Bulletin of the Seismological Society of America. Vol.62, No. 3, pp. 721-750. June, 1972 STRESS ESTIMATES FOR THE SAN FERNANDO, CALIFORNIA, EARTHQUAKE OF FEBRUARY 9, 1971: MAIN EVENT AND THIRTEEN AFTERSHOCKS BY M. D. TRIFUNAC ABSTRACT The strong earthquake ground motion recorded in the center of and above the fault plane is combined with field evidence of faulting and instrumental studies of aftershocks to deduce stresses during and after the San Fernando earthquake of February 9, 1971. Stress computations based on Brune's near-field, shear-wave spectra, peak velocity of ground motion, energy calculated from the strong- motion record, and a model of circular dislocation give mutually consistent stress estimates, which suggest that the effective stress operating during the earthquake was approximately 100 bars, while during the earthquake it dropped several tens of bars. The energy of the main event is estimated to be 1022 dyne cm. Thirteen aftershocks, recorded during the first 6 rain, were associated with stress drops ranging from 10 to 500 bars, these events clustering along the north- eastern end of the dislocation surface. The strong-motion accelerograms provide invaluable data for detailed investiga- tions of the pattern of earthquake energy release during and immediately after an earthquake. Used for the first time in this study, strong-motion accelerograms gave an excellent picture of stress history and migration of seismic activity during the first 6 min. ]INTRODUCTION Dislocation models have frequently been used in studies of long-period waves and permanent deformations caused by earthquakes (Knopoff, 1958; Maruyama, 1963; Burridge and Knopoff, 1964; Haskell, 1964; Haskell, 1969; Press, 1965; Savage and Hastie, 1966; Aki, 1966; Berckhemer and Jacob, 1968; Archambeau, 1968; Burridge, 1969). Aki (1968) and Haskell (1969) successively applied this dislocation model theory to study the near-field ground displacements recorded close to the causative fault (Housner and Trifunac, 1967). Brune (1970) pointed out that most dislocation models employ somewhat arbitrary time dependence for dislocation motion, and he suggested a physically more realistic time function directly related to effective stress. In Brune's model the effective stress a is represented by the difference a 0 -a:, where a 0 is the initial stress and a: is a friction stress which acts to resist fault slippage. Although assumptions like spontaneous stress drop over the whole fault plane (Brune, 1970) may only lead to rough first-order approxima- tions, it appears that this simple theory may realistically portray the details of source mechanism (Trifunac, 1972). Although the mechanism of earthquakes is expected to be governed by a subsonic rupture growth, spontaneous faulting propagating at supersonic velocities is also possible (Weertman, 1969). In this paper, Brune's model is applied to interpret the Pacoima Dam strong-motion accelerogram (Trifunac and Hudson, 1971), which recorded ground motion centered above the fault fracture. The results suggest that strong-motion accelerograph records have great potential value for detailed analysis of the earthquake source mechanism and strong-motion seismology in general. 721