Evenly-spaced columns in the Bishop Tuff as relicts of hydrothermal convection

Abstract:

A few square km of the Bishop Tuff in eastern California, USA have evenly spaced erosional columns. These columns are more resistant to erosion due to the precipitation of the low-temperature zeolite (120-200 ºC), mordenite, which is not found in the surrounding tuff. Similar features observed in the Bandelier Tuff were hypothesized to form when cold water from above infiltrated into the still-hot tuff interior. This water would become gravitationally unstable and produced convection with steam upwellings and liquid water downwellings. These downwellings became cemented with mordenite while the upwellings were too dry for chemical reactions. We use two methods to quantitatively assess this hypothesis. First, scaling that ignores the effects of latent heat and mineral precipitation suggests the Rayleigh number (Ra, a measure of convective vigor) for this system is ~10³ well above the critical Ra of 4π². Second, to account for the effect of multiphase flow and latent heat, we use two-dimensional numerical models in the finite difference code HYDROTHERM. We find that the geometry of flow is consistent with field observations and confirm that geometry is sensitive to permeability and topography. These tests suggest a few things about low-pressure hydrothermal systems. 1) The geometry of at least some convection appears to be broadly captured by linear stability theory that ignores reactive transport, heterogeneity of host rock, and the effects of latent heat. 2) Topographic flow sets the wavelength of convection meaning that these columns formed somewhere without topography—probably a lake. Finally, these observations imply a wet paleoclimate in the Eastern Sierra namely that, in the aftermath of the Long Valley eruption, either rain or snow was able to pool in the caldera before the tuff cooled on the order of a hundred years after the eruption.