Uncertainty of the GPS-Based Tsunami Forecasting: A Case Study for the April 2014 M8.0 Pisagua, Chile Earthquake

Abstract:

Real-time GPS is nowadays considered as a valuable component of next-generation near-field tsunami early warning systems. Whereas the idea of using GPS for tsunami early warning is clear, its practical implementation still requires extensive studies and validation against real events. The northern Chile 1 Apr. 2014 M8.0 Pisagua earthquake and resulting tsunami was extensively recorded by large number of land- and ocean-based sensors including real-time GPS. Earthquakes of this magnitude range (M7.0-M8.0) are especially challenging for tsunami warning centre since, in contrast to the mega-earthquakes, they do not necessary trigger devastating tsunamis. GPS-fingeprints of such earthquakes may also approach the limit of real-time detect ability. Thus, earthquakes of such a magnitude range fall in "a blurred zone", a zone of higher uncertainty in sense of tsunami early warning.

We take the opportunity and consider the 2014 Pisagua event as a case study to explore possible sources and magnitudes of uncertainty in the GPS-based real-time tsunami forecasting. In particular, we test uncertainties related to derivation of real-time source model: (1) Uncertainty of co-seismic displacements retrieved by different real-time GPS-processing strategy: precise relative positioning and precise point positioning; by different software platform: RTKLIB, PANDA, EPOS-RT, Bernese (post) (2) Uncertainty of source inversion by different approaches: single Okada fault, distributed slip in homogeneous and layered half space. We consider individual uncertainties as well as their propagation to the final tsunami forecasting. As a result, “up streams” and “down streams” are combined together, in which case a “close loop” criteria is illustrated for evaluation of uncertainty of the GPS-based tsunami forecasting.