Exploiting infrasound to probe the dynamics of the middle atmosphere

Abstract

This doctoral thesis examines the middle atmosphere (30 - 80 km altitude) using ground-based observations of low-frequency sound waves (infrasound). By exploiting infrasound waves, the thesis aims to increase observations in this challenging region and improve climate and weather prediction models. Four papers are included in the thesis and cover a broad range of topics. One study compares observations and simulations of ocean-generated infrasound waves called microbaroms using an array processing recipe. Analyzing the directional distribution of microbaroms over time reveals biases in the atmospheric models used for simulation. Identifying and addressing these biases can improve the representation of stratospheric dynamics in the models. In another study, vespa analysis investigates the role of ocean-generated infrasound in generation of high-speed polar mesosphere winter echoes. Three radar-observed echoes observed in Sweden are analyzed, proposing and discussing possible origins. A separate study uses observations from high-latitude infrasound arrays to characterize circumpolar stratospheric circulation in 60° – 90° latitudes. A machine learning model maps time series of infrasound observations to zonal wind at around 50 km altitude, providing a near real-time wind characterization. This is the first time when infrasound data and machine learning are combined to characterize the wind in the middle atmosphere. Additionally, the thesis demonstrates the presence of gravity-wave signatures in ground-based infrasound observations using explosion-generated signals. Analysis of sound speed fluctuations and their spectra associates them with atmospheric gravity waves. Comparisons with radar wind measurements show agreement, suggesting that infrasound can extend the observational range of atmospheric gravity-wave scales. Overall, this thesis highlights the potential of infrasound observations to improve the representation of middle atmosphere in climate and weather prediction models. Analyzing infrasound data, identifying biases, and exploring relationships with other atmospheric phenomena can provide valuable insights for improved understanding and forecasting of the middle atmosphere.