A variability-based framework for soil monitoring and early warning of CO2 leakage at CCUS demonstration sites

In Carbon Capture, Utilization and Storage (CCUS) demonstration sites, reliable monitoring of CO₂ leakage is essential for operational safety and risk management. However, practical CCUS projects face challenges in selecting soil monitoring indicators that are both sensitive to leakage signals and robust against environmental interference, making early-warning threshold definition difficult. In this study, a variability-based monitoring framework integrating the coefficient of variation (CV), incremental variability (ΔCV), and an environmental interference coefficient (EC) is proposed to systematically screen soil indicators for CO₂ leakage detection. Indoor and outdoor CO₂ leakage simulation experiments were conducted under CCUS demonstration site conditions, with leakage rates ranging from 0.005 to 0.04 L/min, and dynamic monitoring of 15 soil indicators. The results show that with increasing leakage duration and rate, soil CO₂ concentration, ammonium nitrogen, and bicarbonate (HCO₃⁻) increase, while soil pH decreases. Based on the proposed framework, soil CO₂ concentration and pH are identified as a minimal and robust monitoring indicator set, exhibiting the highest leakage sensitivity and the lowest susceptibility to environmental interference. Conductivity and HCO₃⁻ are classified as secondary indicators, while ammonium nitrogen serves as a tertiary indicator. Furthermore, operational early-warning thresholds and leakage stages are established based on indicator response transitions, with soil CO₂ concentration exceeding 1400 ppm and pH decreasing below 5.4 indicating significant leakage. The proposed framework and monitoring strategy provide a practical and cost-effective approach for early detection and risk management of CO₂ leakage in CCUS demonstration sites.