DEEP FLUID CIRCULATION IN ALPINE SHEAR ZONES, PYRENEES, FRANCE - FIELD AND OXYGEN ISOTOPE STUDIES

A combined field, stable isotope, and whole-rock chemical study was made on late Cretaceous to Tertiary metasomatic shear zones cutting Hercynian gneisses in the Aston Massif, Pyrenees, France. Mylonitisation occurred during the early stages of Alpine compression under retrograde conditions at 400-450°C and about 10 km depth. Whole-rock δ18O values of (+11 to +12‰ in the gneisses) was lowered to +5 to +9‰ in the shear zones, with the quartz-muscovite 18O/16O fractionations of about 2 to 4‰ essentially unchanged. These 18O/16O systematics, together with δD muscovite=-40 to-50‰ indicate that large volumes of formation waters or D-rich meteoric waters passed through the shear zones during deformation. The same fluids also redistributed major elements, as shown by the correlation of δ18O shift with muscovitisation and albitisation reactions in granitic wall rocks. However, even though δ18O was universally lowered within the shear zones, the 18O/16O ratios were not homogenised, nor do they correlate in detail with the presence or absence of muscovitisation, suggesting that fluid flow was probably fracture-controlled and episodic. Field mapping shows that, along the length of a particular shear zone, muscovitisation of granite gneiss dies out 150m above the contact with underlying sillimanite gneiss. Thus, muscovitisation and albitisation of granite gneiss in shear zones and their wall rocks probably occurred during re-equilibration of acidic, chloride-rich, aqueous fluids that had previously moved upward within the shear zones through underlying sillimanite gneiss. Extremely high material-balance fluid-rock ratios (∼103) are required to explain the extent of muscovitisation along this shear zone, implying integrated fluid mass fluxes of about 108 kg/m2; this is probably close to the maximum value for other shear zones in the network. Similar volumes of a more chemically evolved fluid must have passed through the unmuscovitised mylonites, showing that the absence of alteration cannot necessarily be used to infer low values of fluid flux. For reasonable pressure gradients and time scales of fluid movement, effective permeabilities of 10-15 to 10-17 m2 are required. Such values can be accounted for by short-lived, widely-spaced cracks produced during seismic activity. A model is presented in which formation waters were seismically pumped down an underlying, shallow, southward-dipping decollement and then upward through the steeply-dipping shear zone network. © 1990 Springer-Verlag.