Earthquakes - Extreme Events: A Physical Reconstruction and Risk Assessment - page 172

present exhibited a reduction in annual tree-ring width at the beginning of
1960. Although most of these trees showed signs of physical damage such as
root breakage and tilting, the main response consisted of markedly narrower
rings for 1--3 years after the event (Carrera and O'Neill, 2002).
Reaction wood is formed on the underside of a tree when a tree is tilted.
The tilting results in asymmetric ring growth that is expressed as wide annual
rings of reaction wood on the underside and narrow annual growth rings on
theupper side of the tilted tree. Reaction wood in conifers is darker and denser
than late wood and is usually characterised by a reddish-yellow colour and small,
thick-walled cells. Although the formation of reaction wood generally dates the
exact time of tilting, there is evidence that the formation of reaction wood can
be delayed for several years after the event. This can occur where the tree is
severely damaged and narrow annual growth rings form, as an initial response,
on both the upper- and underside (Carrera and O'Neill, 2002).
Scars occur when the bark is removed and the underlying growth tissue
(cambium tissue) is damaged. The tree will initially be unable to create annual
growth rings in the wounded area. Over time, the growth tissue from the sides
of the wound will overlap that wound and form a scar. At times, wounds on old
trees can heal completely and as a consequence be difficult to detect (Carrera
and O'Neill, 2002).
Sheppard and Jacoby (1989)notethat it is important that only trees clearly
affected by an event are used as event-response trees. It is best that trees are
sampled based on age, size, and topographic and geologic setting. The proximity
of tree responses to faulting varies with the nature of that movement. Dip-
slip faulting results in wide disturbance zones whereas areas of disturbance
near strike-slip faulting are restricted to within a few metres of the fault. Both
disturbed trees and undisturbed trees must be sampled as the former provide a
control for the latter. The use of controls is important in order to eliminate all
other potential causal factors. Events such as wet or dry seasons can also affect
thegrowth of annual rings. These effects can be distinguished from seismic
events because seismic events in the tree-ring record display an extended period
of recovery whereas responses due to climatic changes or influences are evident
only for that particular growth season (Sheppard and Jacoby, 1989).
Multiple cores are normally taken from trees by non-destructive measures
such as hand- or power-driven corers in order to help eliminate misinterpreting
climatic or other non-seismic responses. Multiple cores are essential as non-
uniformity of tree rings can occur in trees. Ring-width patterns are matched
across different trees and all rings are assigned exact year dates. This use of cross
dating accounts for any growth anomalies. After cross dating, the ring-width
patterns are checked for disturbances. Disturbances other than the hypothesised