DEEP PROBE SEISMIC OBSERVATIONS AND INTERPRETATION The Deep Probe active source experiment consisted of 10 shots detonated at 7 shot points, which were recorded by 710 portable seismographs deployed twice at ~1200 sites. Nominal instrument spacing was 1.25 km (Figure 1; Gorman et al., 1997). Shot sizes varied from 2400 to 17 000 kg of chemical explosives. The recording arrays, employing most portable refraction seismographs in North America, extended across 19 degrees from northern New Mexico to central Alberta. Just prior to our experiment, the Canadian Lithoprobe program conducted the crustal-scale Southern Alberta Refraction Experiment (SAREX), along the Canadian portion of Deep Probe. Three shot records from SAREX were included in the analysis presented here (S1, S6, and S11; Figure 1). Seismic Observations: Three Province-Related Seismic Signatures The fundamental results from the experiment are well illustrated by the records from the shot point in central Wyoming, SP43, just north of the Cheyenne belt (Plate 1). Markedly different crust and upper mantle signals are seen north and south of the shot point, indicating that profound changes occur in the upper 150 km of the Earth over a distance of no more than 250- 300 km. The primary crustal and mantle seismic waves observed in these records are: Pg, upper crustal refractions; Pi, refracted within a lower crustal layer, PmP, reflected from the Moho; and Pn and related phases, refracted beneath the Moho. Their travel times and amplitudes constrain the thickness of the crust and the seismic velocities in the crust and mantle. For example, the distance at which Pn becomes a first arrival, known as the crossover distance, increases with crustal thickness. South of SP43, the seismic data sample the southern Rockies-Colorado Plateau (Plate 1). Pn becomes a first arrival at ~200 km offset, with velocities of 7.9-8.0 km/s, and is very weak from the crossover at ~200 km to ~425 km. Beyond 425 km, the Pn amplitude strengthens to offsets of 800 km. The weak Pn from 200 to 425 km in the Proterozoic terranes is also seen on records from SP33 and SP37 in New Mexico and Colorado. In the Archean Wyoming province north of SP43 the record is dramatically different (Plate 1): the Pn crossover is at ~260 km, indicating a thicker crust, and the Pn phase with high amplitude and high velocity (8.1-8.4 km/s) to offsets of ~800-1000 km results from a very different upper mantle structure. Similar Pn is seen south of SP49 (Plate 1), from the US- Canadian border to central Wyoming. A less obvious difference in the records from the Wyoming province (SP43 north and SP49 south) is the presence of lower crustal refractions, Pi, with phase velocities of 7.0-7.3 km/s that are first arrivals from 180-260 km and clear second arrivals between distances of 260-400 km (Plate 1). No lower crustal events are apparent in the data from either the Proterozoic terranes to the south, or the Hearne province to the north. North of SP49, in the Hearne province, Pn becomes a first arrival at ~210 km with velocities of 8.1-8.2 km/s. The shorter Pn crossover distance (210 km) is also observed on the other shots in the Hearne province, indicating a thinner crust than in the Wyoming province. Cross-Section of Western North America To interpret the seismic data, we used reflectivity modeling to estimate the average one- dimensional velocity structures of the three distinct geologic provinces (Fuchs and Muller, 1971) and 2-D ray-tracing and traveltime inversion to estimate 2-D crust and upper mantle structure from all Deep Probe shots (Luetgert, 1992; Zelt and Smith, 1992). The structure of the complex sedimentary basins along the profile was determined using published basin studies and velocity- depth information from 71 well logs (Snelson, 1998; Snelson et al., 1998). The main features of the seismic data can be interpreted in terms of three simple (but very different) 1-D velocity models. These models explain the main amplitude relations among the various seismic phases and represent the average crustal and mantle structure within each province (Figure 2 and Plate 1). The Proterozoic crust of the Rocky Mountains and Colorado Plateau is 40-45 km thick, with a linear increase in velocity with depth. Most variation in crustal thickness takes place in the region near the boundary with the Wyoming province. In the Archean Wyoming province, crustal thickness increases from ~40 km at the Cheyenne belt to an average of ~50 km, with a high velocity layer (Vp ~ 7.05-7.30 km/s) in the lower ~25 km of the crust. Continuing north across the Vulcan structure, the Moho depth in the Hearne province shallows to ~40 km, and the crustal velocity is again characterized by a simple vertical velocity gradient, with higher crustal velocities than in the Proterozoic crust. The mean crustal velocity in the Southern Rockies-Colorado Plateau region is 6.3 km/s compared to 6.6 km/s and 6.45 km/s in the Wyoming and Hearne provinces, respectively. Contrasting these with global average thicknesses and velocities of 42 km and 6.4-6.5 km/s for shields and platforms and 46-50 km and 6.4 km/s for orogens (Christensen and Mooney, 1995; Rudnick and Fountain, 1995), the southern Rockies is thin and slow compared to orogens or shields. The Wyoming province is considerably thicker and faster than the averages for shields and it is faster but equivalent in thickness for orogens. The Hearne province is somewhat thinner but average in velocity for shields. Our starting model for 2-D ray-tracing used estimates of average crustal structure along the profile from the 1-D interpretations, previous seismic studies, and other geophysical data (Prodehl and Lipman 1989; Pakiser, 1989; and Schneider and Keller, 1994). Two-dimensional ray-tracing provides a more detailed picture of lithospheric structure (Plate 1), particularly in places where changes in Moho depth of 10 km occur over lateral distances of ~100 km. Gradual thinning of the crust leads to thicknesses of only 30-35 km at both the southern and northern ends of the profile. Amplitude variations of the Pn phase, and its velocity, indicate extreme differences in the upper mantle along the profile (Figure 2 and Plate 1). North of SP43, the mantle which underlies the two Archean provinces has a velocity of 8.1 km/s just below the Moho, a value which increases with depth. The seismic data show no evidence for a low velocity zone. South of SP43, the mantle that underlies the Proterozoic province has a thin lid with velocities of 7.95-8.0 km/s. Below this is a thick low velocity zone which decreases to 7.75 +/- 0.1 km/s at 60 km depth. The high-amplitude Pn phase observed in the offset range 425 to 800 km is a turning/reflected phase from depths of ~90 km and greater. Constraints on Lateral Transitions Due to the large distance between the shotpoints, travel-time modeling concentrated on the seismic phases that are consistent between different shots. The short scale variations in the arrival times were matched well by the estimated sedimentary basin structure. From the top of the basement downward, only small lateral variation within each province is required by the data. The long-offset Pn and crustal arrivals give mean velocities within a province to +/-0.1 km/s and mean depths to the Moho to ~2 km in the Pn crossover regions (Plate 1). More difficult to constrain are the transitions between the three terranes: horizontal resolution is limited by the shot spacing (~400-600km), but fortuitous positioning of the shots and geologic information allow us to make inferences over shorter distances. The strong asymmetry of the arrivals from SP43 implies the transition on the south edge of the Wyoming province is less than ~200 km broad. At the northern edge of the Wyoming province, SP49/S1 shows first arrival refractions from the lower crust to both the north and south; whereas 250 km to the north, S6 does not show lower crustal events as first arrivals and is strongly asymmetric. This implies that the transition between the Wyoming and Hearne provinces is only 100 km wide.