Experimental study of caldera formation

Scaled experiments have been carried out on caldera collapse mechanisms, us ing silicone as analogue magma and dry sand as analogue rock. Experiments w ere carried out in two and three dimensions using a range of roof aspect ra tios (thickness/width 0.2 to 4.5) appropriate for caldera collapse. They re veal a general mechanism of collapse, only weakly dependent on the shape of the reservoir. For low roof aspect ratios (less than or equal to 1), subsi dence starts by flexure of the roof and the formation of outward dipping, r everse ring faults, which in turn trigger formation of peripheral inward di pping, normal ring faults. The subsidence always occurs asymmetrically. In cross section the reverse faults delimit a coherent piston, bounded on each side by an annular zone of inwardly tilted strata located between the reve rse and normal ring fault sets. The surface depression consists of a nondef ormed area (piston) surrounded by an annular extensional zone (tilted strat a). For high aspect ratios (>1), multiple reverse faults break up the roof into large pieces, and subsidence occurred as a series of nested wedges (2- D) or cones (3-D). The extensional zone dominates the surface depression. I n the case where preexisting regional faults do not play a major role, the collapse mechanics of calderas probably depends strongly on the roof aspect ratio. Calderas with low roof aspect ratios are predicted to collapse as c oherent pistons along reverse faults. The annular extensional zone might be the source of the large landslides that generate intracaldera megabreccias . Collapse into magma reservoirs with high roof aspect ratios may be the or igin of some funnel calderas where explosive reaming is not dominant.