Time-Lapse HVSR Analysis for Shallow Subsurface Monitoring at the CaMI.FRS CO2 Sequestration Site

Seismic monitoring is a critical component of Carbon Capture and Storage (CCS) projects, ensuring the containment security of injected fluids and assessing the risks associated with induced seismicity. While fluid injection is known to alter effective stress and pore pressure—potentially inducing velocity changes or fault reactivation—distinguishing these deep subsurface signals from near-surface environmental variations remains a significant challenge. This study utilizes the passive source Horizontal-to-Vertical Spectral Ratio (HVSR) method to investigate the spatiotemporal variations of site response at the CO2 Containment and Monitoring Institute Field Research Station (CaMI.FRS) in Alberta, Canada, providing a robust baseline for long-term integrity monitoring. We analyzed continuous ambient noise data collected between September 2019 and October 2020 from a dense array of short-period seismic stations deployed around the injection well. The injection targets the Basal Belly River Formation at a depth of 300 m. Data processing involved dividing daily records into 150-second windows with 50% overlap, followed by bandpass filtering (0.2–20 Hz) and Konno-Ohmachi smoothing to calculate daily stability-weighted HVSR curves. The results reveal a consistent fundamental resonance frequency (f0) centered at approximately 2 Hz across the study area, corresponding to a soft sediment thickness of 100–150 m overlying the bedrock. While f0 remained relatively stable throughout the monitoring period, the H/V peak amplitude (amplification factor) exhibited significant seasonal time-varying characteristics. Specifically, a strong positive correlation was observed between the amplification factor and environmental variables, including atmospheric temperature, precipitation, and groundwater levels. The amplification factor reached its annual maximum (~2.5–2.6) during the warm, wet summer months (June–August) and dropped to its minimum (~1.5–1.8) during the frozen winter months. These findings suggest that variations in near-surface saturation and soil properties, driven by seasonal climate cycles, significantly modulate seismic site response. Consequently, for effective HVSR-based monitoring of deep CO2 plumes or leakage pathways, it is imperative to decouple these shallow environmental effects from the signals of deep geological alterations. This study demonstrates the efficacy of time-lapse HVSR as a low-cost, non-invasive tool for characterizing site response dynamics and highlights the necessity of multi-physics environmental calibration in CCS monitoring protocols.