124
This suggests that stream water entering from above, or
ground water circulating along joints or faults below the
valley floor, provided the source for the large amounts of gas
required to eject the juvenile rocks and wall-rock debris.
During the eruption, the hydrogeological conditions changed
with the nature of the country rocks and the wall rock col-
lapsed in the diatreme zone. Moreover, abundant basement
clasts in the upper part of the deposit (P4) could indicate that
the Pavin maar formed during the downward migration of
foci explosions as the depression cone formed in a shallow
aquifer according to the Lorenz’s model.
Another interpretation of the Pavin eruption dynamics
can use the results of the scaled subsurface blast experiments
with variable explosion depths, and presence or absence of
preexisting crater (Valentine and White 2012 ; Valentine et al.
2012 , 2014 ; Graettinger et al. 2014 ). In recent experimental
subsurface explosions (Goto et al. 2001 ; Valentine et al.
2014 ), the crater morphology and ejecta dynamics are largely
determined by scaled depth, Dsc = d. E-1/3, where d is the
physical depth of the explosion site and E is the mechanical
energy (J) produced by the explosion. For the optimal scaled
depth (Dsc ≈ 0.004 m J-1/3), the crater has its largest value.
With an increase of the scaled depth relative to the optimal
depth of burial for a given explosion energy, the resulting
craters are smaller until Dsc ≈ 0.008, which marks the approx-
imate transition to non-eruptive explosion (Graettinger et al.
2014 ; Valentine et al. 2014 ). Because most phreatomagmatic
explosions have energies between 10^9 and 10^13 J, the non-
eruptive depth calculated from the equation is in the range
between 8 and 172 m. So, the explosions mostly occur at
depth < 200 m and explosions that contribute most to tephra
ring deposits are likely to occur at depths < 100 m (Valentine
et al. 2014 ). This is consistent with our field data because (1)
the depth of the gneissic basement is 170–200 m below the
highest top of the Pavin tuff ring and near 100 m below the
lowest top of the tuff ring, (2) the gneissic lithics are rare or
absent in P1 and (3) the shallow-derived lava lithics are
between 60 and 80 % of the lithic fraction of P2, P3 and P4
deposits at Clidères (Fig. 6.7). However, note that the pres-
ence of granitic gneissic lithics in the tephra beds does not
mean that the explosion occurred at the basement depth, but
rather, according to Valentine and White ( 2012 ), that some
prior explosions occurred at depth and mixed the lithics
upward until they could be ejected by shallow explosions.
In the field-scale analog experiments, the explosion jets
have heights and shapes that are strongly controlled by
scaled depth and by the presence or absence of a crater. Jet
properties influence the distribution of ejecta deposits out-
side the craters. As explosion depth increases from optimal
depth (largest crater) to non-eruptive explosion depth, the
eruption jets become increasingly vertically focused. In this
case, the vertical core of the jet is associated with rapid fall
and sedimentation of coarse material in the crater. The fall
induces the formation of dilute fine-grained pyroclastic den-
sity currents outer the crater (Valentine et al. 2012 ;
Graettinger et al. 2014 ). This mechanism could explain the
P3 deposits characterized by very fine-grained clasts associ-
ated with rare ballistic blocks. In comparison with P1, the
abundance of surges in P3 could result in the deepening of
explosions or in the presence of a deeper crater after the sub-
plinian P2 phase. This mechanism could also explain the
greater relative abundance of granitic gneissic lithics in P4,
because the ballistic blocks, vertically focused, fallback in
the crater during P3 phase.
Finally, in the blast experiments, the deep-seated materi-
als are not ejected as far as shallow-seated materials
(Graettinger et al. 2014 ). This is consistent with the differ-
ence between the two reference sections and the Pavin drill-
ing. The La Liste section, 2.4 km distant from the crater,
contains less granitic gneissic rock in the lithic fraction
(mostly 0–10 %) than the Clidères reference section, local-
ized 1.5 km from the crater, which has between 20 and 33 %
granitic gneissic lithics and the Pavin drilling (500 m from
the crater) with 20–50 % granitic gneissic lithics.6.6.2 Dispersal Area and Volume of the Pavin
Volcanic DepositAfter defining the main characteristics of the deposit, a map-
ping campain was performed in order to map the area of dis-
persion and to build an isopach map for the Pavin deposit
(Fig. 6.4).
Data available from previous studies (Bourdier 1980 ;
Melet 2009 ; Lêvèque and Vaillant 2010 ; Davesne and
Demoulin 2010 ; Batailler and Jallais 2012 ; Jaillard and
Zylberman 2012 ) was collected and synthesized on a map
built with the ArcGIS software (©ESRI, 2010). To this data
were added some thickness measures from the significant
work carried out by Lavina (Lavina 2002 ; Lavina and del
Rosso d’Hers 2006 , 2009 ; del Rosso d’Hers et al. 2008 ) for
the achievement of the geological map of the Pavin area
(Thonat et al. 2015 ). Some thickness values estimated from
ERT sections or from digging were also added to the map
(Lêvèque and Vaillant 2010 ; Davesne and Demoulin 2010 ;
Batailler and Jallais 2012 ; Jaillard and Zylberman 2012 ).
Based on the data, an isopach map is proposed for the
Pavin deposit (Fig. 6.16). The extension of the volcanic
cloud to the south-east is coherent with Bourdier’s result
( 1980 ). This extension axis is also consistent with informa-
tion given by sedimentary structures such as cross-bedding
and impact sags. However, the new map shows a more irreg-
ular and asymmetrical shape, with locally important thick-
ness variations. Those changes are mostly reflecting the
variability of the topography before eruption (Montchal sco-
ria cone and lava flows with rootless cones) but also lastH. Leyrit et al.