Portal de Periódicos da CAPES

Comment on “ D iverse magmatic effects of subducting a hot slab in SW J apan: Results from forward modeling” by J.‐I. K imura et al.

2015; Wiley; Volume: 16; Issue: 9 Linguagem: Inglês

10.1002/2015gc005914

1525-2027

Ivan Pineda‐Velasco, Tai Truong Nguyen, Hiroshi Kitagawa, Eizo Nakamura,

High-pressure geophysics and materials

The recently published paper by Kimura et al. [2014] reports geochemical and isotopic analyses of Quaternary adakitic dacites (ADK) from two volcanic groups, Daisen and Aonoyama, in southwest (SW) Japan (Figure 1). Based on Pb isotope compositions, these authors suggest that crustal assimilation played a major role in the genesis of ADK. This conclusion differs from that of an earlier study by Feineman et al. [2013], who advocated only a minimal crustal effect and concluded that the Pb isotope array for Daisen ADK can be attributed to partial melting of the subducting slab with a significant amount of sediment. The Pb isotope trends for Daisen ADK presented in these two studies are clearly different (Figure 2) and Kimura et al. [2014] did not directly compare the data sets or provide any explanation for the differences. Distribution of the adakitic volcanoes in SW Japan [Nishiki et al., 2012]. Colored triangles show the volcanic peaks in each area: lightblue, Daisen group; black, Sambe group; darkblue, Aonoyama group. The basemap is created using GMT [Wessel et al., 2013]. Pb isotope composition of ADK from Daisen and Aonoyama, obtained by Feineman et al. [2013], and this study, in comparison with those obtained by Kimura et al. [2014]. (a and b) Two volcanic groups plot in separated regions divided at 18.25. Data sets in Feineman et al. [2013] and obtained in this study show a trend shallower than that produced by the data in Kimura et al. [2014]. Note that our studies revealed that two ADK suites share a common linear trend (grey line). (c and d) Comparison of data sets for Daisen ADK. Compositions of upper crustal (UC) [Feineman et al., 2013] and lower crustal (LC) [Moriyama, 2006] lithologies are shown in the insets. The "X" indicates the hypothetical crustal component suggested by Kimura et al. [2014]. (e and f) Comparison of data sets for Aonoyama ADK. The trajectories of instrumental mass bias and error are shown in Figures 2c and 2d as solid and dotted black lines, respectively. Each line passes through the mean of Kimura et al.'s [2014] data. Red tie-lines connect data for the same lava domes in the Aonoyama volcanic group. The slopes of these tie-lines are similar to the slope representing instrumental mass bias (the one exception is for the data for Dakeyama). Analytical uncertainty (150 ppm) in our study is shown as error bars on the left side of each plot. In this comment, we critically examine the differences between the Pb isotope data sets presented by Feineman et al. [2013] and Kimura et al. [2014]. We present new data for Aonoyama ADK and we provide an explanation for the discrepancy between the data obtained in our lab and in this other recently published study. The samples used in Feineman et al. [2013] and Kimura et al. [2014] were collected from the main peaks around Misen at Daisen volcano [see Feineman et al., 2013, Figure 1]. In Kimura et al. [2014], ADK from the Hiruzen mountains, located 10 km southeast of Misen, are also included (n = 2). Data in Kimura et al. [2014] labeled as being for "Daisen" and "Karasugasen" (n = 12) are used in our evaluation. The major-element compositions for Daisen volcanics presented by Kimura et al. [2014] and obtained in our study, show great overlap (Table 1), thus large differences in the isotopic compositions for the sample suites investigated in the two studies would not be anticipated. We present new data for ADK from 14 lava domes in the Aonoyama chain, containing four volcanoes (Aonoyama, Sengokudake, Tokuyamamitakesan, and Dakeyama), allowing direct comparison of our data set with that of Kimura et al. [2014]. Major-element compositions we obtained for the Aonoyama ADK are essentially identical to those by Kimura et al. [2014] (see Table 2). A double spike (DS) method [Kuritani and Nakamura, 2003], using thermal ionization mass spectrometry (TIMS), was applied in Feineman et al. [2013] and to obtain our new data presented here. Kimura et al. [2014] employed a Tl-spiking (TS) technique employing multicollector inductively coupled plasma mass spectrometry (MC-ICPMS) [Kimura et al., 2010]. In our study, samples were leached with 6M HCl (at 100°C) prior to sample decomposition to eliminate potential contamination and secondary alteration. During the analytical campaign, NBS 981 (n = 10) yielded an average of 16.9418 (14) for 206Pb/204Pb, 15.4995 (19) for 207Pb/204Pb, 36.7261 (47) for 208Pb/204Pb (2σ variability on last digits in parentheses). Major-element abundances of the Aonoyama ADK were obtained by X-ray fluorescence spectrometry, as outlined in Feineman et al. [2013]. Figure 2 demonstrates the differences between the Pb isotope data sets obtained in our laboratory [Feineman et al., 2013, this study] and by Kimura et al. [2014]. The data from Kimura et al. [2014] show steeper trends in 206Pb/204Pb– 207Pb/204Pb and 206Pb/204Pb– 208Pb/ 204Pb plots than those observed by Feineman et al. [2013] and in our more recent study. Discrepancies in the Pb isotope data obtained by DS and TS methods have previously been discussed [e.g., Thirlwall, 2002; Baker et al., 2004]. For example, Baker et al. [2004] demonstrated that analyses by DS methods differ by 3000 ppm from those obtained by TS methods for Icelandic picrite [Stracke et al., 2003]. Baker et al. [2004, 2005] suggested that this discrepancy could reflect (1) inadequate correction of instrumental mass bias using the TS methods, (2) unusual isotopic fractionation induced by excessive matrix load, and (3) contamination by Pb added to the samples both in the field and in the laboratory. In both studies, Pb blanks were reported as being less than 50 pg, and therefore negligible compared with the quantities of Pb analyzed (>100 ng). Data sets obtained by Feineman et al. [2013] and in our study form a trend subparallel to the -error line, on a plot of versus , but shows significantly shallower slope in – (Figure 2). In contrast, the data for Daisen ADK from Kimura et al. [2014] follow the trend defined by mass bias. Also, tie lines connecting the data for each dome in Aonoyama group, presented by Kimura et al. [2014] and in our studies, are subparallel to the trend for mass bias (the one expception is for the data for Dakeyama). This observation is suggestive that the Pb isotope data presented in Kimura et al. [2014] are inaccurate due to an inadequate correction of instrumental mass bias. We note that the separation chemistry employed by Kimura et al. [2014] involves a single-column pass [cf. Kimura et al., 2003] that cannot sufficiently eliminate impurities [Baker et al., 2005; Kuritani and Nakamura, 2002]. A number of studies have documented that the Tl-correction for mass bias is compromised by excessive matrix loads inducing different behavior of Pb and Tl during sample introduction through nebulization and desolvation [Thirlwall, 2002; Kamenov et al., 2004]. Complete removal of matrix loads by double-column methods is crucial for obtaining precise and accurate Pb isotope analysis by both MC-ICPMS and TIMS [Woodhead, 2002; Baker et al., 2004]. Kimura et al. [2014] estimated the composition of crustal component by convergence of the Pb isotope arrays for the Daisen and Aonoyama groups ("X" in insets of Figures 2c and 2d). However, the composition estimated by this approach is clearly distinct from the measured compositions for both upper and lower-crustal lithologies collected in the vicinity of Daisen volcano [Moriyama, 2006; Feineman et al., 2013]. Based on our critical examination, we conclude that the Pb isotope data sets published by Kimura et al. [2014] are inaccurate due to erroneous correction for instrumental mass bias. The apparent convergence of the Pb isotope arrays [Kimura et al., 2014, Figure 7] to what these authors regard as a "common crustal component" is simply an artifact. The Pb isotope compositions of appropriate crustal rocks, presented by Moriyama [2006] and Feineman et al. [2013] indicate the nonavailability of crustal contaminants to deliver this hypothetical component beneath Daisen volcano. Our evaluation highlights the need for careful evaluation of the methods by which Pb isotope are obtained, as a number of factors can lead to erroneous data sets strongly affecting the conclusions reached. We are grateful to Gray Bebout for improving the manuscript. All data for this paper are properly cited and referred to in the reference list. A map (Figure 1) was created using the Generic Mapping Tools [Wessel et al., 2013].