The atmosphere is a powerful generator of infrasound. And here it turns from an "interesting fact" into practice: low frequencies are already used to warn of real threats. But the same weather poses the most insidious task for monitoring systems.
A tornado "hums" in advance
Strong thunderstorm vortices radiate infrasound. Alfred Bedard already linked low-frequency acoustics to the vortices of thunderstorms2, and modern measurements revealed the characteristic infrasound of a tornado-producing storm — with the signal appearing before the funnel touched the ground.1 This opens the way to extra minutes of warning where radar lags behind.
Avalanches: already in operation
This is not a laboratory but a working system. Infrasound arrays of 4–5 sensors reliably detect the release of large avalanches at a distance of 3–5 km, in any weather and with zero visibility, and determine the speed of the front.34 Commercial installations (for example, IDA) guard dangerous slopes above roads and villages around the clock.5 The same arrays detect approaching lahars and debris flows minutes before impact — already a basis for early warning (Johnson et al., 2023; Marchetti et al., 2019).
A passing atmospheric front produces a spatially coherent pressure change across many sensors at once — exactly what an event-detection algorithm is looking for. Telling geophysical infrasound from weather noise is a real scientific problem, not a trifle. It is solved by combining features: the wave's speed and azimuth, its spectrum, its link to meteorological data. In practice the signal is extracted by array correlation (the PMCC method),6 while large analyses of global IMS network data show how to separate incoherent wind noise from spurious coherent signals.7
- A tornado's infrasound sometimes appears minutes before the funnel touches the ground.
- Avalanche stations work in darkness, fog and blizzards — when both the eyes and radar are helpless.
- From the signal delay between an array's sensors, one can even compute the speed of an avalanche front.
- Songbirds flee tornadoes in advance: golden-winged warblers evacuated 1–2 days before a tornado outbreak, apparently "hearing" the storm's infrasound from hundreds of kilometres away (Streby et al., 2015).
Avalanches prove it: cheap local infrasound networks are already saving lives. We are learning to tell a "real event" from weather doppelgängers — and that is exactly the scientific work the network is being built for.
Sources for this article
- peer-reviewed Elbing B.R., Petrin C.E., Van Den Broeke M.S. (2019). Infrasound from a tornado-producing storm. JASA 146(3). pubs.aip.org
- peer-reviewed Bedard A.J. (2005). Low-frequency acoustic energy associated with vortices produced by thunderstorms. Mon. Wea. Rev. 133(1). journals.ametsoc.org
- peer-reviewed Marchetti E. et al. (2015). Infrasound array detection of snow avalanches. NHESS 15. nhess.copernicus.org
- peer-reviewed Mayer S. et al. (2020). Performance of an operational infrasound avalanche detection system. SLF. slf.ch
- organization Wyssen Avalanche Control. IDA® Infrasound Detection of Avalanches. wyssenavalanche.com
- peer-reviewed Cansi Y. (1995). An automatic seismic event processing for detection and location: the PMCC method. GRL 22(9). doi.org
- peer-reviewed Vergoz J. et al. (2022). IMS infrasound data products for atmospheric studies and civilian applications. Earth Syst. Sci. Data 14. essd.copernicus.org
- peer-reviewed Streby H.M. et al. (2015). Tornadic storm avoidance behavior in breeding songbirds. Current Biology 25(1). doi.org
- peer-reviewed Johnson J.B. et al. (2023). Infrasound detection of approaching lahars. Sci. Rep. 13. doi.org
- peer-reviewed Marchetti E. et al. (2019). Infrasound array analysis of debris flow activity and implication for early warning. JGR Earth Surface 124. doi.org