Calderas are among the most active and dangerous volcanoes. Caldera unrest is defined by enhanced seismicity, gravity changes, surface deformation, and degassing. Although much caldera unrest does not lead to an eruption, every eruption is preceded by an unrest episode. Therefore, the proper description of unrest and the forecast of its possible outcome is a timely and challenging task. Here we review the best known unrest at calderas from 1988 to 2014, building on previous work and proposing an updated database. Where established, the root cause for unrest is always magmatic; none was purely hydrothermal or tectonic. An interpretive classification of unrest invokes two spectra-compositional (mafic to felsic) and the state of magma conduits feeding from the magma reservoir(s) to the surface (from fully plugged, through semiplugged, to open). Magma and gas in open conduits can rise and erupt freely; magma in semiplugged conduits erupts less frequently yet still allows some gas to escape; plugged conduits allow neither magma nor gas to escape. Unrest in mafic calderas is subtler, less pronounced, and repeated, especially with open systems, ensuring the continuous, aseismic, and moderate release of magma. Plugged felsic calderas erupt infrequently, anticipated by isolated, short and seismically active unrest. Semiplugged felsic calderas also erupt infrequently and are restless over decades or centuries, with uplift, seismicity, and degassing and, on the longer-term, resurgence, suggesting repeated stalled intrusions. Finally, the expected advances in better understanding caldera unrest are discussed.

The regional stress field in volcanic areas may be overprinted by that produced by magmatic activity, promoting volcanism and faulting. In particular, in strike-slip settings, the definition of the relationships between the regional stress field and magmatic activity remains elusive.To better understand these relationships, we collected stratigraphic, volcanic and structural field data along the strike-slip Central Aeolian arc (Italy): here the islands of Lipari and Vulcano separate the extensional portion of the arc (to the east) from the contractional one (to the west). We collected >500 measurements of faults, extension fractures and dikes at 40 sites. Most structures are NNE-SSW to NNW-SSE oriented, eastward dipping, and show almost pure dip-slip motion; consistent with an E-W extension direction, with minor dextral and sinistral shear. Our data highlight six eruptive periods during the last 55 ka, which allow considering both islands as a single magmatic system, in which tectonic and magmatic activity steadily migrated eastward and currently focus on a 10 km long x 2 km wide active segment. Faulting appears to mostly occur in temporal and spatial relation with magmatic events, supporting that most of the observable deformation derives from transient magmatic activity (shorter-term, days to months), rather than from steady longer-term regional tectonics (10 2 -10 4 years). More in general, the Central Aeolian case shows how magmatic activity may affect the structure and evolution of volcanic arcs, overprinting any strike-slip motion with magma-induced extension at the surface. Keypoints: Relation between strike-slip kinematics and magmatism at volcanic arcs  Two neighbour volcanic islands share a single magmatic system  Magmatic stress is dominant and generates transient faulting