Earthquake Focal Depth Determination Utilizing Depth-Dependent Phases Recorded at Near Regional Distances

Publication Date


Type of Culminating Activity


Degree Title

Master of Science in Geophysics



Major Advisor

James E. Zollweg


Near-regional seismograms recorded for twenty-nine events, 2.5 ≤ M ≤ 4.5, from the 1981 Elk Lake, Washington, and the 1993 Scotts Mills, Oregon, earthquake sequences, were examined in an effort to identify and then utilize depth-dependent phases to constrain earthquake focal depth. Only data recorded at near-regional epicentral distances ranging from 100 to 500 km were used. At these epicentral distances focal depths are usually poorly constrained using only first arrival P phases in the location solution.

Selected events from both sequences were precisely relocated using master event techniques and used as control sets of "known" focal depths. Results of previous crustal velocity studies were used in a raytracing modeling algorithm to predict the arrival times and sequence of crustal phase arrivals at near regional distances (100-500 km). Most phase arrivals predicted by the forward modeling approach were not observed. This may be attributable to the interference of multiple phase arrivals in short time intervals, insufficient phase amplitudes resulting in phase arrivals being obscured by background noise, and/or inadequate representation of the complex velocity structure by existing models. Of the observed phases PmP and reflections from shallower velocity discontinuities were most prevalent, often exhibiting large amplitude but emergent arrivals. Pg was only visible as a first arrival and Pn was emergent, weak and sometimes not detectable for smaller magnitude events (M < 3.0).

By taking an empirical approach of plotting seismograms at single stations in order of increasing focal depth, depth-dependent phases were found at 37% of the stations examined. Two distinct depth-dependent phase arrivals were observed at only 5% of the stations analyzed, but none of the intrastation phase time differences were found to be focal depth-dependent. Due to this limitation in the observed data a new interstation phase time difference method was developed. It was discovered that in several cases time differences between 2 different P phase arrivals recorded at separate stations along the same source-to- receiver azimuth were either negative or positive linear functions of increasing focal depth. These empirically derived interstation phase time difference vs. focal depth (IPTD-FD) functions may be used to determine focal depth, over at least a 4 km source depth range, to a precision between +/- 0.05 to 0.28 km (depending upon the slope of the function). The estimate of focal depth precision assumes a +/- 0.01 s phase picking precision, that the control set focal depth and phase time difference errors have a normal distribution about their true values and that these errors are largely averaged out by fitting a line to the depth-dependent IPTD-FD plots.

In contrast to forward modeling, the empirically-based interstation phase time difference method used towards focal depth determination is relatively simple. Once a calibration set of focal depths is established this method requires no knowledge of source-to-receiver velocity structure, phase identification or origin time. The interstation phase time difference method may therefore be suitable in areas where velocity structure is not well known and station coverage is sparse.

Files over 30MB may be slow to open. For best results, right-click and select "save as..."