Part I. Near-source acoustic coupling between the atmosphere and the solid earth during volcanic eruptions. Part II. Nearfield normal mode amplitude anomalies of the Landers earthquake
NOTE: Text or symbols not renderable in plain ASCII are indicated by [...]. Abstract is included in .pdf document. <p>This thesis consists of two chapters. In the first chapter the normal mode theory of a spherical Earth model is extended to include the atmosphere and the theory is applied...
| Summary: | NOTE: Text or symbols not renderable in plain ASCII are indicated by [...]. Abstract is included in .pdf document.
<p>This thesis consists of two chapters. In the first chapter the normal mode theory of a spherical Earth model is extended to include the atmosphere and the theory is applied to understand the observation of air-ground acoustic coupling during volcanic eruptions and to construct synthetic ground motions. In chapter II, the fully developed normal mode theory of 3D Earth is applied to the nearfield amplitude anomalies of the surface waves of the Landers rearthquake. Synthetic seismograms for the recently-available three dimensional seismic global Earth models are constructed using the normal mode theory and compared with observations. The horizontal scale and the location of lateral seismic velocity variations which caused the amplitude anomalies are examined in detail.</p>
<p>Part I:</p>
<p>Long-period harmonic Rayleigh waves were observed by worldwide seismographic networks during the eruption of Mt. Pinatubo in 1991. It has been suggested that these Rayleigh waves were excited, through atmospheric-solid Earth coupling, by atmospheric oscillations set off by the eruption. We investigated this problem using the Earth's normal modes computed for a spherically symmetric Earth model with the solid (elastic) Earth, ocean and atmosphere. These normal modes represent Rayleigh waves in the solid Earth, tsunamis in the ocean, and Lamb waves, internal acoustic waves and internal gravity waves in the atmosphere. Since the atmosphere has a low sound velocity channel below the thermosphere (altitude 90 km), two characteristic acoustic modes with periods of 230 and 270 s exist. The energy coupling between atmospheric acoustic waves and Rayleigh waves is efficient because of the proximity of the horizontal phase velocities of these waves. The energy distribution suggests that a low altitude volcanic eruption would excite the 230 s mode more strongly than the 270 s mode. This is consistent with the observation for the Pinatubo eruption. In contrast, the internal gravity mode has a period of 300 s. The barographic oscillation at a period of 300 s observed for the 1980 Mt. St. Helens eruption is probably this mode. However, because of its slow phase velocity, it would not couple to Rayleigh waves efficiently, and cannot be detected with seismographs.</p>
<p>Part II-A:</p>
<p>The 1992 Landers earthquake ([...]=7.3) occurred in the middle of the TERRAscope network. Long-period Rayleigh waves recorded at TERRAscope stations [...] after travelling around the Earth show large amplitude anomalies, one order of magnitude larger than spherical Earth predictions up to a period of about 600 s. The ground motions over the epicentral region at and after the arrival of R4-5 are in phase at all stations. These observations are inconsistent with the nearly vertical strike slip mechanism of the Landers earthquake. Synthetic seismograms for a rotating, elliptic and laterally heterogeneous Earth model calculated by the variational method agree well with the observed waveforms. Calculations for various 3D Earth models demonstrate that the amplitudes are very sensitive to the large scale aspherical structure in the crust and the mantle. The anomalies for modes shorter than 300 s period can be explained by lateral heterogeneity shallower than the upper mantle. Rotation of the Earth and lower mantle heterogeneity are required to explain mode amplitudes at longer periods. Current whole mantle seismic tomographic models can fully explain the observed amplitudes longer than 300 s. To assess the effect of the high order lateral heterogeneity in the mantle, more precise estimate of the crustal correction is required.</p>
<p>Part II-B:</p>
<p>We modeled the interaction of the source mechanism and the station location with large-scale lateral heterogeneity using the splitting matrix of an isolated multiplet and the 'source-receiver function' whose spherical harmonic coefficients are given by [...] where s and t are angular and azimuthal order numbers respectively. For a short period of time waveform perturbation is proportional to the integral of products of the splitting function with harmonic coefficients [...] and the 'source-receiver' function. For the Landers earthquake and TERRAscope stations source-receiver geometry, the 'source-receiver function' is dominated by the low-order components, paticularly l = 2, m = ±2 in the epicentral coordinates. This beach-ball like pattern is the same for all the near-source stations located in different quadrants of the strike-slip mechanism. The two maxima of the 'beach ball' pattern coincide with the locations of the degree 2 maxima of the splitting functions; western Pacific and east of South America. These features explain the weak dependence of the waveforms on higher order lateral heterogeneity and similarity of waveforms over the epicentral region. The location and the source mechanism of the Landers earthquake relative to the large scale lateral heterogeneity l = 2, including the variations of the cruatal structures, are responsible for the cause of amplitude anomalies near the epicenter. However, the amplitude near the epicenter of an earthquake with a thrust fault type mechanism, for example the Northridge earthquake, is explained well with a spherical Earth model.</p>
|
|---|