Item

Abstract

Inland water bodies are essential sources of freshwater for several applications, making their level monitoring crucial for understanding the impact of climate change and human activities. The Global Ecosystem Dynamics Investigation (GEDI) [1,2] mission utilizes a spaceborne LiDAR altimeter to produce high-resolution measurements of Earth's canopy. GEDI has recently been added in Google Earth Engine (GEE), a cloud-based platform integrating different datasets and providing efficient analysis tools. This study aims to evaluate the accuracy of GEDI altimetric data over water bodies, comparing the levels recorded by GEDI with ground truth measurements. Furthermore, a robust procedure to remove outliers is proposed and implemented within GEE. Specifically, a number of lakes all over Italy were selected and analyzed for the period from April 2019 to June 2022. The proposed outlier detection methodology consists of two stages. The first employs two flags provided by GEDI metadata. The "quality_flag" and "degrade_flag" provide information about the footprint validity, based respectively on GEDI signal anomalies and the degradation state of pointing or positioning information. Secondly, the methodology employs a robust implementation of the standard 3σ-test, utilizing the Normalized Median Absolute Deviation (NMAD) to identify outliers: measurements not within the range of -/+3*NMAD from the median are removed. After the outlier removal, we obtain an average NMAD of 0.11m over the ten lakes for the considered period.[1] Dubayah et al., 2021. GEDI L3 gridded land surface metrics.[2] University of Maryland, 2022. GEDI ecosystem lidar. www.gedi.umd.edu.