Extending the North Atlantic Hurricane Record Using Seismic Noise

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Extending the North Atlantic Hurricane Record Using Seismic Noise

Department of Earth and Planetary Sciences

Northwestern University

American Geophysical Union

Fall MeetingDecember 15, 2009

S22B-04

Carl Ebeling ([email protected]) and Seth Stein



Energetic ongoing debate: Are rising sea-surface temperatures in the North

Atlantic resulting in trends in hurricane frequency and energy?

(Figure after P. Klotzbach; map source NOAA)

Or, alternatively:

2005

1933



It’s difficult to answer with the existing hurricane record• Length: It’s short• Completeness: Undercount in historical record is likely

(Image source: NOAA)

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

The Earth’s Noise Spectrum: An Example

Secondary microseismic peakGenerated 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 peakGenerated by pressure variations due to vertical fluctuations of waves over shallowing seafloor (Hasselman, 1963) and through interaction with coastlines.

Seasonal variability in frequency and amplitude of primary and secondary peaks—related to storm energy

2005 monthly mean power spectral densities for SANAE station

(Antarctica)



Atmosphere (Wind)-Ocean Wave Link



Depression: < 18 m/sTropical Storm: 18-

32 m/sHurricane: > 33 m/sMajor hurricane: >

50 m/s

Wind speed and wave frequency are linked

Spectra of ocean waves for different wind

speeds (after Moskowitz, 1964)

Secondary microseismic

peak frequency

Primary microseismic

peak frequency

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)

• 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

Hurricane Andrew

HRV

Tropical Depression

Tropical Storm

Hurricane

(Image source: NOAA)



HRV (Harvard, Mass.) seismic station:

• Streckeisen STS-1, 1 Hz sampling

• Long-lived seismic station relatively near path of N. Atlantic hurricanes

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)! Q3 1992 raw pseudo-power recorded at HRV

Andrew?

Andrew: Preliminary Results

Take advantage of shift of secondary microseism peak to longer periods with greater storm energy

and filter appropriately

Andrew: Preliminary Results(Pseudo-power and maximum wind

speed)

Pseudo-power (raw)

Pseudo-power (bp filtered, 5-7 s)

Andrew: Preliminary ResultsEnergy shifts to longer periods with increasing intensity

Andrew: Preliminary Results

Solid circle:Meteorological characterization of hurricane, scaled by 5-7 s mean spectral amplitude

Open circle:Empirical seismological hurricane discriminant applied: No hurricane = green Hurricane = red

Andrew can be “seen” in HRV seismic data

Pseudo-power, 6-hr mean with seismically-identified hurricane (red circles)

Discriminant

Tropical Depression

Tropical Storm

Hurricane



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).



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



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

Extending the North Atlantic Hurricane Record Using Seismic Noise

Documents

hurricane frequency

hurricane signals

msmajor hurricane

green hurricane

seismic station

power signal

wave frequency

hrv seismic datapseudopower