Principal component analysis was performed on displacements of GPS stations in and around Izu Peninsula in 1997. As the first component, displacements due to earthquake swarm activity in March was extracted. The second component had high factor loading in height component of all stations. The second component is supposed to be due to estimation error of range between antennas and satellites. Pattern of temporal change of the component is similar to that of barometric pressure in winter and to atmospheric temperature in summer . It suggests the component be related to atmospheric delay .

View full abstract

We investigate the differences of station velocities obtained by GSFC and IERS from those predicted by the nnr-NUVELIA model. Differences between observed and predicted velocities at stations near plate boundaries are expected to be greater than those at stations far away from any plate boundary since stations near plate boundaries are supposed to change their positions due to the effect of the neighboring plates' motion. We use the velocities in the parallel and transverse directions to the relative motion of the neighboring plate for the stations near plate boundaries. The differences of velocities in the parallel direction clearly shift in the direction of the neighboring plate's motion, while those in the transverse direction show a Gaussian-like distribution with no systematic bias. This indicates that velocities at almost all stations near plate boundaries reflect a deformation in the direction of the neighboring plates' motion. The differences of velocities in the parallel direction at stations within 300 km from the plate boundary are about 10-20% of the velocities predicted by the relative motion of the neighboring plate. On the other hand, the velocities at stations far away from any plate boundary generally agree with the nnr-NUVELIA model. The observed westward velocities analyzed by GSFC at stations on the Pacific plate disagree with the nnr-NUVELIA model by about 7 mm/year. This disagreement suggests some error in the relative motion between the Pacific and North American plates in the nnr-NUVELIA model . However, the disagreement does not appear in ITRF velocities.

View full abstract