Topographically induced height errors in predicted atmospheric loading effects

Atmospheric pressure variations are known to induce vertical displacements of the Earth's surface with magnitudes large enough to be detected by geodetic observations. Estimates of these loading effects are derived using global reanalysis fields of surface pressure as input. The input surface pressure has a minimum spatial sampling, which does not capture true surface pressure variations due to high topographic variability in some regions. In this paper, we investigate the effect that unmodeled topographic variability has on surface pressure estimates and subsequent estimates of vertical surface displacements. We find that the estimated height changes from the topographic surface pressure can be significant (2-4 mm) for sites in regions of high topographic variability. When we compare the estimated height changes to Global Positioning System residuals from the 2005 International Terrestrial Reference Frame Realization, we find that the heights derived from the topographic surface pressure, versus those from the normal surface pressure, perform better at reducing the scatter on the height coordinate time series.