We inverted local earthquake arrival times to estimate spatial variations of seismic velocity. Our model consisted of near-surface station corrections and local perturbations to a standard crustal velocity model. We found a zone of relatively high-velocity trending southeast from the Salton Sea. This zone corresponds to the region of thickest sediments. We compared our results with those of teleseismic studies by Savino et al. (1977). The agreement was excellent, suggesting that the teleseismic delays are caused primarily by crustal velocity variations. Residual delays between the teleseismic observations and predicted crustal delays imply crustal thinning of 3 or 4 km along the axis of the valley.
Known geothermal resource areas at Salton Sea (or Obsidian Buttes), Brawley, and East Mesa, lie on the axis of a zone of thin crust, and they may be intimately related to the Brawley fault. The crust beneath these areas may be locally even thinner than elsewhere along the Imperical Valley trough, or the crust may have locally higher seismic velocities, or both. Neither local earthquake nor teleseismic arrival times can discriminate between these hypotheses, but the issue might be resolved by combining both types of data. Known geothermal resource areas at Heber, Dunes, and Glamis, lie away from the projected trace of the Brawley fault. These areas are nearly aseismic, and overlie crust with apparent seismic velocities only mildly higher than the regional average. These apparent velocity anomalies could be related to crustal thinning, but because of the very mild Bouguer gravity anomalies in these areas, it seems more likely that the velocity anomalies occur entirely within the crust.