Effects of Heterogeneous Permeability on Surface Heat Flow Near Parkfield, CA

Surface heat flow near Parkfield, CA exhibits substantial scatter that is not observed in other portions of the California data set, with differences as large 15 mW/m2 over lateral distances of 5-10 km. This scatter has been an important limitation on interpretations of regional heat flow in terms of geodynamic processes, but to date has not been explained. Several processes have been hypothesized to explain the scatter, including conductive refraction, topographic refraction, and advection of heat by groundwater flow. Previous studies designed to investigate the effects of groundwater flow have focused primarily on the role of advection in perturbing a thermal anomaly associated with frictional heat production along the San Andreas Fault (SAF); these studies have considered only simple permeability architectures for the crust, and the results do not readily explain the observed scatter in the data. Notably, in the California Coast Ranges, distinct lithologic units with widely varying permeabilities have been juxtaposed by Mesozoic subduction and displacement along the fault, suggesting that complex groundwater flow paths are likely. Here, we test the hypothesis that heat advection through an upper crust characterized by heterogeneous permeability can generate the magnitude and spatial characteristics of the scatter in the heat flow data set. We created a 2D coupled groundwater flow-heat transport model along a cross-section perpendicular to the SAF, constrained by data from geologic maps, wells, and geophysical surveys, in order to explore and quantify the relationships between topography, permeability, and simulated near-surface heat flow. We assign a constant temperature and atmospheric pressure at the topographic surface; the lateral model boundaries correspond to regional groundwater divides and are designated as no-flow for both heat and fluid flow, and we prescribe a constant heat flux of 78 mW/m2 and a no-flow hydrologic boundary at the model base 10 km beneath sea level. We evaluate the role of the 1-3 km thick package of Tertiary sediments overlying basement by assigning an impermeable basement and considering sediment permeabilities from 10-15 - 10-20 m2. We also consider the effects of the fault zone itself, and model it as a low permeability barrier, a high permeability conduit, and as a combined conduit-barrier. The standard deviation in heat flow ranges from 16 mW/m2 to 0 mW/m2 for sediment permeabilities of 10-15 m2 to 10-20 m2, respectively. This is ~60% lower than the standard deviation predicted by models with homogeneous permeability. Additionally, including the fault zone itself in simulations results in significant localized variations in heat flow. For example, a high permeability damage zone results in heat flow extremes confined to the width of the damage zone, whereas a high permeability damage zone with a 200 m wide central barrier (combined conduit-barrier) reduces these extremes. Based on our results, we suggest that heat advection by groundwater flow mediated by heterogeneous permeability in the upper crust is one plausible mechanism for generating the observed scatter in heat flow.