The thermal expansion effect exhibited by the GNSS monument and their nearby bedrock due to surface temperature variations is an important factor affecting the nonlinear variation of GNSS height. However, current thermal expansion models either consider only the above-surface GNSS monuments or only non-seasonal temperature variations of the subsurface bedrock, and lack a comprehensive thermal expansion model. Furthermore, previous studies on the contribution of thermal expansion effects to nonlinear variations in GNSS height have been analyzed only against one set of GNSS time series products per single research, but in fact the GNSS time series provided by each agency varied considerably. In this study, we use a refined comprehensive thermal expansion model (TEVD_FSD) to evaluate its contribution to the nonlinear variations from several GNSS height products (SOPAC, JPL, and COMBINED) obtained using different data processing strategies. The results show that the most GNSS stations (about 95%) using the TEVD_FSD model exhibit an annual amplitude increase and phase lag, with an amplitude increase of up to 0.5 mm and phase lag of up to 13° compared with the finite element model, especially for inland and those with deeper GNSS monument stations. This phase lag improves its correlation with the GNSS height, which reduces the GNSS height value to improve the geophysical interpretation. The TEVD_FSD model estimates an annual amplitude of up to 7.5 mm, explaining at most 13.6% of the nonlinear variation in the COMBINED height. The COMBINED product exhibits a further WRMS reduction of up to 20% and 18.7% compared with the SOPAC and JPL products, respectively, which are likely due to the higher accuracy of the combined GNSS solution than of the independent data processing strategy. Our work indicates that differences in data processing strategies for GNSS height time series products significantly affect the interpretability of thermal expansion effects to nonlinear variations.

GNSS纪念碑及其附近基岩因表面温度变化而表现出的热膨胀效应是影响GNSS高度非线性变化的重要因素。然而，目前的热膨胀模型要么只考虑地表以上的GNSS纪念碑，要么只考虑地下基岩的非季节性温度变化，缺乏全面的热膨胀模型。此外，以往关于热膨胀效应对 GNSS 高度非线性变化影响的研究，每次研究仅针对一组 GNSS 时间序列产品进行分析，但事实上，每个机构提供的 GNSS 时间序列差异很大。在本研究中，我们使用精炼的综合热膨胀模型 (TEVD _FSD ) 来评估其对使用不同数据处理策略获得的多个 GNSS 高度产品（SOPAC、JPL 和 COMBINED）的非线性变化的贡献。结果表明，使用TEVD _FSD模型的大多数GNSS站（约95%）表现出逐年振幅增加和相位滞后，与有限元模型相比，振幅增加高达0.5 mm，相位滞后高达13° ，特别是对于内陆地区以及那些拥有较深 GNSS 纪念碑站的地区。这种相位滞后改善了其与 GNSS 高度的相关性，从而降低了 GNSS 高度值以改善地球物理解释。TEVD _FSD模型估计年振幅高达 7.5 毫米，最多解释了组合高度非线性变化的 13.6%。与 SOPAC 和 JPL 产品相比，组合产品的 WRMS 进一步降低了 20% 和 18.7%，这可能是由于组合 GNSS 解决方案比独立数据处理策略具有更高的精度。我们的工作表明，GNSS 高度时间序列产品的数据处理策略的差异显着影响热膨胀效应对非线性变化的可解释性。