The best guide available to geologists for identifying volcanic hazards and for forecasting future eruptions is the eruptive
history of a volcano. By applying a fundamental concept of geology in reverse -- the past is a key to the present and future --
geologists estimate the types and magnitudes of potential eruptions from a volcano based on the rocks and deposits
preserved in its geologic record. The areal extents of different types of deposits around a volcano -- deposits of debris
avalanches, pyroclastic flows, lava flows, lahars ("mudflows"), and tephra ("ash") falls -- delineate areas likely to be affected by
futures volcanic activity. The frequency of past activity can also be inferred by determining when various deposits were
emplaced. Ages of deposits may reveal a pattern of recurrence useful for forecasting future activity. Not all deposits are
preserved or recognized in the geologic record, however, so the apparent frequency of past eruptions is a minimum
The ages of recent but prehistoric deposits are commonly determined by radiocarbon dating. The uncertainty of individual radiocarbon dates may be large, however, especially relative to the ages of young deposits. For example, a radiocarbon date of 10,000 years may have an error range of 150 to 300 years; a radiocarbon date of 300 years commonly has the same uncertainty. Additional ambiguity arises in calibrating carbon-14 ages with calendar years. Therefore, it is impossible, using radiocarbon dating, to determine the specific year or years of recent activity.
A more precise method of dating volcanic deposits of recent age is to identify anomalous growth patterns among the annual rings of trees growing at the time the deposits were emplaced. Trees that were injured but not killed by tephra or lahars may show a sequence of narrow rings beginning at the time of impact. "Cross-dating," the matching of ring-width variation patterns in one tree with corresponding ring patterns in another, should be used to ensure that dating errors are not introduced by missing rings (see section on cross-dating below). A minimum age can also be determined for deposits from the ages of new trees growing on them. Such minimum ages are approximate because they do not account for the period of time required for trees to colonize newly emplaced deposits. Depending on the tree species, climate, and type of deposit, tree establishment might take 1 to 20 years. In some cases, the ring patterns of buried dead trees can be cross-dated with those of living trees. In this manner, deposits responsible for killing the trees can be dated exactly. Tree-ring dating is helping to determine the ages of deposits and eruptive frequencies of several volcanoes in the Cascade Range.
Among the effects on pine trees of the 1943-52 eruptions of Paricutin Volcano, Mexico, was the absence of some annual growth rings. Willis Eggler, a geologist at Dartmouth College, cut cross sections from several pines near Paricutin and found that one to three rings were missing along selected radii in several disks. This showed that although annual rings are formed every growing season, rings may not form completely along all radii in years of limited growth. If a core is collected from a tree where one or more rings are missing, the age of the tree or the dates of volcanically induced narrow rings sequences will be inaccurate. Missing rings can be detected by "cross-dating", or matching, the ring-width variation patterns of sampled trees with the ring patterns of other trees growing nearby (Stokes and Smiley, 1968; Fritts, 1976). Cross-dating is possible in most of North America for several reasons:
Throughout much of North America, annual tree growth is limited by low precipitation; in Alaska it is limited by cold temperatures. Though ring-width variation patterns will be similar from tree to tree, the actual widths of rings in different trees will vary.
The ring-width variation patterns of different trees will match year for year up
to the point where rings are missing in some trees because of volcanic activity
or drought. Beyond this point, ring counts will be one or more years off unless
corrections are made by "inserting hypothetical rings" at the proper places
until ring-width patterns again match those of control trees. Because of the
possible occurrence of missing rings, tree samples should be cross-dated rather
than dated by ring counting.
[Report Menu] ...