Stability maintenance of gravity comparison sites (2017–2024): Environmental factors and data processing strategies

To ensure the sustained stability of absolute gravity benchmark points from 2017 to 2024, this work comprehensively examined observational records from superconducting gravimeters (SG) and absolute gravimeters, while quantitatively assessing environmental effects on gravitational acceleration. The annual fluctuation of the SG (iGrav-012k) scale factor reached up to 2.68 nm/s ² /V, with a weighted average of (–928.702 ± 0.003) nm/s ² /V (relative accuracy of 0.3‰), offering precise calibration parameters for long-term SG monitoring. By eliminating step discontinuities in SG data using FG5-X249 absolute gravimeter measurements, the residual fitting error decreased to 6.3 µGal. Additionally, SG drift was estimated as 1.0 µGal/year through international comparison datasets and FG5 measurements, considerably improving time series consistency. Further investigation indicated that SG residuals exhibited clear seasonal oscillations, mainly attributed to local hydrological processes and ground deformation near the benchmark sites. By integrating groundwater level, rainfall, and deformation monitoring data, and applying a neural network model to separate hydrological load components, the peak-to-peak residual amplitude was reduced from 15 µGal to 7 µGal. Quantitative analysis revealed that hydrological effects contributed roughly 10 µGal to the seasonal variation, whereas surface deformation exerted only a minor impact (< 2 µGal). The findings confirm that careful data correction and isolation of environmental effects are effective in sustaining the long-term stability of gravity benchmarks. The developed workflow provides a reproducible framework for high-precision gravity site maintenance and supports future dynamic monitoring of regional environmental load responses.