Environmental Impact of Controlled-Source Explosions in Ethiopia (Project EAGLE): Surface Shaking, Ground Velocities, and Effects on Buildings

In January 2003, as part of the Ethiopia-Afar Geoscientific Lithospheric Experiment (EAGLE) we conducted a refraction and wide-angle reflection survey of the Main Ethiopian Rift. 757 RefTek "Texan" seismographs with vertical geophones were deployed in 400 km-long axial and cross-rift lines, with another 231 in a central 3D array 100 km in diameter. An 80-instrument passive array of intermediate and broadband sensors was active during our experiment. We recorded 19 borehole shots loaded in nominal 50-meter boreholes, 2 quarry shots, and 2 lake shots. The shots ranged in size from 50-5750 kg, with the most common shot size being 1 tonne. Prior to loading each shot-hole, we measured distances between shots and the nearest structure, typically un-reinforced mud-and-wood houses, occasionally concrete irrigation ditches and aqueducts. We then used semi-empirical formulae derived by Oriard (Hendron and Oriard, 1972) to calculate expected maximum and minimum bounds on ground velocity at these structures, and selected an appropriate shot size to keep the predicted velocity below the "threshold for cosmetic damage", or 2 inches per second, at the most vulnerable structure. The Oriard formulae are derived from measurements associated with blasting for mining and civil engineering purposes and may not accurately predict the ground velocity from the source depths and explosive type used in the EAGLE and other controlled-source experiments. A detailed, trace-by-trace analysis of maximum ground velocities at our closest seismographs can provide data that will be useful in planning future large-scale seismic experiments. Preliminary results from traces within 20 km of our borehole shots suggest that maximum recorded ground velocities were within or below the maximum-minimum range predicted by Oriard, and hence that larger shot sizes could have been used with acceptable risks. A lake shot fired at the optimum depth (84 m for a 1 tonne shot) produced ground velocities that exceeded the predicted maximum at a few recodrers. However, optimum-depth shots are typically a significant distance offshore (c. 2.3 km for our shot) because of the required depth, so are unlikely to present a hazard to onshore structures. A lake shot fired in a shallower lake at half the optimum depth did not produce ground-velocities that exceed the Oriard maximum. Although we fired shots within 100 m of an unreinforced concrete aqueduct, and within 200 m of poorly engineered native buildings in poor structural condition, no damage was recorded. Our "Texan" seismometers recorded only vertical component velocity, using 4.5 Hz geophones. After removal of the geophone response the peak vertical velocity is typically measured at about 3 Hz and occurs shortly after the first arrival, presumably due to surface waves (ground roll). We are currently extending our analysis to include data from broadband, three-component recorders.