The long-term (decade or longer) seismic potential of the Peninsular San Andreas fault, the 125-km-long segment southeast of San Francisco, has been assessed by comparing the slip accompanying the 1906 Mw=7.7 earthquake with the amount of this slip deficit made up since 1906. Of particular importance is the longstanding factor-of-2 inconsistency between 1906 surface offsets (≤1.5 m) and geodetically determined coseismic slip (2.7 m) on the southern 90 km of the Peninsular San Andreas. In reexamining this issue we find that the discrepancy cannot be attributed to either large observational uncertainties in the geodetic data or shortcomings in fault modeling. Rather, the problem lies in the complexity of this 90-km-long fault reach, where the San Andreas trend changes by 6¿, the zone of recent deformation is 0.5--2 km wide, and much of the fault is heavily vegetated and inaccessible. Farther northwest on the 1906 rupture, where the fault geometry is much simpler, evidence from deformed fences shows that right-lateral coseismic slip was commonly distributed across a zone ~20-600 m wide, with only about 70% of the total slip occurring at the main fault trace.

The general importance of distributed coseismic deformation and the evidence for fault zone complexity indicate that the best estimate of 1906 slip on this 90-km fault segment is the geodetically constrained value of 2.6¿0.2 m. Farther northwest on the Peninsular San Andreas the coseismic slip ranges from 3.4 to 4.4 m, and geodetic slip and surface offsets show considerably better agreement. Following previous work, we have assessed the risk posed to the southern Peninsular San Andreas by two broad classes of earthquakes. These appraisals use the geodetic slip distribution, apply a long-term slip rate of 15 mm/yr suggested by geologic and geodetic studies, and assume the approximate correctness of the time-predictable recurrence model. The possibility of an M≂6 1/2 earthquake on the 30-km-long segment northwest of San Juan Bautista cannot be completely precluded, but the evidence bearing on its likelihood is contradictory. In contrast with several previous assessments we find that the risk of an earthquake of M≂7 rupturing the southern ~90 km of the Peninsular San Andreas during the next several decades is quite low. ¿American Geophysical Union 1987