Rupture of the large (M (sub w) 7.8) deep earthquake of 1973 beneath the Japan Sea with implications for seismogenesis

The earthquake of 29 September 1973 beneath the western edge of Japan Sea is the largest deep earthquake along the western Pacific (M (sub w) 7.8, focal depth 575 km). The source region is near the tip of the Japan Wadati-Benioff zone, where a continuous belt of seismicity clearly defines a unique, shallow-dipping ( approximately 30 degrees ) slab down to depths about 600 km. This unique geometry of slab predicts a limited vertical extent available for earthquake rupture, in direct conflict with previous reports of a large, subvertical fault for the 1973 earthquake. Using a comprehensive data set of nearly 20 P and SH waveforms recorded at teleseismic distances, we investigate rupture associated with this event in some detail by inversion of waveforms. In particular, we use wave trains with all relevant depth phases (pP, sP, and sS) that are crucial for resolving the vertical extent of rupture. In addition to a northward component of rupture as in previous reports, we find clear evidence for a significant component of eastward rupture propagation that cannot fit within a subvertical, north-south striking nodal plane. Meanwhile, all regions of major slip are at essentially the same depth. Thus the true fault plane appears subhorizontal, with the entire region of slip extending only about 15+ or -6 km across the slab's thickness. The spatio-temporal distribution of slip is well explained by a simple, circular rupture front, propagating at a speed of 3-4 km/sec. A comparison among the largest deep earthquakes shows that most rupture planes are more or less dimensionally equal, with high aspect ratios ranging approximately from 0.5 to 1. While some events have fairly restricted source regions, others seem to extend more than 30 km into the thickness of the slabs.