Accuracy of locating earthquakes and landslides in sparse seismic network with ambient noise Empirical Green’s Functions

Earthquakes and landslides threaten the safety of human life and property. Fast and accurate location of earthquakes and landslides is critical to disaster mitigation and early warning of the secondary hazards. Hazards locating is a difficult issue in a remote region with sparse seismic network due to low resolution of the velocity structure model. Green’s functions from ambient seismic noises contain the information characterizing the anomalous velocity structure along the propagation path, and it can be used to calibrate effects due to the uncertainty of velocity structure, thereby improving the hazard location accuracy. In this study, we select the 2008 Wudu Ms5.5 earthquake, China, and a large landslide occurred in Nuugaatsiaq, Greenland, on 17 June 2017 as examples, to assess the accuracy of the relative location method based on Green’s functions from ambient seismic noises. The location result of the landslide is about 2.5 km away from the site given by satellite image, which is better than the result based on traditional location method, with a deviation up to ~17 km. Subsequently, we test some impact factors of the location accuracy via the 2008 Wudu earthquake, such as the epicentral distance of the reference stations and the networks with different sparseness. It shows that using a reference station within 30 km and about 4 remote stations for relocation, the relocation accuracy is about 5 km. Our results demonstrate that this algorithm can provide accurate location of earthquake and landslide with seismographic stations in global and regional networks, thus providing timely assistance to early warning of secondary hazards.