Understanding fluid flow in subduction zones

Subduction zones are the loci for a number of phenomena that impact human well-being. The largest earthquakes on record, as well as some of the most destructive have been generated within subduction zones. The most explosive and destructive volcanoes are also found in volcanic arcs over subduction zones and are fed by melting processes associated with subduction. But subduction does not just generate natural hazards. Some of the richest ore deposits in the world, such as those in Chile, are associated with volcanic arcs as well. What makes subduction so uniquely capable of pushing the extremity of natural phenomena? During subduction, cold, hydrated oceanic crust is carried deep into the lithospheric mantle where fluid is released sparking reactions that are difficult to produce through other tectonic processes.

            However, even though these fluids have been linked to important processes, their behavior is still poorly understood. Exhumed high-pressure/low-temperature (HP/LT) metamorphic rocks that show evidence of seismic activity (pseudotachylytes, brecciation) also record significant fluid movement associated with these events that is focused along planes of failure. However, it has yet to be determined if the fluid release accompanying earthquakes is a cause or a result. Trace element signatures distinct from those generated by partial melting of the mantle have been identified in volcanic arcs and interpreted as evidence of chemical input from the subducting slab. However, the proportion of the various subduction inputs (lithospheric mantle, oceanic crust, oceanic sediments) and the method of transportation: partial melt versus aqueous fluid, have yet to be definitively established.

Fig. 1: Transect across a fluid alteration zone showing changes in mineralogy.

            Constraining the behavior of fluids released by dehydration reactions in the subducting slab, and their mass transfer potential is key to understanding their role in seismic and volcanic activity. I propose a geochemical and petrological study of identified fluid pathways in exhumed HP/LT rocks to constrain the source, periodicity, duration, spatial distribution, and mass transfer potential of subduction zone fluid events. Samples will be collected in transects across identified fluid pathways (e.g. shear zones, veins, metasomatic alteration zones) to allow analysis on the meter to micron scale. Geochemical analysis will be carried out on whole rock samples on the millimeter to meter scale to constrain the penetration and mass transfer potential of passing fluids. Diffusion modeling across any measured chemical gradients identified in these whole rock transects will be used to constrain the duration of fluid pulses.

Fig. 2: Complex Mn zoning in garnet from the Western Alps showing the importance of intra-mineral analysis 

Fig. 3: Lithium concentration and δ⁷Li traverse showing evidence for fluid interaction in low-δ⁷Li zone. 

In situ analysis of mineral zoning will be used to constrain the periodicity of fluid movement as well as any disequilibrium between mineral phases associated with preferential alteration during fluid interaction. Geochemical analyses will include major and trace elements as well as lithium and oxygen isotopes. Isotopic signatures will be used to differentiate between fluid sources since both oxygen and lithium have well constrained values for reservoirs such as: altered upper and lower oceanic crust, subducted sediments and serpentinite. The use of both Li and O isotopes provides an opportunity to interrogate the isotopic record of fluid on a range of timescales. Oxygen diffuses slowly in common HP/LT minerals and can retain isotopic heterogeneity over hundreds of millions of years. Lithium on the other hand, has been shown in experimental studies to diffuse much faster than other elements. Thus O isotopes would be expected to robustly record long-lived events while Li isotopes would record more ephemeral fluid interaction. The only other application of Li and O isotopes to metamorphic rocks identified an order of magnitude difference (60m versus 4m) in diffusive exchange across the boundary between a layered mafic complex and a sedimentary diapir. This suggests that pairing O and Li can reveal geochemical nuance not available with a single isotopic system. By constraining fluid behavior within subduction zones on a range of scales we can better understand how it relates to generation of earthquakes and arc volcanism, and improve our models for both natural hazards.