Next-generation marine instruments to join plume debate

Whether hot spot volcanism is the consequence of plate tectonics or has a deep origin in a mantle plume is debated. G.~Foulger (Geol.~Soc.~London Lett.~Online, accessed 9/3/2003), writes that carefully truncated cross sections, with color scales cranked up, give noisy images the illusion of strong anomalies traversing the mantle. Don Anderson, the big daddy of non-plume hypotheses (R.~Kent, Geol.~Soc.~London Lett.~Online, accessed 9/3/2003) has written that the resolution of regional tomography experiments must be improved in order to successfully determine whether (...) the deep mantle is the controlling factor in the formation of proposed hot spots (Keller et al., GRL 27 (24), 2000). In particular for Iceland, at issue is the inherently limited aperture of any land-based seismometer array on the island: (...) the resolution of such images could be increased (...) by using ocean bottom seismometers (...) (ibidem). These problems are not unique to the plume debate. Coverage, resolution and robustness of models of the wave speed distribution in the interior of the Earth obtained by seismic tomographic inversions are limited by the areal distribution of seismic stations. Two thirds of Earth's surface are virtually inaccessible to passive-source seismometry, save indeed for expensive ocean-bottom seismometers or moored hydrophones. Elsewhere at this meeting, Montelli et al. describe how an improved theoretical treatment of the generation and survival of travel-time anomalies and sophisticated parameterization techniques yield unprecedented resolution of the seismic expression of a variety of ``plumes'' coming from all depths within the mantle. On the other hand, the improved resolution required to settling the debate on the depth to the seismic origin of various hot spots will also result from the collection of previously inaccessable data. Here, we show our progress in the development of an independent hydro-acoustical recording device mounted on SOLO floats. Our instrument is able to maintain a constant water column depth below the sound channel and will surface only periodically for position determination and satellite data communication. Using these low-cost, non-recovered floating sensors, the aperture of arrays mounted on oceanic islands can be increased manifold. Furthermore, adding such instruments to poorly instrumented areas will improve the resolution of deep Earth structure more dramatically than the addition of stations in already densely sampled continental areas. Our progress has been made in the design of intelligent algorithms for the automatic identification and discrimination of seismic phases that are expected to be recorded. We currently recognize teleseismic arrivals in the presence of local P, S, and T phases, ship and whale noise, and other contaminating factors such as airgunning. Our approach combines continuous time-domain processing, spectrogram analysis, and custom-made wavelet methods new to global seismology. The lifespan and cost of the instrument are critically dependent on its ability to limit its power consumption by using a minimum amount of processing steps. Hence, we pay particular attention to the numerical implementation and efficiency of our algorithms, which are shown to be accurate while approaching a theoretical limit of efficiency. We show examples on data from ridge-tethered hydrophones and expect preliminary results from a first test deployment in October.