Page 141 - Fundamentals of Physical Volcanology

9780632054435_4_008.qxd 12/10/2007 12:29PM Page 118

118 CHAPTER 8

fountains. Other mechanisms that have been ob-
served in historic eruptions include the formation
of “directed blasts” by the partial disruption of vis-
cous lava domes growing over vents, the complete
collapse of viscous lava domes, and the explosive
disintegration of the fronts of viscous lava flows.

COLUMN COLLAPSE AND PYROCLASTIC FOUNTAINING
Considerable potential for confusion exists in the
literature in connection with the term “column col-
lapse” as a way of forming pyroclastic density cur-
Fig. 8.11 Diagram showing the formation of a pyroclastic
rents. The actual collapse of what had previously
fountain over a vent when insufficient air is entrained into
been a stable eruption column is a discrete, fairly
the eruption column to cause convection. The arrows show
short-lived event. If an eruption column ceases to the direction in which gas and particles are moving, and the
be stable because the bulk density of the material thin solid lines show contours of constant pressure.
within it becomes greater than that of the sur-
rounding air, the material at the top of the column
will take a similar length of time to reach the be influenced just by gravity. The height, h, reached
ground as a stone dropped from that height. The by an object thrown vertically upward at speed u
time needed to fall a distance s when the accelera- when there is no interaction with the atmosphere is
2
tion due to gravity is g and there is no drag force given by [u /(2g)]. Table 8.1 showed that magma
from the surrounding atmosphere is [(2s)/g] 1/2 , so eruption speeds in Plinian eruptions range from
−1
the collapse times for eruption columns 20, 30 and 200 to ∼500ms for magma water contents in the
40 km high are ∼63, 77 and 89 seconds. In practice range 1 to 5 wt%. If this range of values is used, the
there will be some interaction of the outer part of corresponding fountain heights are predicted to
the collapsing column with the surrounding air, but be ∼2km to ∼13 km!
it is still clear that the process will take only a few to In fact this analysis is too simple. The material
at most several minutes. falling down the outer part of the fountain exerts
There is no reason why the collapse of an erup- a pressure on the ground surface where it lands

tion column should cause the eruption to stop; (imagine standing under a waterfall), and in fact it is
and it is clear from well-studied deposits that post- the distribution of high pressure around the vent
collapse eruptions may continue for hours or even that is responsible for deflecting the fountain mate-
days. We have no observational evidence to tell us rial from its near-vertical fall into a ground-hugging
what happens at the vent during and after a large- flow (Fig. 8.11). The high pressure at the level of
scale column collapse event because no one who the vent has another effect. It reduces the speed
has ever been close to one has lived to describe it with which the mixture of pyroclasts and magmatic
(Chapter 11). A mixture of pyroclasts and volcanic gas emerges from the vent because it reduces the
gas is still emerging from the vent at high speed, but amount of gas expansion in the shallow part of
the mixture is no longer able to entrain enough air the dike system, as discussed in section 6.4.2. The
to become buoyant. Theory suggests that a fountain erupting mixture speeds up as the gas expands
would form over the vent (Fig. 8.11). On reaching above the vent, but some of the energy released by
the ground, gas and pyroclasts falling down the the expansion must be used to give the mixture a
outer edge of this fountain would move away as a lateral velocity as well as a vertical velocity, and so
pyroclastic density current. We can get some idea the effective upward speed that we should use to
of the height of the fountain because, although the find the fountain height is less than it would be in an
material in its outer part would interact with the eruption column that has not collapsed. Numerical
atmosphere, the material in the core would mainly simulations of this process (Table 8.2) suggest that