The Multi-Channel Maximum-Likelihood (MCML) method: a multiple source extension for infrasound detection and wave parameter estimation

Summary Infrasound provides a unique capability to monitor remote atmospheric sources over thousands of kilometers, playing a key role in global monitoring. Extracting wavefront parameters, such as the direction of arrival (DOA) and trace velocity, remains challenging in low signal-to-noise ratio (SNR) conditions and in the presence of interfering sources. The Multi-Channel Maximum-Likelihood (MCML) algorithm proved to outperform correlation-based detectors like the Progressive Multi-Channel Correlation (PMCC) in terms of detection and estimation capabilities in these low SNR scenarios. Observations from the International Monitoring System (IMS) show that signals from multiple sources may interfere in the time-frequency domain. However, MCML is able to detect only one source for each time-frequency bin. Several extensions of the MCML algorithm have been developed in order to distinguish multiple sources, including a multiple maximum MLE estimator, a spectral decomposition approach, and an iterative signal subtraction method. Comparing detection and estimation capabilities through simulations having ground truth information with state-of-the-art multiple source algorithms, like MUSIC, shows that the iterative signal subtraction method is the most effective method while maintaining a low computational cost. One major finding is that the monosource detector can still be used by the multiple source extension of MCML avoiding costly computation for a multiple source detector. Processing real events such as rocket launches and explosive meteors highlights the advantages and limitations of each algorithm. In particular, sources masked by impulsive events can still be detected when the SNR contrast remains moderate. This extension leads to a refined analysis for a better understanding of the environmental noise, needed for the operational monitoring of the Comprehensive Nuclear-Test Ban Treaty (CTBT).