Hydrothermal convection and mordenite precipitation in the cooling Bishop Tuff, California, USA

We present field observations of erosional columns in the Bishop Tuff and then use laboratory results and numerical models to argue that these columns are evidence of relict convection in a cooling ignimbrite. Many square kilometers of the Bishop Tuff have evenly-spaced, vertical to semi-vertical erosional columns, a result of hydrothermal alteration. These altered regions are more competent than the surrounding tuff, are 0.1-0.7 m in diameter, are separated by ~ 1 m, and in some cases are more than 8 m in height. JE Bailey (U. of Hawaii, dissertation, 2005) suggested that similar columns in the Bandelier Tuff were formed when slumping allowed water to pool at the surface of the still-cooling ignimbrite. As water percolated downward it boiled generating evenly spaced convection cells similar to heat pipes. We quantify this conceptual model and apply it the Bishop Tuff to understand the physics within ignimbrite-borne hydrothermal systems. We use thin sections to measure changing porosity and use scanning electron microscope (SEM) and x-ray diffraction (XRD) analyses to show that pore spaces in the columns are cemented by the mineral mordenite, a low temperature zeolite that precipitates between 120-200 oC (Bish et al., 1982), also found in the Bandelier Tuff example. We then use scaling to show 1) that water percolating into the cooling Bishop Tuff would convect and 2) that the geometry and spacing of the columns is predicted by the ignimbrite temperature and permeability. We use the computer program HYDROTHERM (Hayba and Ingebritsen, 1994; Kipp et al., 2008) to model 2-phase convection in the Bishop Tuff. By systematically changing permeability, initial temperature, and topography we can identify the pattern of flows that develop when the ignimbrite is cooled by water from above. Hydrothermally altered columns in ignimbrite are the natural product of coupled heat, mass, and chemical transport and have similarities to other geothermal systems, economic ore deposits, and mid-ocean ridge hydrothermal systems. The columns allow direct observation to constrain complex models of multiphase convection, reactive transport, and permeability. Our results also have paleoclimate implications, implying a large and stable source of water in the SE/SSE Long Valley area immediately after the ~760,000 ka caldera-forming eruption.