Mantle reservoir geochemistry from statistical analysis of ICP-MS trace element data of equatorial mid-Atlantic MORB glasses

Ž . We report Inductively Coupled Plasma Mass Spectrometer ICP-MS trace element concentration data for 21 elements in Ž 51 basaltic glasses recovered from 58 stations along a 1500-km segment of the equatorial mid-Atlantic Ridge MAR; 58N to . Ž 58S . Isotope systematics of these glasses revealed mixing between two isotopically distinct Schilling, J.-G., Hanan, B.B., McCully, B., Kingsley, R.H., 1994. Influence of the Sierra Leone mantle plume on the equatorial MAR: a Nd–Sr–Pb . isotopic study. J. Geophys. Res., 99, 12005–12028. HIMU and DM mantle sources. Trace element data presented here describe three distinct mantle reservoirs. Based on our new trace element data, we identified three, and possibly four, end-member mantle reservoirs beneath the equatorial MAR. HIMU-type signatures dominate the glasses from the northern Ž . Ž . equatorial MAR 58N to 08 with LarBa, LarTh and rare earth element REE values similar to P-type mid-Ocean Ridge Ž Ž . . Ž . Basalts MORB; 0.1, 9.11, LarYb s5.5, respectively . The basaltic glasses from the southern region 08 to 58S are N Ž mildly radiogenic in Pb and light REE depleted with N-type MORB trace element values for LarBa, LarTh and REE 0.8, Ž . . 40.0, LarYb -0.6, respectively . We estimate the mantle magma-source characteristics and identify the relative N contributions of the mantle end-member components in the glasses erupted along the 1500-km transect using the trace element concentration data. Specifically, we recognize the presence of a statistically and geochemically distinct mantle reservoir depleted in Th and Ba relative to average depleted MORB between the Romanche and the Chain fracture zones. To reach the full potential of this data set, we model mantle reservoir compositions using a modification of Q-MODE factor analysis. Multivariate classification and reduction of these data clearly distinguish trace element groups each reflective of mantle source heterogeneities present in this region of the MAR. Statistical analysis of these new trace element and available Nd–Sr–Pb isotope data for the same samples defines four statistically significant distinct end-member mantle components. The binary mixing model developed previously from geochemical data that invoked the mixing of two mantle reservoirs, HIMU and N-MORB, cannot adequately explain the variance in these trace element data. Incompatible element ) Corresponding author. Fax: q1-716-244-5689. Ž . Ž . E-mail addresses: [email protected] R.E. Hannigan , [email protected] A.R. Basu . 1 Fax: q1-870-972-3089. 2 Current address: Gen-Keyes Str. 22, A5020 Saltzburg, Austria. 0009-2541r01r$ see front matter q 2001 Elsevier Science B.V. All rights reserved. Ž . PII: S0009-2541 00 00335-1 ( ) R.E. Hannigan et al.rChemical Geology 175 2001 397–428 398 Ž . concentrations and ratios best describe the modeled end-members reservoirs, which are HIMU HPM , transitional N-MORB Ž . Ž . Ž . TNM , depleted N-MORB DNM and an fourth incompatible element-enriched component EPM . The statistically determined end-members explain 100% of the geochemical variance in the population but their modeled compositions may be related to process controlled variation. We clarify the relationship between statistical and geochemical reservoirs and conclude that a multiple mantle component model involving three of the identified reservoirs is clearly necessary to describe, specifically, the trace element characteristics of the equatorial mid-Atlantic MORBs. Our statistical methods yield tabulated trace element concentration data for each of these the modeled mantle components and provide insights into the MAR system, unavailable by any other technique, by allowing an estimation of the relative contributions of these end-member components in each of the 51 basaltic glass samples. q 2001 Elsevier Science B.V. All rights reserved.