Teleseismic deep sounding of the velocity structure beneath the Rio Grande rift
— P. M. Davis, E. C. Parker, J. R. Evans, H. M. Iyer, and K. H. Olsen

A 1,000-km-long linear array of 20 seismic stations between Moab, Utah, and Odessa. Texas, crossing the Rio Grande rift near Santa Fe, recorded 40 useful teleseisms during December 1982 and January 1983. The orientation of this array was chosen to be at an angle of 45° to the strike of the rift because teleseismic arrivals commonly occur along this azimuth. P wave arrival times show a clear delay of as much as 1.5 sec centered on the rift with measurable delays extending several hundred kilometers to either side of it. This delay pattern correlates well with the broad negative Bouguer gravity anomaly in this area which is in isostatic compensation with the topographic uplift. Variation in crustal velocities is too small to fully explain the delay pattern. Preliminary inversions of the data require a low-velocity, low-density body situated well below the Moho. Constraining it to lie completely within the crust significantly degrades the fit of the model to the data and rules out an Airy type of isostatic compensation. Both a dampedleast- squares block inversion and fitting of a bowed "lithosphere—asthenosphere interface" to the data indicate a rift-centered delaying structure at upper-mantle depths (70-200 km), at least 500 km wide. We compare velocity, gravity, and topography across the rift.

The Rio Grande rift, which separates the Great Plains to the east from the Colorado Plateau and Basin and Range province in the west, lies just to the east of the continental divide in a region of broad uplift, high heat flow, and low gravity. Figure 1 (from Cordell, 1978) illustrates the topographic uplift and regional gravity low extending from the Colorado Plateau into the Great Plains. This profile and several similar ones at 10 intervals lying between 33° and 37°N latitude, suggest that a broad, linear low-density zone several hundred kilometers wide is associated with the uplift, and has given rise to the rift, a low-lying block (some 50 km wide). Ander (1980) has presented a wide range of gravity models which fit the observations. We have used the teleseismic P-wave delay technique to provide extra constraints on the models in order to better identify the position and shape of low-density zone.

Spence et al. (1982) reported differential travel times of 0.6 sec between within-rift stations of a northern New Mexico seismic net. This pattern was interpreted as due to a deep-seated, low-velocity body beneath the Valles caldera. They also reported a northeast-trending, low-velocity zone aligned in the same direction as the Jemez lineament. Superimposed on this pattern was up to a 1.8 sec difference between the rift stations and the distant WWSSN station at Lubbock, Texas. Our experiment was designed to investigate this larger, longer-wavelength anomaly, which necessarily required installation of stations over a much longer baselength than that covered by the northern New Mexico net.