Global L-band equivalent AI-based vegetation optical depth dataset

Olya Skulovich; Xiaojun Li; Jean-Pierre Wigneron; Pierre Gentine

doi:10.1038/s41597-024-03810-2

Global L-band equivalent AI-based vegetation optical depth dataset

Sci Data. 2024 Aug 28;11(1):936. doi: 10.1038/s41597-024-03810-2.

Authors

Olya Skulovich¹, Xiaojun Li², Jean-Pierre Wigneron², Pierre Gentine³

Affiliations

¹ Columbia University, New York, NY, 10027, USA. [email protected].
² NRAE, UMR1391 ISPA, University of Bordeaux, F-33140, Villenave d'Ornon, France.
³ Columbia University, New York, NY, 10027, USA.

Abstract

The L-band vegetation optical depth data garners significant interest for its ability to effectively monitor vegetation, thanks to minimal saturation within this frequency range. However, the existing datasets have limited temporal coverage, constrained by the start of the respective satellite missions. Global L-band equivalent AI-Based Vegetation Optical Depth or GLAB-VOD is a global long-term consistent microwave vegetation optical depth dataset created using machine learning to expand the SMAP-IB VOD dataset temporal coverage from 2015-2020 to 2002-2020. The GLAB-VOD dataset has an 18-day temporal resolution and 25 km spatial resolution on the EASE2 grid and covers 2002-2020. An auxiliary consistent daily brightness temperature product, called GLAB-TB, is developed in parallel and ensures the consistency of the VOD product across time periods with different microwave satellites. As a result of its temporal consistency, this dataset can be used to study long-term global and regional trends in vegetation biomass and utilized in any other applications where long-term consistency is necessary. The GLAB-VOD dataset shows excellent spatial correlation globally when compared with biomass (up to R = 0.92) and canopy height (R = 0.93), outperforming its target dataset, SMAP-IB VOD.