oklahoma.crustal.models  updated 2000 AUG24  Jim Lawson, Oklahoma Geological
Survey Observatory.

In order to use models in earthquake phases Pb (critically refracted just below 
the Conrad discontinuity, I have marked the Conard as the top of the layer
in which P-wave velocity first reaches 6.7 km/sec. The less arbitrary Moho
is marked at the top of the half-space (Pn waves are critically refracted
just below the moho) which is also where P-wave velocity first exceeds 8.0
km/sec. Note added 1998 OCT29.

Note: Beginning with Jan 1998, all epicenters located by the Oklahoma
Geological Survey have used the CHELSEA model.


********************************************************************************
TRYGGVASON-QUALLS MODEL-based on a 1960s reverse shot refraction profile
from Chelsea (NE Oklahoma) to Manitou (SW Oklahoma). 
Tryggvason, E. and B. Qualls, Seismic refraction measurements of crustal study
in Oklahoma, in Journal of Geophysical Research, 72,3738-3740, 1967>
Refer to Qualls University of Tulsa MS thesis "CRUSTAL STUDY OF OKLAHOMA" 
for tables of arrival times and other details. Because of the high density of 
recording points along this long reversed profile, this data is never likely
to be equalled.
*******************************************************************************

The model is three layers (with an assumed sedimentary verneer) over a half-
space (mantle). The first arrivals are P1, P2, P3, P4.


ARRIVAL RANGE  TRAVEL TIME IN SEC FOR A SURFACE SOURCE
        in km  WHERE D=SOURCE-RECEIVER DISTANCE IN km.
_______________________________________________________________________________
P1    000-110   TP1=0.168+D/(5.96+-0.01)
P2    110-205   TP2=2.132+D/(6.66+-0.01)
P3    205-245   TP3=4.454+D/(7.20+-0.02)
P4       >245   TP4=9.064+D/(8.32+-0.02)


00.00km --------------------------------
            Pv=4    km/sec (assumed)    
00.54km --------------------------------



            P1v=5.96 km/sec




13.70km --------------------------------CONRAD





           P2v=6.66 km/sec





29.60km -------------------------------





          P3v=7.20 km/sec





50.90km ------------------------------MOHO


          Pv4=8.32 km/sec

Half-space (mantle)


*******************************************************************************
MITCHELL-LANDISMAN MODEL "Geophysical measurements in the southern great plains
", in "The Structure and Physical Properties of the Earth's Crust", AGU
Monograph 14, 1971, pages 77-93. Using Tryggvason's original seismograms, and
taking into account arrival amplitudes, Mitchell and Landisman made two
models for both P and S velocities. It is possible that they over-interpreted
Tryggvason's data, and that such complicated models are not justified.

Note added 2000 AUG24. In 1998 the OGS began epicenter location with
Johannes Schweitzer's HYPOSAT program. See NORSAR (Norwegian Seismic Array)
Science Report 1-97/98 pp94-102. All 1998 epicenters were located with free
and fixed depths using Nuttli, Tryggvason-Qualls, Chelsea, and Manitou
models. In ALL cases the CHELSEA model appeared to give equally good or
better locations. Beginning in the 1998 Oklahoma Earthquake Catalog, ALL
hypocenters have been located with HYPOSAT using the CHELSEA model. Note that
below the "Conrad" discontinuity, CHELSEA and MANITOU are identical. My
statement (see previous paragraph) about the data of Tryggvason and Qualls
not supporting complicated models like Chelsea is incorrect.
*******************************************************************************

CHELSEA (NE Oklahoma)                 MANITOU (SW Oklahoma)

Thickness  Pv    Sv                   Thickness  Pv    Sv
km          km/sec                    km          km/sec
00.6       4.00  2.31                 1.00       5.50  3.18
00.4       6.05  3.49                 
02.1       5.50  3.18        
10.3       6.08  3.51                 9.50       6.08  3.51
03.0       6.49  3.75                 05.1       6.49  3.75
01.5       6.20  3.58                 02.5       6.20  3.58
---------------------CONRAD           ---------------------CONRAD
08.2       6.72  3.88                 08.2       6.72  3.88
09.1       7.05  4.07                 09.1       7.05  4.07
11.1       7.36  4.25                 11.1       7.36  4.25
---------------------MOHO             ---------------------MOHO
half-space 8.18  4.72                 half-space 8.18  4.72

Mitchell and Landisman's paper shows some of Tryggvason's seismograms on pages
79, 80, and 81.


*******************************************************************************
NUTTLI MODEL-Earthquake Notes, Vol.XL No. 4 Dec 1969. Nuttli made a two layer
over half space (with a discontinuity for P but not S within the half space).
He said the model fit "Central US arrivals from Central US earthquakes". This
model has been generally used for location of Oklahoma and regional earthquakes
by the Oklahoma Geological Survey Observatory.
*******************************************************************************
One widespread but unpublished change which has been made in Nuttli's model is
to increase the S velocity in the upper layer from 3.5 km/sec to 3.6 km/sec to
better fit Lg travel times. This change to 3.6 km/sec is used by the Oklahoma
Geological Survey Observatory. OKGEOSURVEY does not use the P only 
discontinuity at 97.0 km depth, as we have never seen arrivals which require
it's presence.




00.0km --------------------------------



       Pv=6.15 km/sec  Sv=3.60 km/sec
                       (originally 3.50)



20.0km ---------------------------------CONRAD



      Pv=6.70 km/sec  Sv=3.67 km/sec



40.0km ---------------------------------MOHO









      Pv=8.18 km/sec  Sv=4.68 km/sec









97.0km ................................


     Pv=8.37 km/sec  (Sv=4.68 km/sec





*******************************************************************************