S12B-01 10:20h
The use of ocean microseisms for monitoring time-dependent stress-induced crustal properties changes
Microseisms are a neglected but rich source of continuous energy observed by the worldwide seismic network of broadband seismometers. They are generated by ocean wave action and are potential energy sources for global 4D tomography of the upper 20km of the crust where earthquakes nucleate. 4D tomography can detect small velocity changes in rocks caused by fluid migration in oil fields. This level of change is on the same order as that expected due to tectonic stress build-up and fluid migration in the crust. The inversion of seismic energy observations for stress change in the crust is dependent on sufficient knowledge of the seismic source locations and timings. Ocean microseisms could be suitable sources for 4D tomography. However, their usefulness for this application is limited by a lack of sufficient source function characterization and a thorough understanding of the temporal and spatial characteristics of the ocean wave sources. Space-based observations may be able to provide critical constraints on the non-linear wave-wave interactions at the ocean surface which generate ocean microseisms. These wave-wave interactions cause pressure pulses that are un-attenuated with depth and generate Rayleigh waves at the seafloor. The seismic record reveals a continuous ocean-generated signal, which is correlated with swell direction, magnitude, and period. If the microseismic energy sources could be properly characterized for crustal tomography, they could enable the monitoring of stress-induced crustal property changes, providing fundamental insights into the nature of stress accumulation and release in tectonically active regions.
S12B-02 INVITED 10:35h
Observations and Modeling of Microseisms in the Santa Clara Valley, California
Previous studies of the 3D velocity structure in the Santa Clara Valley (SCV) showed that teleseismic, local, and microseism data recorded by the 41-station Santa Clara Valley Seismic Experiment (SCVSE, 6/98-12/98) are all sensitive to basin structure and that they may be used to refine the velocity model of the basins. In our recent work we focused on constraining the source of the microseisms and used this for modeling the microseism observations in the SCV. We used an f-k array method on microseisms recorded during the SCVSE to determine if they are monodirectional and to see if their source can be localized. Our results showed that at low frequencies (0.1 to 0.3 Hz), wavefield observations in the SCV display directionality. At higher frequencies (0.3 to 0.5 Hz), wavefield directionality is lost, which may be due to scattering of the waves by the 3D structure in the SCV basins. The important result of these observations is that the source of the microseisms can be localized and can therefore be used in numerical simulations. We used the 3D finite-difference code E3D (Larsen and Schultz, 1995) and the UCB 3D velocity model (Stidham et al., 1999) to simulate the microseism wavefield. A vertically oriented CLVD source located about 27 km offshore was used to generate isotropic Rayleigh waves. We used the source time function that was a superposition of sine waves at discrete periods over the observed microseismic band. The f-k analysis of simulated waveforms agrees with observations in terms of directionality at low frequencies, and the loss of directionality at higher frequencies. We will attempt to refine the method to simulate microseism wavefield by including the source spectrum derived from the ocean wave data recorded at the Santa Cruz buoy. One objective of this research is to use the obtained results to develop a simultaneous inversion of the teleseismic, local, and microseism observations to constrain the internal velocity structure of the SCV basins.
S12B-03 10:50h
Seasonal Variations in Particle Motion of Microseisms and monitoring of water content at shallow depths
There is a long history of study on microseisms, historically regarded as an annoying seismic noise for frequencies between about 0.05 and 0.3 Hz. With the emergence of dense, broadband seismic networks in the world, however, there is clearly an opportunity to carefully analyze microseisms and learn about their sources and the crustal structure in which they travel. The main point of this paper is our discovery of seasonal variations in particle motion of microseisms. We report this result from our analysis of seismic data from California Integrated Seismic Network (CISN) that has more than 150 broadband stations. But we also show some examples from other networks that seem to possess similar characteristics. It was pointed out more than 40 years ago that microseisms consist of Rayleigh waves (e.g. Haubrich et al., 1963). We first confirmed this feature by checking the phase shift between the maximum horizontal motion and the vertical motion. Indeed, we can identify Rayleigh-wave signal by performing this analysis. There exist some other type of energy in microseisms but, in this paper, we focus on the Rayleigh-wave type energy for further analyses. Once we identify the Rayleigh wave signals in microseisms, we can easily determine the direction(s) of energy propagation at a given station. We confirmed that the sources, determined at each station, point toward coasts, which was certainly expected and not surprising. We then discovered, to our surprise, that the ratio of the horizontal amplitude to the vertical (hereafter referred as HZ-ratio) displays seasonal variations. Particle motion is relatively flat in winter (northern hemisphere) and becomes closer to a circle in July. This feature is found at basically all stations with good signal-to-noise ratio that we analyzed. The manner they change with season, especially their frequency dependence, differ from one station to another. This feature is NOT related to the changes in sources of excitation because the HZ-ratio of Rayleigh waves, as it is the ratio for the horizontal and vertical amplitude of the eigenfunction of a local seismic structure, should not change with variations of excitation sources. We argue that this is caused by the changes in water content below each seismic station, especially through changes in groundwater level and the water content in the vadose zone. We have developed theoretical modeling technique and confirmed that it is possible to match data and theory using reasonable numbers for porosity for shallow crust. Apparently, variations of water content at shallow depths, typically within 10-50 m from the surface, amount to changing of the surface boundary conditions for Rayleigh waves. If this model is correct, this feature in microseisms can be used to monitor water content in the shallow crust.
S12B-04 11:05h
Preliminary Application of Microseisms into Groundwater Contamination Monitoring
Microseisms, one scientist's annoying noise are another's diagnostic tool. We are conducting a controlled field experiments with the aim of detecting the infiltration of a contaminant - a biosurfactant - into groundwater. Three sets of instruments are placed 3m, 13m and 32m respectively from a 50m by 50m irrigation site. Each set of instruments consists of a 3-component seismometer and a tilt meter. We are seeking to detect temporal changes in local station corrections that are caused by the irrigation. We use natural signals, such as microseisms as seismic sources and solid Earth tides as sources for the tilt signals. Seasonal changes in the amplitude ratios (horizontal to vertical HZ) of signals from microseisms have been found in California. These seasonal changes are likely to be caused by rather shallow changes in the water table as well as a partial saturated level in the vadose zone. In our field experiment we control the influx of water and monitor it as it percolates down to the ground water. This represents a near ideal arrangement to experimentally check if the HZ ratio can indeed be changed by changes in the local groundwater, or if the cause for the observed seasonal variations has to be found elsewhere. In the laboratory we have found that small additions of some chemicals to water can drastically change the surface energies and thus the wettability of solid surfaces. Surface energy changes in a partially saturated porous rock lead to large changes in complex elastic moduli. In the field experiment we are changing the wettability of the subsurface and are analyzing seismic and tilt data at varying distance from the irrigation site for contaminant caused changes in the moduli. Tilt data show a pronounced change between the three stations during the summer months, probably caused by the differential heating that occurs between the covered irrigation site and the bare ground surrounding it. The observed effect trails off as the instrument's distance from the irrigation site increases. In the seismic data we clearly see the microseismic energy and are now looking for changes in the HZ ratios at the three stations. We keenly anticipate getting and analyzing the seismic records from before, during and following the irrigation of biosurfactants which commences in mid October, 2004.
S12B-05 11:20h
A Comparison of Downslope Propagation, Rough Boundary Effects, and Shear in T-phase Excitation
We compare and contrast different aspects of T-phase excitation. Employing a modal description of the seismic wavefield, it is apparent that if the Earth were a plane-layered, semi-infinite halfspace or a radially symmetric sphere, T-phases would not exist. Even shallow earthquakes are too deep to directly excite the low order modes carrying the T-phase signal. Some mechanism is required to break the strict modal orthogonality and scatter energy into the low order modes. Downslope propagation and a rough ocean bottom can accomplish this. These are in a sense the same process. If the bottom is treated purely as a random rough surface with some characteristic spatial spectrum, the continental or near-island slopes are just long wavelength components of the roughness spectrum. They also have a large correlation length scale. Rough surface scattering is weaker for longer correlation length scales. The actual contribution to the T-phase signal depends on the steepness and length of the slope, and may provide a greater or lesser contribution to the T-phase energy depending on the overall roughness. An additional feature is that if the Earth did not support shear, T-phases would not exist. This is not because a fluid Earth would have smooth boundaries. Rather a fluid Earth will not support interface (Stoneley/Scholte) waves, which seem to be essential for T-phase excitation.