A Study of the Relation between Ocean Storms and the Earth's Hum

Junkee Rhie and Barbara Romanowicz


It had been shown that long period surface waves, especially Rayleigh wave type, are continuously and primarily radiating from the northern oceans during the northern hemisphere winter and from the southern oceans during the summer (Rhie and Romanowicz, 2004; Nishida and Fukao, 2004; Ekström and Ekström, 2005). In this study, we investigate a four day time window, which is free of large earthquakes but during which the long period seismic amplitude due to the "hum" is unusually high. By doing this, we infer that the "hum" events occur close to the shore rather than the deep ocean and the details of the generation mechanism. We also compare the variation in seismic amplitude at long period and short period bands for three years and show that there is a strong correlation during winter periods. It indicates that the "microseisms" and the "hum" events have a common generation mechanism.

Earthquake "free" interval 2000.031-034

The time interval between 2000.031 and 2000.034 is important to infer the "hum" mechanism because it is not contaminated by large earthquakes but shows the high level of long period seismic amplitudes. We considered stacks of vertical velocity seismograms recorded at very broadband STS-1 seismometers of two regional arrays in Japan (F-net) and California (BDSN). For each array, we stacked Gausian filtered seismograms with various center periods according to the dispersion of Rayleigh waves, assuming plane wave propagation from an arbitrary azimuth. We apply a 6 hour running average with a time step of 1 hour to the stacked data at BDSN and F-net respectively. The results of a mean stack amplitude for center period of 150 and 240 s as a function of time and back-azimuth are shown in Figure 25.1. During given time interval, two noise events, probably due to the "hum" of the Earth, are clearly seen on 2000.031 and 2000.033 at both arrays. However, the arrival time of each event at each array is quite different. It is clear that BDSN records the event about 8-10 hours earlier than F-net does. 8-10 hour time difference is too large to be explained by seismic wave propagation, but it can be explained by the propagation speed of "free" infragravity waves at $\sim $ 220 m/s.

We also compare the seismic observation with a direct ocean buoy measurement near the coast of California and Japan. Unfortunately, many buoys near the Japanese eastern coast are not available. However, we can see very strong correlation between the variation in seismic amplitude and significant wave height recorded at buoys near the California coast.

Figure 25.1: (a) Mean stack amplitude (MSA) with 6 hour time window lagged by 1 hour as a function of time and back azimuth for F-net. A Gaussian filter with center period of 150 s was applied before stacking. (b) Same as (a) for BDSN. (c) Same as (a) for center period of 240 s. (d) Same as (c) for BDSN.
\epsfig{file=rhie06_1_1.eps, width=8cm}\end{center}\end{figure}

Comparison with Microseisms

It is well known that non-linear interaction of ocean swells generates microseisms, especially double frequency microseisms, which is dominant seismic noise at short periods (2-25 s, Longuet-Higgins, 1950). We have shown that there is a strong correlation between the variation in ocean wave height and long period seismic energy fluctuation. It indicates that both seismic noises at short (microseisms) and long (hum) periods are probably generated from common sources. Therefore, we compared amplitude levels in the two frequency bands for long time interval (e.g., a whole year). To do that, we need to reduce the contamination from the earthquakes. We developed a data processing method that avoids eliminating time windows contaminated by earthquakes (for details see Rhie and Romanowicz, 2006).

We compare the processed "hum" and microseism amplitude time series for BDSN over 3 years (e.g., Figure 25.2). The level of low frequency (hum) amplitude does not vary significantly with time (Figure 25.2a). However, there is a seasonal variation in the short period (microseism) amplitude, and it shows a strong correlation with significant wave height at buoys near California. During the winter period, short period amplitudes are maximum and the correlation between long and short period amplitude variations is high (Figure 25.2b).

Figure 25.2: (a) A comparison between a processed long period (hum) amplitude (gray) and short period Fourier amplitude (black) for BDSN. (b) Detailed comparison of (a) during first 90 days of year 2000. Time window strongly contaminated by earthquakes are shaded in gray. Corresponding correlation coefficient is shown in the plot.
\epsfig{file=rhie06_1_2.eps, width=8cm}\end{center}\end{figure}


By comparison between the variations in long period seismic amplitudes for two regional arrays in Japan and California, we observed that there is a time gap between arrivals of the common "hum" event on both sides of Pacific Ocean. This observation and direct comparison between seismic amplitude and significant wave height measurements at buoys lead us to a detailed scenario of generation of the "hum" comprising three steps: 1) short period ocean waves interact non-linearly to produce infragravity waves as the storm-related swell reaches the coast of North America; 2) infragravity waves interact with the seafloor locally to generate long period Rayleigh waves, which can be followed as they propagate to seismic stations located across North America; 3) some free infragravity wave energy radiates out into the open ocean, propagates across the north Pacific basin, and couples to the seafloor when it reaches distant coasts north-east of Japan, giving rise to the corresponding low frequency seismic excitation observed on the Japanese F-net array.

We have assembled and processed 3 years of microseism data at stations of the BDSN and F-NET arrays, and show a strong correlation of amplitude fluctuations in the microseismic band with that in the "hum" band during the northern hemispheric winter, but not during summer months, suggesting that in the winter, the microseisms and hum have a common "regional" or "local" origin, whereas in the summer, the origin of the hum is indeed distant (southern hemisphere) while the microseisms remain local and have smaller amplitudes


We wish to thank BDSN, F-net, JMA, and NOAA for providing high quality and continuous seismic and ocean buoy data.


Ekström G., and S. Ekström, Correlation of Earth's long-period background seismic radiation with the height of ocean waves, Eos Trans. AGU, 86, Fall Meet. Suppl., Abstract S34B-02, 2005.

Longuet-Higgins, M. S., A Theory of the origin of microseisms, Philos. Trans. R. Soc. London, Ser. A, 243, 1-35, 1950.

Nishida, K., and Y. Fukao, Spatial variations of excitation amplitudes of Earth's background free oscillations, Eos Trans. AGU, 84, Fall Meet. Suppl., Abstract S31B-1068, 2004.

Rhie, J., and B. Romanowicz, Excitation of Earth's continuous free oscillations by atmosphere-ocean-seafloor coupling, Nature, 431, 552-556, 2004.

Rhie, J., and B. Romanowicz, A Study of the relation between ocean storms and the Earth's hum, Geochem. Geophys. Geosyst., 2006 (in press).

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail: www@seismo.berkeley.edu
© 2006, The Regents of the University of California