Large-scale monitoring of inland water surface levels with GEDI data: an operational cloud-based approach in Google Earth Engine

This paper demonstrates the feasibility of large-scale, reliable monitoring of inland water surface levels through the analysis of the data collected by the Global Ecosystem Dynamics Investigation (GEDI) altimeter. In particular, we propose an automated and worldwide operational workflow, implemented within Google Earth Engine (GEE), benefiting from the availability of the whole GEDI time series in this geospatial cloud platform. Leveraging the massive computational capabilities of GEE, we were able to analyze millions of GEDI footprints and assess the potential of the sensor – in terms of precision and accuracy – to serve as an efficient and reliable remote hydrometer. The workflow is based on a rigorous spatio-temporal outlier rejection procedure and a spatial aggregation of the remaining high-quality footprints, for the robust estimation of a per-epoch median water level and its precision for the considered lake surface. A comprehensive precision and accuracy assessment was performed by comparing the GEDI retrieved water-level time series with in situ gauge data for 11 lakes of variable extent (from tens to several thousand km2) located in three continents. A rather homogeneous precision of GEDI water levels across the lakes was found, with a mean value of 14 cm. Additionally, a good agreement with the reference gauge stations was observed, showing an overall accuracy of 35 cm, a slight overestimation bias (6 cm), and a correlation of 0.76. It is important to note that these results are affected by the uncertainties of the transformation among GEDI reference frame and gauge stations reference frames. The proposed workflow can be easily applied to provide reliable inland water-level time series for all those lakes for which GEDI data is available, offering a generally higher temporal resolution than other altimeters. This approach lays the foundations for integrating GEDI within the set of remote-sensing instruments for water cycle monitoring on a large scale, enhancing our understanding of water storage dynamics in lakes, particularly in remote areas where it is not possible to install and maintain hydrometric gauges.