Response Function Estimation of Marine Magnetotelluric Data Measured at Tristan de Cunha Hot Spot

Abstract:

According to classical plume theory, the Tristan da Cunha hotspot is thought to have played a major role in the rifting of the South Atlantic margins and the creation of the aseismic Walvis Ridge by impinging at the base of the continental lithosphere shortly before or during the breakup of the South Atlantic margins. To understand the tectonic processes of the opening of the South Atlantic, the formation of the Walvis ridge and to understand, whether Tristan da Cunha is the cause or the consequence of the rifting, a network of 26 ocean-bottom magnetotelluric stations (OBMT) (cooperation between GEOMAR Germany and the Earthquake Research Institute at the University of Tokyo) and 24 broadband ocean-bottom seismometers was in operation during February 2012 and January 2013 within the framework of the project ISOLDE during the second phase of the SPP 1375. To complement the ocean-bottom network and the existing stations on Tristan da Cunha, one magnetotelluric land station and two land seismometers were installed on Nightingale Island.

Marine magnetotelluric data is often contaminated with different levels of non-stationary noise originating from water waves, passing ships and tilt movements of the instruments. In traditional data processing, such noisy segments are removed from the time series before processing. As a consequence only a portion of data remains for response function estimation and accordingly no information about longer periods (deeper structure) remains. However, since the focus of this study is melt anomalies in the upper mantle, long period estimators of the response function are vital. To investigate which processing method yields the more reliable estimates between tens to 100 000 seconds (i.e. penetration depth from a few kilometers to approximately 600 to 800 km) for our recorded data, we applied and compared different sophisticated preprocessing and processing methods (e.g., pre-whitening, one-stage BIRRP, advanced two-stage BIRRP and multivariate EMD methods). In the poster, we present a comprehensive comparison and discuss advantages and disadvantages of the different methods.