Using rapid source characterisation to improve ShakeMaps and impact forecasts for large earthquakes (M6.5+) in New Zealand

The New Zealand R-CET Endeavour programme has been developing a suite of tools to characterize the earthquake source and its shaking in near real-time.  A key goal is to provide pathways to improve rapid earthquake impact forecasts.  

Small-to-moderate earthquakes can be reasonably represented by a ‘point source’.  This allows us to generate first shaking maps (GNS Shaking Layers; Horspool et al. 2023) automatically and robustly based on basic earthquake solutions (magnitude and hypocentre). These maps are now routinely available to the public within 10 – 20 minutes: https://www.geonet.org.nz/about/earthquake/shakinglayers#:~:text=What%20is%20Shaking%20Layers%3F,intensity%20anywhere%20in%20the%20country.

For very large earthquakes (M6.5+), a ‘point-source’ is a poor representation of the earthquake, which can rupture tens to hundreds of kilometres of the earth.  First shaking models based on ‘point sources’ could severely underestimate shaking in areas further from the epicentre, but close to the fault rupture. Rapid 3D characterization of the rupture area, even if approximate, has the potential to significantly improve shaking estimates, and allow meaningful first impact forecasts to be generated.

Here we present an overview of rapid source characterization tools implemented for New Zealand under the R-CET programme. These tools include FinDer (Andrews et al. 2023), w-phase (Fry et al. 2022), EQCorrScan (Chamberlain et al. 2017; Warren-Smith & Chamberlain 2022), G-FAST and others . We show examples of tool outputs for large (M6.5+) historical earthquakes in New Zealand and examine their potential to improve rapid shaking models, loss estimates and landslide forecasts. Our results show the importance of including rapid source characterization as a key component of our earthquake response systems, to underpin quality scientific advice for emergency responders.