Page 126 - Caldera Volcanism Analysis, Modelling and Response

Case Study of the Abrigo Ignimbrite, Tenerife, Canary Islands 101

A variety of localised lithic bedforms, such as lithic trails, low- and high-profile
bedforms, discontinuous lenses and stratification, are commonly documented
within ignimbrite depositional units (e.g. Freundt and Schmincke, 1985; Bryan
et al., 1998a; Allen and Cas, 1998; Pittari et al., 2006). These are generally the result
of localised flow processes controlled by topographic features and/or changing flow
regimes. Lithic clasts may also be concentrated into vertical gas escape pipes (Wilson,
1980, 1984) immediately after emplacement (Druitt, 1995; Roche et al., 2002).

2.2. Quantifying spatial variations in lithic assemblages
Grainsize analysis techniques (e.g. sieve analysis, point counting) may be applied to
pyroclastic deposits to estimate average grainsize, degree of sorting and the spatial
variations in the grainsize and relative proportions of the major pyroclastic
components (pumice, lithic clasts, free crystals, ash). The maximum lithic clast size
(ML) of a particular pyroclastic unit, taken as the average of the length of the long
axis of 3–7 largest lithic clasts within a sample area, is a useful quantity which can be
related to the eruption intensity. Isopleth maps, which are contoured representa-
tions of the spatial variation of ML, have been constructed for ignimbrite deposits
to constrain vent locations (e.g. Aramaki, 1984; Smith and Houghton, 1995; Allen,
2001) and to assess the effect of palaeotopography on lithic distributions in
pyroclastic flows (e.g. Giordano, 1998). Similarly, isopleth maps of both maximum
pumice and lithic clast sizes can be constructed for pyroclastic fall deposits to
estimate mass eruption rates and eruption column heights (Carey and Sparks, 1986;
Wilson and Walker, 1987; Fierstein and Hildreth, 1992; Bryan et al., 2000).
Hand specimen identification, further constrained by detailed microscopic
petrographic study, forms the basis for classifying the major lithic compositional
types within pyroclastic units. Major- and trace-element geochemistry may further
refine the classification criteria (e.g. Cole et al., 1998). To assess the relative
proportions of the different lithic types, a variety of sampling methods have been
used, including (a) field or laboratory grid and line point counting, or counting of
only in situ or extracted lithic clasts within a specified sampling area (Heiken and
McCoy, 1984; Druitt, 1985; Potter and Oberthal, 1987; Buesch, 1992; Suzuki-
Kamata et al., 1993; Rosi et al., 1996), or (b) weighing clast populations from
grainsize fractions (Hildreth and Mahood, 1986; Suzuki-Kamata, 1988; Druitt,
1992; Calder et al., 2000). Individual samples generally contain 50 to over 300 lithic
clasts, although desirable sampling statistics are only approached with the latter
number (Suzuki-Kamata et al., 1993). Depending on the classification criteria,
lithic analyses, especially in the field, are generally restricted to lapilli and block
grainsize fractions, which could involve textures or fabrics visible to the naked eye.
In some cases, only a specific grainsize fraction(s) is analysed.
A popular way to visually represent spatial lithic component variations is to
construct a series of pie charts assorted around a locality map, showing the relative
proportion of lithic clast types at each sample site (Figure 1a; Heiken and McCoy,
1984; Potter and Oberthal, 1987; Suzuki-Kamata, 1988; Suzuki-Kamata et al.,
1993). Bar charts and histograms (Suzuki-Kamata et al., 1993; Calder et al., 2000)
or ternary diagrams (Druitt, 1992) can be used to the same effect (Figure 1b).