...GT.¹

We assume that the global time reported by the GPS node is always correct; indeed, the definition of ``global time'' is the FTSP time reported by the GPS node. We verified that the FTSP instability affecting the sensor nodes did not occur on the GPS node, most likely because the MicaZ uses a different processor that is unaffected by the clock driver bug.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... time.²

Unlike acoustic waves, determining seismic wave arrival times is notoriously difficult. The seismograms in Figures 20 and Figure 21 should give the reader some appreciation for this.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... station.³

Note that there is no lower bound on arrival times, since a wave emanating from a deep source could arrive at all stations nearly simultaneously.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.