Estimating Ice Cover on the Great Lakes Using Seismic Ambient Noise

Abstract Lake ice impacts seismic noise on nearby seismometers, and therefore seismic observations can be used to monitor ice changes. However, the transfer function describing how lake‐microseism noise covaries with ice cover has not been quantified. We use seismic data from a long‐operating station near Lake Superior and satellite‐derived lake ice cover from 2009 to 2024 to determine their covariance. A linear model calibrated using annual ice cover averaged during the coldest months (15 February–9 March) explains ∼87% of the variance in 0.5–2 Hz seismic noise averaged during the same period. This 15‐year‐calibrated model is used to predict daily ice cover during the 2014 Polar Vortex with reasonable accuracy, explaining ∼58% variance in the daily 0.5–2 Hz noise. One season of daily lake ice cover data is sufficient to calibrate the model. During unfrozen periods, 0.25–0.5 Hz lake‐wave power explains ∼75% of the variance in the daily averaged 0.5–1 Hz seismic noise, consistent with the secondary microseism mechanism.