It often occurs that to locate an event by computer it is necessary to constrain the
depth at a certain value - like 5, 10 or 18 (Canadians use this) km. This eliminates
one variable from the calculation and prevents the computer from getting stuck in
a loop. The arrival data from the USGS web page shows nearly all arrivals are
emergent. That is to say, there is no distinct sharp arrival. What this does is
prevent a reading of a precise time that the wave arrives. This prevents a good
location from being determined. In this case, it is not surprising that a default depth
would be used. It is odd that all arrivals should be emergent, however. I could see
this for an extended source where arrivals from different parts of the rupture come
at slightly different times and merge to what is seen on the seismograms, however,
for a Ml 4.5 which would rupture only a short distance, emergent phase arrivals
leave some unanswered questions.