Testing natural fracture growth-fracturing resilience feedbacks in rock

The growth of rock fractures enables physical and chemical rock erosion, sets the pace for infrastructure-, rockfall- and landslide-hazards, and influences rock hydrologic, and therefore chemical, processes. Although fracture growth ultimately lowers rock strength, when rocks are subject to stresses lower than that magnitude, laboratory experiments indicate that the growth of fractures can counterintuitively make rock more resilient to subsequent fracture growth through ‘stress memory’ or ‘fatigue-limit’ phenomena such as the Kaiser effect.  Thus, over geologic time scales, all other things being equal, fracturing rates may decrease, which would have important implications for understanding and interpreting a wide range of landscape evolution processes. To date, however, there have been few if any data explicitly showing that fracture-resilience feedback phenomena arise naturally in subaerially exposed rock.

Here we test for a natural stress memory in two ~25 cm diameter boulders for which we have 1-4 years of known environmental exposure history.  The granite boulders were collected from an unvegetated bar in an ephemeral channel issuing from the south flank of the San Bernardino Mountains, California. As such, we infer that natural abrasion in the channel had removed any major cracks or heterogeneities, effectively ‘resetting’ the rock to a relatively pristine state. The rocks were left on the ground in full sun exposure for 1 and 3 years respectively in humid temperate North Carolina and semi-arid temperate New Mexico, USA. Per-minute rock surface and environmental conditions and cracking (using acoustic emissions) were monitored. Prior work (Eppes et al., 2016 & 2020) indicates that thermal stresses were the primary driver of cracking in the rocks during these time periods. The boulders were then cut in half, and 20x40mm cores were collected from various locations within the rock interior, in duplicate and triplicate for locations of varying distance to the rock exterior. We measured core porosity and P-wave velocity in the cores as proxies for initial rock crack composition, as well as thermal conductivity. We then subjected sets of cores collected at different distances from the rock exterior to increasing magnitudes and number of thermal stress cycles in a temperature-monitored oven, beginning with our best approximation of those matching the maximum stresses leading to observed cracking during the 1 and 3 year observation periods. Our preliminary results reveal that initial crack characteristics vary as a function of distance from rock exterior, as might be expected due to the different magnitudes of thermal stresses experienced within these locations within the rock. Thus, we hypothesize that areas starting with the highest porosities and lowest velocities will experience less change following heating cycles than those parts of the rock with few inferred fractures. We hope that these data will help elucidate mechanisms and feedbacks of natural rock fracturing phenomena that occur over geologic time scales.