The Scientist

I recently presented a poster at the 2022 Fall AGU Annual Meeting. The study was titled: Temporal Velocity Variations Associated With The 2020 Eruption Of  Kīlauea Volcano in Hawai'i, Revealed by Ambient Noise Autocorrelation. Contact me if you'd like to see the poster.

• Abstract: Volcano eruptions are some of the most destructive geological phenomena on Earth. Detecting and understanding changes in the magmatic system that occur before/during an eruption is essential for forecasting these environmental and societal hazards. Seismic ambient noise analysis provides a non-traditional approach to observing velocity variations in the crust, and supplies information about perturbations in the interior of the volcano that can be challenging to resolve with ground deformation detection techniques. Cross-correlation of ambient noise has been used in various studies to examine pre-eruptive activity. Here we investigate temporal velocity variations before and during the 2020-2021 eruption of K ̄ılauea Volcano in Hawai‘i, which was the first major activity after the dramatic 2018 eruption and may provide insight into how a volcano readjusts itself after such a massive event. We download one-hour segments of waveform data from the Incorporated Research Institutions for Seismology (IRIS) spanning 60 days before to 30 days after the onset of the eruption (December 20, 2020). We focus on data recorded by 12 broadband seismic stations in and around the summit caldera operated by the U.S. Geological Survey Hawaiian Volcano Observatory. After removal of trend, mean, and instrument response, the data are resampled to a uniform 100-Hz sample rate and bandpass filtered from 1 to 5 Hz. We then calculate the autocorrelation functions of these hourly segments and stack them for each day during the study time period. We apply a time-domain stretching method to the stacked function pairs and solve for the best-fitting set of velocity changes for the entire time period. We observe a velocity decrease of about 0.5% several days before the eruption, which may be related to magma or fluid/gas movement in the crust or an increase in pressure. Our study shows that auto-correlation of ambient noise continues to show promising results for the forecasting of volcanic eruptions.