Against the backdrop of global ageing, a growing empirical literature highlights associations between age-friendly environments (AFEs) and older adults' well-being. However, the evidence is difficult to cumulate because AFE measurement is fragmented, characterized by heterogeneous operational approaches and uneven domain coverage. Importantly, this measurement heterogeneity is not merely a methodological limitation; it acts as a structuring force that shapes inference, comparability, and policy learning across studies. This systematic review aims to map how AFEs are operationalized and how they are connected to older adults' well-being. Following PRISMA, four databases were searched to April 2025; 5305 records were screened, and 58 studies were included. We classified AFE measures as subjective (n = 37), objective (n = 11), and mixed (n = 10) and mapped indicators to the WHO Age-Friendly Cities and Communities (AFCC) domains. Our synthesis reveals that AFE measurement is skewed to physical domains, such as transportation, outdoor spaces and buildings, and health services, with social domains underspecified. Associations are often positive for features such as reliable public transport, proximate everyday services, and opportunities for social participation; however, magnitudes and even directions vary by context and by how AFEs are operationalized. We also observe non-linear patterns (for example, in street connectivity or greenness), and occasional negative associations, particularly in disadvantaged settings where higher service density may signal unmet need rather than support. Some of this heterogeneity likely reflects analytic choices, including denominator selection, buffer size and shape, and the omission of micro-scale features and barriers. We recommend a comprehensive AFE assessment that combines subjective and objective approaches, balances physical and social domains, and improves transparency on validity and cross-cultural equivalence. Emerging tools such as street-view imagery and other fine-scale urban data can help capture micro-spatial features with greater precision.

Can transit-oriented development (TOD) cool the city? A multi-scale geospatial assessment of daytime surface heat exposure in Tokyo

Fine-scale aboveground biomass estimation of restored vegetation in tailings ponds: integrating UAV multi-source data and explainable machine learning

Scalable dynamic origin-destination demand estimation enhanced by high-resolution satellite imagery data

Net primary production (NPP) is a key indicator of the terrestrial carbon cycle, and its response to disturbance and subsequent recovery is important for understanding regional carbon sink dynamics. Conventional region-based statistical approaches have limitations in capturing localized heterogeneous changes. In this study, a typical ecologically fragile region on the northwestern Sichuan Plateau was selected as the study area. Using the Google Earth Engine (GEE) platform, Landsat time-series imagery (2001–2020) and MOD17A3HGF NPP data were integrated. The LandTrendr algorithm was applied to identify vegetation disturbance patches, and two representative disturbance years (2008 and 2014) were selected for long-term analysis. Trend analysis, coefficient of variation, and the Hurst exponent were used to characterize the spatiotemporal dynamics and stability of NPP in disturbed areas. The results show that: (1) NPP declined after disturbance and then exhibited a recovery trend, with significant spatial heterogeneity in recovery rates; (2) recovery trajectories differed between disturbance years, indicating combined effects of disturbance intensity and environmental conditions; and (3) Hurst exponent analysis suggests that although recovery trends are persistent in most areas, some disturbed patches show potential instability. This study establishes an analytical framework integrating disturbance detection and recovery tracking, which improves the representation of NPP dynamics in heterogeneous regions and provides a basis for assessing ecosystem recovery and carbon sink dynamics.

Accurate spatial identification of check dams is a key prerequisite for evaluating soil and water conservation benefits and optimizing dam system planning on the Loess Plateau. Current deep learning models face severe misclassification and omission issues under complex terrain due to the scarcity of check dam samples and the lack of prior geographic knowledge. This study proposes a recognition method based on Faster R-CNN, constrained by suitable areas and microtopography. The Xiliugou watershed in Inner Mongolia was selected as the study area. Based on Google Earth imagery and field survey data, a check dam sample dataset was constructed, integrating the morphological features of “linear dam body with a trapezoidal slope.” Using the construction suitable area constraints defined by the Technical Specifications for Check Dams and microtopography standard deviation (δ) derived from DEM as dual spatial filtering mechanisms, these were deeply embedded into the Faster R-CNN model to limit the search space and enhance geographic plausibility. Experimental results show that the constrained Faster R-CNN model achieved a precision and recall of 92.86% and 96.89%, compared with the accuracy rate of only deep learning model recognition (60.61%), which significantly increased by 32.25%, indicating that geographical constraints have an enhancing effect. Using this method, a total of 191 embankment dams were identified in the Xiliugou Basin. New 30 unrecorded embankment dams (21 small dams and 9 micro-dams) were discovered. The model’s good generalization ability was verified in the Han Tiechuan geographical isolation area, which contained 153 embankment dam samples, with an accuracy rate of 72.94%. Spatial analysis further revealed the “successive interception along tributaries” distribution pattern and strong spatial aggregation characteristics (box dimension D ≈ 0.36) of check dams in the Xiliugou watershed. This study confirms the critical role of suitable area and microtopography constraints in improving the accuracy and reliability of deep learning models and provides a transferable technical paradigm for automated, high-precision surveys of regional soil and water conservation projects.

The southeastern Tibetan Plateau (SETP) hosts the highest concentration of temperate glaciers on the Tibetan Plateau, which have experienced accelerated melting in recent decades. However, comprehensively quantifying the spatiotemporal variations in these glacier elevation changes remains a challenging task. To address this, we developed a framework integrating multi-source satellite observations—including ASTER stereo imagery, ICESat, ICESat-2, and CryoSat-2 CryoTEMPO-EOLIS data—to estimate elevation changes across the region. By deriving yearly time-series elevation changes for each 0.5° × 0.5° tile, we calculated a region-wide mean glacier elevation change rate of −0.710 ± 0.046 m/yr from 2000 to 2022. Our analysis reveals significant spatiotemporal heterogeneity. Temporally, the glacier thinning rate accelerated by 31.4% during 2011–2022 compared to 2000–2011. Spatially, among glaciers larger than 4 km2, 12.6% experienced slight thickening, 56.5% showed slight thinning, and 30.9% underwent rapid thinning. The maximum observed thickening and thinning rates were 0.81 m/yr and −1.68 m/yr, respectively.

ABSTRACT Crop classification serves as a fundamental component for agricultural resource monitoring and refined cropland management. However, in regions characterized by complex crop types and diverse planting patterns, traditional remote sensing methods frequently struggle to maintain classification stability across multiple years and varying geographic scales. This study focuses on a typical irrigated agricultural area in Changji Prefecture, Xinjiang, where a multi-source time-series feature set was constructed by integrating Sentinel-2 optical imagery and Sentinel-1 radar data. To address existing limitations, we propose a dual-branch crop classification model, termed Multi-source Spatial-Temporal-Phenological Integration (MSTPI), which synergizes the spatial feature extraction capabilities of SegFormer with the temporal phenological recognition strengths of the Time-Weighted Dynamic Time Warping (TWDTW) algorithm. Additionally, refined farmland boundaries were delineated using the CLCFormer model to establish unified spatial analysis units, thereby mitigating classification uncertainties caused by mixed pixels. In a 2022 benchmark experiment, the TWDTW branch integrating the kernel Normalized Difference Vegetation Index (kNDVI) and VH features achieved an overall accuracy (OA) of 87.91% under a decision-level fusion strategy. With the introduction of the SegFormer branch, the OA increased to 93.56%, significantly outperforming models based on single features or feature-level fusion. The average OA for temporal transfer testing from 2020 to 2024 was 88.17%, while spatial transfer validation across seven counties yielded an average OA of 86.24%, highlighting the model’s robust generalization potential in both temporal and spatial dimensions. Parcel-level maps revealed distinct spatial patterns, and the identified crop rotation patterns demonstrated strong correspondence with local agricultural practices. Ultimately, the MSTPI model exhibited exceptional performance in terms of classification accuracy and regional adaptability, offering a highly effective framework for crop type identification in complex agricultural landscapes

Google Earth Engine (GEE) has substantially expanded the scope of geospatial analysis by providing access to petabyte-scale satellite imagery and geospatial data together with cloud-based computation. This accessibility enables researchers to study large-scale environmental and socio-economic processes with high spatial and temporal coverage. In recent years, numerous R packages have emerged to leverage GEE’s functionalities. However, constructing and managing complex spatio-temporal databases for continuous monitoring of remotely sensed data remains challenging and often requires advanced coding skills. To bridge this gap, we introduce the geeLite R package (available on CRAN), which supports the construction and updating of local databases for GEE-derived outputs, enabling users to track their evolution over time. By storing results in SQLite format — a serverless, self-contained database solution requiring no additional setup — geeLite simplifies data collection. It also streamlines conversion to native R formats and provides functions to aggregate and process the resulting databases for downstream analysis. • Google Earth Engine provides vast satellite imagery and processing power. • The geeLite R package connects Google Earth Engine to local spatial data workflows. • Users can create and manage databases for real-time spatial tracking. • Data are stored in SQLite format for portability and offline accessibility. • The package streamlines aggregation and analysis of time series data.

Climate change is restructuring ecological communities globally, yet the impacts are often underestimated or poorly resolved due to the lack of historical baselines. In temperate oceans, biologically diverse and socioeconomically important kelp forests are the marine ecosystem most threatened by climate change. However, long-term historical baselines for kelp forests are lacking and the processes driving community-level changes remain poorly resolved. Here, using recently discovered aerial imagery and subtidal quadrat data from 1972, we recreated historical baselines for kelps and associated benthic macroalgae in a global hotspot within the northern Salish Sea (British Columbia, Canada). We resurveyed the same sites in 2023 to quantify community shifts, showing that a half-century ago, bull kelp (Nereocystis luetkeana) formed expansive kelp forests in the region (>550 ha), none of which remain today. Satellite time series of bull kelp show that the majority was lost between 1972 and 1984. These data increase baselines of bull kelp canopy extent in this area by more than 10-fold. Changes to the benthic kelp forest assemblage were mainly driven by loss of the dominant kelp, Saccharina latissima (-78%), across all depths. Historically abundant species of red algae also decreased substantially (e.g., Mazzaella splendens [-98.5%] and Plocamium pacificum, [-62.1%]), largely above three meters depth. Applying the community temperature index (CTI) to this half-century comparison, we show that CTI of the kelp forest community (+1.4°C; 95% CI: 0.43-2.37°C) had tracked increases of summer SST (+1.66°C; 95% CI: 1.20-2.13°C) more closely than winter SST (+0.65°C; 95% CI: 0.46-0.84°C), indicating that temperatures during the hottest summer months are likely driving community shifts. The abundance of cold-affinity species decreased more than warm-affinity species abundance had increased, indicating that the subtidal kelp forest community was predominantly restructured by deborealization, rather than tropicalization. Community deborealization may be prevalent in temperate hotspots that are disjunct from areas with similar climatology, creating colonization barriers for warm-affinity species. Our study underscores the importance of historical data for understanding the true magnitude of climate change impacts and suggests that deborealization of temperate kelp forest communities may be more common than has previously been recognized.

Hydrothermal alteration mapping and mineralization potential assessment in the Taznakht region (Morocco) through integrated ASTER imagery, airborne radiometric and magnetic data

In precision agriculture, the assessment and estimation of key crop parameters are crucial aspects for the optimisation of input usage and, as an ultimate goal, for the improvement of yield quality and quantity. In this context, a reliable prediction of yield by remotely sensed imagery is an enabling technology for optimisation. In this work, an innovative method for estimating yield in maize cultivation is presented, which exploits multi-temporal and multispectral Sentinel-2 satellite imagery with supervised Machine Learning (ML) techniques. For model training and validation, yield ground truth experimental data from combine harvesters was used, enabling the yield estimation at sub-field scale. The investigation, which was conducted on five case study plots, involved a preliminary comparison of four ML-based algorithms, trained with raw spectral bands. An assessment of the effect of the training dataset on the yield prediction accuracy was then performed. A set of Vegetation Indices (VIs) and Two Band Indices (TBIs) was also considered for this purpose. Finally, a multi-temporal analysis was conducted, in which the temporal evolution of crop spectral data over the maize growing season was exploited using imageries acquired in different epochs. The obtained results proved that an accurate estimation of maize yield can be reached using a Gaussian process regression model, exploiting multi-temporal features directly provided by the raw spectral bands. The model showed a high accuracy in the estimation of maize yield, even when fed with data acquired during only the maize vegetative phase, thus proving its capacity as a prediction tool. • Supervised machine learning techniques are used to estimate maize yield. • Combine harvester ground truth data enables prediction at subfield scale. • Multi-temporal imageries from Sentinel-2 improve the estimation. • Gaussian Process Regression algorithms reach accuracies up to R 2 > 0.9

This study investigates the sedimentation of volcanic particles from low-altitude (< 2km a.s.l.), near-daily ash plumes and clouds at Sakurajima volcano (Japan). Plume dynamics were monitored using imagery (visible wavelength) and geophysical (ash discharge rates) data. Ash fallout was characterized by using ground-based (disdrometer, particle electrical charge sensor, and sampling) and drone-mounted (optical particle counter, atmospheric sensor, and sampling) instruments. A comparison of particle size distributions and aggregate proportions between samples collected by drone 500m above the take-off sites and those collected on the ground shows that aggregation develops rapidly during sedimentation. This process involves collisions between coarse ash (up to 1mm) and fine ash particles (< 63µm). Particle binding is promoted by electrostatic attraction (forming particle clusters) or high atmospheric humidity (forming accretionary pellets). These results provide innovative in-situ evidence of ash aggregation, offering new insights into its dynamics in natural settings, crucial for improving volcanic ash dispersion forecasting.

Traffic emissions contribute disproportionately to exposure inequities in dense cities. Delivery fleets and public buses can intensify these burdens in vulnerable communities, yet most assessments overlook such dynamics by aggregating fleets into broad categories and averaging across the day. This obscures when and which vehicles drive inequities. We developed a vehicle-class and hour-resolved approach for Hong Kong to estimate traffic emissions. We integrated high-resolution traffic counts, street imagery, and detector data with machine learning and computer vision to model hourly NOx and PM2.5 for all road segments during typical daily activity hours. The framework explains over 95% of the variance in NOx and PM2.5 emissions. Results show substantial traffic-related emission inequities in Hong Kong, with low-income residents experiencing 8%-9% higher NOx levels than high-income residents, and Chinese residents experiencing 40%-52% higher NOx levels than White residents. The dominant contributors shift over the day, with delivery fleets driving daytime inequities and franchised buses amplifying evening inequities. Across all hours, light-duty goods vehicles contribute 31-35% of disparities, franchised buses 25-35%, and heavy-duty goods vehicles 19-23%, varying by population group. This study provides one of the first data-driven analyses of vehicle-specific impacts, revealing when and which vehicles drive inequities and guiding equity-focused interventions.

The present study investigated the detection performance of the YOLOv8s, YOLO11s, and YOLO12s models, implemented within convolutional neural network architectures, for identifying floricane raspberry (Rubus idaeus L.) shrubs using RGB imagery and multispectral data acquired in the near-infrared, red-edge, red, and green spectral bands with a DJI Mavic 3 Multispectral drone. Model training and validation were conducted to evaluate both within-modality detection performance and cross-modality transferability. Under all training scenarios, the YOLO-based detectors reached near-saturated accuracy levels. However, cross-domain assessments demonstrated substantial variability depending on the spectral configuration of the input imagery. Overall, the combination of UAV-based multispectral sensing with convolutional neural network detection frameworks establishes a technological basis for automated shrub monitoring and constitutes a meaningful advancement toward intelligent raspberry production systems. This integration further creates new prospects for the technological development of cultivation practices for this crop within the rapidly evolving landscape of artificial intelligence-driven agriculture.

Accurate multi-temporal terrain change detection in mountainous regions is crucial for disaster monitoring and environmental management; however, it remains challenging due to the complex topography, dense vegetation, and occlusion. This study proposes an integrated, automated framework for terrain change detection that combines 2D semantic segmentation and 3D geometric analysis, utilizing UAV imagery and point cloud data. This method uses the DeepLabV3 neural network to perform multi-class semantic segmentation of orthophotos, combining Fast Point Feature Histograms (FPFH) and Random Sample Consensus (RANSAC) algorithms, followed by Iterative Closest Point (ICP) based local refinement. Geometric changes are subsequently quantified using the Multiscale Model-to-Model Cloud Comparison (M3C2) distance metric. When applied to 0.06 km2 of mountainous terrain in the Guanziling area, Tainan City, Taiwan, between January 2024 and January 2025, the framework achieved a mean Intersection over Union (mIoU) of 87.05% in semantic segmentation and a root-mean-square error (RMSE) of 4.2cm in geometric registration for rigid structures. The model's capacity to generalize was evaluated by validation against 100 independent, manually annotated tiles, which showed a strong coefficient of determination (R2 = 0.9251) between the predicted and actual change proportions. Comparisons in complex "Medium Change" zones showed that while 2D-only methods identified 9% of changes due to shadows and occlusion, the integrated 3D analysis successfully detected physical displacements averaging 2.12m. Overall, the proposed 2D-3D fusion framework enables reliable terrain change detection and supports post-disaster assessment, landslide monitoring, and infrastructure management in complex mountainous environments.

The integration of heterogeneous geospatial data, specifically low-cost unmanned aerial vehicle (UAV) imagery and mobile light detection and ranging (LiDAR) system point clouds, presents a significant challenge due to the significant radiometric and structural discrepancies between the two modalities. This study proposes a novel air-to-ground semantic feature matching framework to achieve precise geometric registration between these data sources by effectively incorporating semantic-constraint deep learning-based matching. The methodology transformed the cross-sensor alignment challenge into a robust two-dimensional image matching problem. This was achieved by first using YOLOv11 for semantic segmentation of common road markings in both the UAV orthoimage and the converted LiDAR intensity image to generate highly consistent feature references. Subsequently, the SuperPoint detector and a graph neural network matcher, SuperGlue, were applied to these semantic images to establish reliable geomatics information correspondence points. Experimental results confirmed that this semantic-guided strategy consistently outperformed traditional feature-based matching (i.e., scale-invariant feature transform + fast library for approximate nearest neighbors), particularly by converting the noisy LiDAR intensity image into a stabilized semantic representation. The explicit application of semantic constraints further proved effective in eliminating false matches between geometrically similar but semantically distinct objects. The final object-specific analysis demonstrated that features with clear, complex geometric structures (e.g., pedestrian crossings and directional arrows) provide the most robust matching control. In summary, the proposed framework successfully leverages semantic context to overcome cross-sensor heterogeneity, offering an automated and precise solution for the geometric alignment of mobile LiDAR data.

Abstract Mangunharjo Beach, located on the northern coast of Central Java, plays a vital ecological and socio-economic role, supporting fisheries and coastal livelihoods. However, the region faces severe coastal erosion due to wave action, land subsidence, and human activities. This study assesses the effectiveness of various coastal protection strategies, including hard structures and nature-based solutions by comparing the existing conditions with three intervention scenarios: breakwater installation, mangrove restoration, and a hybrid approach that combines both measures. First, spatial analysis is conducted using satellite imagery and historical shoreline change data to assess erosion patterns. Then, numerical simulations are carried out using Delft3D, involving coupled Delft3D-Flow and Delft3D-Wave modules, to evaluate wave attenuation under each scenario. Furthermore, a literature-based assessment is performed to analyze the ecological functions and feasibility of mangrove-based solutions. It is found that the success of mangrove restoration depends heavily on local ecological conditions and is most suitable in areas that naturally support mangrove growth. The findings of this study are expected to inform policymakers in developing adaptive and sustainable coastal protection strategies for Mangunharjo Beach and other vulnerable coastal zones.

Environmental monitoring is essential for understanding and minimizing human impact on ecosystems. Traditional methods like manual sampling and laboratory testing, while accurate, are often costly, time-consuming, and difficult to scale, especially in low-resource settings. Artificial intelligence (AI) is increasingly addressing these limitations by enabling automated data collection, real-time analysis, and predictive modeling. Techniques such as machine learning (ML) and deep learning (DL) are being applied to monitor air and water quality, track climate patterns, and support biodiversity efforts. Hybrid AI models further improve accuracy by integrating various analytical approaches. Key applications include species identification, habitat assessment, wildlife tracking, and anti-poaching, utilizing tools such as drone imagery, camera traps, and GPS data. This review explores the latest advancements in AI-based environmental monitoring, emphasizing technologies like explainable AI (XAI), edge computing, and the Internet of Things (IoT), which improve transparency and reduce processing costs. It also addresses ongoing challenges, including data quality, computational demands, and the need for interpretable models. By evaluating practical limitations and proposing interdisciplinary strategies, this article highlights the transformative potential of AI for sustainable environmental management. Successful implementation will depend on ethical frameworks, policy alignment, and cross-sector collaboration to fully realize AI's role in global ecological stewardship.

A novel spatially enabled neural model for flood susceptibility in Northern Morocco