Transcrustal, interconnected, or isolated magma reservoirs, can we tell the difference?

The organization of melts below active volcanoes plays a large role on the size, composition, and character of the ensuing eruptions. In principle it should be possible to determine the distribution of melts with geophysical data, such as inversion of seismic velocity, density, and gravity. In practice, it appears that the size and distribution of the melts have a topology that only allows for very approximate information, typically made of hot big potatoes where melts are likely to reside. This is specially the case for many calderas and large stratovolcanoes from subduction zones, and unfortunately such view is not good enough to address the likely size of future events. Petrological studies of erupted crystals can, also in principle, provide a view of the depth and possible variety storage areas. However, geothermobarometry data tend to have large errors on pressure, and in a similar manner to most geophysical data many calderas and stratovolcanoes show a very wide distribution of depths (transcrustal systems?). Notwithstanding, many systems show a broad agreement between the petrological and geophysical data, with three or more storage areas, roughly at 30 km (or deeper), 7-10 km, and 1-3 km. Moreover, the configuration of the distribution of the melts may change over time, and crystal-kinetic and some geophysical data suggest that in mafic volcanoes new connections and merging of melts from different environments may occur days prior to eruption, whereas large-scale amalgamation melts from multiple reservoirs to a much larger one can occur in years to decades in silicic eruptions from calderas. Yet, data from geophysical and petrological studies of several recent monogenetic dike-fed eruptions, show simpler magma plumbing configurations, with well-defined areas of isolated magma storage and connections between them near-real time, from which magmas of different compositions are erupted. In this presentation we will review the data on magma plumbing structure and dynamics from different datasets and discuss how could progress be made for their use in mitigation of volcano hazards.