Vertical Motions of the Hawaiian Islands and Other Sites on the Pacific Plate and Comparison to Global Loading Models

Existing GPS data shows that all of the Hawaiian Islands are subsiding, with the island of Hawaii subsiding faster than the other islands. While the subsidence of the island of Hawaii might be explained by ongoing loading from the growing volcanoes or a deep deflationary source beneath the active volcanoes, it is more difficult to explain the subsidence of the other islands. We have sought to determine if this subsidence might reflect an error in the reference frame or the observations, or if these observations represent a real geophysical phenomenon. To do so, we computed an updated velocity field for 76 continuous GNSS sites on the island of Hawaii and 57 other continuous sites located across the Pacific plate. Our results show that horizontal motion of Pacific GNSS sites relative to the Pacific plate is well predicted by the GEODVEL model. However, in the vertical a systematic pattern of subsidence is observed across the Pacific, with sites typically subsiding at rates of 0.1-0.3 mm/yr. We then compare these observations to predicted vertical deformation patterns for two existing global loading models from Riva et al. (2017) and Coulson et al. (2021). Both of these models compute the global vertical deformational response to present-day ice and water mass change of the Greenland and Antarctic ice sheets, as well as ice caps and mountain glaciers; they do not include the viscoelastic response to past load changes. Both models show subsidence of the Pacific plate, which we compare to our observed pattern of subsidence of Pacific GNSS sites. The models are expressed in a reference frame in which the origin is the center of mass of the (deformed) solid Earth, while the GNSS velocities, expressed in the ITRF, are in a reference frame in which the origin is at the center of mass of the Earth system (solid earth, plus fluid loads), and this frame difference must be accounted for. Preliminary results show that the model subsidence is on the same order as the observed subsidence, indicating that present-day (and past) loading deformation must be accounted for in evaluating even far-field uplift and subsidence.