrothermally altered Mauna Loa lavas.
|
The relatively high
18O values for the summit olivines (Fig. 3), and the disequilibrium between olivine and matrix 18O values in some samples (Fig. 5), indicate that the contamination occurred late, probably just prior to or during eruption. The most likely location for this contamination would be in Klauea's shallow (3–6 km deep; Klein et al., 1987) magma chamber, which is probably located within the flanks of Mauna Loa, or in the conduit during eruption (Fig. 7). The frequent, major collapses of the summit crater floor of 100+ m and the gaps in eruptive activity provide a suitable mechanism for disruption of the roof of the summit magma chamber and the assimilation of Mauna Loa or Klauea hydrothermally altered rocks into Klauea magma. The 1924 collapse resulted in significant changes in lava trace element chemistry, which were explained by 3·5% contamination from partially melted, seawater altered Klauea basalt (Pietruszka & Garcia, 1999a). However, the trends of the lava chemistry changes following this event are consistent with a Mauna Loa source (decreasing Pb and Nd and increasing Sr isotopes). The small drop in the 18O value during this period (0·22; Table 3) would require 5% bulk contamination, if the lowest 18O value from the KS-3 well (1· 86) is used. The results from these two assimilation models are consistent and suggest that the extent of the contamination was small but significant for the 1924 event. The larger drop in matrix 18O value following the 1868 collapse (0·45; Table 3) requires more extensive contamination (12%) using the same assumptions. Unfortunately, no Pb, Sr or Nd isotope data are available for the 1877 sample to test this hypothesis. Additional testing of contamination scenarios using rock major or trace element data is complicated by the probability that the contaminants were only partially digested (e.g. 1924 eruption; Pietruszka & Garcia, 1999a). Mauna Loa rocks have lower incompatible elements than Klauea rocks, although during partial melting and assimilation, the incompatible element concentrations of the resulting melts will be increased relative to the host Klauea magma. An alternative process to bulk contamination for the summit rocks with more typical Klauea geochemistry (higher Pb and lower Sr isotope ratios; e.g. sample 1954-1) is O isotope exchange with groundwater or altered lavas, which would leave no geochemical signature other than lowering the 18O value (e.g. Klauea groundwater and steam condensates have 18O values of –1 to –14; Hinkley et al., 1995). Regardless of the contamination source, the rapid change in matrix 18O values suggests that Klauea's shallow magma reservoir is small, which is consistent with some geophysical estimates (e.g. Decker, 1987) and geochemical modeling (Pietruszka & Garcia, 1999b). The Klauea situation is in marked contrast to the consistency of matrix 18O values for Grimsvötn volcano over the last 800 years. The consistency of Grimsvötn's matrix 18O values was related to its large crustal magma chamber, which is supported by the 15 km3 flood basalt flow it produced in 1783 (Bindeman et al., 2006). In contrast, Klauea's summit reservoir is thought to have been small (only 2–3 km3) over the last 200 years (Pietruszka & Garcia, 1999b). The crustal contamination of Klauea magmas is superimposed on a longer-term geochemical variation that reflects source heterogeneity (Pietruszka & Garcia, 1999a). The 18O composition of this source is discussed in the next section.
Klauea source oxygen isotope values
The origin of the diverse 18O values observed in mantle-derived basalts has been of considerable interest. Although a limited range of 18O values typifies MORB glasses (5·4–5·8), 18O values correlate with source geochemical parameters (Eiler 2001) suggesting a mantle origin for at least some of the observed 18O diversity in mantle-derived basalts (Harmon & Hoefs, 1995). However, concern has surrounded the origin of lower O isotopes (< 5·4, below normal MORB values), which are observed in many ocean island basalts. Some studies relate lower values to crustal contamination (e.g. Garcia et al., 1998; Thirlwall et al., 1997), whereas others have suggested they are an inherent feature of the mantle (e.g. Garcia et al., 1993; Harmon & Hoefs, 1995). For example, Wang et al. (2003) argued for a ‘primary’ mantle 18O value of 5·5 for HSDP2 Mauna Kea lavas, with lower values resulting from contamination (e.g. Wang et al., 2003). The new results presented here for historical Klauea summit lavas support the contamination explanation for lower 18O values.
The consistency of the olivine 18O values in historical Klauea summit lavas (5·1 ± 0·1 ; Fig. 3), despite the significant range in Pb and Sr isotopes during this period (Table 3), is strong evidence in favor of the interpretation that there is a limited range in 18O values in the mantle source (5·45–5·75; the value depends on the magnitude of the equilibrium matrix–olivine fractionation), despite the heterogeneous source producing Klauea magmas (e.g. Pietruszka & Garcia, 1999a). This range is consistent with the prevailing value for MORB mantle (Ito et al., 1987; Eiler 2001), suggesting that the 18O value of the Hawaiian mantle plume feeding Klauea is typical of the upper mantle. However, it should be noted that diversity exists within the plume. In particular, olivines from some Ko’olau volcano lavas have much higher 18O values (5·5–6·0 vs 5·0–5·2), which are inferred to result from melting eclogite blocks within the Hawaiian plume (Eiler et al., 1996).
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
A wide range in matrix
18O value (4·9–5·6) was determined for Klauea summit lavas spanning 310 years of eruptive history. These results are in marked contrast to the essentially constant but much lower matrix 18O values (3·1 ± 0·1) reported for the last 800 years of eruptive activity for the Icelandic hotspot volcano Grimsvötn. Klauea matrix values do not vary systematically with time, although lower values were found in rocks erupted following major summit collapses or eruptive hiatuses. The difference between the matrix 18O values for these volcanoes probably reflects the size of the magma reservoir system, with the Klauea system being much smaller. Olivine 18O values in most of the Klauea lavas are nearly constant (5·1 ± 0·1), whereas those from Grimsvötn volcano show substantial variation (2·3–5·2). The disequilibrium between matrix and olivine 18O values in Klauea samples indicates that the lower matrix values were acquired by the magma after olivine grew, probably just before or during eruption. The sources of the contamination of Klauea historical lavas are probably the host rocks for the magma reservoir system. These include Mauna Loa and Klauea rocks, based on O, Pb and Sr isotope data, whereas oceanic crust contamination is not indicated. The extent of crustal contamination of Klauea historical magmas was probably minor (5%) for the 1924 collapse, depending on the assumed contaminant. The contamination signature of Klauea magma is superimposed on a larger, longer-term cyclic geochemical variation that reflects source heterogeneity. Klauea's heterogeneous source, which is well represented by the analyzed summit lavas, probably has 18O values within the normal MORB mantle range (5·4–5·8), based on the new olivine 18O values.
|
ACKNOWLEDGEMENTS |
|---|
This paper is dedicated to David Green on the occasion of his 70th birthday. David has been an inspiration to all who study basaltic magmas. Thanks go to the many friends who have assisted with this study, including Sorena Sorensen for lending critical samples from the Smithsonian collection; Nancy Baker, Rachel Konishi, Joann Romano, Kate Bridges, Kristina Garcia and Joe Horrell for help in the field; Chad Shishido and Kelly Kolysko for sample preparation; to Rhea Workman for assistance with the laboratory work; and Reed McEwan and John Tacinnelli for matrix sample oxygen isotope analysis. Analytical work in the Caltech stable isotope laboratories was supported, in part, by NSF grant EAR-0337736 and a grant from the EAR Technician support program. Our thanks go to journal reviewers Amy Gaffney, J. M. Rhodes and Ilya Bindeman for their helpful comments. This paper is SOEST Contribution 7149, and was supported by NSF grants (EAR03-36874) to MG and (EAR-0345905) to JE.
*Corresponding author. E-mail: morgarcia{at}hawaii.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONKiLAUEA ERUPTIVE HISTORY SUMMARYSAMPLESMETHODS AND RESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Bindeman IN, Sigmarsson O, Eiler JM. Time constraints on the origin of large volume basalts derived from O-isotope and trace element mineral zoning and U-series disequilibria in the Laki and Grimsvotn volcanic system. Earth and Planetary Science Letters (2006) 245:245–259.
Borthwick J, Harmon RS. A note regarding ClF3 as an alternative to BrF5 for oxygen isotope analysis. Geochimica et Cosmochimica Acta (1982) 46:1665–1668.[CrossRef][ISI]
Brigham WT. The Volcanoes of Kilauea and Mauna Loa on the Island of Hawaii, Their Variously Recorded History to the Present Time. In: Memoirs of the Bernice Pauahi Bishop Museum (1909) 2. Honolulu, HI: Bishop Museum Press. [Reprinted 1974, Kraus Reprint Company, Millwood, NY, 608 pp.].
Casadevall TJ, Dzurisin D. Stratigraphy and petrology of the Uwekahuna Bluff Section, Kilauea Caldera. US Geological Survey, Professional Papers (1987) 1350:351–375.
Clayton RN, Mayeda TK. The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochimica et Cosmochimica Acta (1963) 27:43–52.[CrossRef][ISI]
Cooper K, Eiler JM, Asimow P, Langmuir C. Oxygen isotope evidence for the origin of enriched mantle beneath the mid-Atlantic ridge. Earth and Planetary Science Letters (2004) 220:297–316.[CrossRef][ISI]
Coplen TB. Normalization of oxygen and hydrogen isotope data. Chemical Geology (Isotope Geoscience) (1988) 72:293–297.[CrossRef]
Dana JD. Characteristics of Volcanoes (1891) New York: Dodd, Mead. 399.
Decker RW. Dynamics of Hawaiian volcanoes: an overview. US Geological Survey Professional Paper (1987) 1350:997–1018.
Dvorak JJ, Dzurisin D. Variations in magma supply rate at Kilauea Volcano, Hawaii. Journal of Geophysical Research (1993) 98:22255–22268.
Eiler JM. Oxygen isotope variations of basaltic lavas and upper mantle rocks. In: Stable Isotope Geochemistry. Mineralogical Society of America, Reviews in Mineralogy—Valley JW, Cole DR, eds. (2001) 43:319–364.
Eiler JM, Farley K, Valley JW, Stolper EM, Hauri E, Craig H. Oxygen isotope evidence against bulk recycled sediment in the source of Pitcairn Island lavas. Nature (1995) 377:138–141.[CrossRef]
Eiler JM, Farley KA, Valley JW, Hofmann A, Stolper EM. Oxygen isotope constraints on the sources of Hawaiian volcanism. Earth and Planetary Science Letters (1996) 144:453–468.[CrossRef][ISI]
Eiler JM, Schiano P, Kitchen N, Stolper EM. Oxygen isotope evidence for recycled crust in the sources of mid-ocean ridge basalts. Nature (2000) 403:530–534.[CrossRef]