Rayleigh Wave Ellipticity Measurements for Imaging the Crust and Upper Mantle Beneath Portugal

Mainland Portugal has experienced several past large earthquakes both offshore and onshore. Examples of significant past events include the great Lisbon earthquake in 1755 (M ~8.5-8.7) that generated a devastating tsunami, the 1858 M ~7.1 event in Setúbal, and the Benavente earthquake of M ~6.0 in 1909. To better characterize seismicity that pose a potential risk to this region, accurate models of regional Earth structure are required. The lack of a region-wide high quality broadband (BB) seismic network in the past however, has hindered a quantitative characterization of the region’s Earth structure. Since 2006 the permanent Portuguese seismic network has greatly expanded, which resulted in a continuously accumulating large volume of high quality BB data. Within the scope of the FCT-funded project AQUAREL (Accurate QUAntification of Regional Earthquakes and earth structure: application to western Iberia, PTDC/CTE-GIX/116819/2010) project, we utilize this new dataset to measure Rayleigh wave ellipticity from teleseismic events that will be inverted for the seismic structure beneath Portugal.

The horizontal-to-vertical amplitude ratio (H/V ratio) at a given station is defined as the Rayleigh wave ellipticity, which in theory depends only on the structure immediately beneath that particular station. We are measuring H/V ratios from BB seismograms of teleseismic events that occurred between 2009 and 2013 that have hypocentral depths less than 25 km (to reduce the effects of higher modes) and within epicentral distance range 50° – 120° (to minimize interference of other waves). These measurements are made both in time domain at discrete central periods from 25s through 120s and frequency domain using a multitaper method. Theoretical predictions of H/V ratios are calculated using the normal mode summation code MINOES, taking in to account the effects of both fundamental mode and higher mode energy accurate up to 0.1 Hz. With our semi-automated amplitude picking procedure along with standard quality control steps, we are able to measure synthetic H/V ratios with little scatter. Our initial H/V ratios measured from BB data indicate systematic variations with respect to the reference model PREM. We will present initial results from forward and inverse modeling of these measurements.