The role of rock fracture in erosion

Before rock can be eroded, intact bedrock must be broken into smaller pieces that can be detached from the landscape. This detachment process can occur at the scale of grains, thereby promoting grain-by-grain attrition from a rock surface, or at scales that enable landslides. In all cases, fractures serve to decrease overall rock strength by reducing cohesion and potentially by lessening the effective angle of internal friction (Terzaghi, 1962). The creation of fractures results from two major classes of processes: geomorphic and tectonic. Geomorphic processes encompass the broad suite of physical, chemical, and biotic processes that serve to break down rock masses. In almost all cases, these processes are most intense at the surface, and they diminish in effectiveness with depth, typically in some poorly known manner (Fig. 1). Tectonic fracturing results from myriad stresses within tectonic plates, but in actively deforming landscapes, the most likely cause of fracturing is the transport of rocks above irregular fault surfaces. Any kink or change in dip of a fault surface causes a concentration of stresses in the rocks of the hangingwall as they are moved and folded above it (Molnar et al., 2007). The spatial cloud of aftershocks that follows coseismic slip along a fault plane testifies to the dispersed fracturing that commonly occurs as accumulated stress is released in the hangingwall.

Whereas recognition of these fracturing processes and their importance is not new, quantification of their magnitude has been a persistent challenge. How deep is the geomorphically fractured layer? How does fracture density vary with depth or with rock type? Within any mountain range, how diverse is the degree of tectonic fracturing? How does the degree of fracturing (or lack thereof) influence erosion processes? Our inability to see into the shallow subsurface restricts our insights on these questions.

In the early 1980s, the Japanese geomorphologist Suzuki (1982) proposed that seismic velocities could be used as a proxy for local rock strength. More recently, new insights on fracturing of the upper 10 to 20 meters of the rock column have been derived from backpack-able portable seismic arrays (Fig. 2). First, strings of geophones are evenly spaced in 1- to 5-m intervals along linear arrays that stretch for 30 to 60 meters across bedrock surfaces. These arrays are then used to record the first arrival times and the velocity of compressional seismic waves (p-waves) that are created by hitting a strike plate with a sledge hammer at one end of the string of geophones.

In comparison to the velocity of compressional waves in intact bedrock, such as velocities measured on solid rock samples in a lab, the reduction of p-wave velocities in the subsurface of field sites can be attributed to fracturing that impedes transmission of seismic waves. Time-distance plots of the first p-wave arrivals at each geophone can, therefore, reveal key aspects of the pattern of subsurface fracturing. When fracture densities are uniform with depth (consistent with tectonic fracturing), time-distance plots of p-wave arrivals are linear and velocity is uniform with depth (Fig. 3A). In contrast, fracture densities that are greater near the surface and decrease with depth (consistent with geomorphic fracturing), produce slow seismic velocities at that surface that increase with depth. This gradient in velocity results in first p-wave arrivals in time-distance plots that define a curved line (Fig. 3B).

Recent studies on the South Island of New Zealand (Clarke and Burbank, in review) reveal both linear and curved p-wave time-distance plots that correspond with tectonic and geomorphic fracturing models, respectively. Most sites show two layers: an upper geomorphically fractured layer overlying a uniformly fractured lower layer (Fig. 3C). In New Zealand's Fiordland, the base of this geomorphic zone ranges from 2 to 16 meters, but averages 7 meters (Fig. 4). Where this geomorphically fractured layer is absent, apparently due to its removal by erosive processes, the fracture density is nearly uniform with depth and is attributed to tectonic fracturing. A provocative feature of these data is that the geomorphically fractured upper layer tends to be present only when the rock beneath it has been only weakly fractured by tectonic processes. Where tectonic fracturing is more intense (Fig. 4), geomorphic fracturing is commonly not seen. Such highly fractured rocks are interpreted to be sufficiently weak that they fail by bedrock landslides with such frequency that development of a geomorphically fractured layer at the surface is inhibited.

These observations underpin a simple conceptual model (Fig. 5) in which geomorphic fracturing propagates downward into a rock column through time, thereby increasing fracture density and reducing rock strength. Such fracturing should affect the erosional efficacy of many surface processes. In sites where landsliding is a major contributor to overall denudation, this fracturing can weaken formerly strong bedrock to the point of failure, such that most landslides are shallow and occur within this geomorphically weakened zone. In contrast, where the underlying bedrock has already been considerably weakened by pervasive tectonic fracturing, landslides of any depth can occur, and the formation of a highly fractured, geomorphic surface layer is uncommon.

• Clarke, B., and Burbank, D.W., in review, Quantifying bedrock fracture density in the shallow subsurface: Implications for bedrock landslides and erodability: Journal of Geophysical Research.
• Molnar, P., Anderson, R. S., and Anderson, S. P., 2007, Tectonics, fracturing of rock, and erosion: Journal of Geophysical Research, v. 112, p. F03014, doi:10.1029/2005JF000433.
• Suzuki, T., 1982, Rate of lateral planation by Iwaki River, Japan: Transactions of the Japanese Geomorphological Union, v. 3, p. 1-24.
• Terzaghi, 1962, Stability of steep slopes on hard unweathered rock: Geotechnique, no. 12, p. 51-270.