DAS full-resolution cable networks in seismology

The rapid adoption of distributed acoustic sensing (DAS) retrofitted on submarine cables has enabled sensor coverage for seismology applications over vast oceanic regions that were previously lacking real-time sensors. The data integration into operational use in seismic networks often involves significant data downsampling and independent processing of single-cable data streams. However, a sparse set of pre-processed measurements from virtual seismic stations along the cable can be very valuable for earthquake early warning and location. Going forward, significantly scaling up data coverage and density is possible with recently introduced DAS technologies. Real-time DAS data from full-resolution cable networks will greatly enhance measurement geometries to improve location accuracy, and advance advanced processing and analysis making use of amplitude information.

When instrumenting multiple cables and jointly processing the data, the limitations of linear cable routing can be overcome to improve location accuracy. This is supported by DAS multiplexing that allows a single instrument to connect to fibres from multiple cables that may be available at a cable landing station. Further range extension of the DAS system will also contribute to geometrical coverage, in particular when reaching the non-straight routing segments tracking seabed topologies off the continental shelf. We will discuss the implications for seismology of recently introduced DAS technologies that extend the spatial coverage.

Scaling up DAS coverage to cable networks and transferring full-resolution data significantly increases requirements to bandwidth capacity and transmission protocols for high-datarate streaming. Users of DAS technologies in the energy sector have faced this challenge and suitable implementations can be adapted for seismology applications. These capacity expansions will support centralized processing at data centres for events recorded on multiple cables, even involving disparate interrogator locations. We describe how the requirements to timing concurrency and data contract management is supported.

While advanced protocols and increased network bandwidth are important enabling technologies for full-resolution DAS monitoring on cable networks, the significant increase in data volumes must also be tackled by developing edge processing techniques that can compute compact pre-processed data products. Recently introduced techniques for sub-array beamforming on pre-defined linear segments of the cable can considerably compress the data. These data products are then used for earthquake location processing at a centralized computing facility with access to the data from all cables in the DAS cable network. We discuss numerically efficient implementations of such edge processing techniques and the potential for relieving data transmission requirements.