Thus, predictions of ground shaking intensities are not based on distances from possible epicenters, but on distances from known faults, or segments of faults, on which large earthquakes are anticipated.

Intensity decreases ("attenuates") with distance from the fault. (See Ref. 28.) But the critical distance is not simply the nearest distance to the fault. Seismologists have come to realize that earthquake sources radiate energy at depth; thus, the distance used to attenuate expected shaking must be measured between the site and this underground source. (See Refs. 25, 33, 34, 35, and 38.) However, rupture propagates both upward from this underground source and along the fault axis. (This "directivity" effect is described in the next paragraph.) Thus, there is significant amplification of shaking within a mile of these major fault zones.

The final factor affecting the change of intensity with distance from the fault is the magnitude of the earthquake. The intensity boundaries extend further from the fault source for larger magnitude earthquakes. Thus, a site 20 miles from the fault source will experience stronger and longer shaking from an earthquake with a moment magnitude of 7 than from an earthquake with a moment magnitude of 6. Even though the energy released in an earthquake is over thirty times as great in a magnitude 7 quake than a magnitude 6 quake, the shaking is not 30 times as intense. Rather, a larger area is exposed to strong shaking.

Note: One additional factor in the recent Loma Prieta earthquake was the reflection of the seismic waves from the Moho. (The "Moho" is short for the Mohorovicic discontinuity, the boundary between the earth's crust and mantle, and is named for the Croatian scientist who discovered it.) This "bounce" resulted in stronger shaking which ranged from 45 to 60 miles from the fault trace and amounted to somewhere between one-half and one intensity increment level increase over what might have been expected. (See, for example, Ref. 45.) Both Oakland and San Francisco were within this distance band. However, there are insufficient data to reliably calculate such increases for future earthquakes. Because the Loma Prieta earthquake began deeper than is typical for Bay Area earthquakes, this Moho-related increase was probably closer to the fault source than would be expected in future Bay Area earthquakes. Thus, the increase, if it occurs, will be in areas with lower baseline shaking levels and should result in small or insignificant increases in damage.