2005 Volume 50 Issue Special Pages S59-S76
One of the outstanding targets of physical volcanology is searches for the real nature of magma reservoirs. Calderas and their formation have been long disputed in relation with magma reservoirs. Calderas must be defined by their subsurface structures not by their localities. From such a standpoint, calderas are classified into two: high gravity-anomaly type and low gravity-anomaly type because gravity anomalies reflect basement structures and density contrasts of caldera fills. The hypotheses of caldera formation in the early 20th century had not taken consideration of the following factors: Their subsurface structures, depths of magma reservoirs and strength of the crust. Some models of caldera formation depend on inappropriate analogy between small-scale geologic events and large-scale ones. First, strength of the earth crust is reviewed in laboratory and field measurements, with the distinction between shallow and deep parts of the crust. The two mechanical models of caldera formation are criticized from the standpoint of strength of the upper crust: a) “Model of cap rock”: After a simple calculation, it is proved that calderas smaller than 3km in diameter may be interpretable by this model. b) "Model of collapse into an emptied magma reservoir": Stress distribution due to collapse of a magma reservoir is estimated under some assumptions. The stress decreases in inverse proportion to square of the distance from the origin. Then, depths of magma reservoirs have much importance to the stresses at the earth surface caused by their collapses. Then, reliable geophysical data of locations of magma reservoirs are collected, and their average depth is provisionally determined as 10km. After the above estimates, it may be said that calderas can be scarcely formed by collapses of magma reservoirs measuring a few kilometers in diameter and locating at depths deeper than 5km. The hypothesis that calderas were formed by collapses of pre-caldera volcanoes in a body into magma reservoirs, is examined by their seismic effects in some historical eruptions resulting calderas or larger craters. A few eruptions were accompanied with earthquakes of M 7; However, they can be interpreted as magmatotectonic earthquakes. Many eruptions were usually accompanied with earthquakes of M 5 or smaller than M 6, and the earthquakes were not always simultaneous with the large eruptions. The hypothesis proved to be unreasonable. To exemplify the actual state of understanding for caldera structure, the three typical calderas in Japan, Hakone, Aso and Aira calderas are briefly described. Some volcanological data newly acquired on these calderas are introduced and additionally interpreted. Volcanic processes confirmed on such Japanese calderas should be substantially common to all calderas on the earth even if they may take some different features. It is finally concluded that calderas are composites (or amalgamations) of plural explosion craters which are often accompanied with violent eruptions of pyroclastic material and that calderas are not produced by collapses of volcanoes into emptied magma reservoirs.
