Skip to content
Open Data Finder
data.gov.au

Probabilistic tsunami inundation assessment from a near-field earthquake source

Published by Geoscience Australia Data

About this dataset

Tsunamis are a destructive and dangerous hazard for coastal communities. To prepare for future events, Probabilistic Tsunami Hazard Assessment (PTHA) can be used to estimate the frequency and severity of tsunamis. Large-scale PTHAs have been developed by modelling a suite of tsunami events to robustly represent all possible earthquake-generated tsunamis (Behrens et al., 2021). The 2018 Probabilistic Tsunami Hazard Assessment (PTHA18) is a global PTHA that serves as a database of offshore tsunami frequency and wave data (Davies & Griffin, 2018). In this presentation, we will demonstrate the efficient conversion of offshore PTHA18 data to onshore probabilistic inundation data (Figure 1) through the modelling of a stratified random sample of tsunami events. The town of Vanimo in Papua New Guinea (PNG) is used as a case study. We consider the tsunami threat to the north coast of PNG from earthquakes generated in the New Guinea subduction zone. PNG does not currently have a comprehensive tsunami warning system. However, in the case of the New Guinea trench, ground shaking from a large earthquake may provide communities with some warning and, should this occur, PNG government advice is to evacuate to locations 800 m from the coastline (Kawata et al., 1999). Results suggest that this directive is a safe one and inundation further than 800 m inland is estimated to occur with a very low frequency. Evacuation to this distance from the coast may, however, not be achievable within the short period of time between a large earthquake on this trench and the resulting tsunami. Tsunami arrival times are estimated to be <10 min at an average frequency of roughly once in 300 years (Annual Exceedance Probability of 0.333). For Vanimo specifically, this guidance may be best followed in combination with evacuation to at least 5 m elevation. This allows for evacuation to the high elevation Vanimo peninsula (Figure 1) which is close to population centres on the isthmus, and is estimated to be inundated less frequently than other sites within 800 m of the coastline. The offshore PTHA maximum water levels above Mean Sea Level (MSL) show a positive linear correlation with onshore maximum water levels. Linear regression data for every 10x10m cell in the model domain shows how the elevation features influence tsunami wave amplification. The linear regression coefficient at the coastline (0 m elevation) is between 1 and 1.6 with R2 varying between 0.8 and 0.95. Shoaling is apparent in the nearshore and amplification is greater on the exposed, seaward side of the Vanimo peninsula than in more sheltered locations. Such variation makes a good case for applying comprehensive probabilistic inundation as opposed to simpler methods. Conclusions may be applicable to other locations with a near-field source. Presented at the 24th International Congress on Modelling and Simulation (MODSIM2021)

Probabilistic tsunami inundation assessment from a near-field earthquake source · Open Data Finder