Explosive volcanic eruptions are essentially controlled by the physical properties of the rising magma like temperature, density, viscosity or degree of silica polymerization and gaseous solubility or volatile content. These essential properties control magma behavior and mainly affected by siliceous content in the melt.
Magmas with elevated gas content and high viscosities which have andesitic, dacitic and rhyolitic compositions are liable to erupt extremely violently. This is because the dissolved gases like H2O, SO2, CO2 and others released during pressure decrease as magma rising and approaches the surface cannot escape gradually. Experiments have shown that the amount of a dissolved gas in magma (its solubility) at atmospheric pressure is zero, but rises with increasing pressure.
For example, in an andesitic magma saturated with water and six kilometers below the surface, about 5 percent of its weight is dissolved water. As this magma moves toward the surface, the solubility of the water in the magma decreases, and so the excess water separates from the magma in the form of bubbles. As the magma moves closer to the surface, more and more water exsolves from the magma, thereby increasing the gas/magma ratio in the conduit. When the volume of bubbles reaches about 75 percent, the magma disintegrates to pyroclasts (partially molten and solid fragments) and erupts explosively.
An extreme example of this the formation of pumice. This is a rock which consist entirely of very thin glass-walled vesicles; there are so many bubbles (vesicles) relative to glass that pumice can flout on water. the escape of gas can be so violent that the magma breaks into tiny glassy fragments called volcanic ash.
Magmatic flare-up are controlled by geodynamic processes that in most cases should imply extensional processes. It needs long time residence of primitive magma to realize a complicated system to reach long time fractionational crystallysation processes (to reach the rhyolite composition) that most surely imply a large mushy system and/or crustal anatexys. However there are only suppositions up to now how this kind of processes works to allow magma triggering processes in the mantle, as in the already recommended articles. It is indeed one of the most intriguing among all the geodynamic processes and still it will be a lot of work to understand it better.
Factors involved in simple flux‐melting models cannot not fully explain magma generation during magmatic “flare‐ups”. This is strictly related to the geodynamic processes, e.g., delamination of both lower crust and lithospheric mantle during shortening and thickening, or mantle upwelling especially those linked with volcanism in back‐arc basins, and etc.