Earthquake History of the Northern Imperial Fault, Imperial Valley, California, since the last Lake Cahuilla Highstand, circa A.D. 1680

The Imperial fault (IF) is the only fault in southern California to have ruptured in two major earthquakes in the 20th century. In 1940, it ruptured end-to-end (both north and south of the international border) in an M 6.9 earthquake, and in 1979, the northern segment of the fault (north of the border) ruptured again in an M 6.4 event. Slip in 1940 was highest (5-6 m) along the central portion of the fault and lowest (<1 m) along the northern portion, with a high slip gradient between these two segments just north of the border. The 1979 earthquake involved surface rupture along only the northern 30 km of the fault, with dextral offsets being <1 m and being nearly identical to 1940 offsets along the northern 20 km of the rupture. The similarities and differences of the two events led Sieh (1996) to propose a "slip-patch model" for the Imperial fault, whereby the fault ruptures with frequent moderate earthquakes along its northern end, like in 1979, and with less frequent larger events like 1940 along its entire length. According to the model, the central patch, which experienced high slip in 1940 and did not rupture in 1979, would rupture with relatively infrequent events (roughly every 260 years) with typically 5-6 m of slip per event; meanwhile, the northern patch, which corresponds to the 1979 rupture, would rupture more frequently (roughly every 40 years) with up to 1 m of slip per event. This model is consistent with the slip distribution observed in 1940 and in 1979. Paleoseismic investigations along the central patch also support this model, as the penultimate event there occurred shortly after the last Lake Cahuilla (LC) highstand at around A.D. 1680 (Thomas and Rockwell, 1996). Prior to the present investigation, however, there were no data on events prior to 1940 on the northern patch, which could serve to either support or refute the slip-patch model. We have opened a trench across the IF south of Harris Road, adjacent to Mesquite Basin, where the fault has both dextral and normal slip components. On the downdropped side of the fault, a laminated clay unit (inferred to be the most recent LC clay, at ca. A.D. 1680; dating results are pending) is exposed in the trench and is overlain by nearly 2 m of younger deposits; the overlying material consists of bedded fine sands and silts (inferred to be overbank deposits from a nearby channel), which interfinger with massive silts and clays (inferred mostly to be colluvium). Unfortunately, the normal component of slip for all earthquakes in the trench was almost exclusively restricted to a single east-dipping plane or set of closely spaced planes, so that the amount of dip slip per event cannot be resolved; nonetheless, a series of fissures and flower structures adjacent to the main fault in the hangingwall block permit the distinction of individual events. There is good evidence for 4, and possibly 5 events since the last LC highstand, based on filled-in fissures and abrupt upward terminations of multiple fault strands and liquefaction cracks. The youngest of these events are inferred to be 1979 and 1940; the oldest, which produced liquefaction and ruptured to the top of the most recent lake deposits, probably occurred at or very soon after the highstand, based on the arguments that no lake deposits post-date the event, and that the ground was still saturated at the time of the earthquake. This event may have been the penultimate (ca. A.D. 1680) event seen on the central patch of the IF.