Simplify your processing with the WorldCover annual composities

#Sentinel, #Copernicus

Published on 22 June 2023

During the development of ESA WorldCover 2020 and 2021, the first global land cover maps at 10 meters resolution based on Sentinel-1 and Sentinel-2, the team at VITO Remote Sensing processed approximately four petabytes of data deriving a series of annual 10 m median composites. These cloud-free global mosaics served as input for the land cover classification process but have the potential to be used for many additional applications as well.

They are therefore pleased to share the WorldCover Sentinel-1 and Sentinel-2 annual mosaics. One of the experts behind WorldCover, Daniele Zanaga, explains more about these mosaics and the growing interest expressed by various users, showcasing their great potential for research and business applications.

Processing Sentinel-1 and Sentinel-2 yearly archives

The Copernicus Sentinel-1 and Sentinel-2 satellites capture our earth's surface with high temporal frequency. Monitoring our globe since 2014 and 2015, they provide several petabytes of images annually. An enormous amount of raw data that can provide insights into the state of Earth's surface and enables research of human impact and climate change at an unprecedented level. But processing this vast volume of raw data into analysis ready data (ARD), suitable for downstream applications, presents significant challenges.

Running the pre-processing workflows at a global scale is technically intricate and financially demanding. Sentinel-1 for example requires radiometric calibration, terrain correction, speckle de-noising, reprojection, and mosaicking. For Sentinel-2 data, clouds screening, cleaning, reprojection, and mosaicking are needed to make the data usable for modelling.  

For the development of WorldCover 2020 and 2021, the first global land cover maps at 10 meters resolution based on Sentinel-1 and Sentinel-2 data, prepared in the framework of the European Space Agency WorldCover project, we ran these workflows at scale to process and produce a set of annual 10 m temporal composites for the years 2020 and 2021. 

After the release of the WorldCover products, many users expressed their interest and requested access to these composites. We therefore decided to share the WorldCover Sentinel-1 and Sentinel-2 annual composites for public use. These mosaics are interesting to kickstart several projects without the daunting task and costs of processing extensive amounts of data.

Continue reading