Applications of time series analysis to geophysical data

Abstract

This thesis consists of three papers applying the techniques of time series analysis to geophysical data. Surface wave dispersion along the Walvis Ridge, South Atlantic Ocean, is obtained by bandpass filtering the recorded seismogram in the frequency domain. The group velocity is anomalously low in the period range of 15-50 s, and formal inversion of the data indicates both crustal thickening to 12.5 km and low shear velocity (4.25-4.35 km/s) to depths of 40-50 km. The electromagnetic induction fields at a deep ocean site northeast of Hawaii were used to determine the electrical conductivity of the earth to 400 km depth. Singular value decomposition of the data matrix indicates three degrees of freedom, suggesting source field complications and a two dimensional conductive structure. Inversion of one of the principal terms in the response function shows an abrupt rise in electrical conductivity to 0.05 mho/m near 160 km with no resolvable decrease below this. A model study suggests that moving source fields influence the induction appreciably in the other principal response tunction. A set of piston cores from the northeast Atlantic Ocean are used to construct paleomagnetic time series covering the interval 25-127 kybp. Stratigraphic control is provided by counts of planktonic toraminifera, and empirical orthogonal function analysis shows a significant decrease in sedimentation rate at the interglaciai/glacial transition. The sediments are magnetically stable and reliable relative paleointensity measurements could be obtained. Spectral analysis of the directions reveals a predominant 10 ky periodicity and no dominant looping direction.