Fluid pressure is a key parameter in earthquake mechanics, controlling seismic failure and plate coupling in convergent zones. Yet fluid pressure is also extremely difficult to quantify at seismogenic depth, which limits our knowledge of the stress state in accretionary prisms. Here, we show that the geochemical record of exhumed hydrothermal quartz veins may be used to place quantitative bounds on fluid pressure variations in subduction zones. The studied veins come from sediments accreted and exhumed by plate convergence in southwestern Japan. Quartz in veins displays growth rims of contrasted bright blue/dark brown cathodoluminescence (CL) colors, high/low Al concentrations, and low/high fluid inclusion abundance. Because Si-Al substitution (and charge compensation by Li) strongly depends on the rate of quartz precipitation and Si solubility, Al-Li concentrations must be sensitive to fluid pressure. This is confirmed by fluid inclusions, the density of which, converted into trapping pressures, record fluid pressure drops by up to ∼70 MPa from CL-brown, Al-Li-poor rims to CL-blue, Al-Li-rich quartz rims. CL-blue rims grow at a fast rate, high Si supersaturation and low fluid pressure whereas CL-brown rims grow at a slower pace, lower Si supersaturation, and higher fluid pressure. Quartz trace element chemistry thus offers a promising tool to quantify deep fluid pressure variations and their relationships to earthquakes.
Plain Language SummaryRocks at depth contain a small proportion of cavities filled with a fluid.The pressure of this cavity-filling fluid plays a major role in earthquakes generation. Nonetheless, measuring fluid pressure at depth is extremely difficult. One method relies on the analysis of fluid inclusions trapped at depth, but it is plagued by the possible modifications of their properties (i.e., their reequilibration) during exhumation. In this study we have conjointly analyzed, in japanese deformation zones, methane-rich fluid inclusions and the chemical composition of their quartz crystal hosts, which grew contemporaneously. We observed, in the quartz, growth rims, pointing to a succession of growth increments, with either a large or a low concentration in aluminum, a trace element in quartz. The Al-richer/poorer growth rims contain fluid inclusions that in average record lower/higher fluid pressure, respectively. As a consequence of the correlation between fluid pressure and aluminum content in quartz, the latter signal appears as a new proxy of fluid pressure at depth. It bears the large advantage, with respect to fluid inclusions, not to be affected by reequilibration processes that erase the original message from the depth. The temporal variations in aluminum/fluid pressure are interpreted as the result of the seismic cycle.