Evaluating and Improving Earthquake Early Warning in Central America

Date: 4/25/2019

Time: 06:00 PM

Room: Grand Ballroom

El Salvador, Nicaragua and Costa Rica are exposed to frequent damaging earthquakes (23 M7+ earthquakes occurred in these countries last 100 years). Tsunamigenic thrust events occur at the Middle America Trench where the Cocos plate subducts by 72-81 mm/yr northeastward beneath the Caribbean plate. Extensive damage is also expected from on-shore shallow crustal events, as demonstrated by the 1972 Managua M6.2 earthquake. This seismotectonic setting is challenging for Earthquake Early Warning (EEW) because relevant events include both moderate-sized on-shore seismicity that is often in close proximity to densely urbanized areas, where early warning alerts for areas experiencing strong shaking would at best be very short and often late; and tsunamigenic offshore seismicity, where large lead times are possible but the finite source must be accounted for, and hypocentral distances are uncertain. Since 2016, the Swiss Development and Cooperation Fund has funded a collaboration between Swiss and regional partners to build and evaluate prototype EEW systems in Central America. The EEW algorithms Virtual Seismologist (VS) and Finite fault rupture Detector (FinDer) can be operated by any network running the popular SeisComP3 network processing and management suite. Currently, these algorithms are being run at MARN in El Salvador, RSN in Costa Rica and INETER in Nicaragua. In this contribution, we assess how fast and accurate EEW could be at these networks. We estimate the current and potential performance of EEW alerts by modeling alert time delays using various existing and possible regional network configurations, taking into account communication and processing delays. Models are compared with the on-going observed performance obtained by the VS and FinDer algorithms. We further highlight current capabilities and challenges for providing EEW alerts in Central America. We also discuss how to get a robust approach to EEW from combining different algorithms and from combining class A accelerographs with large numbers of low cost accelerographs.

Presenting Author: John Clinton

Authors