Abstract:
The low-lying islands in the Western Tropical Pacific are vulnerable to a faster sea
level variation in comparison to the global mean sea level change. Until the Satellite
Altimetry technique came into operation, the primary mode of assessing the Sea
Surface Heights (SSH) was limited to Tide Gauge observations. The tide gauge
inferred SSH consists of absolute sea level variations and Vertical Land Motions
(VLM). Therefore, VLM trends are essential for accurate sea level prediction and
coastal impact assessment. In this study, the VLM rates are computed in the Western
Tropical Pacific Island using the Global Positioning Satellite (GPS) time series
obtained from the Nevada Geodetic Laboratory. The length of the GPS duration and
the completeness of the time series recordings served as selection criteria. Vertical
position time series of the selected GPS stations were post-processed to remove the
outliers and offsets and adjusted for the Post-Seismic Deformation. The VLM linear
rates for each station were computed using the Least Squares Estimation method by
fitting the observations to a parametric model. Estimated VLM rates showed moderate
land subsidence in 26 stations (~87 %) out of the 30 GPS stations considered, and the
fastest subsidence rate of 9.01 mm/yr was observed in the Tafuna island, American
Samoa. In another approach, the VLM rates were deduced by the difference between
Altimetry (ALT) and Tide Gauge (TG) sea level records. Similar criteria as for GPS
were used for the selection of the tide gauge stations. In order to maintain equal
geophysical conditions in both altimetry and tide gauge observations, corrections were
applied to the tide gauge records using the ocean loading tidal model. The ALT-TG
rates were then calculated for three options based on the correlation of the SSH of
altimetry grids around the corresponding TGs. 19 (95 %) out of 20 TGs showed
moderate subsidence. The highest correlated VLM rates obtained from the ALT-TG
were then compared with the co-located GPS stations within a 50 km radius. The
results of 28 combinations showed a close agreement between the two methods, with
a mean rate difference of 0.08 mm/yr and an RMSE of 1.68 mm/yr.
