Estimating crustal thickness in the central South Island

SKU:
SR_1998-027-pdf
$0.00
(Inc. GST)
$0.00
(Ex. GST)
Write a Review

Wilson, D.; Eberhart-Phillips, D. 1998 Estimating crustal thickness in the central South Island. Lower Hutt: Institute of Geological & Nuclear Sciences. Institute of Geological & Nuclear Sciences science report 98/27 61 p.

Abstract: The goal of this study is to map crustal thickness in the central South Island, New Zealand. Crustal thickness variations are of fundamental importance to understanding the tectonic evolution and current state of the lithosphere. This study falls under the ''mantle'' of the Southern Alps Passive Seismic Experiment (SAPSE). Earthquake waveform data are processed and analysed to determine the crustal thickness and velocity of the lithosphere. Earthquake P and S arrival time data are typically used to determine hypocentres and 1-D and 3-D velocity models of the upper crust. For active-source 2-D profiles with dense receiver arrays, later reflected arrivals also can be used to model crustal velocity and crustal thickness because the arrivals can be appropriately identified by correlating across the array. It is not easy to reliably identify the later arrivals from earthquakes recorded on sparsely distributed receivers, although such strong later arrivals are frequently observed in the SAPSE data. In this study we use earthquakes recorded on both the distributed SAPSE stations and the 2-D dense arrays. Thus, we aim to provide estimates of crustal thickness over a larger region than has been examined with 2-D models of active-sources recorded on 2-D arrays. Crustal velocity models could be determined by fitting sets of earthquake phase arrival times. However, in practice, it is difficult to determine direct, refracted, and reflected phase arrivals at source receiver ranges close to the critical ''cross-over'' point. By sorting a large set of seismic traces into bins with common bouncepoints these phase arrivals should stack coherently and aid interpretation. Thickness and velocities are obtained for each bin by matching arrivals from the stacked dataset with those predicted by either a 1-D forward model, or an inversion algorithm (Richards-Dinger and Shearer, 1997). (auth)