Origin of garnet-gedrite-grunerite-cordierite rocks from garnet-biotite-sillimanite psamopelitic gneisses during melting and granulite grade metamorphism (Osor Complex, CCR, NE Iberia)

MIREIA TRAVERIA; SEBASTIÁN OSVALDO VERDECCHIA; FRANCISCO JOSÉ MARTÍNEZ; MERITXELL RABANILLO; JOAN RECHE; GISELA LEOZ

Ullapool, Escocia

Congreso; Granulites and Granulites 2018; 2018

Mineralogical Society of Great Britain & Ireland

g-oam-cam-crd-q gneisses (COR) occur in lenses within g-bi-sil-q semipelitic gneisses (QSP) in the sil zone of the LP-HT Variscan Osor Complex in the Catalan Coastal Ranges (NE of Barcelona). Both show similar q contents (≈ 35%), are affected by an S2 foliation, contain syn-tectonic q ± pl ± g veins with big subhedral g porphyroblasts near the veins and smaller, anhedral syn-tectonic g crystals far from the veins. In both, g show rims almost devoid of inclusions (q, op and MI?s) and inclusion-rich inner zones. The outcrop relations also suggest a genetic link among the two rock types. Here we support a derivation of COR bulk composition (BC) from QSP bulk during granulite-grade metamorphism at ca. 5.5±0.5 kbar and 750±50 ºC. Geochemistry of major and trace elements suggests clear bulk composition trends going from melt-rich QSP towards Fe-Mg-Mn-Ti richer and Na-Ca-K-Al poorer, melt-depleted COR. NCKFMASTO Theriak-Domino (de Capitani and Petrakakis, 2010) closed and open system thermodynamic models of COR and QSP suggests the necessity of losing melt ± hydration. Both geochemical and thermodynamic models predict that the lost melts should have been first Na-rich and evolving towards a K-richer granitic melt, which fits with observed g-pl-q (throndhjemitic) and leucogranitic leucosomes. The closed system modelling show some inconsistencies between predicted and observed mineralogy. In QSP the observed assemblage (g-bi-sil-pl-q-op) is predicted to be supersolidus (with Na-rich liq) at around 700-750ºC and 5.5 kbar. Some observed features (decreasing bi and pl, increasing sil and crd) require higher T (≈ 800ºC), but at these conditions the model also predict kfs, which is not observed. In addition, the model predicts high melt proportions that should have been clearly unstable according to rheological critical melt percentages RCMP proposed by different authors (Rosember and Handy, 2005). On the other hand, in the improbable case that melts where not lost, retrogression would have given rise to an almost g-absent bi-sil-ploam-q-op assemblage, which is also not compatible with the observations. In COR the model predicted assemblage at 800 ºC and 5.5 kbak is g-q-liq-op and the genesis of cam and oam clearly requires retrograde T (