Dependence of Moment-tensor Solutions on Source Location Observed at Pacaya Volcano, Guatemala

Abstract:

Synthetic modeling aimed at measuring the capability of a seismic network to resolve source mechanisms can provide a guide to the deployment of sensors on volcanoes. Recovering the source mechanisms of events is especially challenging because at frequencies of about 1 Hz, which are common for volcanic sources, scattering strongly influences seismic recordings. The focus of this research is to explore the trade off between the number and location of seismic stations and the accuracy of seismic source reconstructions in the presence of heterogeneous structures. We investigate this relationship at Pacaya volcano, Guatemala. During a fieldwork campaign in October-November 2013, four 3-component broadband seismometers were installed around the central vent at distances between 0.6 and 1.5 km. In addition to tremor, the network recorded a long-period event that repeated thousands of times each day. In order to determine the optimal deployment strategy for the next field campaign, we conduct a sensitivity analysis using synthetic seismograms. The repetitive nature of the source and the accessibility of the volcano will facilitate deployment of a spatially-dense seismic network, in which a subset of stations is moved around the cone to enable records from dozens of sites. We used then different subsets of stations and velocity models to test the expected capability of the network to extract a reliable moment-tensor. Preliminary results highlight a strong dependence of position on the moment tensor solutions. The source mechanism changes from a sill to a dyke as the source becomes deeper. It is therefore critical to get an accurate location to better reconstruct the source mechanism. The results of this study have broad implications for volcano seismic source studies, which often involve repetitive events, but typically face the same challenges of heterogeneous, but poorly constrained structure and weak, shallow sources.