Temporal evolution and decline of seismicity associated with cyclic operations at the Serajeh underground gas storage, Iran

Underground gas storage (UGS) operations can perturb reservoir conditions and modulate local seismicity, but the temporal evolution of this response under cyclic injection–withdrawal remains poorly constrained. We investigate seismicity around the Serajeh UGS in central Iran by combining 29,781 deep-learning-based single-station earthquake detections, multi-station relocation, centroid moment tensor inversion, and operational records. The relocated hypocentres form a compact cluster near the Serajeh structure, primarily at depths of 3–8 km, overlapping with and extending below the 2.4–3.0 km storage interval. Independent comparison with relocated earthquakes confirms that S–P time at the nearby GHVR station provides a useful first-order proxy for local source distance. Cross-correlation between monthly gas injection and seismicity shows a distinct maximum at an offset of approximately 8 months. This feature is reproduced by the phase-randomized ReshuffleCorr analysis, while additional structure-preserving tests indicate that its global statistical strength and exact timing depend on the treatment of temporal variability. An objective change-point analysis identifies a marked decrease in seismicity centred on April 2015, nearly four years after storage operations began, and the decrease remains evident using a conservative magnitude threshold. Seismicity subsequently remains lower despite continued cyclic operations. The selected focal mechanisms show predominantly strike-slip to oblique source kinematics across pre-operational, operational, and post-operational periods. We interpret these observations as evidence for an evolving temporal and spatial association between storage operations and a pre-existing fault–reservoir system. These results highlight the value of integrating high-resolution seismic monitoring with independently constrained locations and operational records for seismic-hazard assessment at UGS sites.