In a message dated 01/01/2010, skmort@............ writes:

Another FYI, in case you didn't  notice in the notes on the USGS pages,  
the IASP91 is based on what they  term a spherically-symmetric  model, the 
ellipsoid is not taken into account.  Those pages are  apparently not going for 
ultra-high accuracy, just for a simple average.   I guess, they assume, 
those pages are close enough for government work, for  the public,  HA!
Stephen



Hi Stephen,
 
    The ellipsoidal error is small. What does vary the  track time is the 
actual track between the observer and the quake. The IASPI91  model uses 
average velocity values and note that the times are also dependant on  the quake 
depth.
 
    Just a bit of background. If you want highly  accurate triggering, the 
usual method is to correlate the outputs of two or  three sensors, one of 
them a vertical. This gives a much more reliable  result than you can obtain 
from a single sensor.
 
    Regards,
 
    Chris Chapman






In a message dated 01/01/2010, skmort@............ writes:
Another FYI, in case yo=
u didn't=20
  notice in the notes on the USGS pages,  the IASP91 is based on what=
 they=20
  term a spherically-symmet=
ric=20
  model, the ellipsoid is not taken into account.  Those pages are=20
  apparently not going for ultra-high accuracy, just for a simple average.=
 =20
  I guess, they assume, those pages are close enough for government work,=
 for=20
  the public,=20
HA!
 Stephen


Hi Stephen,
 
    The ellipsoidal error is small. What does var=
y the=20
track time is the actual track between the observer and the quake. The IAS=
PI91=20
model uses average velocity values and note that the times are also depend=
ant on=20
the quake depth.
 
    Just a bit of background. If you want highly=
=20
accurate triggering, the usual method is to correlate the outputs of two=
 or=20
three sensors, one of them a vertical. This gives a much more re=
liable=20
result than you can obtain from a single sensor.
 
    Regards,
 
    Chris Chapman