Comparing Methods and Defining Practical Requirements for Extracting Harmonic Tidal Components from Groundwater Level Measurements

Abstract The groundwater pressure response to the ubiquitous Earth and atmospheric tides provides a largely untapped opportunity passively characterize quantify subsurface hydro-geomechanical properties. However, this requires reliable extraction of closely spaced harmonic components with relatively subtle amplitudes but well-known tidal periods from noisy measurements. minimum requirements for suitability existing records analysis are unknown. This work systematically tests compares ability two common signal processing methods, discrete Fourier transform (DFT) least squares (HALS), extract component First, realistic conditions simulated by analyzing large number synthetic data sets variable sampling frequencies, record durations, sensor resolutions, noise levels gaps. Second, model real-world different characteristics is validated. results reveal that HALS outperforms DFT in all aspects, including handle While there clear trade-off between frequency duration, rates should not be less than six samples per day shorter 20 days when simultaneously extracting constituents. accuracy detection degraded increasing decreasing resolution. resolution same magnitude as expected amplitude sufficient absence excessive noise. provide practical framework determine level can optimize future monitoring strategies improve passive characterization using signatures.