THE GENESIS OF PYROXENITE-RICH PERIDOTITE AT CABO ORTEGAL (NW SPAIN). INFERENCES FROM GEOCHEMICAL, MINERAL AND PB-SR ISOTOPIC DATA

Abstract

Within the allochthonous Cabo Ortegal mafic-ultramafic complex of NW Spain abundant pyroxenite occurs in km size peridotite massifs of Limo, Herbeira and Uzal. In the Herbeira massif, pyroxenite forms a lens-shaped body ca. 3 km long and 300 m thick, essentially composed of websterite bands separated by minor dunite. Towards the top of the succession clinopyroxenite and orthopyroxenite alternate locally with garnet-rich mafic rocks. Petrography and field data allow to distinguish three main rock-types: (i) peridotites, essentially harzburgite and dunite, (ii) common pyroxenites, including massive olivine-websterite to dm to cm thick orthopyroxenite and garnet-bearing pyroxenites, and (iii) mafic rocks rich in either clinopyroxene or amphibole together with abundant garnet and zoisite, which associate garnet-rutile clinopyroxenite. The age of the pyroxenites has been established at ca. 390 Ma (U-Pb zircon, rutile; Santos et al., 1996). In binary diagrams (MgO vs. elements) peridotites plot at the Mg-richer end (> 30 wt% MgO), while mafic rocks and associated pyroxenite have low Mg contents (< 18 wt% MgO). The remaining common pyroxenites, expand between these two poles and show limited overlapping with peridotites. Displayed trends are continuous, but some bent shapes (SiO2, CaO) or enrichments (Na2O, Cr), mostly for pyroxenites, give an indication for different processes involved in the genesis of these rocks. Normalization to primitive mantle shows variable patterns not correlated with the main rock-types. These are characterized by important LILE enrichment for all samples with negative anomalies for Rb, Th, Nb, P and Zr, and relative enrichment in Cs, Ba, U and Sr. Ti anomalies are not systematic. Chondrite normalized REE patterns also exhibit variable LREE enrichment, Eu anomaly and total REE contents. The observed difference in major and trace element composition point to minor variations at the source area for the different rock-types, and the most probable scenario in all cases appears to be that of an arc-related mantle-wedge whose chemical composition had been affected by the introduction of melts or fluids. Limited enrichment in LILE and LREE indicate subduction of an oceanic crust poor in sediments (a relatively young oceanic basin?) and limited incorporation of external material and/or metasomatic processes within the source region previous to crystallization. Cr/(Cr+Al) ratios of spinel and XMg ratios of olivine from dunite and harzburgite plot in separate areas and do not follow the so called “olivine-spinel mantle array”. While harzburgites might correspond to ocean-floor or arc peridotites, an origin of the dunites as mantle restites appears to be excluded. Pyroxenites have distinctly lower Fo contents akin to that of volcanic rocks from different environments, MORBs to island arc tholeiites (Arai, 1994). Major element composition of clinopyroxene shows that dunite and harzburgite overlap in the region of clinopyroxenes from island arc peridotites (cf. Koloskov and Zharinov, 1993), while common pyroxenites and mafic rocks extend well into the island arc pyroxenite region. These data are comparable to those obtained for a similar peridotite-pyroxenite association from the Bragança complex (north Portugal) interpreted as a fragment of supra-subduction zone depleted mantle related to an island arc-continent collision (Bridges et al., 1995). Pb isotopic composition of pyroxene, garnet and amphibole from pyroxenites corrected for 390 Ma yield restricted range values similar to those reported for Beni Boussera and Lherz massifs where contribution or contamination with recycled crustal materials has been suggested (Mukasa et al., 1991; Pearson et al., 1993). In the present case, measured compositions are intermediate between those of present day MORBs and that of crustal derived granitic injections within the peridotites (Uzal massif, Santos et al., 1996), which would lend support to the crustal participation assumption. Zoisite exhibits a markedly different composition suggesting isotopic disequilibrium that might reflect a different (metamorphic?) origin for this mineral. Likewise, Pb isotopic composition of minerals from mafic rocks and associated garnet-rutile pyroxenites are lower than those of other pyroxenites which points to isotopic disequilibrium between the two rocks-types. 87Sr/86Sr initial ratios of analyzed minerals are mostly comprised between 0.7033 and 0.7049. These are very similar to data from orogenic massifs elsewhere (e.g., Balmuccia, Lherz, Lanzo), and close to those of present day OIBs suggesting participation of a component with high 87Sr/86Sr ratio in the genesis of these rocks. Nevertheless, it should appear that the minerals did not equilibrate for Sr isotopic composition in some rocks as Sri ratios display in cases significant variations within the same sample. It is concluded that the Cabo Ortegal ultramafic massifs have originated within a mantle-wedge environment variably modified by fluid infiltration. Harzburgite and dunite reflect minor modifications during partial melting within the mantle-wedge. Instead, pyroxenites with disrupted REE and LILE enriched patterns might correspond to areas of such mantle-wedge that have been modified by recycling in variable amounts of oceanic lithosphere and sediments. The possibility of continental crust being involved can not be excluded in view of the occurrence of granite injections within peridotite. Furthermore, the present data suggest that incompatible trace elements and isotopic tracers may record modifications on the sub-arc mantle chemistry by fluids or melts derived from sediments and altered oceanic crust.