VI. CONCLUSIONS

Mining explosions detonated at or near the earth's surface provide a source of events that will have to be identified during the monitoring of a CTBT. In an attempt to understand the seismic waveforms generated by these sources, a series of techniques that combine multiple data sets and models of the blasting practices have been developed. These same tools may be useful to the blasting community as diagnostics of blasting practices.

Ground motion records, video, timing and spatial data have been used to constrain critical physical processes in the blasts. The first step in the synergy of these different data sets and combination with models of the processes for interpretation is conversion of all data sets to digital form. A series of processing steps have been developed for converting standard Hi-8 video to digital form. Important in this process is consideration of the large size of the resulting data sets, development of maximum time resolution, and removal of artifacts of the explosive environment such as the ground motion effects on the camera. De-interlacing of the video data after digitization provides 16.67 ms resolution of near surface source processes with standard Hi-8 video hardware.

Combined images that include video, ground motion and models representing timing information from the source provide the opportunity for improved interpretation of the blasting process and its effect on the ground motion. The video and the models allow inclusion of two and three dimensional spatial effects in the interpretation. The time varying nature of the problem is replicated by creating movies of these visualizations .

Examples displayed in this paper demonstrate a new technique for data interpretation as well as provide constraints on the seismic source functions from mining explosions. Data from the explosive array indicate that, as a result of downhole delays, significant differences between design and measured detonation times can occur. In cases where all explosions in a single row are detonated with the same delay as in the Russian example (Figure 10 and 11), this variation may not destroy the constructive and destructive interference introduced by the energy from each row of explosions. In cases where each hole in the blast is detonated at a different time (Figure 15 and 16), the variance in the detonation times can result in important degradation in the spectral scalloping that might be used to identify such sources. Even in the case where the scatter results in quite different individual detonation times, the total duration of source processes remains relatively unchanged and thus spectral interference representative of the total duration of the explosive array may still provide a useful discriminant for this event type.

Comparison of the videos of the explosive processes with ground motion records indicates that within the bandwidth of the observations (2 Hz seismometers) that secondary source processes, such as material cast or spalled, have little contribution to near-source waveforms. The correlation in time of the video and ground motion records supports this conclusion.

The time varying aspects of this problem, coupled with digital data quantifying this aspect of the phenomenology, has motivated this exploration of new techniques in data display and comparison to models. The World Wide Web make possible easy dissemination of these interpretative products in conjunction with a standard journal article.

ACKNOWLEDGMENTS

This work was made possible by the Department of Energy and the CTBT Research and Development Program at Los Alamos National Laboratory. Initial support for the video image processing was provided by AFOSR under Grant F49620-93-1-0146 at SMU. John Smith, John Wiggins and R. Frank Chiappetta are thanked for their support in the field. D. Craig Pearson, Meredith Ness and Ben Smith were responsible for the data acquisition. Xiaoning Yang helped with preparation of some of the ground motion data.

Go to Table of Contents or Next Section