Ground motion and building damage from underground nuclear detonations

Following the Rulison detonation (43 kt at 2570 m), the damage sustained by houses in the communities surrounding that experiment showed a marked correlation with the acceleration spectrum in the .05 to .2 sec period range. This correspondence was used to predict the damage from the Rio Blanco experiment (three 30 kt explosives detonated simultaneously at depths of 1780, 1900, and 2040 meters). Data from Rio Blanco agree substantially with predictions, and have been used to extend the range of the motion-damage correlation. Since Rio Blanco caused damage to only 66 buildings in communities where ground motions were measured (compared to 167 for Rulison), statistics are poor, and the data show considerable scatter. Nevertheless, with the aid of data from explosions at the Nevada Test Site, the damage correlation curve can now be extended to motions equivalent to the limit of perception. The Rio Blanco data also give additional confidence in the ground motion prediction procedure which scales both yield and depth of burial. The method shows that peak accelerations increase with both yield and depth, and response spectra for deeper explosions are enhanced in the damage-causing frequency interval. Motions from Gasbuggy, Rulison, and Rio Blanco--all detonated in sedimentary basins--are essentially consistent with this theory. Conversely, detonations at shallow depths and in more compactible media such as alluvium would cause lower-amplitude accelerations