The last eruption appears to have been the result of the sinking of barren island into a graben during the Rebound/long duration mercalli X earthquake (r/lx) of 26/12 2004. The relaxing crust and sinking into a graben appears to have compressed the volcano at that time creating an overpressure in the high level magma chamber that erupted through the pre 2004 vent cinder cone approximately doubling the lava cover in the caldera.
The 2004 r/lx & accompanying bloody great big tsunami had one other effect. It caused fractures to form @ the base of the crust which incorporated magma from the base crustal magma discs which had accumulated magma derived from the descending plate. These fractures formed perpendicular to the vector of crustal relaxation (~30m SW). As the backarc baisin recompresses (for approximately the next 500 years or so), the magma filled fractures will be squeezed sideways at base crustal level into the least compressed part before ascending via eruption when they meet more compressed crust at all sides.
This eruption & its initial high ash cloud indicates that the first of these fractures has now been squeezed into the base of Barren island volcano.
Will be awaiting further reports with interest. The newly incorporated magma tends to be rather gassy as the magma discs are near equilibrium pressure with the base of the crust (60-90Km down).
Another volcano of interest here was Seulawah Agam, but this one appears to have merely sunk, and steamed a bit post 2004. Other reasonably large mid crustal quakes occurred in the region of Puet sague volcano, near the E. end of the 2004 r/lx footprint.
Historic records and the shape of the South Andaman backarc basin indicates that the bulk of the magma filled fractures will be squeezed into barren island.
have a nice day: Ag