Dynamic fracturing: ®eld and experimental observations Amir Sagy * , Ze'ev Reches, Itzhak Roman Institute of Earth Sciences, and School of Applied Sciences, Hebrew University, Jerusalem 91904, Israel Received 28 January 2000; accepted 7 December 2000 Abstract We analyzed a system of complex joints in thick dolomite layers that are exposed within the western margins of the Dead Sea basin. These joints display two dominant features: `tree-like' branching and a gradual increase of density that leads to local fragmentation. The development of this joint system is investigated in laboratory experiments with samples of brittle/ductile layered composites. The samples were subjected to layer-parallel extension and displayed three styles of fracturing: planar fractures, known from previous tests; branching fractures and clustering fractures, observed here for the ®rst time in layered composites. Based on fracture morphology, we deduced that the branching and clustering fractures in the experiments, and the tree-like, closely spaced joints in the ®eld, propagated at dynamic, high- velocity growth rates. It is proposed that the morphological features described here could be used as ®eld criteria to recognize dynamic rates of rock fracturing. q 2001 Elsevier Science Ltd. All rights reserved. Keywords: Fractures; Rock mechanics; Joints; Dynamic; Experimental; Tensile 1. Introduction Countless fractures appear in the earth's crust. Many of them grow unstably and at high velocities, as evidenced by earthquakes and rockbursts. However, there is no systematic method to deduce the growth velocity of a fracture from its appearance in the ®eld. Engelder and Fischer 1996) stated that ªEven though propagation velocity is so important in identifying the joint-driving mechanism, little is known about natural joint velocityº. Knowing the growth velocity of fractures is also important for applied problems such as seismic hazard evaluation, rockburst risk and the evaluation of fracture density in layered rocks. Fast growth of fractures is termed `dynamic fracturing', a process in which tensile fractures grow at velocities comparable with Rayleigh wave velocity of the host material Freund, 1990; Ravi-Chandar, 1998; Fineberg and Marder, 1999). In this study, we propose ®eld criteria for dynamic fracturing of rocks and relate these criteria to new experimental results reported here. Estimating joint propagation velocity in the ®eld is not a simple task. One may deduce from explosive experiments that rock fractures associated with catastrophic events, like earthquakes, impacts or rockbursts, are dynamic fractures. Another approach is based on the experimental observations that slow growing fractures are smooth and simple, whereas dynamic ones are rough and complex e.g. Bahat, 1991; Lawn, 1993; Sharon and Fineberg, 1996). Using this approach, some fractographic features associated are suspected as indicating fast propagation. For example, Wallner lines and velocity hackles could indicate high propagation velocity Kulander et al., 1979, pp. 26±34). The ®rst part of the present analysis describes the detailed geometry of joints within an exposure of dolomite layers close to the Dead Sea basin. The second part presents a series of experiments with brittle±ductile layered samples that were subjected to extension, and in which new fractur- ing styles have been observed. Finally, we suggest that the studied ®eld joints grew at dynamic velocities and discuss the relations between dynamic fracturing and joint density. One should note that the term `joint' is used here for an extension fracture observed in the ®eld, whereas the more general term `fracture' is used for a tensile or shear) frac- ture observed in laboratory experiments. 2. Dynamic fracturing and fracture morphology During slow, quasi-static propagation of fractures, the input of mechanical energy, U M , is suf®cient only for generation of new surface area of the growing fracture, U M 1 U S  0 where U S is the new surface energy. During fast, dynamic fracture propagation, the input of mechanical energy, U M , Journal of Structural Geology 23 2001) 1223±1239 0191-8141/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved. PII: S0191-814100)00190-5 www.elsevier.nl/locate/jstrugeo * Corresponding author. Fax: 1972-2566-2581. E-mail address: reches@earth.es.huji.ac.il A. Sagy).