Estimates for energy fluxes between land and atmosphere improved
Scientists from Max Planck Institute for Biogeochemistry (MPI-BGC), together with international collaborators, have developed an ensemble of machine learning-based energy flux products and made them available to the broader scientific community. Using complementary datasets from local measurements, satellite remote sensing and meteorology, the scientists determined global energy fluxes between the land surface and the atmosphere, and their respective uncertainties. The results have been published today in the nature/scientific data journal.
Estimating latent heat (evapotranspiration), sensible heat or net radiation fluxes over the global land is still challenging and uncertain. To provide improved land-atmosphere energy flux estimates and uncertainty quantification, the scientists from MPI-BGC teamed up with international colleagues. They used a variety of data from 224 FLUXNET eddy covariance towers, remote sensing and meteorological measurements together with various machine learning methods to generate a large ensemble of energy flux products. These data products provide time-varying energy flux estimates and their uncertainties for every location over land for the last several decades. The dataset is expected to provide an excellent benchmark for the global Earth System Models.
The computation and archiving of the data are carried out in the high-performance computer cluster at MPI-BGC. The FLUXCOM data are freely available (CC4.0.BY license) from the MPI-BGC’s data portal.