Beyond Accuracy: Assessing Calibration of Geospatial Foundation Models and Their Sensitivity to Distribution Shifts

Geospatial Foundation Models (GeoFMs) are most commonly ranked and selected by accuracy on standard benchmark conditions via averaged ranks. We show that this protocol is too narrow: the promised deployment in critical EO tasks requires further angles of analysis, mainly calibration, the agreement between a model's confidence and its correctness. Across 16 frozen encoders, four classification and five segmentation datasets, and two orthogonal stress axes, every encoder degrades as corruption intensifies, and the ranking changes as well. Across the four classification benchmarks, EO-pretrained and ImageNet-pretrained encoders are indistinguishable on clean accuracy and clean calibration, and EO pretraining provides no more stability under shift than ImageNet pretraining. Under shift the GeoFMs drift further into overconfidence than the ImageNet-pretrained encoders, at every grade and in every corruption family. A centered kernel alignment (CKA) analysis ties this to representational rigidity: EO-pretrained embeddings move less under corruption while losing just as much task information and remaining overconfident. We apply three commonly explored uncertainty quantification methods and find that temperature scaling and deep ensembles cannot counteract the degradation, while a Gaussian-process probe roughly halves ECE under severe cloud only by tripling it on clean data. In selective prediction experiments, we find that confidence-based abstention cannot defer around confidently wrong predictions, and advocate that benchmark rankings and evaluations should therefore operate across a multitude of conditions and metrics to more holistically evaluate model development progress and close the gap to real world deployment scenarios.

HeatCast: A Benchmark for Neighborhood-Scale LST Forecasting across 124 U.S. Cities

Land Surface Temperature (LST) is a widely used satellite-derived measure of urban surface heat, but there is no shared benchmark for forecasting it at 30 m. Prior studies usually cover one to three cities, use kilometer-scale products, or do not release data and code. We introduce HeatCast, a Landsat-based benchmark for monthly LST forecasting across 124 U.S. cities from 2013 through June 2025. HeatCast contains 30 m monthly tiles with LST, elevation, surfacereflectance RGB, three spectral indices, broadband albedo, quality masks, and Local Climate Zone (LCZ) labels, together with a fixed temporal split, LCZ-stratified metrics, and a reference evaluation harness. We evaluate a CNN+LSTM and Earthformer on next-month forecasting, where Earthformer reaches 7.74 K RMSE against 10.42 K for the CNN+LSTM. Forecasting from the eight nonLST channels alone reaches 7.72 K, against 8.15 K from LST history and 8.68 K from RGB. The data, code, and weights are released under MIT at https://doi.org/10.57967/hf/9889.

Fields of the Planet: Field Boundary Mapping Beyond 10m

Field-boundary maps support crop monitoring, irrigation planning, and yield estimation, but many smallholder parcels span only a few 10 m Sentinel-2 pixels. We introduce Fields of the Planet (FTP), a 3 m PlanetScope companion to Fields of The World (FTW) that pairs the same polygons, seasonal windows, and train/test splits with 133,168 co-registered PlanetScope patch-window targets across 24 countries. FTP evaluates field delineation as parcel recovery by vectorizing predictions before scoring panoptic quality (PQ), object F1, size-stratified PQ, and meter-scale matched-boundary error. Under matched architectures and training recipes, 3 m imagery raises PQ from 21.0 to 35.5, raises PQ on sub-0.5 ha fields from 5.8 to 15.7, and cuts matched-boundary error from 18.6 m to 7.4 m.

No One Knows the State of the Art in Geospatial Foundation Models

Geospatial foundation models (GFMs) have been proposed as generalizable backbones for disaster response, land-cover mapping, food-security monitoring, and other high-stakes Earth-observation tasks. Yet the published work about these models does not give reviewers or users enough information to tell which model fits a given task. We argue that nobody knows what the current state of the art is in geospatial foundation models. The methods may be useful, but the GFM literature does not standardize evaluations, training and testing protocols, released weights, or pretraining controls well enough for anyone to compare or rank them. In a 152-paper audit, we find 46 cross-paper disagreements of at least 10 points for the same model, benchmark, and protocol; 94/126 papers with extractable pretraining data use a configuration no other paper uses; and 39% of GFM papers release no model weights. This lack of community standards can be solved. We propose six concrete expectations: named-license weight release, shared core evaluations, copied-versus-rerun baseline annotations, variance reporting, one shared evaluation harness, and data-vs-architecture-vs-algorithm controls. These gaps are a coordination failure, not a fault of any individual lab; the authors of this paper, like many others in the GFM community, have contributed to them. Rather than just critiquing the community, we aim to provide concrete steps toward a shared understanding of how to innovate GFMs.

Are Pretrained Image Matchers Good Enough for SAR-Optical Satellite Registration?

Cross-modal optical-SAR (Synthetic Aperture Radar) registration is a bottleneck for disaster-response via remote sensing, yet modern image matchers are developed and benchmarked almost exclusively on natural-image domains. We evaluate twenty-four pretrained matcher families--in a zero-shot setting with no fine-tuning or domain adaptation on satellite or SAR data--on SpaceNet9 and two additional cross-modal benchmarks under a deterministic protocol with tiled large-image inference, robust geometric filtering, and tie-point-grounded metrics. Our results reveal asymmetric transfer--matchers with explicit cross-modal training do not uniformly outperform those without it. While XoFTR (trained for visible-thermal matching) and RoMa achieve the lowest reported mean error at $3.0$ px on the labeled SpaceNet9 training scenes, RoMa achieves this without any cross-modal training, and MatchAnything-ELoFTR ($3.4$ px)--trained on synthetic cross-modal pairs--matches closely, suggesting (as a working hypothesis) that foundation-model features (DINOv2) may contribute to modality invariance that partially substitutes for explicit cross-modal supervision. 3D-reconstruction matchers (MASt3R, DUSt3R), which are not designed for traditional 2D image matching, are highly protocol-sensitive and remain fragile under default settings. Deployment protocol choices (geometry model, tile size, inlier gating) shift accuracy by up to $33\times$ for a single matcher, sometimes exceeding the effect of swapping matchers entirely within the evaluated sweep--affine geometry alone reduces mean error from $12.34$ to $9.74$ px. These findings inform both practical deployment of existing matchers and future matcher design for cross-modal satellite registration.

PRUE: A Practical Recipe for Field Boundary Segmentation at Scale

Large-scale maps of field boundaries are essential for agricultural monitoring tasks. Existing deep learning approaches for satellite-based field mapping are sensitive to illumination, spatial scale, and changes in geographic location. We conduct the first systematic evaluation of segmentation and geospatial foundation models (GFMs) for global field boundary delineation using the Fields of The World (FTW) benchmark. We evaluate 18 models under unified experimental settings, showing that a U-Net semantic segmentation model outperforms instance-based and GFM alternatives on a suite of performance and deployment metrics. We propose a new segmentation approach that combines a U-Net backbone, composite loss functions, and targeted data augmentations to enhance performance and robustness under real-world conditions. Our model achieves a 76\% IoU and 47\% object-F1 on FTW, an increase of 6\% and 9\% over the previous baseline. Our approach provides a practical framework for reliable, scalable, and reproducible field boundary delineation across model design, training, and inference. We release all models and model-derived field boundary datasets for five countries.

From Pixels to Patches: Pooling Strategies for Earth Embeddings

As geospatial foundation models shift from patch-level to pixel-level embeddings, practitioners must aggregate thousands of pixel vectors into patch representations that preserve class-discriminative signal while matching downstream label resolution. The default choice, mean pooling, discards within-patch variability and can drop accuracy by more than 10% under spatial shift. To evaluate this effect, we introduce EuroSAT-Embed: 81,000 embedding GeoTIFFs derived from three foundation models: AlphaEarth, OlmoEarth, and Tessera. We benchmark 11 training-free and 2 parametric pooling methods under both random and geographically disjoint test splits. Our results show that richer pooling schemes reduce the geographic generalization gap by up to 40% relative to mean pooling and increases accuracy by up to 5% on spatial splits. We recommend Generalized Mean Pooling (GeM) as a drop-in replacement for mean pooling: it improves accuracy without increasing embedding dimensionality. For maximum accuracy, Stats pooling (concatenation of min/max/mean/std pooling) performs best at 4x the embedding size. We further find that pooling effectiveness varies across embedding sources and that higher-dimensional embeddings benefit most from distributional statistics.

Fields of The World: A Field Guide for Extracting Agricultural Field Boundaries

Field boundary maps are a building block for agricultural data products and support crop monitoring, yield estimation, and disease estimation. This tutorial presents the Fields of The World (FTW) ecosystem: a benchmark of 1.6M field polygons across 24 countries, pre-trained segmentation models, and command-line inference tools. We provide two notebooks that cover (1) local-scale field boundary extraction with crop classification and forest loss attribution, and (2) country-scale inference using cloud-optimized data. We use MOSAIKS random convolutional features and FTW derived field boundaries to map crop type at the field level and report macro F1 scores of 0.65--0.75 for crop type classification with limited labels. Finally, we show how to explore pre-computed predictions over five countries (4.76M km\textsuperscript{2}), with median predicted field areas from 0.06 ha (Rwanda) to 0.28 ha (Switzerland).

Earth Embeddings as Products: Taxonomy, Ecosystem, and Standardized Access

Geospatial Foundation Models (GFMs) provide powerful representations, but high compute costs hinder their widespread use. Pre-computed embedding data products offer a practical "frozen" alternative, yet they currently exist in a fragmented ecosystem of incompatible formats and resolutions. This lack of standardization creates an engineering bottleneck that prevents meaningful model comparison and reproducibility. We formalize this landscape through a three-layer taxonomy: Data, Tools, and Value. We survey existing products to identify interoperability barriers. To bridge this gap, we extend TorchGeo with a unified API that standardizes the loading and querying of diverse embedding products. By treating embeddings as first-class geospatial datasets, we decouple downstream analysis from model-specific engineering, providing a roadmap for more transparent and accessible Earth observation workflows.

HydroChronos: Forecasting Decades of Surface Water Change

Forecasting surface water dynamics is crucial for water resource management and climate change adaptation. However, the field lacks comprehensive datasets and standardized benchmarks. In this paper, we introduce HydroChronos, a large-scale, multi-modal spatiotemporal dataset for surface water dynamics forecasting designed to address this gap. We couple the dataset with three forecasting tasks. The dataset includes over three decades of aligned Landsat 5 and Sentinel-2 imagery, climate data, and Digital Elevation Models for diverse lakes and rivers across Europe, North America, and South America. We also propose AquaClimaTempo UNet, a novel spatiotemporal architecture with a dedicated climate data branch, as a strong benchmark baseline. Our model significantly outperforms a Persistence baseline for forecasting future water dynamics by +14% and +11% F1 across change detection and direction of change classification tasks, and by +0.1 MAE on the magnitude of change regression. Finally, we conduct an Explainable AI analysis to identify the key climate variables and input channels that influence surface water change, providing insights to inform and guide future modeling efforts.