An alluvial succession on the northeast side of the San Bernardino strand of the San Andreas fault includes distinctive aggradational and degradational features that can be matched with correlative features on the southwest side of the fault. Key among these are (1) a terrace riser on the northeast side of the fault that correlates with an offset channel wall on the southwest side of the fault and forms a basis for slip estimates for the period ca. 35 ka to the present, and (2) a small alluvial fan on the southwest side of the fault that has been matched with its most likely source gullies on the northeast side of the fault and forms a basis for slip estimates for the last 10.5 k.y. Slip-rate estimates for these two separate intervals are nearly identical. The rate for the older feature is most likely between 8.3 and 14.5 mm/yr, with a 95% confidence interval of 7.0–15.7 mm/yr. The rate for the younger feature is most likely between 6.8 and 16.3 mm/yr, with a 95% confidence interval of 6.3–18.5 mm/yr.
These rates are only half the previously published slip rate for the San Andreas fault 35 km to the northwest in Cajon Pass, a rate that traditionally is extrapolated southeastward along the San Bernardino section of the fault. Results from Plunge Creek suggest that about half of the 25 mm/yr rate at Cajon Pass transfers southeastward to the San Jacinto fault, as proposed by other workers on the basis of regional geologic relations. These results indicate that the discrepancy between latest Quaternary slip rates and present-day rates of strain accumulation across the San Bernardino section of the San Andreas fault from geodesy can be largely explained by slip transfer between faults, leading to spatial variation in rate along the San Andreas fault. Nonetheless, the latest Pleistocene and Holocene slip rate at Plunge Creek is still somewhat faster than rates inferred for the San Bernardino section of the San Andreas fault based on elastic block modeling of geodetic data and may be more appropriate than those rates for hazard estimation.