Intrinsic Uncertainty of Noise Cross-Correlation Functions and Its Implications for Source Properties

Abstract This article presents a quantitative approach for evaluating the intrinsic uncertainty (IU) of a noise cross-correlation function (NCF) and applies it to realistic NCFs obtained from broadband seismic arrays in Taiwan. The evolution of the standard deviation (STD) of the NCF waveform with increasing correlation time closely aligns with theoretical predictions, confirming the validity of the method. The derived IU provides an objective metric for assessing the NCF quality, particularly for coda waves, which are widely used to monitor temporal structural variations. Through an analytical procedure applied to a synthetic experiment, the results demonstrate that the distribution and excitation strength of noise sources strongly influence the IU. Consequently, combining IU estimates with NCF amplitudes derived from different normalization approaches enables a comprehensive assessment of ambient noise source characteristics. Using the proposed approach, this study reveals the excitation mechanisms of short-period secondary microseisms, local primary microseisms, and primary microseisms.