Geophysical imaging of a tilted mantle plume extending at least 500 km beneath the Yellowstone caldera provides compelling support for a plume origin of the entire Yellowstone hotspot track back to its inception at 17Ma with eruptions of flood basalts and rhyolite. The widespread volcanism, combined with a large volume of buoyant asthenosphere, supports a plume head as an initial phase. Estimates of the diameter of the plume head suggest it completely spanned the upper mantle and was fed from sources beneath the transition zone, We consider a mantle–plume depth to at least 1,000 km to best explain the large scale of features associated with the hotspot track. The Columbia River–Steens flood basalts form a northward-migrating succession consistent with the outward spreading of a plume head beneath the lithosphere. The northern part of the inferred plume head spread (pancaked) upward beneath Mesozoic oceanic crust to produce flood basalts, whereas basalt melt from the southern part intercepted and melted Paleozoic and older crust to produce rhyolite from 17 to 14 Ma. The plume head overlapped the craton margin as defined by strontium isotopes; westward motion of the North American plate has likely “scraped off” the head from the plume tail. Flood basalt chemistries are explained by delamination of the lithosphere where the plume head intersected this cratonic margin. Before reaching the lithosphere, the rising plume head apparently intercepted the east-dipping Juan de Fuca slab and was deflected ~250 km to the west; the plume head eventually broke through the slab, leaving an abruptly truncated slab. Westward deflection of the plume head can explain the anomalously rapid hotspot movement of 62 km/m.y. from 17 to 10 Ma, compared to the rate of ~25 km/m.y. from 10 to 2 Ma.

A plume head-to-tail transition occurred in the 14-to-10-Ma interval in the central Snake River Plain and was characterized by frequent (every 200–300 ka for about 2 m.y. from 12.7 to 10.5 Ma) “large volume (>7000 km³)”, and high temperature rhyolitic eruptions (>1000 °C) along a ~200–km-wide east–west band. The broad transition area required a heat source of comparable area. Differing characteristics of the volcanic fields here may in part be due to variations in crustal composition but also may reflect development in differing parts of an evolving plume where the older fields may reflect the eruption from several volcanic centers located above very large and extensive rhyolitic magma chamber(s) over the detached plume head while the younger fields may signal the arrival of the plume tail intercepting and melting the lithosphere and generating a more focused rhyolitic magma chamber.

The three youngest volcanic fields of the hotspot track started with large ignimbrite eruptions at 10.21, 6.62, and 2.05 Ma. They indicate hotspot migration N55° E at ~25 km/m.y. compatible in direction and velocity with the North American Plate motion. The Yellowstone Crescent of High Terrain (YCHT) flares outward ahead of the volcanic progression in a pattern similar to a bow-wave, and thus favors a sub-lithospheric driver. Estimates of YCHT-uplift rates are between 0.1 and 0.4mm/yr.Drainage divides havemigrated northeastwardwith the hotspot. The Continental Divide and a radial drainage pattern nowcenters on the hotspot. The largest geoid anomaly in the conterminous U.S. is also centered on Yellowstone and, consistent with uplift above a mantle plume.

Bands of late Cenozoic faulting extend south and west from Yellowstone. These bands are subdivided into belts based both on recency of offset and range-front height. Fault history within these belts suggests the following pattern: Belt I — starting activity but little accumulated offset; Belt II — peak activity with high total offset and activity younger than 14 ka; Belt III—waning activitywith large offset and activity younger than 140 ka; and Belt IV — apparently dead on substantial range fronts (south side of the eastern Snake River Plain only). These belts of fault activity have migrated northeast in tandem with the adjacent hotspot volcanism. On the southern arm of the YCHT, fault activity occurs on the inner, western slope consistent with driving by gravitational potential energy, whereas faulting has not started on the eastern, outer,more compressional slope. Range fronts increase in height and steepness northeastward along the southern-fault band.

Both the belts of faulting and the YCHT are asymmetrical across the volcanic hotspot track, flaring out 1.6 times more on the south than the north side. This and the southeast tilt of the Yellowstone plumemay reflect southeast flow of the upper mantle.