Experimental birth of a maar–diatreme volcano

Abstract

Maar–diatreme eruptions are hazardous to people and infrastructure, and are also linked to the formation of the kimberlitic variety of diatremes, which is important economically. Processes occurring in the subsurface diatreme and their relation to surface eruptions are not yet well understood. We conducted field-scale experiments using analog materials to shed more light on these processes, especially the formation of the proto-diatreme during the first explosions of a maar eruption. Specifically, a series of buried explosions in a prepared, layered substrate (pad) produced craters, extra-crater deposits and sub-crater deposits analogous to volcanic maar craters, tephra rings and incipient diatremes. Post-explosion substrate excavation revealed that single large explosions produce sub-crater deposits extending nearly to the crater-rim crest. The same energy divided into three blasts, either co-located or at different depths with the same epicenter, produced narrower and sometimes deeper sub-crater deposits even though the final sizes of the craters were similar to that produced by the single large blast. The sub-crater deposits have an upper zone with domains from different substrate depths, and an underlying zone distinguished primarily by being more loosely packed than the original substrate. Videos show surface motion extending beyond the post-shot crater rim, and largely vertical ejection and fallback of material into the footprint of these deposits, especially for the explosions that occurred below optimal depth of burial. We infer that much of the loosely packed material was disassembled, vertically transported to different heights during the explosions, then fell back without significant relative lateral movement of grains. However, subvertical fallback did produce apparent cross-cutting structures in shallow sub-crater deposits. One explosion ejected material from the deepest substrate horizon, but it was redeposited only within the crater and is unrepresented in the ejecta rim. Implications of the experiments for maar–diatreme volcanoes, including some kimberlite pipes, are as follows: (1) vertical focusing of deep explosions in the diatreme explains the deficit of deep wallrock lithics observed at maar volcanoes; (2) direct vertical fallback is possibly an important process forming diatreme deposits, especially during the earliest stages; and (3) even in our limited simulation the number and scaled depth of explosions clearly affect proto-diatreme size and structure.