Between approximately 8 and 10 million years ago, the Yellowstone Hotspot was beneath Craters of the Moon. This time was characterized by violent rhyolite eruptions and caldera formation. Many geologists think the Yellowstone Hotspot formed just 17 to 18 million years ago; a few geologists think that it is much older. More and more evidence points to the hotspot having formed in the Earth's upper mantle at a depth of about 125 miles, rather than being a mantle plume from the core/mantle boundary. The hotspot has a plume shape, but the plume is probably not completely molten. It is a column of hot rock, which may have been produced by radioactive decay, in which some of the molten rock flows upward. The column flows upward until it hits the overlying North American Plate, which consists of the crust and the uppermost mantle, and is colder than the upward-flowing magma. Periodically, blobs of iron-rich basaltic magma rise up into the crust from a depth of about 50 miles. In the crust, these molten blobs melt overlying silica-rich rocks and form sponge-like magma chambers of partially-molten rhyolite.
Catastrophic eruptions of huge volumes of rhyolitic magma have taken place along the Eastern Snake River Plain about 100 times in the past 16.5 million years. These eruptions often produced huge craters called calderas; some are 10 to 40 miles wide. Many of the approximately 100 calderas overlapped and may be associated with 7 to 13 volcanic centers. Although some of the mountain ranges that existed on the Eastern Snake River Plain before the hotspot may have been blown away by the eruptions, it is more likely that they were swallowed up as the floor of the caldera sank during the violent explosions.
The Yellowstone hotspot itself is stationary, while the North American Plate has been moving in a southwesterly direction over it. The plate's movement has produced a progressively younger trend of rhyolitic eruptions to the northeast. Between 6 million and 15,000 years ago, numerous basaltic eruptions produced a 4,000-foot-thick sequence of lava flows in the vicinity of Craters of the Moon. Between 15,000 and 2,000 years ago, the Craters of the Moon Lava Field formed during eight major eruptive periods. During this time the Craters of the Moon lava field grew to cover 618 square miles. The Wapi and Kings Bowl lava fields formed contemporaneously about 2,200 years ago.
Recent seismic data suggest that the Yellowstone Hotspot left behind a slab of basalt 6 to 10 miles thick. This slab is poised in a mid-crustal position and some of it is thought to contain partial melt. It is believed that this slab represents the slag left in the bottom of the numerous magma chambers spawned by the hotspot. This region is experiencing basin and range type faulting, which is stretching or pulling apart the crust. The Lost River Range north of the town of Arco is good evidence that these forces are still active. In 1983 these forces caused a magnitude 7.3 earthquake, during which Mount Borah rose about 1 foot and the entire Lost River Valley in that vicinity dropped about 8 feet. On the Eastern Snake River Plain, rather than producing mountain ranges, the tensional forces have caused decompression melting, which results in dike emplacement and periodic eruption of molten rock onto the surface. As long as these forces continue to act, more eruptions will eventually occur.
The recurrence interval for eruptive activity in the Craters of the Moon Lava Field averages 2,000 years and it has been more than 2,000 years since the last eruption. The constancy of most recent lava output rates suggest that slightly over one cubic mile of lava will be erupted during the next eruption period. In the past, eruptions in the Craters of the Moon Lava Field have generally shifted to the segment of the Great Rift with the longest repose interval. Therefore, the next eruptive period is expected to begin along the central portion of the Great Rift in the Craters of the Moon Lava Field, but may well propagate to the northern part of the monument in the proximity of the loop road. Initial flows, based on past performance, will probably be relatively non-explosive and produce large-volume pahoehoe flows. Eruptions from potential vents on the northern part of the Great Rift may be comparatively explosive and may produce significant amounts of tephra (airfall material ejected from a volcano), destroy cinder cones by both explosion and collapse, and build new ones.