This study examines the integration of artificial intelligence (AI) into the processing of Synthetic Aperture Radar (SAR) imagery by the Brazilian Air Force to strengthen the strategic monitoring of the Amazon. Although the literature on the Revolution in Military Affairs and Network-Centric Warfare is well established, how AI-based automation transforms state capacity for territorial governance and the projection of sovereignty in contexts of non-traditional threats remains underexplored. Methodologically, the research adopts a qualitative and exploratory approach, based on document analysis and a systematic review of specialized literature (2022–2025). The analysis is conducted in light of the concepts of Data-Centric Warfare and Cyber Statecraft, examining how AI–SAR integration reconfigures data-mediated territorial governance and strengthens informational and technological sovereignty. The results demonstrate significant operational gains: reduced analysis time, increased target-detection accuracy, resource optimization, and compression of the Observe–Orient–Decide–Act cycle. Long-term sustainability requires robust ethical governance, algorithmic transparency, and accountability to mitigate risks of bias and discriminatory surveillance. It is concluded that AI–SAR integration represents a strategic reconfiguration of the relationship between the state, technology, and sovereignty, offering an adaptable model for Global South countries facing analogous challenges of territorial control.

Ship detection in Synthetic Aperture Radar (SAR) imagery remains challenging due to complex backgrounds, scale variations, and limited semantic discrimination in conventional detectors. To address these critical challenges, we propose SAR-SwinX (SAR Swin Transformer-enhanced YOLOX): a novel hybrid lightweight one-stage detection model composed of anchor-free Exceeding You Only Look Once (YOLOX) as a baseline and enhanced with Swin transformer modules based on cross-stage partial connections (CSP) to improve contextual representation. This hybrid design combines the local feature extraction strengths of Convolutional Neural Networks (CNNs) with the global semantic modeling of visual transformers, enabling effective multiscale ship detection in cluttered maritime scenes. Extensive experiments conducted on two public SAR datasets, including SSDD and HRSID, consistently validate the superiority of SAR-SwinX over the baseline YOLOX-s and existing state-of-the-art methods. A key result of our approach is that SAR-SwinX improves mAP@50:95 by 2.03% and 3.46%, enhances recall by 0.73% and 0.52%, and boosts the F1-score by 0.56% and 0.16% for SSDD and HRSID, respectively. These results highlight SAR-SwinX as an efficient and robust solution for SAR ship detection in complex environments, with favorable computational efficiency.

Synthetic aperture radar (SAR) imagery is widely used for target detection in complex backgrounds and adverse weather conditions. However, high-precision detection of rotated small targets remains challenging due to severe speckle noise, significant scale variations, and the need for robust rotation-aware representations. To address these issues, we propose SAR-DRBNet, a high-precision rotated small-target detection framework built upon YOLOv13. First, we introduce a Detail-Enhanced Oriented Bounding Box detection head (DEOBB), which leverages multi-branch enhanced convolutions to strengthen fine-grained feature extraction and improve oriented bounding box regression, thereby enhancing rotation sensitivity and localization accuracy for small targets. Second, we design a Ck-MultiDilated Reparameterization Block (CkDRB) that captures multi-scale contextual cues and suppresses speckle interference via multi-branch dilated convolutions and an efficient reparameterization strategy. Third, we propose a Dynamic Feature Weaving module (DynWeave) that integrates global–local dual attention with dynamic large-kernel convolutions to adaptively fuse features across scales and orientations, improving robustness in cluttered SAR scenes. Extensive experiments on three widely used SAR rotated object detection benchmarks (HRSID, RSDD-SAR, and DSSDD) demonstrate that SAR-DRBNet achieves a strong balance between detection accuracy and computational efficiency compared with state-of-the-art oriented bounding box detectors, while exhibiting superior cross-dataset generalization. These results indicate that SAR-DRBNet provides an effective and reliable solution for rotated small-target detection in SAR imagery.

Optical and synthetic aperture radar (SAR) imagery are highly complementary in terms of texture details and structural scattering characterization. However, their imaging mechanisms and statistical distributions differ substantially. In particular, pseudo-high-frequency components introduced by SAR coherent speckle can be easily entangled with genuine optical edges, leading to texture mismatch, structural drift, and noise diffusion. To address these issues, we propose WEMFusion, a wavelet-prior-driven framework for frequency-domain decoupling and discrepancy-aware state-space fusion. Specifically, a multi-scale discrete wavelet transform (DWT) explicitly decomposes the inputs into low-frequency structural components and directional high-frequency sub-bands, providing an interpretable frequency-domain constraint for cross-modality alignment. We design a hybrid-modality enhancement (HME) module: in the high-frequency branch, it effectively injects optical edges and directional textures while suppressing the propagation of pseudo-high-frequency artifacts, and in the low-frequency branch, it reinforces global structural consistency and prevents speckle perturbations from leaking into the structural component, thereby mitigating structural drift. Furthermore, we introduce a discrepancy-aware gated Mamba fusion (DAG-MF) block, which generates dynamic gates from modality differences and complementary responses to modulate the parameters of a directionally scanned two-dimensional state-space model, so that long-range dependency modeling focuses on discrepant regions while preserving directional coherence. Extensive quantitative evaluations and qualitative comparisons demonstrate that WEMFusion consistently improves structural fidelity and edge detail preservation across multiple optical–SAR datasets, achieving superior fusion quality with lower computational overhead.

Synthetic Aperture Radar (SAR) is a powerful observation system capable of delivering high-resolution imagery under variable sea conditions to support target detection and tracking, such as for ships. However, conventional optical target detection models are typically engineered for complex optical imagery, leading to limitations in accuracy and high computational resource consumption when directly applied to SAR imagery. To address this, this paper proposes a lightweight shape-aware and direction-weighted algorithm for SAR ship detection, SADW-Det. First, a lightweight streamlined backbone network, LSFP-NET, is redesigned based on the YOLOX architecture. This achieves reduced parameter counts and computational burden by incorporating depthwise separable convolutions and factorized convolutions. Concurrently, a parallel fusion module is designed, leveraging multiple small-kernel depthwise separable convolutions to extract features in parallel. This approach maintains accuracy while achieving lightweight processing. Furthermore, addressing the differences between SAR imagery and other imaging modalities, a direction-weighted attention was devised. This enhances model performance with minimal computational overhead by incorporating positional information while preserving channel data. Experimental results demonstrate superior detection accuracy compared to existing methods on three representative SAR datasets, SSDD, HRSID and DSSDD, while achieving reduced parameter counts and computational complexity, indicating strong application potential and laying the foundation for cross-modal applications.

The performance of deep learning approaches for Synthetic Aperture Radar (SAR) target detection is often limited by the scarcity of annotated data. While Self-Supervised Learning (SSL) has emerged as a powerful paradigm to mitigate data dependence, its potential in SAR target detection remains largely underexplored. In this study, we propose SARDet-MIM, a comprehensive framework based on Masked Image Modeling (MIM), to enhance SAR target detection. The approach consists of two stages. In the self-supervised pre-training stage, we propose an innovative Structural and Scattering Masked Autoencoder (SSMAE) method for SAR imagery. Unlike conventional MIM methods, which typically reconstruct raw pixels, SSMAE employs a physics-aware reconstruction target comprising multi-scale gradient and SAR-Harris features. This strategy explicitly guides the network to capture discriminative structural contexts and intrinsic scattering features that benefit SAR target detection. For downstream detection, we construct a Maximally Pre-trained Detector (MPD), which integrally transfers the pre-trained ViT encoder–decoder architecture to the detection network to fully exploit pre-trained representations. Extensive experiments on three SAR target detection datasets demonstrate that SARDet-MIM consistently outperforms competing methods.

The coherent speckle noise in synthetic aperture radar (SAR) imagery, together with complex sea clutter and large variations in ship target scales, poses significant challenges to accurate and robust ship detection, particularly under strict lightweight constraints required by satellite-borne and airborne platforms. To address this issue, this paper proposes a high-precision lightweight detection network, termed High-Lightweight Net (HLNet), specifically designed for SAR ship detection. The network incorporates a novel multi-scale backbone, Multi-Scale Net (MSNet), which integrates dynamic feature completion and multi-core parallel convolutions to alleviate small-target feature loss and suppress background interference. To further enhance multi-scale feature fusion while reducing model complexity, a lightweight path aggregation feature pyramid network, High-Lightweight Feature Pyramid (HLPAFPN), is introduced by reconstructing fusion pathways and removing redundant channels. In addition, a lightweight detection head, High-Lightweight Head (HLHead), is designed by combining grouped convolutions with distribution focal loss to improve localization robustness under low signal-to-noise ratio conditions. Extensive experiments conducted on the public SSDD and HRSID datasets demonstrate that HLNet achieves mAP50 scores of 98.3% and 91.7%, respectively, with only 0.66 M parameters. Extensive evaluations on the more challenging CSID subset, composed of complex scenes selected from SSDD and HRSID, demonstrate that HLNet attains an mAP50 of 75.9%, outperforming the baseline by 4.3%. These results indicate that HLNet achieves an effective balance between detection accuracy and computational efficiency, making it well-suited for deployment on resource-constrained SAR platforms.

With the advancement of synthetic aperture radar (SAR) sensor technology, linear structures such as building facades have become increasingly discernible in SAR imagery. Accurate detection of these line features is critical for object recognition and 3D model reconstruction. To the best of our knowledge, few existing methods explicitly address the problem of detecting lines composed of point scatterers. In this paper, we analyze the characteristics of such lines and propose a novel point scatterer-driven growth-based approach, termed PSG-Line, for their detection. Point scatterers are first extracted by combining the ordered-statistics constant false alarm rate (OS-CFAR) algorithm with non-maximum suppression and Harris corner response thresholding. Line segments are then initiated from these scatterers and iteratively extended by incorporating subsequent points that satisfy a set of geometric constraints. Finally, the detected line segments are validated based on the Helmholtz principle. Local principal orientations of point scatterers are estimated and incorporated into the line segment growth and validation stages. Both simulation and real-life SAR data experiments demonstrate that the PSG-Line algorithm outperforms existing line detection methods in accurately detecting lines composed of point scatterers.

Traditional field-based techniques are primarily used to monitor and assess the potential maize harvest. This study aimed to integrate Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 optical imagery to map maize fields and monitor their growth stages in the Kasisi area of Chongwe District, Zambia. Multi-temporal satellite data (biweekly from November 2019 to April 2020) were analysed to capture phenological changes of maize. Dual-polarised Sentinel-1 SAR (VV and VH) backscatter was combined with vegetation indices (Normalised Difference Vegetation Index – NDVI, and Normalised Difference Water Index – NDWI) from Sentinel-2. These features were used to train a Random Forest classifier to delineate maize fields, using field-collected training data. The results show that the classification had a very high overall accuracy of 96.7% (Kappa = 0.95), successfully distinguishing maize from other land covers. Maize fields were mapped, covering about 6721 ha (about 55% of the study area). Temporal analyses of SAR and NDVI identified four critical growth phases: sowing (early November 2019), emergence (mid-December 2019), maturity (mid-January to mid-February 2020), and harvest (March–April 2020). SAR backscatter increased during vegetative growth, stagnated at maturity, and declined after harvest. At the same time, NDVI trends peaked in late March and dropped by the end of April, confirming crop senescence in Zambia. The integration of radar and optical data proved effective for agricultural monitoring in a cloud-prone region, offering a scalable and timely approach to crop mapping. These results demonstrate that remote sensing can provide near-real-time information to support yield prediction, thereby supporting planning, decision-making, and policy interventions in the agriculture sector.

Recognition for Targets Type of Synthetic Aperture Radar Imagery using Deep Learning Methods

Reliable observation of water resources is a major challenge for sustainable development, particularly in the river-centric deltaic countries like Bangladesh, where the data is generally scarce. Leveraging operational satellites, this study presents a real-time capable water level (WL), discharge (Q), and floodplain monitoring framework implemented for the Brahmaputra River in Bangladesh. The multi-satellite approach presented here combined satellite altimetry, synthetic aperture radar (SAR), and optical imagery. A set of WL time series is obtained first from Jason-2/3 and Sentinel-3 altimetry, while a combination of Sentinel-1 SAR and Sentinel-2 optical images is used to extract the floodplain extent. Seasonal Rating Curve (RC) models are then developed to estimate Q from the river WL (altimetry) and width (imagery). The altimetry WL measurement is further complemented by the width-based Q utilizing an inverse RC. Furthermore, the water level is combined with a floodplain map to extract floodplain topography and its evolution. The proposed framework provides consistent and reliable observations in the Brahmaputra River, with a bias, root mean-squared errors (RMSEs), and correlation coefficient of 0.03 m, 0.68 m, and 0.96 for WL, and −168.22 m3/s, 4161.46 m3/s, and 0.97 for Q, respectively, relative to a mean discharge of approximately 30,000 m3/s. The locations of high erosion–accretion across the river reach are also well-captured in the evolving floodplain maps. By integrating multiple satellite altimetry missions with SAR and optical imagery, the multi-satellite approach reduces the effective monitoring interval for both water level and discharge from approximately 10 days (single-mission altimetry) to about 4 days, enabling improved capture of extreme events such as floods. As the operational satellites used in this study are expected to provide long-term observations, the proposed framework supports sustainable monitoring of floodplain dynamics in Bangladesh and other similar data-poor environments, towards informed water management under ongoing climatic and anthropogenic changes.

Abstract Ship classification in synthetic aperture radar (SAR) imagery plays a critical role in many practical applications. However, existing semi-supervised frameworks still suffer from unstable pseudo-label quality and insufficient utilization of unlabeled data. In this article, we propose a dual-teacher network with an adaptive reliable bank (DTRB-Net), a semi-supervised learning framework that integrates a dual-teacher architecture with an adaptive reliable bank to address these challenges. The teacher network and its subnetwork collaboratively generate pseudo-label pairs that are then filtered through a two-stage selection strategy based on pseudo-label confidence and prediction discrepancy to ensure reliability. These high-quality pseudo-label pairs are stored and dynamically updated in a class-wise adaptive reliable bank, providing stable contrastive samples for the student network, and enabling more effective exploitation of unlabeled data. In addition, we design a new loss function that jointly leverages labeled and unlabeled data to enhance the student network’s feature learning capability. Extensive experiments on the FUSAR-Ship and OpenSARShip datasets show that DTRB-Net achieves superior accuracy on both three- and six-class ship classification tasks compared with existing methods, demonstrating the effectiveness and robustness of the proposed framework.

Validation of flood inundation modelling using multi-source SAR imagery.

Owing to recent extreme weather events, flood risk has been rising annually, increasing the demand for fast and accurate flood mapping. Synthetic aperture radar imagery has received considerable attention for flood-mapping applications owing to its all-weather, day-and-night imaging capabilities. Although previous studies have achieved accurate mapping in non-urban areas, challenges remain for urban regions. This study focuses on flood events in Japan by employing a deep learning model and PALSAR-2 imagery to classify non-flooded areas, floods in open areas, and floods in urban areas. To understand the complex spectral characteristics specific to urban areas, this study investigates the integration of geographical features, such as slope and building footprints, into the segmentation process. The experimental results suggest that the inclusion of these supplementary data improves the prediction performance of the trained models.

ABSTRACT The high acquisition cost of synthetic aperture radar (SAR) imagery has been a persistent obstacle to advanced deep learning researches. Recently, image translation techniques have emerged as promising solutions for augmenting SAR datasets by translating readily available optical images into SAR-like representations. However, the substantial stylistic differences between optical and SAR images pose significant challenges in accurately extracting optical image semantics and replicating SAR image styles, especially when co-registered data is unavailable. To address this challenge, we propose an unpaired optical-to-SAR image translation (O2SIT) method, named extract-and-transform generative adversarial network (ET-GAN). First, we introduce cascaded coordinate attention (CA) bottleneck blocks that enhance the positional information of feature maps, thereby precisely extracting optical image semantics. Second, to better capture SAR style characteristics, we employ histograms as auxiliary supervision by constructing a differentiable histogram using kernel density estimation and global average pooling. On this basis, the squared earth mover distance is adopted as an additional loss to guide the generator in producing synthetic images with pixel distributions similar to real SAR images. Experimental results on SEN12, WHU-SEN-City, and GaoFen aircraft detection (GF-AD) dataset demonstrate that ET-GAN achieves competitive SAR image generation performance compared to other state-of-the-art methods, with PSNR of 17.11 on SEN12 and FID of 168.16 on GF-AD. Transfer learning results demonstrate that the images generated by ET-GAN can bring about 3% accuracy improvement to SAR aircraft detection.

Land subsidence constitutes a critical hazard to coastal megacities globally, amplifying flood risks and damaging infrastructure. Taking Tianjin—a major port city underlain by compressible sediments and affected by groundwater over-exploitation—as a case study, we address two key research gaps: the absence of a quantitative framework coupling groundwater extraction with construction land expansion, and the inadequate separation of seasonal and long-term subsidence drivers. We developed an integrated remote-sensing-based approach: high-resolution subsidence time series (2016–2023) were derived via Small BAseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) using Sentinel-1 Synthetic Aperture Radar (SAR) imagery, validated against leveling measurements (R > 0.885, error < 20 mm). This subsidence dataset was fused with groundwater level records and annual construction land maps. Seasonal-Trend Decomposition using Loess (STL) isolated trend, seasonal, and residual components, which were input into a Random Forest (RF) model to quantify the relative contributions of subsidence drivers. Dynamic Time Warping (DTW) and Cross-Wavelet Transform (CWT) were further employed to characterize temporal patterns and lag effects between subsidence and its drivers. Our results reveal a distinct shifting subsidence pattern: “areal expansion but intensity weakening.” Groundwater control policies mitigated five historical subsidence funnels, reducing areas with severe subsidence from 72.36% to <5%, while the total subsiding area expanded by 1024.74 km2, with new zones emerging (e.g., northern Dongli District). The RF model identified the long-term groundwater level trend as the dominant driver (59.5% contribution), followed by residual (23.3%) and seasonal (17.2%) components. Cross-spectral analysis confirmed high coherence between subsidence and long-term groundwater trends; the seasonal component exhibited a dominant resonance period of 12 months and a consistent subsidence response lag of 3–4 months. Construction impacts were conceptualized as a “load accumulation-soil compression-time lag” mechanism, with high-intensity engineering projects inducing significant local subsidence. This study provides a robust quantitative framework for disentangling the complex interactions between subsidence, groundwater, and urban expansion, offering critical insights for evidence-based hazard mitigation and sustainable urban planning in vulnerable coastal environments worldwide.

Ship monitoring using Synthetic Aperture Radar (SAR) data faces significant challenges in detecting small vessels due to low spatial resolution and speckle noise. While ESRGAN (Enhanced Super-Resolution Generative Adversarial Network) has shown promise for image super-resolution, it struggles with SAR imagery characteristics. This study proposes SA/SU-ESRGAN, which extends the SU-ESRGAN framework by incorporating a spatial attention mechanism loss function. SU-ESRGAN introduced semantic structural loss to accurately preserve ship shapes and contours; our enhancement adds spatial attention to focus reconstruction efforts on ship regions while suppressing background noise. Experimental results demonstrate that SA/SU-ESRGAN successfully detects small vessels that remain undetectable by SU-ESRGAN, achieving improved detection capabilities with a PSNR of approximately 26 dB (SSIM is around 0.5) and enhanced visual clarity in ship boundaries. The spatial attention mechanism effectively reduces noise influence, producing clearer super-resolution results suitable for maritime surveillance applications. Based on the HRSID dataset, a representative dataset for evaluating ship detection performance using SAR data, we evaluated ship detection performance using images in which the spatial resolution of the SAR data was artificially degraded using a smoothing filter. We found that with a 4 × 4 filter, all eight ships were detected without any problems, but with an 8 × 8 filter, only three of the eight ships were detected. When super-resolution was applied to this, six ships were detected.

In many land cover classification tasks, the limited precision of individual sensors hinders the accurate separation of certain classes, largely due to the complexity of the Earth’s surface morphology. To mitigate these issues, decision fusion methodologies are employed, allowing data from multiple sensors to be synthesized into robust and more conclusive classification outcomes. This study employs fully polarimetric Synthetic Aperture Radar (PolSAR) imagery and leverages the strengths of three decomposition methods, namely Pauli’s, Krogager’s, and Cloude’s, by extracting their respective components for improved detection. From each decomposition method, three scattering components are derived, enabling the extraction of informative features that describe the scattering behavior associated with various land cover types. The extracted scattering features, treated as independent sensors, were used to train three neural network classifiers. The resulting outputs were then considered as local decisions for each land cover type and subsequently fused through a decision fusion rule to generate more complete and accurate classification results. Experimental results demonstrate that the proposed Multi-Class Bahadur–Lazarsfeld Expansion (MC-BLE) fusion significantly enhances classification performance, achieving an overall accuracy (OA) of 95.78% and a Kappa coefficient of 0.94. Compared to individual classification methods, the fusion notably improved per-class accuracy, particularly for complex land cover boundaries. The core innovation of this work is the transformation of the Bahadur–Lazarsfeld Expansion (BLE), originally designed for binary decision fusion into a multi-class framework capable of addressing multiple land cover types, resulting in a more effective and reliable decision fusion strategy.

Automatic Electrical energy plays a critical role in the economic and technological development of modern societies. High voltage lines are utilized for the efficient distribution of this energy. These lines enhance energy efficiency and sustainability by minimizing energy transmission losses. Ensuring the safe and uninterrupted transmission of electrical energy is of paramount importance for the robustness and maintenance of the infrastructure. Consequently, the role of high-voltage poles in the effective and efficient transmission of energy is of paramount importance. The objective of this study is to utilize remote sensing and various shallow learning techniques to identify high voltage line poles in the vicinity of Batman. The Sentinel-1 Synthetic Aperture Radar (SAR) satellite was utilized in this study. Sentinel-1 synthetic aperture radar (SAR) imagery was processed on the Google Earth Engine (GEE) platform to derive backscatter information for infrastructure and terrain analysis. Dual-polarization data (VV and VH) from both ascending and descending orbits were acquired between January and June 2025. Standard preprocessing steps, including thermal noise removal, radiometric calibration, and terrain correction using the SRTM 30 m digital elevation model, were automatically applied through the GEE Sentinel-1 GRD pipeline. Median composites were generated to minimize speckle noise and temporal variability. The derived VV and VH backscatter coefficients were subsequently used to extract spatial texture and backscatter difference features that characterize electrical pylons and transformer areas. The dataset was composed of VH (Vertical-Horizontal) and VV (Vertical-Vertical) polarization modes. The analysis of these datasets was conducted using algorithms belonging to the supervised learning model, which included the support vector machine, KNN, decision tree, quadratic discriminant, and naive Bayes algorithms. The findings indicate that the support vector machine model demonstrated an 85.0% success rate, the quadratic discriminant model exhibited an 82.5% success rate, the KNN model achieved an 82.2% success rate, the decision tree model attained a 76.8% success rate, and the naive Bayes model registered a 74.0% success rate. The study demonstrates the efficacy of artificial intelligence in facilitating more precise, expeditious, and systematic control of high-voltage poles. In light of the fact that conventional methods tend to be both time-consuming and costly when it comes to the maintenance of high voltage lines, the objective of this study is to enhance the efficiency of the energy infrastructure by means of contributing to the improvement of the monitoring and maintenance processes of power transmission lines.

Abstract. Ice cover of water bodies in the northern high latitudes (NHL) is highly sensitive to the changing climate, and its dynamics exert substantial impacts on the NHL ecosystems, hydrological processes, and the carbon cycle. Yet, operational quantification of ice cover dynamics for smaller water bodies (e.g., ≤25 km2) over vast, remote NHL regions remains limited. Here, we developed an ice fraction dataset for small water bodies (ponds, lakes, and rivers; 900 m2 to 25 km2) across the Arctic Coastal Plain of Alaska (ACP) from 2017 through 2023, using Sentinel-1 Synthetic Aperture Radar (SAR) imagery, texture features, and Daymet air temperature data. The dataset has a spatial resolution of 1 km and a temporal resolution of approximately 6 d. Compared with the Google Dynamic World (DW) product derived from Sentinel-2 optical remote sensing, our dataset shows high consistency with DW (R=0.91, RMSE=0.19) while having enhanced temporal coverage due to less SAR constraints from solar illumination, cloud cover, and atmospheric conditions. Validation against in-situ observations suggests that our dataset is more capable of capturing small water body ice phenology (e.g., freeze-up and break-up dates) relative to DW, with an 11 d reduction in mean absolute error. Our ice fraction dataset reveals high spatial heterogeneity in ice conditions mainly occurring in June for small water bodies across the ACP. The ice phenology analysis over three selected subregions further shows that a warmer transition period generally leads to earlier ice break-up and later freeze-up, while the responses of ice fraction to warming climate vary among and within individual water bodies. The resulting dataset is anticipated to fill a gap in ice phenology studies for small water bodies, improve our understanding on the interactions between ice dynamics and climate change, and enhance the coupled modelling of ice and carbon processes. The S1 ice fraction dataset is publicly available at https://doi.org/10.5281/zenodo.17033546 (Lin et al., 2025).