The oceans are home to many of the world's most significant plate margins, including subduction zones, the monitoring of which is critical for precisely understanding the associated risk of earthquakes and tsunamis. While satellites, laser and GPS observations enable routine millimetre monitoring of plate movement on land, until recently the inability to undertake complementary underwater measurements at the required precision for extended time periods, especially in the outermost subduction zone offshore, has been a critical flaw. This is where much of the elastic strain build-up and release occurs and consequently improving scientific understanding of the key processes in these regions is an important basis for future seismic hazard assessment. In response to this challenge, GPS-Acoustics (GPS-A) techniques have been deployed in a number of projects, but early GPS-A deployments suffered from limited seabed transponder endurance and the requirement for extended periods of costly ship time to position these on the seabed A similar, but different, approach is to use instrumentation originally developed by the offshore industry for precise measurement of seabed deformation for research into submarine earthquakes. Deployments of these seabed transponder networks provide precise relative positioning between the transponders, which has enabled observation of deformation at a fine scale. The largest and most challenging of these projects so far has been the GeoSEA project (Geodetic Earthquake Observatory on the SEAfloor), which GEOMAR first deployed in Nov/Dec 2015 off the west coast of Chile on the Nazca-South American plate boundary as the world's first ever subduction zone acoustic monitoring system. The deployment location last ruptured in an earthquake in 1877 and was identified as a seismic gap prior to the 2014 Iquique/Pisagua earthquake, which triggered a 2.1 metre tsunami. The array consists of 23 Autonomous Monitoring Transponders (AM7), initially deployed for a period of 3.5 years in 3 sub-arrays ranging in depth between 2,746 5,364 metres, and is designed to accurately measure crustal deformation as a proxy for crustal strain build-up. The AMTs run a fully automated logging regime gathering acoustic travel time (range) between neighbouring units, pressure, sound velocity, temperature and tilt data. Bi-directional acoustic communications enable this data to be harvested on demand and/or amendments to the logging regime. Since its initial deployment, data from the array has been periodically recovered by German and US research ships as well as GEOMAR's GeoSURF Wave Glider. This paper outlines the scientific and technological challenges of planning, establishing and sustaining the GeoSEA array and, using initial data from the project, show how it is beginning to meet its scientific goals.
