Geology Reference
In-Depth Information
of well-dated volcanic ashes, and stratigraphic
superposition has permitted correlation among
these turbidites, but such correlation cannot
readily distinguish between one large event and
several events spaced very closely in time. One
clever way to address this issue is to determine
the turbidite record both above and below
confluences of submarine channels. If the
upstream channels host turbidites that were not
synchronous with each other, the turbidite
record below their confluence should reveal the
sum of the two upstream records. Instead, for
almost all channels tested, the records show the
same number of turbidites within equivalent
stratigraphic sections, thereby implying that
singular earthquakes affected both tributary
channels. This test of synchrony lends strength
to the correlation of turbidites along the Cascadia
margin that has yielded a time series extending
from the 1700AD event to the base of the
Holocene and contains 18 events (Goldfinger
et al.
, 2003). A similar methodology has been
applied to the turbidite record offshore from the
northern San Andreas Fault (Goldfinger
et al.
,
2007). There, 15 events over the past 2800 years
define an average recurrence interval of about
200 years (similar to the 230-year recurrence
interval estimated from onshore data for the
northern San Andreas Fault). The distribution of
turbidites argues that at least eight out of the
last 10 earthquakes ruptured over 300 km of the
northern San Andreas Fault. Although acquiring
sea-floor samples is not inexpensive, the data
synthesized along many hundreds of kilometers
of coast provide a potent perspective on
high-magnitude earthquakes.
multiple earthquakes. Offset geomorphic fea-
tures, on the other hand, can be readily observed
and surveyed, and they are often distributed
along the length of a fault. Geomorphic studies
following recent earthquakes have clearly
depicted variations in coseismic slip along a
fault's length (Fig. 4.20). The same metho-
dology provides a means to document the spa-
tial variations in cumulative deformation during
past ruptures and, hence, underpins reconstruc-
tions of longer-term slip rates. In alluvial envi-
ronments, typical displaced geomorphic features
include stream channels, terrace risers, channel
walls, debris flows and their raised levees, small
alluvial fans, ridges, and gullies. In coastal envi-
ronments, displaced features could include
beach ridges, coral platforms, solitary coral
heads, delta plains, and wave-cut notches (Figs
6.2, 6.10, 6.13, and 6.14).
Although a rough estimate of the amount of
displacement of an offset feature can often be
made with a tape measure, detailed topographic
maps of displaced features and the area sur-
rounding them is usually preferable, because
such maps permit a more rigorous geometric
reconstruction (Fig. 6.16A). Typically, such maps
are constructed using a theodolite with a built-in
electronic distance measuring device (an EDM
or “total station”), with differential GPS, or with
terrestrial laser scanners. In many field situations,
displaced features no longer directly intersect
the fault plane because slope processes have
either eroded or buried part of them. In such
cases, the trend of a feature - as represented by
planar surfaces, such as terrace risers, or by
linear features, such as debris-flow levees or
the intersection between a terrace tread and
riser - has to be projected on to the fault plane
(Fig. 6.16A). To the extent possible, a structure
contour map of the fault plane should be
constructed, especially with dipping faults. In
order to measure a horizontal offset, once the
fault plane is specified, the distance is measured
between the projections of linear features on to
the fault plane after any vertical component of
offset is subtracted. With a detailed topographic
map, often the uncertainties in such a projection
can be estimated and incorporated into the
displacement estimate (Fig. 6.16A).
Displaced geomorphological features
Although detailed stratigraphic records in
trenches can provide constraints on multiple
earthquakes in the past, each trench requires a
large investment of time and is most relevant
only to those segments of the fault directly adja-
cent to it. Moreover, some faults have very large
offsets in individual earthquakes, e.g., up to
18 m on some strike-slip faults (Rodgers and
Little, 2006), such that extraordinarily long
trenches would be required to document
