Page 97 - Caldera Volcanism Analysis, Modelling and Response

72 Roberto Sulpizio and Pierfrancesco Dellino

4.2. Development of self-organised pulses in a PDC and stepwise
aggradation
Recently, Sulpizio et al. (2007) proposed a model for small-scale PDCs that joins
together the en masse and aggradation approaches. This model successfully explains
characteristics of small-scale PDCs from the subplinian eruption of Pollena from
Somma-Vesuvius (Rosi and Santacroce, 1983; Rolandi et al., 2004; Sulpizio et al.,
2005), but it can also satisfactorily describe some general characteristics of PDC
deposits.
The model assumes that most PDC deposits originate from stratified flows, in
which the segregation of particles with higher terminal velocities in the lowermost
part of the flow can develop a high-concentration zone (Valentine, 1987; Branney
and Kokelaar, 2002; Dellino et al., 2004). This basal portion of the flow can move
down-slope as a succession of high-concentration pulses whose thickness varies as a
function of turbulence and kinematics of different density waves (Figure 7a, d and f ).
This allows consideration of a single-pulse as a flow-boundary zone for its entire
thickness, in which the interplay among shear-rate, rate of deposition and
concentration of particles determines the depositional regime (fallout, fluid-escape,
granular flow and traction; Figure 7n). Pulse stoppage occurs en masse when
resistance forces overpass the driving ones (e.g. when grain interlocking forces
overpass gravitational forces). The four types of flow-boundary zones are completely
intergradational (Burgissier and Bergantz, 2002; Branney and Kokelaar, 2002)and
mixed regimes are common (Sulpizio et al., 2007).
This implies that, due to the masking effect of the phoenix cloud on the
underflow, the current appears to move as a steady mass flow, but, in reality,
the main body is segmented into different pulses that run very close to each other.
The development of multiple pulses within a gravity current have been observed
during large flume experiments (Major, 1997; Iverson, 1997), small-scale
laboratory experiments (Savage, 1979; Brennen et al., 1983; Huppert et al.,
1986) and hypothesised for both volcaniclastic and pyroclastic deposits on the basis
of field evidence (Chough and Sohn, 1990; Schwarzkopf et al., 2005; Sulpizio
et al., 2006, 2007). The mechanisms responsible for pulsating behaviour in a
gravity-driven current are still not completely understood, but can be related to the
successive release of material during generation (i.e. during column or dome
collapse and failure of soils or slopes; Pierson, 1980; Scott et al., 2001; Cole et al.,
2002; Loughlin et al., 2002) and/or to the inhomogeneous mass distribution within
the current that induces development of kinetic waves (Lee and Leibig, 1994;
Major and Iverson, 1999; Schwarzkopf et al., 2005). Lee and Leibig (1994)
numerically demonstrated how a system with an initially random density field
evolves to a configuration in which neighbouring regions have a high-density
contrast. At the early stage of development, the density contrast between nearby
regions increases linearly with time. In a gravitational field, this results in different
velocities in adjacent portions of the current:

jðr Þ jðr Þ
1
2
Uð1; 2Þ¼ (4)
r r
1 2