Applying GNSS-IR Technique with High-Rate Receiver to Reinforce the Accuracy of Landslide Early Waring in Tawain

GNSS Interferometric Reflectometry (GNSS-IR) exploits the multipath interference between direct GNSS signals and ground-reflected signals received by a geodetic antenna. Variations in the reflected signal phase and amplitude, observed in the Signal-to-Noise Ratio (SNR) data, encode changes in near-surface properties, most critically soil moisture. For landslides, soil moisture is a first-order control on effective stress reduction, shear strength loss, and pore-water pressure buildup. Thus, GNSS-IR provides a physically meaningful proxy for hydrologic preconditioning before slope failure. GNSS-IR was set to detect the hydrologic state around the station, and InSAR was used to obtain regional deformation. GNSS wet-delay data served as in situ rainfall measurements. All these data were combined to observe rapid wetting, sustained saturation, and deformation. This architecture significantly reduced false alarms compared with rainfall-only systems. Several dual-phase GNSS tracking stations have been installed in the mountainous regions of Taiwan to determine the precise location and detect slope stability. This approach collected historical data to train the machine learning model at each station, and the model parameters could predict rapid wetting before reaching the critical point. The preliminary results show an improvement of 20% compared to the traditional empirical method and could issue an early warning of as much as 5-10 minutes with a 20 Hz GNSS receiver.