Nice to see Pavlof is approximately on schedule. I was beginning to think it was going to be late.
The volcano has "near scheduled" eruptions as a result of it squeezing in a very long fracture of dike like morphology @ the base of the crust/upper mantle (somewhere between 40 and 90Km down) created by a Rebound/long duration mercalli X (r/lx) quake in the 1780s (one of a pair in one year) that produced a bloody great big tsunami @ Sanak island and the pair relaxing the crust 1964 style from Unimak (W.) to Kodiak (E.), with a runout vector approximately south.
The recompression convergence being faster W. than E. is pinching the fracture shut in the W. and dilating the fracture E. but it cannot travel past the arc graben and as a result pops up @ Pavlof which has been covered by the products of regular eruptions and is now (like Shishaldin, & Cleveland) a very photogenic cone shape.
The eruptions are semi regular as a result of the continued convergence across the arc which narrows the fracture near the volcano due to compression from the forearc. There is as a result a timespan from magma motion W-E along the fracture to produce enough pressure to dilate the fracture treminus and squirt new gas rich magma into the base of the magma chamber. There is then a small amount of extra time needed for the degassing magma to dilate the vertical conduit of the volcano and pop gas filled magma out of the top.
Big eruptions occur at semi regular volcanoes such as Pavlof when a parallel magma filled fracture breaks into the fracture currently being tapped.@ that point an eruption similar to the initiating event (Vei 4 @ Pavlof) would occur when the new fracture break in magma reaches the volcano.
Mayon volcano has had several fracture break in eruptions in its recorded history which have produced large eruptions overwhelming the magma in the chamber (usually basaltic andesite), and causing it to erupt basalt in a subplinian event with pyro flows and clastogenic lava.
The E.Aleutian chain & Alaska peninsula region is in the early stages of recompression, as a result it is unlikely (so far) that a parallel fracture has broken into the long set squeezing into a semi regular volcano.
Have a nice day: Ag