Extending the North Atlantic Hurricane Record Using Seismic Noise

1
Extending the North Atlantic Hurricane Record
Using Seismic Noise
Carl Ebeling (carl_at_earth.northwestern.edu) and
Seth Stein
Department of Earth and Planetary
Sciences Northwestern University American
Geophysical Union Fall Meeting December 15,
2009 S22B-04
2
Energetic ongoing debate Are rising sea-surface
temperatures in the North Atlantic resulting in
trends in hurricane frequency and energy?
Or, alternatively
2005
1933
(Figure after P. Klotzbach map source NOAA)
3

• Its difficult to answer with the existing
  hurricane record
• Length Its short
• Completeness Undercount in historical record
  is likely

• Seismology can help
• Develop discriminant using digital seismic
  data recorded
• during times of well-characterized hurricanes
• Apply to decades-long archive of ambient
  seismic noise
• records

(Image source NOAA)
4
The Earths Noise Spectrum An Example
Seasonal variability in frequency and amplitude
of primary and secondary peaksrelated to storm
energy
Secondary microseismic peak Generated by
interference between groups of waves of same
frequency traveling in opposite
directions. Energy coupled to sea floor pressure
variation does not decay with depth
(Longuet-Higgins, 1950).
Primary microseismic peak Generated by pressure
variations due to vertical fluctuations of waves
over shallowing seafloor (Hasselman, 1963) and
through interaction with coastlines.
2005 monthly mean power spectral densities for
SANAE station (Antarctica)
5
Atmosphere (Wind)-Ocean Wave Link
Wind speed and wave frequency are linked
Spectra of ocean waves for different wind speeds
(after Moskowitz, 1964)
From seismology Higher energy is indication of
increased storm energy (Astiz and Creager,
1994) Frequency of largest wave is indication of
maximum sustained wind speed (Bromirski et al.
1999)
Depression lt 18 m/s Tropical Storm 18-32
m/s Hurricane gt 33 m/s Major hurricane gt 50 m/s
Secondary microseismic peak frequency
Primary microseismic peak frequency
6
Hurricane Andrew

• August 23-26, 1992 (cat. 4 at landfall)
• 922 mb at landfall
• Sustained winds of 227 km/h
• Gusts to 282 km/h
• 40 dead US 20 billion in losses

(Image source NOAA)
HRV

• HRV (Harvard, Mass.) seismic station
• Streckeisen STS-1, 1 Hz sampling
• Long-lived seismic station relatively near path
  of N. Atlantic hurricanes

7
Andrew Preliminary Results
Power in signal is proportional to amplitude
squared Concerned about relative changes only
(pseudo-power) But non-hurricane signals in raw
pseudo-power signal (only hurricane Andrew in
August)!
Andrew?
Q3 1992 raw pseudo-power recorded at HRV
Take advantage of shift of secondary microseism
peak to longer periods with greater storm energy
and filter appropriately
8
Andrew Preliminary Results
(Pseudo-power and maximum wind speed)
Pseudo-power (raw)
Pseudo-power (bp filtered, 5-7 s)
9
Andrew Preliminary Results
Energy shifts to longer periods with increasing
intensity
10
Andrew Preliminary Results
Andrew can be seen in HRV seismic data
Solid circleMeteorological characterization of
hurricane, scaled by 5-7 s mean spectral
amplitude Open circle Empirical seismological
hurricane discriminant applied No hurricane
green Hurricane red
Discriminant
Pseudo-power, 6-hr mean with seismically-identifie
d hurricane (red circles)
11
Conclusions
Hurricane Andrew can be identified seismically
while offshore by using microseismic power
recorded at a distant seismic station.
Microseismic power must be filtered to recover
this signal (5-7 s passband in the case of
hurricane Andrew).
12
Future Work

• Discriminate local storms by using data from
  additional
• station (San Juan, Puerto Rico SJG)
• Evaluate effect of water depth, tectonic
  boundaries
• Extend analysis to number of storms with
  varying
• intensities
• Convert large numbers of analogue seismograms
  to
• digital records on production basis

13
Acknowledgements

• Incorporated Institutions for Seismology
  (IRIS)
• Data Management Center (DMC)
• Global Seismic Network (GSN)-IRIS/USGS
• Dr. Luciana Astiz, University of California,
  San
• Diego
• Dr. Phil Klotzbach, Colorado State University
• National Science Foundation Graduate Research
  
• Fellowship Program