H31P-02:
Self-Organizing Reactive Fluid Escape from Dehydrating Rocks

Wednesday, 17 December 2014: 8:15 AM
Timm John, Free University of Berlin, Berlin, Germany, Oliver Pluemper, Utrecht University, Utrecht, Netherlands, Yuri Podladchikov, University of Lausanne, Lausanne, Switzerland, Johannes C Vrijmoed, ETH Zurich, Zurich, Switzerland and Marco Scambelluri, University of Genova, Italy, Genova, Italy
Abstract:
Water escape from dehydrating rocks within the Earth’s interior is a key process for long-term global water and element cycles, eg. at subduction zones a fluid escape mechanism must exist that prevents ocean water to be drained into the mantle. Existing fluid flow models require a priori physical assumptions (eg. preexisting porosity) and cannot resolve the evolution from initial fluid production to flow channelization. In order to develop a model of this evolution, we need to unravel natural laboratories that display the incipient dehydration stages and the micro- to macro-scale fluid escape route evolution. The Erro-Tobbio meta-serpentinites (Italy) provide a unique snapshot into these early dehydration stages, recording the breakdown of hydrous antigorite to anhydrous olivine plus fluid and the formation of an olivine-vein network. We find that dehydration, fluid pooling, and flow initiation are controlled by micro-scale compositional rock differences. Our model starts with a rock in which all water is stored in solid and any preexisting porosity is negligible (zero-porosity case). As the rock descents into the mantle increasing T will initiate dehydration reactions, dividing the rock continuously into a dry solid and a fluid-filled porosity. Spatially variable reaction progress results in dynamically evolving porosity/permeability and heterogeneous fluid-pore pressure distributions. Fluid-pressure gradient relaxation causes fluid flow and its thermodynamic feedback triggers reactions to progress, resulting in a self-amplifying process.

Our new thermodynamic-mechanical model for reaction-porosity waves shows that fluid flow occurs solely in the reaction products and self-organizes into channelized fluid escape networks. This holds the key to formulating future quantitative models that address spatiotemporal processes such as the coupling between fluid release at depth and volcanic eruptions and the amounts of structurally bound water transferred into deep Earth.