An Operational Data-Driven Framework For Developing High-Resolution Leaf Area Index Products

Leaf area index (LAI) is a critical biophysical variable that describes vegetation structure and has been recognized as an Essential Climate Variable by the Global Climate Observing System [1] . Accurate characterization of LAI is essential for modeling the exchange of energy, carbon, and water between ecosystems and the atmosphere from regional to global scales [2] - [3] . Over the past four decades, long-term LAI records from moderate-resolution (0.25 – 4 km) satellite sensors such as AVHRR and MODIS have led to significant advancements in understanding the terrestrial carbon and water cycles under a changing climate [4] - [5] . However, emerging applications in climate mitigation and adaptation, agroecosystem, and hydrology require a new generation of high-resolution LAI data products [6] - [7] . In this study, we propose a generic framework to generate high-resolution LAI satellite products (e.g. for Landsat and Sentinel-2) from well-established moderate-resolution LAI datasets (e.g. MODIS) using machine learning and cloud computing. We provide an example of this framework in action for generating 30-m LAI data from Landsat images using MODIS LAI products and Google Earth Engine.