Mohammad-Hadi Rezvani, Christopher S. Watson, Matt A. King

We further developed a space-time Kalman approach to estimate time-variable signals in residual altimeter systematic errors and vertical land motion (VLM) around the Australian coast since the 1990s, through combining multi-mission absolute sea-level (ASL), relative sea-level (RSL) from tide gauges (TGs) and GPS heights records. Our results confirmed continent-wide subsidence and TG-specific VLMs yielding a ~40% reduction in RMSE of geographical ASL variability, compared with rates determined using spatially interpolated GPS velocities that fail to capture localized trends by up to ~1.5 mm/yr. Stacked time series of non-linear deformation at TGs and nearby GPS showed some correlation, suggesting the technique was partially successful in reflecting the surface loading. Site-by-site inspection revealed spurious non-linearity likely caused by residual oceanographic signals present between the TG and altimeter measurement locations. Our average mission-specific error estimates are small but significant, typically within ~±0.5-1.0 mm/yr, with negligible effect implied on the overall rate of ASL. Analysis of the time variability of altimeter errors confirmed stability for most missions except for Jason-2 with an anomaly reaching ~2.8 mm/yr in the first ~3.5 years of operation which is supported by analysis from the Bass Strait altimeter validation facility. Weak correlation with the dominant climate mode suggests potential deficiencies in the resolution of the time-variable gravity field used for orbit determination as a possible cause, yet other drivers cannot be discounted. Our approach advances the ability to estimate TG-specific VLMs and regional altimeter systematic errors, and highlights that residual oceanographic signals remain a fundamental limitation to such techniques.