Extraction and Analysis of Seismic Body Waves from Ambient Noise for Crustal Imaging on Montserrat, with Implications for Temporal Monitoring

Continuously recorded seismic data acquired with closely spaced (ca 100m) stations deployed during the SEA-CALIPSO seismic experiment on Montserrat in December, 2007, were processed using ambient noise techniques to assess their utility for imaging volcanic structures at depth with body waves. Approximately 3 days of data were acquired along 3 quasi-linear arrays during the experiment, primarily to record airgun and microseismic sources. Cross-correlation was used to generate virtual shot gathers that were subsequently processed to enhance coherent events. Direct P wave arrivals were consistently produced by the cross-correlation process, but extraction of credible deeper reflections was less consistent. The dataset was partitioned to isolate recordings that were dominated by coherent airgun and microearthquake arrivals from those that appear to contain primarily ambient noise. The virtual gathers based on data with airgun energy appear to provide better imaging of deeper structure, in particularly subhorizontal reflections that may mark subsurface sill complexes. Imaging deep structure in active volcanic crust has been notoriously difficult with conventional controlled source methods, so the uneven quality of the data recovered by the ambient noise processing here is not unexpected. However, these results suggest that ambient noise techniques are a viable alternative to traditional controlled source techniques, especially in physically inaccessible regions, albeit longer recording times than available with this dataset are needed. Furthermore, by eliminating the need for expensive sources, this approach suggests a more practical approach to monitoring of temporal variations in reflectivity associated with movement of subsurface magma.