Data

PREDICT: Predicting Resilience and Early Detection of Impending Climate Transitions

The PREDICT project will utilize a wide range of Earth observation data, with a primary focus on Essential Climate Variable (ECV) datasets developed within the ESA Climate Change Initiative (CCI) programme. These datasets will be complemented by other satellite products and model outputs.

For Amazon forest dieback analysis:

• Vegetation-CCI CDRs including data from SPOT, MODIS, MERIS, Metop, OLCI, GEDI, and VIIRS
• Vegetation Optical Depth (VOD) from SSMI, TRMM, TMI, AMSRE, AMSR2, SMOS, and SMAP
• Biomass-CCI CDRs including data from Sentinel-1A/B, ALOS-2, PALSAR-2, and GEDI
• LandCover-CCI CDRs including data from MERIS and Sentinel-2
• Climate data from CMIP6 Ensemble

For dryland patterned vegetation analysis:

• Vegetation-CCI CDRs
• LandCover-CCI CDRs
• CHIRPS Rainfall data

For permafrost thaw analysis:

• Permafrost-CCI CDR Extent/Thickness
• Sentinel-1 SAR
• GHG-CCI CDRs from Sentinel-5P, GOSAT
• Sentinel-2 and Commercial (Planet/MAXAR) Optical Imagery
• LST-CCI CDRs including from (A)ATSR, MODIS, SLSTR, SSMI

The project will also utilize observational uncertainty information provided with the CCI datasets, as well as harmonize multi-sensor datasets through a weighted ensemble machine learning fusion framework, employing Bayesian gap-filling techniques and novel scale-aware approaches.

Data products will be made available via the CCI Open Data Portal and CEDA, with all software and analysis tools developed as open-source with appropriate documentation.