Allison A. Price*1, Matthew G. Jackson1, Janne Blichert-Toft2, Mark Kurz3, Jim Gill4, Jerzy Blusztajn3, Frances Jenner5, Raul Brens6, Richard Arculus7
Total Page：16
File Type：pdf, Size：1020Kb
Geochemistry, Geophysics, Geosystems Supporting Information for Allison A. Price*1, Matthew G. Jackson1, Janne Blichert-Toft2, Mark Kurz3, Jim Gill4, Jerzy Blusztajn3, Frances Jenner5, Raul Brens6, Richard Arculus7 1. University of California, Santa Barbara, Department of Earth Science, 1006 Webb Hall, Santa Barbara, CA, 93106 USA (*[email protected]) 2. Laboratoire de Géologie de Lyon, CNRS UMR 5276, Ecole Normale Supériere de Lyon and Université Claude Bernard Lyon 1, 46 Allée d’Italie, 69007 Lyon, France 3. Woods Hole Oceanographic Institution, Woods Hole, MA, USA 4. Department of Earth Sciences, University of California Santa Cruz, Santa Cruz CA 95064, USA 5. Department of Environment, Earth and Ecosystems, The Open University, Walton Hall, Milton Keynes, UK 6. Department of Earth & Planetary Sciences, Macquarie University, Sydney, NSW 2109, Australia 7. Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia *Corresponding author Contents of this file Text S1 Figures S1 to S3 Additional Supporting Information (Files uploaded separately) Data sets S1 to S3 Introduction In Text S1 we provide an overview and interpretation of the mixing models shown in Supporting Information Data set 3 and Supporting Information Figure S3. Figure S1 shows a silica versus total alkali plot that includes the new Rotuman and Fijian OIB (FOIB) lavas in this study, with subdivisions for different rock classifications based on Le Bas et al. [1986]. In Figure S2 we show primitive mantle-normalized trace element patterns for the FOIB and Rotuman lavas in this study. Figure S3 shows mixing models between Samoan components and ambient depleted mantle components shown in various isotopic spaces. Text S1. Introduction Previous studies have provided compelling evidence for the presence of Samoan geochemical components in lavas erupted in the northern Lau Basin (e.g. Volpe et al. [1988], Gill and Whelan [1989], Poreda and Craig [1992], Wendt et al. [1997], Ewart et al. [1998], Pearce et al. [2007], Tian et al. [2008, 2011], Lupton et al. [2009], Hahm et al. [2012], Jenner at al. [2012], Lytle et al. [2012], Price et al. [2014, 2016], Nebel and Arculus [2015]). Recent work suggest that Samoan lavas host a variety of geochemical components, including EM2 (enriched mantle 2), EM1 (enriched mantle 1), HIMU (high µ = high 238U/204Pb), high 3He/4He, and DM (depleted mantle) components [Jackson et al. 2014]. To better evaluate how much of the new isotopic data from the northern Lau and North Fiji Basin lavas presented in this paper can be explained by the addition of these five Samoan components to the depleted backarc basin mantle, we provide a series of mixing models between four geochemically depleted end-member components (identified in lavas from the Lau and North Fiji Basins) and the five components suggested for the Samoan plume. Methods and caveats To generate the two-component mixing models presented in Supporting Information Figure 3, we use four geochemically depleted end-member components from the Lau and North Fiji Basins and mix them with five Samoan lavas, which are representative of the five Samoan components defined in Jackson et al. [2014] (EM2, EM1, HIMU, 3He/4He, and DM). The pertinent data for all end-members are shown in Supporting Information Table 3, and represent data from Samoan lavas and depleted Lau/North Fiji Basin lavas. The four depleted backarc basin endmembers and the five Samoan plume components were selected to generate an array of 20 (i.e., 4 x 5 = 20) mixing lines that encompass a large fraction of the new geochemical data on northern Lau and North Fiji Basin lavas. Of course, selection of additional depleted backarc basin components (and, thus, generation of additional mixing models) can encompass more of the geochemical data in the northern Lau and North Fiji Basins, but mixing models that include just four depleted backarc basin components capture much of the geochemical variability in the region. It is important to note that each of the lavas selected as end-members in the mixing model has complete geochemical dataset (including Hf, Pb, Sr, and Nd isotopic and trace element data measured on the same sample). This approach severely limited the choice of possible end-members in the mixing model, as relatively few lavas from either the northern Lau and North Fiji Basins or the Samoan plume have complete datasets for Sr, Nd, Pb and Hf isotopic ratios and trace element concentrations. Furthermore, we note that other inputs to the northern Lau and North Fiji Basin region (e.g., subducted HIMU Rurutu hotspot EM1 Rarotonga hotspot tracks) are not considered in these models, as our goal is to evaluate how much of the geochemical variability in lavas from the backarc basins can be explained by incorporation of Samoan components. Implications Overall, we find that two-component mixing lines among four geochemically depleted backarc basin end-members with five Samoan components capture much of the geochemical variability in the northern Lau and North Fiji Basin lavas. However, not all of the geochemical variability in the northern Lau and North Fiji Basin can be explained by the mixing models presented here. For example, much of the previously published Sr-Nd-Pb-Hf isotopic data from northeast Lau Basin (NELB) lavas plot outside the mixing models in nearly all isotopic spaces, in particular the Δ207Pb/204Pb - Δ208Pb/204Pb plot (See Supplementary Information Figure S3). Therefore the NELB lavas cannot be explained solely by the mixing of Samoan components with the ambient depleted mantle in the backarc basins. However, it was shown in Price et al. [2016] that NELB lavas likely sample components from both the Rarotona and Rurutu hotspot tracks, which are subducting into the northern Tonga Trench, but the Rarotonga and Rurutu hotspots are not included in our mixing models. Like the NELB samples, the mixing models presented here do not capture all of the radiogenic isotopic variability identified in samples from the FOIB and Rotuman lavas. In particular, the FOIB and Rotuman lavas have Δ208Pb/204Pb that is too low to be explained by the mixing of ambient depleted mantle and Samoan components (See Supplementary Information Figure S3). Similarly, one lava from the West Cikobia Volcanic zone (sample NLTD-9-1), has a strong EM1 signature with 207Pb/204Pb at a given 206Pb/204Pb (i.e., high Δ207Pb/204Pb) that is too high to be explained by mixing between depleted backarc basin and Samoan plume components. As NELB lavas have been shown to sample geochemically enriched hotspot components in the region that are not Samoan (Falloon et al., 2007; Price et al., 2016), it is not unreasonable to suggest that the FOIB and Rotuma lavas, as well as sample NLTD-9-1, might also sample non-Samoan geochemical components. However, this does not exclude the possibility that Samoan components are present in the NELB, FOIB/Rotuma, and NLTD-9-1 lavas: we only utilize two component mixing models here, and it is possible that multi-component mixtures (which include depleted backarc basin, Samoan and non-Samoan hotspot components) may explain the isotopic variability found in these lavas. 7 Basaltic Trachy- Andesite 6 Trachy- Trachy- Basalt Andesite 5 i kal Al te leii 4 Tho O (wt%) 2 3 2 O + Na Basaltic 2 Basalt Andesite K Andesite 1 0 45 50 55 60 SiO2 (wt%) Rotuma Island Fijian OIB in this study with new major element data Fijian OIB in this study with previously published major element data Figure S1. Silica versus total alkali plot, with subdivisions for different rock classifications based on Le Bas et al. [1986]. The alkali-tholeiite line is from Macdonald and Katsura [1964]. Previously published major element data for Fijian OIB examined in this study are from Gill and Whelan [1989] and Pearce et al. [2007]. The one Fijian lava marked with a “+” symbol represents the Type II lava (WQ7b), while all other Fijian lavas are Type I (see section 3.3.7 of the paper for more information). The diamond with a “+” symbol is ROT-11 (the only Rotuman lava that is tholeiitic) and the diamond with an “x” symbol is ROT-6 (the only Rotuman lava that is a trachy-basalt). The dark grey field represents previously published Fijian OIB major element data for lavas not studied here and are from Gill [1984]. The light grey field represents previously published Rotuma major element data for lavas not studied here and are from Price et al. [1990]. A. 100 New Rotuma ICP Trace Element Data ROT-4 ROT-12 ROT-3 10 ROT-8 ROT-11 Average Upolu Average MORB Sample/Primitive Mantle Average BAB 1 Rb Ba Th U Nb Ta K La Ce Pb Pr Nd Sr Zr Hf Sm Eu Ti Gd Tb Dy Ho Y Er Tm Yb Lu 100 B. New Fijian OIB ICP Trace Element Data W251 WQ208 W271a W135 10 FJ-12-5 WQ64 Mago Average Upolu Sample/Primitive Mantle Average MORB Average BAB 1 Rb Ba Th U Nb Ta K La Ce Pb Pr Nd Sr Zr Hf Sm Eu Ti Gd Tb Dy Ho Y Er Tm Yb Lu C. 100 Previously Publ. Fijian OIB ICP Trace Element Data WQ28 WQ7b Average Upolu Average MORB 10 Average BAB Sample/Primitive Mantle 1 Rb Ba Th U Nb Ta K La Ce Pb Pr Nd Sr Zr Hf Sm Eu Ti Gd Tb Dy Ho Y Er Tm Yb Lu Figure S2. Primitive mantle-normalized trace element patterns for the lavas examined in this and previous studies. Panel A shows new data from Rotuma Island lavas, panel B shows new data from young Fijian lavas, and panel C shows lavas previously published young Fijian OIB (FOIB) lavas (Pearce et al [2007]).