A multi-metric methodology for short-term performance assessment of continuous gravimeters

Continuous gravimeters operating at permanent stations are commonly evaluated using long-term background observations; however, quantitative approaches capable of jointly assessing background noise, strong-earthquake responses, and post-seismic recovery over short observation periods remain limited. This study proposes a multi-metric framework for short-term performance evaluation of continuous gravimeters and validates it using 1 Hz observations from six gPhone gravimeters in the Fujian network during 1-20 April 2024. Following time synchronization, solid Earth tide correction, atmospheric adjustment, and hourly Welch power spectral density (PSD) estimation, background quality was characterized across four frequency bands using median power, interquartile range (null), 90th percentile (null), and quiet window probability, which were integrated into a transparent synthetic noise index (null). The temporal stability and parameter independence of background rankings were evaluated through daily leave-one-out tests and sensitivity analyses with varying Welch segment lengths. Strong-earthquake performance was further assessed using theoretical phase-window signal-to-noise ratios (null s), S-wave arrival-time consistency, event-day spectral power enhancement, and post-seismic recovery time during the 2024 M_W7.4 Hualien earthquake. Results show that Putian consistently achieved the best integrated background performance under all weighting schemes, with overall station rankings remaining invariant in both daily leave-one-out and spectral-parameter sensitivity tests. Quanzhou exhibited the highest proportion of quiet observation windows in the 1-30 mHz band, while all six stations recorded clear band-limited earthquake responses with high phase-window null s and post-seismic recovery times ranging from 9.0 to 28.0 h. These findings demonstrate that short-term station performance can be evaluated consistently across the complementary dimensions of background quality, event detectability, and recovery capability. The proposed framework provides a transparent and quantitative methodology for short-term performance assessment of continuous gravimeters, offering practical support for station operation, network management, strong-motion observation, and the evaluation of next-generation high-precision gravimetric instruments.