New Mexico Geological Society Annual Spring Meeting — Abstracts

Local microearthquake seismograms recorded near Socorro, New Mexico often show, in addition to direct P and S arrivals, clear, strong S_zS, P_zP, and S_zP phases representing S-to-S, P-to-P, and mode-converted S-to-P reflections from a sill-shaped magma body near 19 km depth. The reflections have been used previously to confirm that the reflector is magma, map the areal extent of the magma body, study shear velocity (V₃) distribution in the crust, and study the internal structure of the magma body. To better constain locations of local earthquakes recorded with the Socorro seismic network we developed a computer program which simultaneously inverts direct and reflected arrival times to jointly solve for hypocenters and a crustal velocity model.

Using this program we inverted arrival time data from 75 local earthquakes which had between 6 and 11 recording stations per event, an average of 20 picks per event, and an epicentral distribution of 2400 km² surrounding Socorro. At least 3 and as many as 10 reflections per event were included in the inversion. The complete data set consists of 564 P, 485 S,169 S_zS, 77 P_zP, and 160 S_zP arrival times. Before inverting, timing errors of between 0.075 and 0.450 s were assigned each arrival, depending on pick quality. For the results reported below parameter errors are for one standard deviation, and all diagonals of the resolution matrix exceed 0.94.

We found compressional velocity (Vp) for the upper 10 km of crust (5.94 ± 0.02 km s^-1), V_pfor the ductile mid-crust (5.80 ± 0.06 km s^-1), average depth to the reflector (18.77 ± 0.19 km), and Poisson's ratio for the upper (0.256 ± 0.002) and lower (0.229 ±0.006) layers. Thus, V_pprobably decreases slightly in the ductile crust while V₃increases slightly (3.40 to 3.44 km s^-1) with depth. An even areal distribution of positive and negative reflected-phase residuals indicates no regional trend for magma body dip. Small differences between station corrections calculated with all the data and corrections calculated with only direct arrivals also indicate little regional dip or unevenness on the reflecting surface.

Focal depth estimates were much better constrained when reflections were included in the locations. For the 75 events depth error averaged 0.59 ± 0.12 km while depth error averaged 1.90 ± 0.95 km when only direct arrivals were used to locate the events. Origin times were also better constrained, but reflections did not influence epicenter determination. This improvement in hypocenter estimation will allow us to better determine the dimensions of the seismogenic zone and to better understand the behavior of the earthquake swarms which are common to central New Mexico.