Description
Environmental geodesy stands at a transformative moment. Precise space geodetic measurements that once served mainly navigation and reference frames now expose the fingerprints of climate change – from floods/droughts to boundary layer air quality episodes, subsidence in vulnerable coastal regions, and hydrological extremes in transboundary river basins worldwide. Yet our most powerful systems – satellite gravimetry, GNSS, InSAR, and GNSS-RO – remain underused because of coarse resolution, data gaps, and processing assumptions that discard low altitude signals.This keynote explores how artificial intelligence can unlock these latent potentials in space geodetic environmental sensing, turning signals into actionable insights and solutions. First, we show how generative and physics informed AI can transform GRACE/GRACE FO terrestrial water storage from hundreds of kilometres to high-resolution basin scale products, enabling storage aware climate extreme (flood\drought) monitoring in data scarce transboundary basins and vulnerable coastal deltas. Second, we explore how physics-informed AI (geodetic-AI) could exploit water balance physics to produce high resolution, uncertainty aware “digital water memory” fields that directly constrain operational flood forecast models to reduce peak bias and improve warning lead times.
Beyond water, we present AI enhanced GNSS RO frameworks that recover planetary boundary layer height from low altitude signals traditionally discarded by standard processing, sharply reducing errors relative to radiosondes and diagnosing seasonal biases in reanalyses. Additional examples span AI powered elevation model enhancement, landslide segmentation, and GNSS based precipitable water monitoring for extreme weather prediction.
Together, these cases trace an emerging observation to decision pipeline: multi sensor geodetic signals are fused, reconstructed, and constrained by physics aware AI, then ingested into hydrological, atmospheric, and risk models to support climate services. The talk closes by outlining a research agenda for the geodetic community: How can GNSS, GRACE gravimetry, InSAR, and GNSS RO, coupled with trustworthy AI, underpin next generation climate resilience?
| Period | 15 Oct 2026 → 18 Oct 2026 |
|---|---|
| Event title | International Conference on Remote Sensing and Geoinformation (RESG2026) |
| Event type | Conference |
| Location | Beijing, ChinaShow on map |