Evidence of frost-cracking inferred from acoustic emissions in a high-alpine rock-wall

Abstract

Ice formation within rock is known to be an important driver of near-surface frost weathering as well as of rock damage at the depth of several meters, which may play a crucial role for the slow preconditioning of rock fall in steep permafrost areas. This letter reports results from an experiment where acoustic emission monitoring was used to investigate rock damage in a high-alpine rock-wall induced by natural thermal cycling and freezing/thawing. The analysis of the large catalog of events obtained shows (i) robust power-law distributions in the time and energy domains, a footprint of rock micro-fracturing activity induced by stresses arising from thermal variations and associated freezing/thawing of rock; (ii) an increase in AE activity under sub-zero rock-temperatures, suggesting the importance of freezing-induced stresses. AE activity further increases in locations of the rock-wall that are prone to receiving melt water. These results suggest that the framework of further modeling studies (theoretical and numerical) should include damage, elastic interaction and poro-mechanics in order to describe freezing-related stresses.

Introduction

The formation of ice within rock is likely to be an important driver of near-surface frost weathering (Hallet et al., 1991) and rock damage at the depth of several meters (Murton et al., 2006), and in steep terrain, this process may be crucial for the slow preconditioning of rock fall from warming permafrost areas (Matsuoka and Murton, 2008; Gruber and Haeberli, 2007). However, the transfer of corresponding theoretical insight and laboratory evidence to natural conditions characterized by strong spatial and temporal heterogeneity of the rock properties (e.g. fracture state, water content, thermal and hydraulic conductivity) and thermal conditions is nontrivial. To examine rock fracture in natural conditions, we performed a pilot experiment, monitoring acoustic emissions (AE) in a high-altitude rock-face during a 4-day period. In such conditions, the mechanical loading of rock results from the combination of a constant gravity load and fluctuating loads related to (i) thermal stresses, arising from the gradient of the temperature field, (ii) pressure variations in rock pores and cracks, due to water or to ice formation and (iii) short-term external loading such as earthquakes. While large thermal stresses can only occur close to the rock surface, ice formation in pores, cracks and fractures can potentially generate large stresses at greater depths, as suggested by theoretical and lab studies. Reporting a preliminary analysis of the microseismic activity monitored at a high alpine ridge, Amitrano et al. (2010) recently stressed the importance of ice formation in fractures as they observed micro-seismic activity corresponding with particular trends of the temperature that could enhance ice formation. But the lack of details in the spatial and temporal distribution of the seismic events precluded the full understanding of the relationship between temperature evolution and related ice formation at small spatial scale and the triggering of seismic events.

The mechanical loading of rocks involves local inelastic processes that produce elastic wave propagation so-called acoustic emission (AE) at small scales and micro-seismicity (MS) at larger scales. Beside the common physical origin of the elastic wave emission, essentially induced by the propagation or shearing of cracks, these two terms denote differences in the frequency content of the recorded signals corresponding to sources of different size (see Hardy, 2003 for a full presentation). MS relates to the range 1–10³ Hz whereas AE relates to the range 10⁴–10⁶ Hz. The corresponding source size is 1–10³ m for MS and 10⁻³–10⁻¹ m for AE. The material attenuation, which increases with frequency, precludes the detection of AE after approximatively 1 m of wave propagation, whereas MS can be detected at larger distances (up to km).

Measuring AE or MS activity therefore provides a powerful technique to monitor the evolution damage at different scales. Due to their wide frequency range, the simultaneous recording of AE and MS currently is technically not possible. The AE has been extensively used as a tool at the laboratory rock sample scale [e.g. Lockner, 1993] whereas MS has been mostly used at larger scales in order to study seismicity and rockburst in mines, tunnels or quarries (Hardy, 2003). In all these cases, AE/MS are considered to be an indicator of inelastic behavior that can be related to damage increase or to shearing of existing fractures (Cox and Meredith, 1993, Lockner, 1993). Several recent studies report MS monitoring of slope instability (Amitrano et al., 2010, Gaffet et al., 2010). The originality of our study is to apply high-frequency AE monitoring, a technique traditionally used in laboratory experiments, to investigate rock damage during freezing, in field conditions. The main advantage of using high-frequency monitoring is the sensitivity to emissions of relatively small energies. This allows us to obtain a large catalog of events within a short monitoring period, which is crucial to perform statistical analyses. At such high frequencies, acoustic signals are attenuated within about a meter, which determines the spatial scale of our study. This is an advantage as most of the acoustic activity related to freezing can be expected to occur within a meter from the surface. This technique finally offers a high temporal resolution, as event rates up to 10³ per second can be detected.

Fracturing dynamics during mechanical loading usually displays scaling properties in the domains of size, space and time (Alava et al., 2006, Sethna et al., 2001). In the domain of size (magnitude) for example, the seismic events induced by damage processes display a power-law (PL) distribution, $N(s)\sim s^{-b}$, where s is an estimate of the event size (e.g. the maximum amplitude of the AE signal or its energy), N(s) is the probability distribution function (PDF) and b is a constant. This distribution is equivalent to the well-known Gutenberg–Richter relationship observed for earthquakes (Gutenberg and Richter, 1954). Scaling properties in space and time of the events have also been reported, characterizing their spatial and temporal clustering. The emergence of these scaling properties is considered to be a universal feature of the damage dynamics in heterogeneous media (Alava et al., 2006) as it is observed in a very robust manner for various loading conditions, various materials and scales ranging from the micrometer (microcracks) to thousands of kilometers (the Earth's-crust or the sea ice cover). In this letter we report an original in situ experiment of AE monitoring in high altitude thermal conditions. We show that AE activity resulting from natural thermal cycling and induced freezing/thawing shows similar scaling properties, suggesting that the local stress fluctuations encountered are high enough to induce micro-fracturing.