would be capable of making those changes. Successful seismic and geodetic inversions may construct remarkable pictures of the fault rupture surface at depth and the associated complexities in both fault geometry and the dynamic rupture process. Powerful examples of the application of inversion techniques can be found in the studies of recent earthquakes at Imperial Valley and Coalinga, California, and the earthquake at Borah Peak, Idaho.
During the past 20 years efforts to obtain records of strong ground motion have increased, and the data bank has grown with each successive earthquake. The data have been recorded, for the most part, on networks of relatively simple accelerometers with minimal magnification capabilities. These records have been invaluable to the engineering community in designing structures to withstand earthquakes. Within the past few years a new generation of instruments has been developed, and they are now being deployed to supplement the older instruments (Figure 5.14). This new generation of instruments can digitally record a wide spectrum of frequencies over a wide dynamic range. The new data so gathered permit the analysis not only of strong ground motion, as did the older generation of instruments, but also details of the faulting process. Such instruments, and the data recorded by them, are revolutionizing our understanding of rupture mechanics.
Theoretical models have been developed for predicting the shapes of curves that represent earthquake ground-motion intensity, called seismic wave spectra. Although there is still some controversy over the physical interpretation of the models, different investigators have made predictions that are in general agreement with observed records, using data recorded in regions such as western North America where substantial numbers of strong-motion records are available. With that agreement as encouragement, theoretical models