Temporal variations of the geoid and dynamic maintenance of height reference frame using surface mass loading and GRACE/GRACE-FO data

SUMMARY High-precision geoid models have traditionally been static, neglecting temporal variations. However, achieving geoid accuracy within 1–2 cm and maintaining dynamic height reference frames necessitates consideration of geoid spatiotemporal variations. Gravity Recovery and Climate Experiment/GRACE Follow-On (GRACE/GRACE-FO) and surface mass loading models provide means to estimate geoid changes, but their accuracy and reliability require further validation. This study proposes a method for dynamically maintaining regional height reference frames by integrating Global Navigation Satellite System (GNSS) reference stations as core nodes and incorporating time-varying geoid data. This method dynamically corrects station heights by computing normal height variations using GNSS observations and geoid changes. Experiments conducted in Beijing and Shandong derived geoid changes using GRACE/GRACE-FO and surface mass loading, validated against long-term GNSS observations and leveling surveys. Results show a strong correlation (R ≈ 0.9; NSE > 0.4) between geoid changes derived from GRACE and surface mass loading, although amplitude discrepancies of up to 4 mm existed. In 41 experimental cases, accuracy improvement was observed in over 90 per cent of instances following geoid change corrections. In Beijing, 18 out of 26 results achieved accuracy improvements exceeding 20 per cent, five of which surpassed 90 per cent. In Shandong, 11 out of 15 results improved by over 10 per cent, including five exceeding 40 per cent. These findings confirm the feasibility and effectiveness of using GRACE/GRACE-FO and surface mass loading to estimate geoid changes. The proposed method significantly improves the accuracy of dynamic height reference frame maintenance, providing valuable insights for further refinement of geoid models.