Resolving land subsidence contribution to present and future relative sea level change using satellite geodesy

Coastal zones face increased risks from the combined effects of climate-driven sea-level rise and vertical land motion (VLM), which together determine rates of relative sea-level (RSL) change. While oceanic contributions to RSL are increasingly well monitored and projected, land subsidence (i.e., negative VLM) remains one of the least systematically observed and most spatially heterogeneous components of RSL, despite its potential to locally exceed climate-driven ocean rise by an order of magnitude. This observational gap is especially pronounced in rapidly urbanizing and data-limited regions, where sparse tide-gauge and GNSS networks hinder the identification of subsidence hotspots and their evolving impacts on coastal risks.

In this talk, I present a framework that leverages satellite geodesy as a climate observing system to resolve the spatiotemporal dynamics of land subsidence and quantify its contribution to present and future relative sea-level change, using Java Island, Indonesia, as a regional-scale case study. We generated high-spatial resolution (75 m) contemporary VLM fields from using multi-geometry Sentinel-1 interferometric synthetic aperture radar (InSAR), revealing widespread and temporally evolving subsidence patterns with rates exceeding 1 cm per year across multiple coastal and inland urban centers. While Jakarta has dominated the subsidence narrative in Indonesia, we find that several other coastal cities, including Cirebon, Pekalongan, Tegal, and Semarang, are sinking two to three times faster, with localized rates approaching 10 cm per year.

To disentangle the dominant drivers of deformation, we applied unsupervised machine-learning spatiotemporal clustering to InSAR time series, guided by geological and land-use information. This analysis reveals nonlinear and spatially heterogeneous subsidence behaviors primarily associated with groundwater extraction in urban, industrial, and agricultural regions, alongside localized deformation linked to natural processes such as volcanism. Finally, we constructed synthetic tide-gauge records at 5-km spacing along the 1,500 km northern coastline by integrating InSAR-derived VLM with satellite altimetry and probabilistic sea-level projections. These virtual gauges show that neglecting land subsidence leads to systematic underestimation of RSL change by more than 90% in some locations and that subsidence will remain the dominant contributor to RSL rise across much of the coastline through 2050.

This work illustrates how geodetic observing systems can fill critical observational gaps in coastal climate research, enabling spatially explicit, process-informed RSL estimates and providing a transferable framework for improving sea-level risk assessments in vulnerable, data-sparse regions worldwide.