Page 133 - Fundamentals of Physical Volcanology

9780632054435_4_008.qxd 12/10/2007 12:29PM Page 110

110 CHAPTER 8

tion. These are usually called lithic clasts, and are average grain size increases upward the deposit is
far less likely to break on landing than fragile pieces inversely graded. Grading can occur for two rea-
of pumice. Although at any one site the lithic clasts sons. First, the wind speed may change during an
are generally smaller than the lower density pumice eruption. An increase in wind speed will carry all
clasts (recall from eqn 8.1 that it is (dσ), the prod- clasts further from the vent, so at any one location
uct of size and density, that controls where clasts increasingly larger clasts will be deposited with
land), it is common to find that, near the vent, the time and inverse grading will develop. Second,
value of (dσ) for the largest lithic clast is greater there may be a change in the mass flux of the
than the values from any of the pumices, due to eruption. An increase in mass flux causes an increase
pumice breakage, and it is this largest value of (dσ) in the eruption cloud height (see eqn 6.7) and this
that is used in any analysis of the deposit. too causes the dispersal of clasts of a given size to
• The thickness of the deposit progressively increase with time and so leads to inverse vertical
decreases with distance from the vent. While this grading. Both of these processes will act over most
is generally true there are exceptions. Thus, if the or all of a deposit. However, local grading can occur
deposit accumulates on a steep slope (greater than in regions of slumping on steep slopes due to the
∼30°) slumping may occur; and at any location differential movement of clasts of differing sizes.
there may be erosion of the deposit by nonvolcanic
processes after it is emplaced. Furthermore, soil
8.3 The application of eruption
formation by weathering can change the thickness
column models
of a deposit. There may also be effects related to the
dynamics of the eruption itself. Thus, if there is an
unusually large proportion of medium-sized pyro- A model is of value only if it actually reproduces
clasts leaving the vent, relative to the larger and behaviors that we see in reality. Theoretical erup-
smaller sizes, these will fall mainly at intermedi- tion models predict how pyroclast dispersal is
ate distances from the vent, and so a situation may related to the eruption conditions and the atmo-
occur where the downwind thickness of the deposit spheric conditions. Since the development of mod-
increases away from the vent for some distance els of this kind in the 1970s and 1980s there have
before eventually decreasing in the more common been several well-observed eruptions for which
way. Another complication can occur when an erup- the models have been shown to work quite well,
tion column contains a lot of condensing water certainly well enough to justify their use in diagnos-

vapor, perhaps because the eruption occurs during ing conditions during ancient eruptions whose
a rain storm so that both air and water droplets are products are preserved. For prehistoric eruptions,
being entrained in the gas-thrust region. When this which include the largest eruptions in the geolo-
happens, a water film forms on the small pyroclasts gical record, events vastly greater than anything
and they can stick together progressively to form experienced thus far by humans (see Chapter 10),
larger aggregates called ash clusters or ash pel- such techniques provide us with the only way of
lets. In dry eruption clouds, frictional effects can determining the mass fluxes and eruption cloud
cause particles to become electrostatically charged, heights. This is very important in assessing the
and this too can cause aggregates to form. Being possible hazards presented by active volcanoes
larger, aggregates fall to the ground faster than the that have little or no historical record of eruption
small clasts forming them would have done, and (see Chapter 11). Modeling has also enabled vol-
again this can cause unusual patterns in the thick- canologists to develop a long-term view of the
ness of the deposit. effects that large eruptions might have had in the
• A deposit may exhibit vertical grading. This past and could have in the future on our climate
means that the range of grain sizes at any given loca- and environment. This issue is discussed further
tion in the deposit changes with height in the in Chapter 12.
deposit. If the average grain size decreases upward Both the testing and the application of the
in the deposit this is called normal grading; if the models require the same steps to be taken: the fall