Calderas: Structure, Unrest, Magma Transfer and Eruptions

Calderas are subcircular depressions resulting from the withdrawal of magma from the chamber to feed an eruption or an intrusion. Calderas are often associated with large and long-lived felsic (high in silica content) magmatic systems, but may also occur in mafic (poor in silica content) systems. Felsic calderas host the largest and most explosive and destructive eruptions, at times erupting >1000 km3 of magma (Druitt and Sparks, 1984; Lipman, 1997; Branney and Acocella, 2015). For these explosive eruptions we lack geophysical and geochemical observations of shallow magma transfer and eruption, so that they remain poorly understood. In contrast, mafic calderas may feed large, long-lived effusive eruptions, which are often observed, and which also have the potential of global impact (e.g., the 1783–85 Laki-Grimsvötn eruption in Iceland; Thordarson and Self, 1993). Beside representing a source of extreme hazard, calderas are also exciting environments where multiple geological and geophysical processes, including storage and transport of magma and hydrothermal fluids, response of rock to pressurization and depressurization, deposition and compaction of various materials, interact on a wide range of temporal and spatial scales. Over the long-term (hundreds to thousands of years), volcanic activity at calderas reveals unique features. Unlike central volcanoes, where most eruptions occur from a vent or crater on the summit of the edifice, at calderas volcanic activity is often scattered over large areas, with up to tens of monogenic vents found within the caldera, along its rim and on the outer slope of its edifice. Eruptive patterns may also shift over the course of the caldera evolution (e.g., Walker, 1984; Newhall and Dzurisin, 1988; Takada, 1997; Corbi et al., 2015). Such complex distributions of eruptive vents make forecasting the location of any future eruptive vent opening particularly challenging. Also in the shorter-term (weeks to years) volcanic activity at calderas reveals a distinctive behavior. Calderas are in fact frequently, and for long stretches of time, affected by unrest, i.e., activity deviating from an established baseline of geophysical and geochemical parameters, including shallow seismicity, crustal deformation, degassing. In addition, the pressurization of a hydrothermal system often produces inflation that may mask smaller signals due to shallow magma emplacement, making it hard to recognize whether unrest is magmatic (e.g., Newhall and Dzurisin, 1988). Every caldera experiences at least one episode of unrest every few decades, and some calderas, including Yellowstone (Wyoming, United States), Askja (Iceland) or Aira (Japan) have been restless for a century (Newhall and Dzurisin, 1988; Acocella et al., 2015). Most unrest episodes do not culminate in eruptions, especially at felsic (or silicic) calderas. These felsic calderas are also often associated with resurgence, that is the evident uplift of a portion of their floor over centuries or more (Marsh, 1984; Newhall and Dzurisin, 1988; Acocella et al., 2015; Kennedy et al., 2012; De Silva et al., 2015; Galetto et al., 2017; Acocella, 2019). Due to the apparent lack of recognizable temporal patterns in these frequent states of unrest, that often take place over and again with seemingly erratic temporal evolution, forecasting any impending eruption is particularly difficult. Moreover, it is still unclear how observations from individual unrest episodes should be interpreted in the context of the longer-term, cumulative unrest history, as even a minor perturbation may destabilize a system which may have reached its limit during previous unrest episodes, as observed at Rabaul (Acocella et al., 2015; Kilburn et al., 2017). Unrest leading to eruptions (or eruptive unrest) is usually accompanied by significant seismicity and degassing, lasting only a few months; conversely, unrest not culminating in eruptions (noneruptive unrest) shows minor seismicity and degassing, lasting much longer (Sandri et al., 2017). Therefore, in addition to challenging our understanding of the formation and dynamics of large and complex magmatic systems, calderas pose significant complications in the temporal and spatial eruptive forecast. Here we contribute to defining these problems and identifying potential solutions by reviewing recent studies on calderas focusing on the processes affecting magma transfer and eruptions, both on the longer- and shorter-term. In particular, we relate the structure of calderas to their state of stress, highlighting how the latter influences shallow magma transport and the location of the eruptive vents. Then we briefly review the unrest processes, to consider the conditions controlling magma eruptibility.