Toward a Standardized Approach to Processing Time-Lapse Ambient Noise Interferometry at Volcanoes
Abstract
Seismic noise interferometry is becoming an increasingly popular technique for monitoring volcanoes. By cross-correlating continuously recorded ambient noise, we can recover changes in seismic velocity that reflect the evolving behaviour of volcanic systems. Despite this, wider usage beyond academic circles remains limited. This, in part, owes to the large influence that different processing choices can have on the final results and complexities in their interpretation. At volcanoes, in particular, the presence of volcano seismicity poses additional challenges in the construction of stable cross-correlation functions. Extra care must therefore be taken to ensure observed velocity changes reflect real changes in the medium. We address this problem by considering whether a standardized approach exists for monitoring volcanoes through ambient noise interferometry. This divides into three main steps : (1) Understanding the wavefield coherence through network covariance matrix analysis (using the Python package Covseisnet), (2) : Exploring the expected quality of cross-correlation functions at different frequencies through time, and (3) : Unsupervised clustering of cross-correlation functions based on waveform similarity. The last two steps are performed using broadband cross-correlation functions, with the goal of using these to better inform key decisions that can have a large influence on measured velocity changes. Further, we use the results of clustering, alongside the wavefield coherence, to better understand if temporal changes in cross-correlation functions reflect a source or a medium change. Combining these steps, we demonstrate their usage at multiple volcanoes with different characteristics, tectonic settings, and network configurations. Case studies include Piton De La Fournaise (Reunion Island), Mt Ruapehu and Whakaari (New Zealand), Kilauea (Hawaii), Stromboli (Italy), and Volcan de Colima (Mexico). In doing so, we show how such measures can be applied to design an optimal processing scheme for real-time ambient noise monitoring at any volcano.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFM.V24A..08Y