Upper mantle P-wave velocity structure from PASSCAL teleseismic transects across Idaho, Wyoming and Colorado

Ken Dueker,Huaiyu Yuan

doi:10.1029/2004gl019476

Abstract

[1] Inversion of teleseismic P-time residuals along two linear seismic arrays provides well-resolved inverse images of the upper mantle. Low velocity bodies beneath the Yellowstone Hotspot Track and the Grand Mesa volcanic field in western Colorado are similar in shape and magnitude. In the two places where our transects cross the 1.78 Ga Archean-Proterozoic Cheyenne suture, high velocities are imaged to 150 km. Beneath the Leucite Hills volcanic field in Wyoming a small upper mantle low velocity body is imaged. Whether these anomalies are lithospheric or asthenospheric in origin is poorly constrained, but the consistent high velocities beneath the Cheyenne suture suggest a lithospheric origin. Likewise, the lack of extension beneath the western Colorado volcanic fields suggests that a low-solidus lithosphere is being disturbed by a regional heating event.

Highlights

[1] Inversion of teleseismic P-time residuals along two linear seismic arrays provides well-resolved inverse images of the upper mantle
In cases where a low velocity body underlies a suture, we suggest that lowsolidus materials may be preferentially melting during regional heating events
The velocity anomaly beneath the Yellowstone hotspot track (YHT) is consistent with the previous P-wave images produced from the SRP-93 data set [Saltzer and Humphreys, 1997; Schutt and Humphreys, 2004], except that depth extent of the high velocity anomaly to the SE of the eastern Snake River Plain has been resolved to lie above 200 km

Summary

Data and Methods

[4] Our data set is formed from three teleseismic transects (Figure 1): the five month 1993 Snake River Plains (SRP93) transect, the three month 1997 Deep Probe N-S transect (DP-NS) and the three month 1997 Deep Probe NW-SE transect (DP-NW). To improve error estimates and equalize resolution, a summary ray data set is formed by mapping the P-residuals for each station onto a slowness grid. Non-linear inversion, to account for ray bending effects, has not been done because previous work with the SRP-93 portion of our data set shows its effects are minor [Saltzer and Humphreys, 1997]. Regularization has been applied by weighting the column space of the data kernel matrix with an apriori model covariance matrix and the addition of diagonal damping. Choice of an optimal regularization parameter was found by assessing the trade-

Results

Summary Ray Data

Findings

Discussion