Remote sensing data has become an essential tool for forest analysis, enabling observation of forest cover, structure, and change across large areas and over time. Satellite, airborne, and ground based sensors provide complementary views of forest ecosystems, capturing canopy structure, spectral reflectance, and biomass. A wide range of artificial intelligence and machine learning models, from classical statistical approaches to deep learning architectures such as convolutional networks and transformers, extract meaningful information from these data sources individually. Another direction is to combine these sources through multimodal fusion strategies, at the input, feature, or decision level, bringing together optical, radar, LiDAR, hyperspectral, and UAV based data to capture aspects of forest condition that no single sensor can reveal on its own, while environmental information such as climate, soil, topography, and hydrology can further help explain the underlying drivers of forest dynamics. Building on this, advances in deep learning, self-supervised learning, and foundation models offer new opportunities to learn from such large and heterogeneous datasets and transfer information across sensors, regions, and applications.
Several key questions remain around which data sources and combinations are most informative, how different modalities can be effectively integrated, what training and reference data are needed, and how emerging models can achieve robust and transferable performance across forest ecosystems. This session aims to bring these perspectives together, identify current needs and opportunities, and foster approaches that enable more transferable and scalable forest applications.

Topics of interest include, but are not limited to:
• Foundation models and pretraining strategies for forest applications
• Multimodal data fusion for forest monitoring
• Deep learning and novel AI and machine learning architectures for forest analysis
• Integration of remote sensing, environmental, and field data
• Forest health and disturbance detection, such as drought, fire, pests, and logging
• Tree and forest species mapping and classification
• Forest dynamics and change over time

We particularly encourage contributions that explore the use of different data sources, multimodal data fusion, foundation models, or new methodological approaches, and we welcome early career scientists and researchers from diverse backgrounds and institutions.