Title: Cloud based biomass prediction using GEDI and Landsat data in ArcGIS
Abstract:
Forests play a crucial role as terrestrial carbon sinks, contributing significantly to the global carbon budget. Aboveground biomass (AGB) within forests is a key component of the carbon cycle. Advances in cloud computing and remote sensing technology have opened new avenues for estimating forest AGB. AGB can be assessed through remote sensing imagery, providing spectral characteristics of land cover, and altimetry sensors like Global Ecosystem Dynamics Investigation(GEDI), offering a 3D structure of the forest. To facilitate AGB mapping using these multi-source remote sensing data, we have developed machine learning and cloud-based tools and solutions within ArcGIS software. In this context, we review the workflows for estimating AGB, focusing on the state of Oregon in the United States as a study area. Firstly, we introduce a trajectory dataset data model for managing and analyzing altimeter data, demonstrating how to extract point data from the GEDI level 4A product. Secondly, we discuss accessing Landsat and DEM datasets in Microsoft Planetary Computer using STAC. Thirdly, we train a random tree regression model using the extracted AGB points as ground truth, and DEM, Landsat images, and derived indices as explanatory variables. Finally, we generate AGB predictions for the entire state of Oregon. This solution proves valuable and instrumental for biomass mapping in any other state or region, utilizing data from various cloud platforms such as AWS.
Biography:
Hong Xu holds a Master of Science degree in computational mathematics and another Master of Science degree in remote sensing and GIS. Currently serving as a Principal Software Product Engineer at Environmental Systems Research Institute (Esri), she leads the development of software capabilities and tools for the management, processing, and analysis of remotely sensed Earth data. Her focus areas encompass machine learning, image time series analysis, multidimensional data, and altimetry data.
