Effect of thermoelastic strain inferred from a single station ambient noise studies in the context of the Magna M5.7 event in Utah

Abstract

Single station ambient noise analysis provides a practical approach for monitoring seismic velocity changes (dv/v) but its sensitivity to shallow structure makes it susceptible to environmental influences at higher frequencies. We investigate dv/v variations at seismic station UU.NOQ, the nearest three component seismic station to the 2020 M5.7 Magna, Utah earthquake epicenter. We observe a 0.1% coseismic dv/v drop at 1-2 Hz. However at higher frequencies we observe an annual cyclical variations dominating the dv/v signal which obscured the earthquake related dv/v drop. Through cross correlation and coherence analyses comparing dv/v with temperature, precipitation, and GPS derived groundwater proxies, we give evidence that shows temperature as the primary driver of seasonal dv/v variations. Notably the dv/v signal lags surface temperature by approximately 55 days. We apply thermoelastic strain modeling to explain this lag which can be explained due to the presence of an unconsolidated sediment layer that acts as a thermal delay filter. The estimated layer thickness of approximately 1.5 m is independently verified using Horizontal to Vertical Spectral Ratio analysis, which shows a fundamental site frequency above 5 Hz consistent with thin sedimentary cover. Our results demonstrate that thermoelastic strain can significantly influence shallow dv/v measurements in certain scenarios and provide a methodological framework for characterizing temperature induced effects in similar kind of ambient noise monitoring studies.