Crustal thickness from seismic noise correlations in preparation for the InSight mission to Mars
by Gesa Karen Becker
Date of Examination:2018-06-05
Date of issue:2018-08-31
Advisor:Dr. Brigitte Knapmeyer-Endrun
Referee:Prof. Dr. Ulrich Christensen
Referee:Prof. Dr. Laurent Gizon
Files in this item
Name:ThesisBecker.pdf
Size:8.44Mb
Format:PDF
Description:Dissertation
Abstract
English
The formation of terrestrial planets is still poorly understood. Studying the interior of Mars by analysing seismic waves can help to answer questions about that process. In the absence of plate tectonics, Mars has retained much of its crust from early stages of the planet’s evolution. Knowledge of its crustal thickness and structure can therefore contribute to the understanding of terrestrial planet formation and evolution. In preparation for the InSight seismometer installation on Mars, a method is developed to image the reflectivity of the subsurface and extract the crustal thickness from ambient seismic noise. The developed method makes use of passive seismic interferometry by calculating auto- and single-station cross-correlations to obtain the reflection response of the subsurface. It is then tested on seismic stations across Central and Eastern Europe with varying crustal thicknesses, comparable to estimates for the InSight landing site. A processing scheme of spectral smoothing, phase auto- or cross-correlation, time-frequency domain-phase weighted stacking and bandpass filtering, predominantly between 1-2 Hz, is introduced. This processing scheme works well for the vertical and horizontal component data. One autocorrelation result is obtained for each station from the vertical component data. The horizontal component data gives two autocorrelation results, in the North-South and East-West directions, and two cross-correlation results of mixed directions. The obtained correlations are spatially and temporarily stable, but for all components a correlation with the presence of cultural noise can be observed. The vertical component correlations converge to a stable result after ∼6 months of data, whereas the horizontal component correlations require more data, especially the cross-correlations. Prior information, in the form of seismic velocity and the predicted Moho depth, are included as an uncertainty target window of expected reflectivity change. The reflectivity change associated with the Moho is automatically determined inside the uncertainty window in case of the vertical component autocorrelations. For the horizontal component correlations, this needs to be done manually to include all four correlation results. The determined lag times are converted to depths using the prior information about the seismic velocities. The obtained Moho depths compare well with results obtained from other studies using different methods and are consistent across vertical and horizontal component correlations. Furthermore, it is possible to identify additional deeper and shallower reflectors. The combination of the determined lag times of Moho related reflectivity change from the horizontal and vertical component correlations gives the possibility to obtain estimates of the vp/vs-ratios. However, these ratios show large variability and sometimes unreasonable values and therefore need to be treated with caution. This variability can be explained by e.g., non-vertical incidence, and the resulting different paths of the waves. The horizontal component autocorrelations can nevertheless be used to detect possible crustal anisotropy below the station. In general, crustal thickness can be extracted from seismic noise correlations for a single station, for different seismometer types, when only moderate amounts of data are available, without the need for strong seismic sources, and independent of the structure of the subsurface. All this makes the introduced method a promising tool for the InSight mission.
Keywords: seismic noise; Mars; seismic interferometry; Europe; crustal structure