Layered amphibolite sequence in NE Sardinia, Italy: remnant of a pre-Variscan mafic silicic layered intrusion?

A banded amphibolite sequence of alternating ultramafic, mafic (amphibolite) and silicic layers, tectonically enclosed within Variscan migmatites, outcrops at Monte Plebi (NE Sardinia) and shows similarities with leptyno-amphibolite complexes. The ultramafic layers consist of amphibole (75-98%), garnet (0-20%), opaque minerals (1-5%) and biotite (0-3%). The mafic rocks are made up of amphibole (65-80%), plagioclase (15-30%), quartz (0-15), opaque minerals (2-3%) and biotite (0-2%). The silicic layers consist of plagioclase (60-75%), amphibole (15-30%) and quartz (10-15%). Alteration, metasomatic, metamorphic and hydrothermal processes did not significantly modify the original protolith chemistry, as proved by a lack of K2O-, Rb-enrichment, CaO-, MgO- depletion and by no shift in the REE patterns. Field, geochemical and isotopic data suggest that ultramafic, mafic and silicic layers represent repeated sequences of cumulates, basic and acidic rocks similar to macrorhythmic units of mafic silicic layered intrusions. The ultramafic layers recall the evolved cumulates of Skaergaard and Pleasant Bay mafic silicic layered intrusions. Mafic layers resemble Thingmuli tholeiites and chilled Pleasant Bay mafic rocks. Silicic layers with Na2O: 4-6 wt%, SiO2: 67-71wt% were likely oligoclase-rich adcumulates common in many mafic silicic layered intrusions. Some amphibolite showing a strong Ti-, P- and REE- depletion are interpreted as early cumulates nearly devoid of ilmenite and phosphates. All Monte Plebi rocks have extremely low Nb, Ta, Zr, Hf content and high LILE/HFSE ratios, a feature inherited from the original mantle sources. The mafic and ultramafic layers show slight and strong LREE enrichment respectively. Most mafic layer samples plot in the field of continental tholeiites in the TiO2-K2O-P2O5 diagram and are completely different from N-, E- and T-MORB as regards REE patterns and Nd, Sr isotope ratios but show analogies with Siberian, Deccan and proto-Atlantic rift tholeiites. Comparisons with Thingmuli, Skaergaard and Kiglapait rocks and with experimental data suggest that the Monte Plebi intrusion was an open-to-oxygen system with fO2  FMQ. Mafic and ultramafic samples yielded Nd(460) = + 0.79/ + 3.06 and 87Sr/86Sr = 0.702934 - 0.703426, and four silicic samples Nd(460) = -0.53/-1.13; 87Sr/86Sr = 0.703239 - 0.703653. Significant differences in Nd isotope ratios between mafic and silicic rocks prove that both groups evolved separately in deeper magma chambers, from different mantle sources, with negligible interaction with crustal material, and were later repeatedly injected within a shallower magma chamber. The spectrum of Sr and Nd isotope data is consistent with a slightly-enriched mantle metasomatized during an event earlier than 460 Ma. The metasomatising component was represented by alkali-Th-rich fluids of crustal origin rather than by sedimentary materials, able to modify alkali and Sr-Nd isotope systematics. Monte Plebi layered amphibolites might represent the first example of a strongly-metamorphosed fragment of an early Paleozoic mafic silicic layered intrusion emplaced in a thinning continental crust and then tectonically dismembered by Variscan orogeny.