54 © The Containment of Underground Nuclear Explosions re-entered after the test. Such measurements suggest that the explosion only affects rock strength to a distance from the shot point to about three cavity radii (165 (yield) ”). The second observation, obtained from seismic measurements of tectonic release, suggests a larger radius for the volume of rock affected by an explosion. The seismic signals from underground nuclear explosions frequently contain signals created by what is called ‘‘tectonic release.”’ By fracturing the rock, the explosion releases any preexisting natural stress that was locked within the rock. The release of the stress is similar to a smail earthquake. The tectonic release observed in the seismic recordings of underground explosions from Rainier Mesa indicate the loss of strength in a volumeof rock with a minimum radius equal to 500 (yield). Although the drill samples and the seismic data appear to contradict each other, the following explanation appears to accountfor both: The force of the explosion creates a cavity and fractures rock out to the distance of 2 cavity radii from the shot point. Out to 3 cavity radii, existing cracks are extended and connected, resulting in a decrease in seismic shear velocity. Outside 3 cavity radii, no new cracks form. At this distance, existing cracks are opened and strength is reduced, but only temporarily. The open cracks close immediately after the shock wave passes due to the pressure exerted by the overlying rock. Because the cracks close and no new cracksare formed, the rock properties are not changed. Postshot tests of seismic shear velocity and strength are the same as pre-shot measurements. This is consistent with both the observations of surface fractures and the slight disturbances seen along bedding planes at distances greater than 3 cavity radii. The surface fractures are due to surface spall, which would indicate that the rock was overloaded by the shock wave. The disturbances of the bedding planes wouldindicate that fractures are being opened out to greater distances than 3 cavity radii. In fact, the bedding plane disturbancesare seen outto a distance of 600 (yield) *, whichis consistent with the radius determined from tectonic release. The large radius of weak rock derived from tectonic release measurements represents the transient weakening from the shot. The small radius of weak rock derived from the post-shot tests represents the volume where the rock properties have been permanently changed. From the point of view of the integrity of the tunnel system, it is the smaller area where the rock properties have been perma- nently changed (radius = 165 (yield) ”) that should be considered for containment. Becausetheline-ofsight tunnel is located so that the stemming plug region and closures are outside the region of permanently weakened orfractured material, the closure system is not degraded. HOWSAFEIS SAFE ENOUGH? Every nuclear test is designed to be contained and is reviewed for containment. In each step of the test procedurethere is built-in redundancy and conservatism. Every attempt is made to keep the chance of containment failure as remote as possible. This conservatism and redundancyis essential, however: because no matter how perfect the process may be, it operates in an imperfect setting. For each test, the containmentanalysis is based on samples, estimates, and models that can only simplify and (at best) approximate the real complexities of the Earth. As a result, predictions about containment depend largely on judgments developed from past expenence. Most of what is known to cause problems—carbonate material, water, faults, scarps, clays. etc.—was learned through experience. To withstand the consequences of a possible surprise, redundancy and conservatism is a requirement not an extravagance. Consequently, ail efforts undertaken to ensurea safe testing program are necessary, and they must continue to be vigorously pursued. Deciding whether the testing program 1s safe requires a judgementof howsafe is safe enough. The subjective nature of this judgernent is illustrated through the decision-making process of the CEP. which reviews and assesses the containment of each test.>9 They evaluate whethera test will be contained using the categorizations of ‘*high confidence,” **adequate degree ofconfidence,’’ and ‘some doubt.”’ But, the CEP has no guidelines that attempt to quantify or describe in probabilistic terms what constitutes for example, an “‘adequate degree of confidence.’’ Obviously one can never have 100 percent confidence that a test will not release radioactive material. Whether ‘‘adequate confi- 39The ContainmentEvaluation Panel is a group of representatives from various laboratories and technical consulting organizauions who evaluate the proposed containmentpian for each test without regard to cost or other outside considerations (see ch. 2 for a complete discussion).