Extending Multi-Sensor Remote Sensing Approaches for Landslide Timing Analysis in Google Earth Engine

The dating of landslides allows the reconstruction of slope dynamics and link landslide events to the geomorphological processes that triggered them. In many areas, the timing of landslide initiation is uncertain due to limited observational data. In the absence of such information, the analysis of satellite remote sensing time series provides an essential tool for landslide dating, thanks to the availability of multi-sensor observations. In this context, Google Earth Engine (GEE) provides a cloud-based environment capable of processing large volumes of satellite data, enabling the application of methodologies across different study areas. This study presents an integrated methodology for the objective dating of landslides through the combined use of optical and radar satellite data. To improve the temporal coverage of the analyses, Landsat 8 data are incorporated with Sentinel-2 data. The optical data are used to assess changes in vegetation indices associated with landslide triggers, allowing the identification of a preliminary time window. Simultaneously, ALOS PALSAR-2 radar data are complemented with Sentinel-1 data to fully exploit the radar datasets available in GEE and expand the information provided by backscattering analysis, which is sensitive to changes in the physical properties of the ground surface. The use of radar data improves the detection of surface changes even under cloudy conditions or in the absence of sunlight. To strengthen the analysis, the landslide area is compared with an adjacent undisturbed reference area, defined using a variable buffer around the landslide body. The integration of optical and radar data allows estimation of landslide initiation timing, reducing possible errors in analyses based on a single sensor. Furthermore, the implementation of the methodology in GEE ensures high reproducibility, allowing scripts to be easily shared and the analysis to be applied to different study areas.