%0 Journal Article
%A Szwarcman, Daniela
%A Roy, Sujit
%A Fraccaro, Paolo
%A Gíslason, Orsteinn Elí
%A Blumenstiel, Benedikt
%A Ghosal, Rinki
%A De Oliveira, Pedro Henrique
%A Almeida, Joao Lucas de Sousa
%A Sedona, Rocco
%A Kang, Yanghui
%A Chakraborty, Srija
%A Wang, Sizhe
%A Gomes, Carlos
%A Kumar, Ankur
%A Gaur, Vishal
%A Truong, Myscon
%A Godwin, Denys
%A Khallaghi, Sam
%A Lee, Hyunho
%A Hsu, Chia-Yu
%A Asanjan, Ata Akbari
%A Mujeci, Besart
%A Shidham, Disha
%A Balogun, Rufai Omowunmi
%A Kolluru, Venkatesh
%A Keenan, Trevor
%A Arevalo, Paulo
%A Li, Wenwen
%A Alemohammad, Hamed
%A Olofsson, Pontus
%A Mayer, Timothy
%A Hain, Christopher
%A Kennedy, Robert
%A Zadrozny, Bianca
%A Bell, David
%A Cavallaro, Gabriele
%A Watson, Campbell
%A Maskey, Manil
%A Ramachandran, Rahul
%A Moreno, Juan Bernabe
%T Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications
%J IEEE transactions on geoscience and remote sensing
%V 64
%@ 0018-9413
%C New York, NY
%I IEEE
%M FZJ-2025-05777
%P  4400120 
%D 2025
%X This paper presents Prithvi-EO-2.0, a new geospatial foundation model that offers significant improvements over its predecessor, Prithvi-EO-1.0. Trained on 4.2 million global time series samples from NASA’s Harmonized Landsat and Sentinel-2 data archive at 30-m resolution, the new model incorporates temporal and location embeddings for enhanced performance across various geospatial tasks. Through extensive benchmarking with GEO-Bench, the model outperforms the previous Prithvi-EO model by 8% across a range of tasks. It also outperforms six other geospatial foundation models when benchmarked on remote sensing tasks from different domains and resolutions (i.e. from 0.1 m to 15 m). The results demonstrate the versatility of the model in both classical Earth observation and high-resolution applications. Early involvement of end-users and subject matter experts (SMEs) allowed constant feedback on model and dataset design, enabling customization across diverse SME-led applications in disaster response, land cover and crop mapping, and ecosystem dynamics monitoring. Prithvi-EO-2.0 is available as an open-source model on Hugging Face and IBM TerraTorch, with additional resources on GitHub. The project exemplifies the Trusted Open Science approach embraced by all involved organizations.
%F PUB:(DE-HGF)16
%9 Journal Article
%R 10.1109/TGRS.2025.3642610
%U https://juser.fz-juelich.de/record/1050065