Paolo Fraccaro

Johannes Jakubik, Sujit Roy, C. E. Phillips et al.

Significant progress in the development of highly adaptable and reusable Artificial Intelligence (AI) models is expected to have a significant impact on Earth science and remote sensing. Foundation models are pre-trained on large unlabeled datasets through self-supervision, and then fine-tuned for various downstream tasks with small labeled datasets. This paper introduces a first-of-a-kind framework for the efficient pre-training and fine-tuning of foundational models on extensive geospatial data. We have utilized this framework to create Prithvi, a transformer-based geospatial foundational model pre-trained on more than 1TB of multispectral satellite imagery from the Harmonized Landsat-Sentinel 2 (HLS) dataset. Our study demonstrates the efficacy of our framework in successfully fine-tuning Prithvi to a range of Earth observation tasks that have not been tackled by previous work on foundation models involving multi-temporal cloud gap imputation, flood mapping, wildfire scar segmentation, and multi-temporal crop segmentation. Our experiments show that the pre-trained model accelerates the fine-tuning process compared to leveraging randomly initialized weights. In addition, pre-trained Prithvi compares well against the state-of-the-art, e.g., outperforming a conditional GAN model in multi-temporal cloud imputation by up to 5pp (or 5.7%) in the structural similarity index. Finally, due to the limited availability of labeled data in the field of Earth observation, we gradually reduce the quantity of available labeled data for refining the model to evaluate data efficiency and demonstrate that data can be decreased significantly without affecting the model's accuracy. The pre-trained 100 million parameter model and corresponding fine-tuning workflows have been released publicly as open source contributions to the global Earth sciences community through Hugging Face.

Julia Anna Leonardi, Johannes Jakubik, Paolo Fraccaro et al.

Geospatial Foundation Models (GFMs) typically lack native support for Hyperspectral Imaging (HSI) due to the complexity and sheer size of high-dimensional spectral data. This study investigates the adaptability of TerraMind, a multimodal GFM, to address HSI downstream tasks \emph{without} HSI-specific pretraining. Therefore, we implement and compare two channel adaptation strategies: Naive Band Selection and physics-aware Spectral Response Function (SRF) grouping. Overall, our results indicate a general superiority of deep learning models with native support of HSI data. Our experiments also demonstrate the ability of TerraMind to adapt to HSI downstream tasks through band selection with moderate performance decline. Therefore, the findings of this research establish a critical baseline for HSI integration, motivating the need for native spectral tokenization in future multimodal model architectures.

Michiaki Tatsubori, Takao Moriyama, Tatsuya Ishikawa et al.

When providing the boundary conditions for hydrological flood models and estimating the associated risk, interpolating precipitation at very high temporal resolutions (e.g. 5 minutes) is essential not to miss the cause of flooding in local regions. In this paper, we study optical flow-based interpolation of globally available weather radar images from satellites. The proposed approach uses deep neural networks for the interpolation of multiple video frames, while terrain information is combined with temporarily coarse-grained precipitation radar observation as inputs for self-supervised training. An experiment with the Meteonet radar precipitation dataset for the flood risk simulation in Aude, a department in Southern France (2018), demonstrated the advantage of the proposed method over a linear interpolation baseline, with up to 20% error reduction.

Blair Edwards, Paolo Fraccaro, Nikola Stoyanov et al.

The application of models to assess the risk of the physical impacts of weather and climate and their subsequent consequences for society and business is of the utmost importance in our changing climate. The operation of such models is historically bespoke and constrained to specific compute infrastructure, driving datasets and predefined configurations. These constraints introduce challenges with scaling model runs and putting the models in the hands of interested users. Here we present a cloud-based modular framework for the deployment and operation of geospatial models, initially applied to climate impacts. The Climate Impact Modelling Frameworks (CIMF) enables the deployment of modular workflows in a dynamic and flexible manner. Users can specify workflow components in a streamlined manner, these components can then be easily organised into different configurations to assess risk in different ways and at different scales. This also enables different models (physical simulation or machine learning models) and workflows to be connected to produce combined risk assessment. Flood modelling is used as an end-to-end example to demonstrate the operation of CIMF.

Johannes Jakubik, Felix Yang, Benedikt Blumenstiel et al.

We present TerraMind, the first any-to-any generative, multimodal foundation model for Earth observation (EO). Unlike other multimodal models, TerraMind is pretrained on dual-scale representations combining both token-level and pixel-level data across modalities. On a token level, TerraMind encodes high-level contextual information to learn cross-modal relationships, while on a pixel level, TerraMind leverages fine-grained representations to capture critical spatial nuances. We pretrained TerraMind on nine geospatial modalities of a global, large-scale dataset. In this paper, we demonstrate that (i) TerraMind's dual-scale early fusion approach unlocks a range of zero-shot and few-shot applications for Earth observation, (ii) TerraMind introduces "Thinking-in-Modalities" (TiM) -- the capability of generating additional artificial data during finetuning and inference to improve the model output -- and (iii) TerraMind achieves beyond state-of-the-art performance in community-standard benchmarks for EO like PANGAEA. The pretraining dataset, the model weights, and our code are open-sourced under a permissive license.

Julian Kuehnert, Hiwot Tadesse, Chris Dearden et al.

Geospatial observations combined with computational models have become key to understanding the physical systems of our environment and enable the design of best practices to reduce societal harm. Cloud-based deployments help to scale up these modeling and AI workflows. Yet, for practitioners to make robust conclusions, model tuning and testing is crucial, a resource intensive process which involves the variation of model input variables. We have developed the Variational Exploration Module which facilitates the optimization and validation of modeling workflows deployed in the cloud by orchestrating workflow executions and using Bayesian and machine learning-based methods to analyze model behavior. User configurations allow the combination of diverse sampling strategies in multi-agent environments. The flexibility and robustness of the model-agnostic module is demonstrated using real-world applications.

Muhammad Sohail Danish, Muhammad Akhtar Munir, Syed Roshaan Ali Shah et al.

While numerous recent benchmarks focus on evaluating generic Vision-Language Models (VLMs), they do not effectively address the specific challenges of geospatial applications. Generic VLM benchmarks are not designed to handle the complexities of geospatial data, an essential component for applications such as environmental monitoring, urban planning, and disaster management. Key challenges in the geospatial domain include temporal change detection, large-scale object counting, tiny object detection, and understanding relationships between entities in remote sensing imagery. To bridge this gap, we present GEOBench-VLM, a comprehensive benchmark specifically designed to evaluate VLMs on geospatial tasks, including scene understanding, object counting, localization, fine-grained categorization, segmentation, and temporal analysis. Our benchmark features over 10,000 manually verified instructions and spanning diverse visual conditions, object types, and scales. We evaluate several state-of-the-art VLMs to assess performance on geospatial-specific challenges. The results indicate that although existing VLMs demonstrate potential, they face challenges when dealing with geospatial-specific tasks, highlighting the room for further improvements. Notably, the best-performing LLaVa-OneVision achieves only 41.7% accuracy on MCQs, slightly more than GPT-4o, which is approximately double the random guess performance. Our benchmark is publicly available at https://github.com/The-AI-Alliance/GEO-Bench-VLM .

Sagar Soni, Akshay Dudhane, Hiyam Debary et al.

Automated analysis of vast Earth observation data via interactive Vision-Language Models (VLMs) can unlock new opportunities for environmental monitoring, disaster response, and {resource management}. Existing generic VLMs do not perform well on Remote Sensing data, while the recent Geo-spatial VLMs remain restricted to a fixed resolution and few sensor modalities. In this paper, we introduce EarthDial, a conversational assistant specifically designed for Earth Observation (EO) data, transforming complex, multi-sensory Earth observations into interactive, natural language dialogues. EarthDial supports multi-spectral, multi-temporal, and multi-resolution imagery, enabling a wide range of remote sensing tasks, including classification, detection, captioning, question answering, visual reasoning, and visual grounding. To achieve this, we introduce an extensive instruction tuning dataset comprising over 11.11M instruction pairs covering RGB, Synthetic Aperture Radar (SAR), and multispectral modalities such as Near-Infrared (NIR) and infrared. Furthermore, EarthDial handles bi-temporal and multi-temporal sequence analysis for applications like change detection. Our extensive experimental results on 44 downstream datasets demonstrate that EarthDial outperforms existing generic and domain-specific models, achieving better generalization across various EO tasks. Our source codes and pre-trained models are at https://github.com/hiyamdebary/EarthDial.

Akashah Shabbir, Muhammad Umer Sheikh, Muhammad Akhtar Munir et al.

Recent progress in multimodal reasoning has enabled agents that can interpret imagery, connect it with language, and perform structured analytical tasks. Extending such capabilities to the remote sensing domain remains challenging, as models must reason over spatial scale, geographic structures, and multispectral indices while maintaining coherent multi-step logic. To bridge this gap, OpenEarthAgent introduces a unified framework for developing tool-augmented geospatial agents trained on satellite imagery, natural-language queries, and detailed reasoning traces. The training pipeline relies on supervised fine-tuning over structured reasoning trajectories, aligning the model with verified multistep tool interactions across diverse analytical contexts. The accompanying corpus comprises 14,538 training and 1,169 evaluation instances, with more than 100K reasoning steps in the training split and over 7K reasoning steps in the evaluation split. It spans urban, environmental, disaster, and infrastructure domains, and incorporates GIS-based operations alongside index analyses such as NDVI, NBR, and NDBI. Grounded in explicit reasoning traces, the learned agent demonstrates structured reasoning, stable spatial understanding, and interpretable behaviour through tool-driven geospatial interactions across diverse conditions. We report consistent improvements over a strong baseline and competitive performance relative to recent open and closed-source models.

Carlos Gomes, Benedikt Blumenstiel, Joao Lucas de Sousa Almeida et al.

TerraTorch is a fine-tuning and benchmarking toolkit for Geospatial Foundation Models built on PyTorch Lightning and tailored for satellite, weather, and climate data. It integrates domain-specific data modules, pre-defined tasks, and a modular model factory that pairs any backbone with diverse decoder heads. These components allow researchers and practitioners to fine-tune supported models in a no-code fashion by simply editing a training configuration. By consolidating best practices for model development and incorporating the automated hyperparameter optimization extension Iterate, TerraTorch reduces the expertise and time required to fine-tune or benchmark models on new Earth Observation use cases. Furthermore, TerraTorch directly integrates with GEO-Bench, allowing for systematic and reproducible benchmarking of Geospatial Foundation Models. TerraTorch is open sourced under Apache 2.0, available at https://github.com/IBM/terratorch, and can be installed via pip install terratorch.

Benedikt Blumenstiel, Paolo Fraccaro, Valerio Marsocci et al.

Large-scale foundation models in Earth Observation can learn versatile, label-efficient representations by leveraging massive amounts of unlabeled data. However, existing public datasets are often limited in scale, geographic coverage, or sensor variety. We introduce TerraMesh, a new globally diverse, multimodal dataset combining optical, synthetic aperture radar, elevation, and land-cover modalities in an Analysis-Ready Data format. TerraMesh includes over 9~million samples with eight spatiotemporal aligned modalities, enabling large-scale pre-training. We provide detailed data processing steps, comprehensive statistics, and empirical evidence demonstrating improved model performance when pre-trained on TerraMesh. The dataset is hosted at https://huggingface.co/datasets/ibm-esa-geospatial/TerraMesh.

Benedikt Blumenstiel, Nassim Ait Ali Braham, Conrad M Albrecht et al.

This technical report presents SSL4EO-S12 v1.1, a multimodal, multitemporal Earth Observation dataset designed for pretraining large-scale foundation models. Building on the success of SSL4EO-S12 v1.0, the new version addresses the previous challenges of data misalignment and a limited data structure for low-barrier, analysis-ready EO processing. SSL4EO-S12 v1.1 covers the world's 10,000 largest cities and its surroundings within a 50 km radius across four seasons, resulting in a diverse collection of nearly one million patches. SSL4EO-S12 v1.1 packages the data in Zarr file format for cloud-efficient loading and representation of meta-information such as including cloud masks and geolocation. Released under the CC-BY-4.0 license, SSL4EO-S12 v1.1 facilitates open research and provides a robust foundation for future advancements in self-supervised learning and geospatial analysis. The dataset is available online through https://datapub.fz-juelich.de/ssl4eo-s12, and we provided additional resources at https://github.com/DLR-MF-DAS/SSL4EO-S12-v1.1.

Francesc Marti-Escofet, Benedikt Blumenstiel, Linus Scheibenreif et al.

Earth observation (EO) is crucial for monitoring environmental changes, responding to disasters, and managing natural resources. In this context, foundation models facilitate remote sensing image analysis to retrieve relevant geoinformation accurately and efficiently. However, as these models grow in size, fine-tuning becomes increasingly challenging due to the associated computational resources and costs, limiting their accessibility and scalability. Furthermore, full fine-tuning can lead to forgetting pre-trained features and even degrade model generalization. To address this, Parameter-Efficient Fine-Tuning (PEFT) techniques offer a promising solution. In this paper, we conduct extensive experiments with various foundation model architectures and PEFT techniques to evaluate their effectiveness on five different EO datasets. Our results provide a comprehensive comparison, offering insights into when and how PEFT methods support the adaptation of pre-trained geospatial models. We demonstrate that PEFT techniques match or even exceed full fine-tuning performance and enhance model generalisation to unseen geographic regions, while reducing training time and memory requirements. Additional experiments investigate the effect of architecture choices such as the decoder type or the use of metadata, suggesting UNet decoders and fine-tuning without metadata as the recommended configuration. We have integrated all evaluated foundation models and techniques into the open-source package TerraTorch to support quick, scalable, and cost-effective model adaptation.

Daniela Szwarcman, Sujit Roy, Paolo Fraccaro et al.

This technical report presents Prithvi-EO-2.0, a new geospatial foundation model that offers significant improvements over its predecessor, Prithvi-EO-1.0. Trained on 4.2M global time series samples from NASA's Harmonized Landsat and Sentinel-2 data archive at 30m resolution, the new 300M and 600M parameter models incorporate temporal and location embeddings for enhanced performance across various geospatial tasks. Through extensive benchmarking with GEO-Bench, the 600M version 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 0.1m to 15m). 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) are among the key factors that contributed to the project's success. In particular, SME involvement allowed for constant feedback on model and dataset design, as well as successful customization for diverse SME-led applications in disaster response, land use and crop mapping, and ecosystem dynamics monitoring. Prithvi-EO-2.0 is available on Hugging Face and IBM terratorch, with additional resources on GitHub. The project exemplifies the Trusted Open Science approach embraced by all involved organizations.

Carlos Gomes, Isabelle Wittmann, Damien Robert et al.

Over the past decades, there has been an explosion in the amount of available Earth Observation (EO) data. The unprecedented coverage of the Earth's surface and atmosphere by satellite imagery has resulted in large volumes of data that must be transmitted to ground stations, stored in data centers, and distributed to end users. Modern Earth System Models (ESMs) face similar challenges, operating at high spatial and temporal resolutions, producing petabytes of data per simulated day. Data compression has gained relevance over the past decade, with neural compression (NC) emerging from deep learning and information theory, making EO data and ESM outputs ideal candidates due to their abundance of unlabeled data. In this review, we outline recent developments in NC applied to geospatial data. We introduce the fundamental concepts of NC including seminal works in its traditional applications to image and video compression domains with focus on lossy compression. We discuss the unique characteristics of EO and ESM data, contrasting them with "natural images", and explain the additional challenges and opportunities they present. Moreover, we review current applications of NC across various EO modalities and explore the limited efforts in ESM compression to date. The advent of self-supervised learning (SSL) and foundation models (FM) has advanced methods to efficiently distill representations from vast unlabeled data. We connect these developments to NC for EO, highlighting the similarities between the two fields and elaborate on the potential of transferring compressed feature representations for machine--to--machine communication. Based on insights drawn from this review, we devise future directions relevant to applications in EO and ESM.

Akashah Shabbir, Muhammad Akhtar Munir, Akshay Dudhane et al.

Recent progress in large language models (LLMs) has enabled tool-augmented agents capable of solving complex real-world tasks through step-by-step reasoning. However, existing evaluations often focus on general-purpose or multimodal scenarios, leaving a gap in domain-specific benchmarks that assess tool-use capabilities in complex remote sensing use cases. We present ThinkGeo, an agentic benchmark designed to evaluate LLM-driven agents on remote sensing tasks via structured tool use and multi-step planning. Inspired by tool-interaction paradigms, ThinkGeo includes human-curated queries spanning a wide range of real-world applications such as urban planning, disaster assessment and change analysis, environmental monitoring, transportation analysis, aviation monitoring, recreational infrastructure, and industrial site analysis. Queries are grounded in satellite or aerial imagery, including both optical RGB and SAR data, and require agents to reason through a diverse toolset. We implement a ReAct-style interaction loop and evaluate both open and closed-source LLMs (e.g., GPT-4o, Qwen2.5) on 486 structured agentic tasks with 1,773 expert-verified reasoning steps. The benchmark reports both step-wise execution metrics and final answer correctness. Our analysis reveals notable disparities in tool accuracy and planning consistency across models. ThinkGeo provides the first extensive testbed for evaluating how tool-enabled LLMs handle spatial reasoning in remote sensing.

Ruben Gonzalez Avilés, Linus Scheibenreif, Nassim Ait Ali Braham et al.

Hyperspectral imaging provides detailed spectral information and holds significant potential for monitoring of greenhouse gases (GHGs). However, its application is constrained by limited spatial coverage and infrequent revisit times. In contrast, multispectral imaging offers broader spatial and temporal coverage but often lacks the spectral detail that can enhance GHG detection. To address these challenges, this study proposes a spectral transformer model that synthesizes hyperspectral data from multispectral inputs. The model is pre-trained via a band-wise masked autoencoder and subsequently fine-tuned on spatio-temporally aligned multispectral-hyperspectral image pairs. The resulting synthetic hyperspectral data retain the spatial and temporal benefits of multispectral imagery and improve GHG prediction accuracy relative to using multispectral data alone. This approach effectively bridges the trade-off between spectral resolution and coverage, highlighting its potential to advance atmospheric monitoring by combining the strengths of hyperspectral and multispectral systems with self-supervised deep learning.

Naomi Simumba, Nils Lehmann, Paolo Fraccaro et al.

Geospatial Foundation Models (GeoFMs) are transforming Earth Observation (EO), but evaluation lacks standardized protocols. GEO-Bench-2 addresses this with a comprehensive framework spanning classification, segmentation, regression, object detection, and instance segmentation across 19 permissively-licensed datasets. We introduce ''capability'' groups to rank models on datasets that share common characteristics (e.g., resolution, bands, temporality). This enables users to identify which models excel in each capability and determine which areas need improvement in future work. To support both fair comparison and methodological innovation, we define a prescriptive yet flexible evaluation protocol. This not only ensures consistency in benchmarking but also facilitates research into model adaptation strategies, a key and open challenge in advancing GeoFMs for downstream tasks. Our experiments show that no single model dominates across all tasks, confirming the specificity of the choices made during architecture design and pretraining. While models pretrained on natural images (ConvNext ImageNet, DINO V3) excel on high-resolution tasks, EO-specific models (TerraMind, Prithvi, and Clay) outperform them on multispectral applications such as agriculture and disaster response. These findings demonstrate that optimal model choice depends on task requirements, data modalities, and constraints. This shows that the goal of a single GeoFM model that performs well across all tasks remains open for future research. GEO-Bench-2 enables informed, reproducible GeoFM evaluation tailored to specific use cases. Code, data, and leaderboard for GEO-Bench-2 are publicly released under a permissive license.

Mustansar Fiaz, Hiyam Debary, Paolo Fraccaro et al.

Recent advances in reinforcement learning (RL) have delivered strong reasoning capabilities in natural image domains, yet their potential for Earth Observation (EO) remains largely unexplored. EO tasks introduce unique challenges, spanning referred object detection, image or region captioning, change detection, grounding, and temporal analysis, that demand task aware reasoning. We propose a novel post training framework that incorporates task aware rewards to enable effective adaptation of reasoning based RL models to diverse EO tasks. This training strategy enhances reasoning capabilities for remote sensing images, stabilizes optimization, and improves robustness. Extensive experiments across multiple EO benchmarks show consistent performance gains over state of the art generic and specialized vision language models. Code and models will be released publicly at https://mustansarfiaz.github.io/GeoVLM-R1/ .

Geoffrey Dawson, Remy Vandaele, Andrew Taylor et al.

Artificial Intelligence (AI) Foundation models (FMs), pre-trained on massive unlabelled datasets, have the potential to drastically change AI applications in ocean science, where labelled data are often sparse and expensive to collect. In this work, we describe a new foundation model using the Prithvi-EO Vision Transformer architecture which has been pre-trained to reconstruct data from the Sentinel-3 Ocean and Land Colour Instrument (OLCI). We evaluate the model by fine-tuning on two downstream marine earth observation tasks. We first assess model performance compared to current baseline models used to quantify chlorophyll concentration. We then evaluate the FMs ability to refine remote sensing-based estimates of ocean primary production. Our results demonstrate the utility of self-trained FMs for marine monitoring, in particular for making use of small amounts of high quality labelled data and in capturing detailed spatial patterns of ocean colour whilst matching point observations. We conclude that this new generation of geospatial AI models has the potential to provide more robust, data-driven insights into ocean ecosystems and their role in global climate processes.

Ruben Gonzalez, Conrad M Albrecht, Nassim Ait Ali Braham et al.

Hyperspectral imaging provides detailed spectral information, offering significant potential for monitoring greenhouse gases like CH4 and NO2. However, its application is constrained by limited spatial coverage and infrequent revisit times. In contrast, multispectral imaging delivers broader spatial and temporal coverage but lacks the spectral granularity required for precise GHG detection. To address these challenges, this study proposes Spectral and Spatial-Spectral transformer models that reconstruct hyperspectral data from multispectral inputs. The models in this paper are pretrained on EnMAP and EMIT datasets and fine-tuned on spatio-temporally aligned (Sentinel-2, EnMAP) and (HLS-S30, EMIT) image pairs respectively. Our model has the potential to enhance atmospheric monitoring by combining the strengths of hyperspectral and multispectral imaging systems.

Michal Muszynski, Levente Klein, Ademir Ferreira da Silva et al.

Global vegetation structure mapping is critical for understanding the global carbon cycle and maximizing the efficacy of nature-based carbon sequestration initiatives. Moreover, vegetation structure mapping can help reduce the impacts of climate change by, for example, guiding actions to improve water security, increase biodiversity and reduce flood risk. Global satellite measurements provide an important set of observations for monitoring and managing deforestation and degradation of existing forests, natural forest regeneration, reforestation, biodiversity restoration, and the implementation of sustainable agricultural practices. In this paper, we explore the effectiveness of fine-tuning of a geospatial foundation model to estimate above-ground biomass (AGB) using space-borne data collected across different eco-regions in Brazil. The fine-tuned model architecture consisted of a Swin-B transformer as the encoder (i.e., backbone) and a single convolutional layer for the decoder head. All results were compared to a U-Net which was trained as the baseline model Experimental results of this sparse-label prediction task demonstrate that the fine-tuned geospatial foundation model with a frozen encoder has comparable performance to a U-Net trained from scratch. This is despite the fine-tuned model having 13 times less parameters requiring optimization, which saves both time and compute resources. Further, we explore the transfer-learning capabilities of the geospatial foundation models by fine-tuning on satellite imagery with sparse labels from different eco-regions in Brazil.