The Seismic Hazard from Major Earthquakes in the Central and Eastern U.S.

1
The Seismic Hazard from Major Earthquakes in the
Central and Eastern U.S.
Martitia P. Tuttle M.P. Tuttle and
Associates Georgetown, ME
John E. Ebel Weston Observatory Department of
Geology and Geophysics Boston College
2
The questions being posed here What is the
probability of a major (Mgt7) earthquake in the
Central and Eastern US (CEUS)? Where can a major
earthquake occur in the CEUS?
3
Probabilistic Seismic Hazard of the CEUS
This map depicts earthquake hazard by showing,
by contour values, the earthquake ground motions
that have a common given probability of being
exceeded in 50 years. (USGS web site)
4
The USGS Probabilistic Seismic Hazard Maps

• The ground motions being considered at a given
  location are those from all future possible
  earthquake magnitudes at all possible distances
  from that location. (USGS web site)
• Probabilistic seismic hazard calculations assume
  that earthquakes occur randomly in time with
  known (assumed) spatial and magnitude
  distributions.
• Smaller earthquakes are more probable than large
  earthquakes, so probabilistic seismic hazard maps
  primarily reflect the hazard from smaller and
  less potentially damaging earthquakes.

5
Potential Economic Losses from Large Earthquakes
Small magnitude earthquakes have only a local
effect, while large magnitude earthquakes can
affect a large area.
Rasmusson, 2003
6
Potential Economic Losses from Large Earthquakes
(Cont.)
Small magnitude earthquakes have only a minor
economic consequences, while large magnitude
earthquakes can major economic consequences.
Probabilities of losses in 50 years.
Rasmusson, 2003
7
Major Earthquakes and Seismic Hazard in the CEUS

• For seismic hazard calculations in the CEUS, we
  need to know how often and where strong (i.e.,
  Mgt7.0) earthquakes can occur in eastern North
  America. We know strong earthquakes can occur
  (i.e., 1811-1812, 1886, 1663, 1929), but
  estimating their rate of occurrence and likely
  future sites is highly uncertain.

8
Current Methods of Estimating the Rates of Strong
Earthquakes

• For the CEUS, the rates of strong earthquakes are
  estimated currently using
• -- Extrapolations from the rates of smaller
  seismicity using Gutenberg-Richter (log N versus
  M) curves
• -- Geologic data on the history of past strong
  earthquakes (paleoseismological investigations)

9
Problems with the Current Approach

• In the CEUS, more strong earthquakes have been
  observed historically than expected from log N
  versus M analyses

10

• For Eastern North America, the same problem is
  encountered

11
The Characteristic Earthquake Model and Major
Earthquake Recurrence

• The characteristic earthquake model also
  implies that the rate of large earthquakes cannot
  be extrapolated from log N versus M curves of
  smaller earthquakes

Large earthquakes may occur more often than
expected from the routine smaller seismicity
12
New Madrid as an Example of the Problem

• Evidence from New Madrid illustrates the
  difficulty of estimating the rates of large
  magnitude events even in a well-studied area
• -- Studies of liquefaction and other earthquake
  induced features indicate that there have been as
  many as 4 M7 earthquake episodes at New Madrid
  in the last 2000 years (characteristic(?) strong
  earthquakes more frequent than expected from log
  N versus M data)
• -- Geodetic data from GPS observations show no
  resolvable neotectonic deformation, suggesting
  that the past rate of strong earthquakes at New
  Madrid may not be sustained there in the future

13
Small Earthquake Activity as Aftershocks of
Earlier Major Earthquakes

• The paleoseismicity model of Ebel et al. (2000)
  postulates that many or even most of the small,
  annual seismicity of the CEUS is aftershock
  activity of large earthquakes from long ago.
  This model has implications for the recurrence
  times of strong earthquakes in the CEUS.

14
The Dilemma Posed by the Paleoseismicity Model

• If many or even most of the small, annual
  earthquakes in the CEUS are aftershocks of large
  earthquakes from long ago, then it may not be
  possible to extrapolate the rates of occurrence
  of strong earthquakes from the smaller earthquake
  activity of the region.
• How, then, can the rates of large (say Mgt7)
  earthquakes be estimated?

15
Paleoseismicity Model Tutorial
16
Implications of the Paleoseismicity Model for
Seismic Hazard Analyses

• Most or all earthquakes below some magnitude
  (5.5? or 6.0?) are aftershocks of past strong
  events. The rates of these shocks can be
  extrapolated accurately from Gutenberg-Richter
  recurrence relations.
• The rates of strong earthquakes (Mgt7.0) cannot
  be extrapolated from the rates of smaller
  magnitude earthquakes.

17
Estimating the Rate of Mgt7.0 Events with the
Paleoseismicity Model

• Try looking at the number of active spatial
  earthquake clusters and estimate the magnitude
  and time of each possible past strong earthquake
• Get an estimate of the rate of past strong
  earthquakes from the total number of active
  clusters that may represent strong earthquakes
  within some specific time window (i.e., 1000
  years)

18
Spatial Earthquake Clusters in the CEUS
The green arrows show the locations of the 1663
Charlevoix and 1811-1812 New Madrid earthquakes.
The seismicity rates in these areas can be used
to constrain the aftershock activity rate
parameters in Omoris Law.
Contours show average rates of M0 earthquakes in
60 years (Frankel, 1995).
19
Possible Major Paleoearthquake Sites in the CEUS
The red arrows show CEUS locations where the
average rate of M0 events in 60 years exceeds
8.0. These are suspected sites of past large
earthquakes. (Map from Frankel, 1995)
20
Estimated Rates of Past Strong Earthquakes in the
CEUS
If all the paleoseismic main shocks have M7.0
Rate M0 EQs in 60 Years 16 8
Time Window (years) 1118 2124
Observed of Clusters 8 15
(Assuming average values for the a, b, and p
parameters in Omoris Law)
21
If the paleoseismic main shocks have a
Gutenberg-Richter Distribution between M7.0 and
M7.5 (Using Nishenko and Bollinger (1990) CEUS
Relations)
Rate M0 EQs in 60 Years 16 8
Time Window (years) 1118 2124
(Assuming average values for the a, b, and p
parameters in Omoris Law)
22
Results of the Analysis

• If the spatial clusters represent the locations
  of Mgt7.0 earthquakes during the past 1,000-2,000
  years, then the rate of strong earthquakes in the
  CEUS has been greater than extrapolations from
  Gutenberg-Richter recurrence relations would
  suggest.
• If some past strong earthquakes were followed by
  less active aftershock sequences, then the number
  of spatial clusters identified here may reflect a
  lower bound on the number of Mgt7.0 events in the
  past 1,000-2,000 years.

23
Geologic Evidence of Major Prehistoric Earthquakes

• Geologic studies at a number of sites in the CEUS
  have demonstrated the occurrences of major
  earthquakes during the past several thousand
  years e.g., the Meers Fault (Oklahoma), the
  Wabash Valley (Indiana), coastal New Hampshire
  (possible tsunami deposit), northern South
  Carolina. These discoveries suggest that there
  are potentially many places in the CEUS where
  Mgt7 earthquakes could take place.

24
Conclusions

• The paleoseismicity model suggests that the
  seismicity of Mlt6 in the CEUS is predominantly
  aftershocks of past strong earthquakes.
  Seismicity rates computed from Gutenberg-Richter
  recurrence relations for Mlt6 can be used to
  accurately compute the seismic hazard from
  earthquakes of this size. Thus, the shorter term
  seismic hazard values for the CEUS, which reflect
  primarily the ground motions from lower magnitude
  events, are relatively unbiased estimates of the
  hazard.

25
Conclusions (cont.)

• Under the paleoseismicity model, seismicity
  rates for Mgt7.0 events cannot be computed from
  Gutenberg-Richter recurrence relations of the
  smaller magnitude seismicity. Seismicity
  clusters in the CEUS suggest that the rate of
  past Mgt7.0 events may be greater than G-R
  recurrence relations indicate. Thus, the longer
  term seismic hazard values in the CEUS, which
  reflect primarily the ground motions from higher
  magnitude events, may under-represent the seismic
  hazard.

26
(No Transcript)