Atmospheric and Oceanic Sciences

Abstract North America is retreating from carbon pricing at the precise moment when its necessity is most acute. In March 2025, Canada eliminated its federal consumer carbon tax, and the United States has dismantled the regulatory architecture for greenhouse gas regulation, including revoking the foundational 2009 EPA endangerment finding. This paper asks three questions: (1) How does the North American retreat constitute an ethical failure when evaluated through ethical security economics? (2) What does international comparative evidence reveal about the conditions under which carbon pricing succeeds? (3) What structural reforms would render carbon pricing ethically grounded and politically durable? We apply ethical security economics, a values-based framework grounding economic evaluation in five interrelated principles (sustainability, justice, peace, compassion, and authenticity with accuracy), to analyse these concurrent policy reversals. We contrast North America’s retreat with European jurisdictions where carbon pricing has been associated with substantial emissions reductions alongside economic growth, and find that the retreat constitutes a comprehensive ethical failure across all five dimensions. We introduce the concept of reactive economics, situated within established scholarship on carbon lock-in and fiscal path dependency, to characterise the structural pattern of subsidising harm-generating activities while dismantling instruments designed to address their consequences. The paper concludes with institutional recommendations, including an independent Carbon Price Board mechanism, for reintegrating carbon pricing within values-explicit governance frameworks.

Introduction Seasonal freeze–thaw processes in black soil critically regulate soil structural stability, hydrothermal migration, and aggregate integrity, thereby influencing water retention, root-zone stability, and agricultural productivity and ecological stability. In cold-region agricultural ecosystems, freeze–thaw-induced changes in pore structure and permeability also strongly influence contaminant migration, retention, and the effectiveness of soil remediation measures by altering water flow pathways and mass transport processes. However, the multi-scale coupled mechanisms linking ice lens growth, hydrothermal transport, and structural degradation under varying initial water content and freezing intensity remain poorly understood. Methods To address this, a thermo–hydro–mechanical phase-field coupled model (THM-PF), specifically adapted for black soil, was developed to simulate interactions among temperature variation, moisture migration, stress redistribution, and pore structural evolution. Gradient cooling experiments (−15 °C to −3 °C), combined with microscopic observations, were conducted to validate the model and elucidate coupled freeze–thaw mechanisms and structural degradation processes under controlled conditions. Results The results indicate that freeze–thaw fragmentation in black soil is primarily controlled by ice lens-induced segregation stress, exhibiting threshold behavior defined by the critical separation void ratio (e sep ). When this threshold is exceeded under strong freezing, the system rapidly destabilizes, with pore pressure reaching 107 kPa and porosity increasing by up to 546.47% relative to the initial state. Initial water content further regulates degradation pathways: at high water content (>32.3%), ice lens growth dominates and causes severe aggregate breakage (up to 99% after 30 cycles), whereas at low water content (<3.4%), degradation is mainly driven by mechanical friction, resulting in limited structural change (porosity increase of 21.62%). In addition, permeability heterogeneity alters heat transfer pathways by 15°–23°, leading to discontinuous ice lens distribution and enhanced spatial heterogeneity. Discussion Structural modification (e.g., gravel–sand incorporation) can reduce temperature gradients by up to 27%, thereby suppressing ice lens development. These results collectively reveal a cascade mechanism of “ice lens segregation → hydrothermal heterogeneity → structural degradation” and highlight the potential risks to soil stability, crop productivity, and contaminant transport in high-intensity freeze–thaw regions. Based on this mechanism, targeted mitigation strategies are proposed, including optimal water content control (<28%), permeability regulation through graded fillers, and layered structural design, providing theoretical guidance and technical support for soil conservation, agricultural management, ecological restoration, and contaminant transport management in black soil regions.

This study offers a new approach to probabilistic earthquake hazard assessment (PEHA) in the densely populated regions of Southern Sumatra and West Java, Indonesia. While much attention is given to powerful, offshore megathrust earthquakes, this research focuses on a different yet equally dangerous threat: shallow, moderate-magnitude earthquakes (4.5 ≤ Mw ≤ 6.5) that occur on land. These events, often caused by unmapped faults, pose a significant risk due to their proximity to major cities and infrastructure. To develop a more reliable model, a best-fit earthquake rate model was estimated using declustered shallow earthquake events as a reference. This model enhances existing methods by offering a more precise depiction of where these shallow, damaging earthquakes are likely to occur. We accomplished this by analyzing a comprehensive probability of exceedance (PoE) of earthquakes with magnitudes up to 6.5 and depths up to 50 km that occurred between 1963 and 2022, mapping and modeling both the known active faults and the historical seismic activity in the region, and using advanced statistical methods to create a highly reliable, integrated seismicity rate model. The final product, the Integrated Most Reliable Spatial Seismicity Rate Model (ModelIMRSSR), is proposed as a useful tool for government authorities and urban planners. It can be used to create detailed seismic hazard maps that highlight areas of highest risk, especially those with unmapped faults. By guiding development away from these high-risk zones and identifying specific locations for physical reinforcement, this research provides a framework for sustainable investment. The proactive use of these findings can lead to more resilient communities and a significant reduction in potential damage and loss of life from future earthquakes.

Accurate large-scale mapping of coastal dune types is critical for coastal management but remains challenging due to the dunes’ high spatial heterogeneity and the spectral complexity in coastal environments. Therefore, this study aims to develop a coastal dune mapping approach and demonstrate its applicability across Australia and New Zealand. Firstly, we leveraged geological maps to delineate the spatial extent of coastal dune fields, effectively masking out non-dune areas. Secondly, we applied an object-based image analysis (OBIA), followed by a random forest classifier that integrated multi-source features to classify six coastal dune types within the delineated coastal dune fields, including active dunes, foredunes, sandy beaches, semi-stabilized dunes, stabilized dunes, and wetlands (dune slacks). Based on this approach, we generated a 10 m fine classification map of coastal dune types across Australia and New Zealand, covering a total area of 27,564 km 2 . Accuracy assessment yielded an overall classification accuracy of 90.68 %. Most categories achieved satisfactory performance, with stabilized dunes showing the highest accuracy (96.81 %). We also found that approximately 498 km 2 of coastal dune area has been converted to urban development land. However, this development pressure was mitigated by conservation efforts, as 15,781.80 km 2 (59.44 %) of the total coastal dune area was situated within protected areas. This study provides a replicable methodology for large-scale fine classification of coastal dune types. The resulting map offers scientific support for monitoring coastal dune systems state and evolution under human intervention.

Phytoplankton and zooplankton dynamics in the western Ross Sea are strongly shaped by interactions among sea-ice retreat, water-column structure, and spatial environmental gradients. During the ROSSMIZE expedition (November 1994–January 1995), a multidisciplinary survey across four contrasting regions of the western Ross Sea captured the transition from early-season ice influence to summer stratification. Mesozooplankton were sampled using a sensor-equipped BIONESS system, enabling high-resolution coupling of biological patterns with temperature, salinity, fluorescence, and depth. The region emerged as a mosaic of subsystems structured by latitude, ice history, and hydrographic conditions. Depth was the dominant driver of community composition and diversity: generalized additive models indicated that diversity peaked at intermediate depths, reflecting a balance between surface-driven production and deeper, more stable water masses. Temperature, salinity, and fluorescence further modulated these patterns, underscoring the sensitivity of pelagic communities to fine-scale physical gradients. Together, these results demonstrate that spatial structuring of zooplankton in the western Ross Sea is governed not only by seasonal ice dynamics but also by depth-dependent habitat features and latitudinal environmental transitions during the spring–summer period.

With an ever-growing recognition of Land Surface Temperature (LST) as a key Essential Climate Variable (ECV), it becomes utmost important to have such a variable at the fine spatial and temporal scales of urban spaces and dynamics. Sentinel-3 provides coarse LST (1 km, daily) based on thermal imagery acquired by its Sea and Land Surface Temperature Radiometer (SLSTR) as well as fine Spectral Directional Reflectances (SDRs, 300 m, every two days) synergically inferred from both SLSTR and the Ocean and Land Colour Instrument (OLCI), which gives the opportunity for using the latter as a predictor in the downscaling of the former. Herein, two scale-invariance-based architectures were developed: a single-timestamp (STS) model, trained with coarse data of the timestamp whose fine target it infers; and a multi-timestamp (MTS) one, trained with multiple timestamps. Note that while several Machine Learning (ML) models besides Linear Regression (LR) were considered for the MTS architecture, only LR was used for the STS one due to the limited amount of available data which the former require for hyperparameter tuning. The models were developed over four Danish Functional Urban Areas (FUAs) using SRD-derived indices and seasonal and geospatial predictors and validated against Landsat data. While Gradient Boosting (GB) achieved the best coarse-scale performance it corresponded to the worst fine-scale performer together with Random Forest (RF), indicating scale invariance breakdown. Tree-based models performed poorly due to extrapolation limitations, whereas Neural Net (NN) and LR proved more robust. After residual correction, single-timestamp LR achieved the best fine-scale performance, making it the most reliable and recommended architecture for operations. The overall results showed that, although ML models may better predict the target at their training scale, their performance may not significantly generalise at others, therefore revealing scale specificity. Furthermore, the results suggested that usage of the more general multi-timestamp architecture instead of the single one may deteriorate performance.

Quantifying landscape ecological risk (LER) using multi-period land use data and ecological indicators is essential for understanding regional ecological dynamics. However, LER assessment is sensitive to spatial delineation, introducing uncertainty. This study developed an integrated LER model that incorporates the remote sensing ecological index and abundance index, and evaluated spatial unit effects through comparative analyses of fishnet, hexagonal, sub-watershed, and county units. LER dynamics in the Shandong Peninsula Urban Agglomeration (SPUA) from 2004 to 2024 were analyzed, and a boosted regression trees model was applied to quantify the relative importance of influencing factors and their nonlinear effects. The results indicate that: (1) sub-watershed units showed greater robustness and stability across multiple evaluation indicators, supporting their suitability for LER assessment; (2) LER in the SPUA exhibited a fluctuating but overall slightly increasing trend over the past two decades, with a persistent west-low and east-high spatial pattern; and (3) relief degree (29.13%) and nighttime light (17.78%) were the dominant factors shaping LER, showing an inverted U-shaped response and a saturating nonlinear increase, respectively. This study supports the use of sub-watershed units as an appropriate spatial unit for LER assessment and provides insights into terrain-sensitive conservation and sustainable land-use management in urbanizing regions.

To explore a new approach to reducing the use of external carbon sources and phosphorus removal chemicals in conventional wastewater treatment, this study developed an anaerobic–oxic–anoxic sequencing batch reactor (AOA-SBR) system (Rf) with iron shavings addition (180 g, 60 g/L), using a blank reactor (R0) as the control. Synthetic wastewater with a C/N ratio of 7.5 was used as the influent. The operating cycle of the AOA-SBR reactor consisted of a 120 min anaerobic phase, a 120 min aerobic phase, and a 60 min anoxic phase, with a hydraulic retention time (HRT) of 12 h. Results showed that the SVI30 of Rf remained at approximately 35 mL/g. The average removal efficiencies of TN and TP in Rf reached 70% and 96%, respectively, which were higher than those of the control. The addition of waste iron shavings improved sludge settleability and nitrogen and phosphorus removal performance of the biochemical system. Fe-C microelectrolysis significantly enriched Candidatus_Competibacter and Candidatus_Nitrocosmicus while inhibiting nitrite-oxidizing bacteria (NOB). This triggered persistent low-level nitrite accumulation within the system, diversified nitrogen-removal pathways, and ultimately improved the total nitrogen-removal efficiency. The extended anaerobic period in the anaerobic–oxic–anoxic (AOA) mode enriched phosphate-accumulating organisms, achieving synergistic chemical and biological phosphorus removal. This study provides a novel strategy for advanced wastewater treatment without external carbon sources or phosphorus additives.

In optical remote sensing 3D reconstruction, high-resolution satellite stereo matching is a critical task, yet it is challenged by extreme imaging geometries, texture-less and repetitive patterns, occlusions, and scene variations caused by spatio-temporal heterogeneity. To address these issues, we propose IFMA-Stereo, an innovative binocular disparity estimation method that leverages a monocular depth foundation model. Our approach constructs a multi-scale spatial information pyramid to jointly integrate the foundation model with a disparity extraction network. At the feature level, an attention interaction mechanism captures multi-dimensional contextual dependencies and transforms general scene understanding priors into long-range associative features suitable for stereo cost volume construction. At the pixel level, a cyclic iterative refinement module embeds depth information from the foundation model throughout the iteration process and performs joint optimization, enhancing the model’s adaptability in geometrically complex regions. Experiments on the US3D and GaoFen-7 datasets demonstrate that IFMA-Stereo achieves superior performance in challenging areas (texture-less regions, disparity discontinuities, repetitive patterns) and effectively mitigates prediction errors caused by spatio-temporal heterogeneity, albeit at the cost of increased inference time compared to baseline methods. Quantitatively, the method achieves an end-point error (EPE) of 1.347 and a D1 error of 7.26% on the US3D dataset, and an EPE of 1.585 and a D1 error of 13.41% on the GaoFen-7 dataset. Notably, the method also yields precise predictions for unseen urban areas, indicating strong generalization. These results confirm that IFMA-Stereo achieves state-of-the-art accuracy in remote sensing disparity estimation.

Accelerated biodiversity loss due to anthropogenic factors has led some researchers to suggest Earth may be approaching a 6 th mass extinction event. While artificial intelligence (AI) cannot counteract this biodiversity loss directly, AI has seen an increased adoption as a powerful tool for marine biodiversity monitoring, particularly by enabling the rapid and efficient collection and processing of large volumes of highly accurate biodiversity data. Such monitoring can, in turn, inform conservation strategies and evidence-based marine policy. This review aims to increase the accessibility of AI-based imaging techniques for marine biodiversity research and to help researchers make informed decisions on how these tools can effectively be applied to marine biodiversity studies. This review is intentionally framed as an introductory primer rather than a novel conceptual synthesis, aiming to consolidate and clarify foundational AI imaging concepts for a marine biodiversity audience. We compare traditional survey methods with AI-assisted approaches to marine data collection, explain the fundamental principles behind AI imaging workflows, and discuss common limitations and sources of bias associated with their implementation. By establishing this baseline knowledge, we review current applications of AI imaging in marine biodiversity monitoring, and outline practical pathways for integrating these systems into marine research programs. Ultimately, we aim for readers to gain a deeper understanding not only of how complex AI systems support marine biodiversity monitoring, but how best to deploy them responsibly and effectively to address the growing data challenges facing marine science today.

This study explores women’s subjective resilience to climate change in informal settlements in Nairobi, Kenya, focusing on the lived experiences of women who face heightened vulnerability. Informal settlements, characterized by overcrowding, inadequate infrastructure, and insecure tenure, are disproportionately affected by extreme weather events such as flooding and heatwaves. While existing literature highlights climate resilience at the socio-ecological systems level, there is limited attention on women’s personal experiences and adaptive strategies. This research fills that gap by investigating how women perceive and respond to climate challenges, contributing valuable insights into how individual experiences of resilience interact with broader systems of adaptation. Using qualitative methods, the study examines the roles women play in household and community-level adaptation, emphasizing their agency and the systemic barriers they encounter, including poverty, political marginalization, and limited access to resources. The findings reveal that women’s resilience is shaped by interactions between personal assets and strategies and external resources at every level of the social-ecology. These interactions both reinforce and challenge broader socio-ecological resilience frameworks, highlighting the need for integrated, context-specific climate adaptation. The study calls for more inclusive approaches to climate adaptation that build on mutual aid and community-level initiatives in informal settlements to recover, adapt and transform in the face of climate change. Ultimately, this research offers a foundation for designing more effective, community-driven climate strategies that center women’s experiences and promote sustainable, system-level resilience.

Plant genotype plays a critical role in shaping root-associated microbiota and in modulating plant tolerance to soilborne diseases such as Fusarium root rot (FRR). In this study, we investigated how four wheat (Triticum aestivum) varieties, with differing tolerance to FRR, influence the composition and structure of bacterial communities in the rhizosphere and root endosphere. In the current study evaluated root traits that may contribute to the genotype-specific assembly of bacterial communities across the four wheat genotypes. The variety Concret exhibited the highest FRR tolerance, whereas Pilier was the most susceptible. Analyses of root morphology revealed significant genotype-dependent differences in root length and volume. Notably, traits associated with the tolerant genotype were positively correlated with the abundance of key beneficial bacterial genera in the rhizosphere, including Bacillus, Lysobacter, and Sphingomonas. Untargeted metabolomics identified 879 features, with 20 key metabolites distinguishing the wheat genotypes, including alkaloids, benzoate derivatives, and benzoxazinoid-derived compounds. Correlation analysis revealed significant relationships between these root metabolites and key bacterial taxa. This findings demonstrate that wheat genotypes influence the assembly of the root microbiota through genotype-based morphological and metabolic traits, providing valuable insights into the specific root traits that wheat genotypes can leverage to modulate the plant microbiome and enhance disease resistance.

Geological lithology interpretation in multi-source remote sensing imagery still faces challenges due to the spatial heterogeneity of multi-source features and the fragmented spatial distribution of geological elements. To address these issues, we propose the complementary feature self-integration network (CFSI-Net), a unified framework for multi-source feature fusion and lithology interpretation. To mitigate spatial heterogeneity in multi-source features, we developed the cross-domain feature extraction and fusion module (CDFEFM), a dual-branch structure that adaptively extracts and fuses complementary features across channels while reducing redundant lithology information. Furthermore, we introduced the frequency-spatial feature integration module (FSFIM) and a segmentation module (SegM) to alleviate the problem of boundary ambiguity in segmentation. FSFIM captures boundary structural details and spatial contextual dependencies from the frequency and spatial domains, whereas SegM further refines spatial dependencies across channels, enriching textural details and enhancing the discrimination of adjacent elements. Experimental results on the CUG-LithXZ dataset demonstrate that CFSI-Net consistently outperforms state-of-the-art models, achieving an OA of 80.13% and mIoU of 66.18% in Geo-GZ, and an OA of 72.49% and mIoU of 50.26% in Geo-LZ. These results confirm that the synergistic utilization of frequency and spatial information provides richer details for accurate geological lithology interpretation in complex terrains.

Temperate phages play crucial ecological roles in engineered microbial communities, yet their adaptive strategies under antibiotic stress remain unclear. Here, metagenomic analysis was used to investigate how temperate phages facilitate host adaptation in activated sludge acclimated to chloramphenicol (CAP). Antibiotic stress markedly reshaped bacterial and temperate phage communities, with dominant degraders (e.g., Sphingomonas and Caballeronia ) reaching relative abundances of 6.5–42.0%. Temperate phages exhibited specific adaptive responses by significantly enriching antibiotic resistance genes, including multidrug ( arlR and mtrA ) and peptide ( bcrA ) resistance genes, resulting in a 1.56–4.15-fold increase in the phage-derived resistome relative to the control. They also mediated general adaptive responses by encoding auxiliary genes involved in oxidative stress mitigation, DNA repair, biofilm formation, and antiviral defense. Host-phage linkage prediction identified 1045 phage–bacteria interactions, including 11 ARG-harboring viral operational taxonomic units associated with dominant CAP-degrading hosts. Collectively, our findings reveal that temperate phages facilitate microbial resilience in antibiotic-stressed environments by delivering mutualistic genetic traits, encompassing both specific (antibiotic resistance genes) and general (antiviral defense, metabolic, and stress mitigation) adaptive responses, highlighting their ecological significance and potential for enhancing the stability and performance of wastewater treatment systems under pharmaceutical stress.

To control the quality of mineral-medicinal waters and ensure their therapeutic benefits, spas often rely on periodic discrete sampling to analyze the physico-chemical properties of their pools. The AQUAPRED project aims to digitize this process by deploying IoT systems within the spa facilities, enabling real-time data acquisition via calibrated multi-parameter probes. Using data collected by these pilot systems, we develop and validate a predictive machine learning model capable of forecasting the short-term evolution of the thermal water properties. Historical data from each facility allow the model to learn the specifics dynamics of each spa. As a practical application, we propose an anomaly detection module based on residual analysis from predicted and observed values. Significant discrepancies signal events of interest and emergent trends, such as anomalous readings, contamination or sensor drift. The methodology is evaluated using real data from six spas associated with the AQUAPRED project. The results demonstrate the model’s effectiveness and support its feasibility for deployment in other thermal establishments.

Few-shot remote sensing scene classification aims to accurately recognize unseen scene categories using only a scarce number of labeled samples, which has emerged as a research hotspot in the field of remote sensing image interpretation. However, remote sensing images intrinsically suffer from large intra-class variations, high inter-class similarities, and complex background interferences. Traditional few-shot learning methods typically perform feature metric learning in Euclidean space, making it difficult to capture the non-Euclidean geometric distribution characteristics of remote sensing features, and they often neglect the spatial structural information embedded in feature maps. To address these issues, this paper proposes a novel few-shot remote sensing scene classification method, termed ZSFS-RGBN, which integrates a Zigzag Scanning Feature Sequence with a Riemannian Geometric Barycenter Network. Specifically, ResNet12 is first employed as the backbone to extract deep convolutional feature maps from both the support and query sets. Second, a Zigzag scanning strategy is introduced to reorganize the two-dimensional feature maps into one-dimensional feature sequences, thereby effectively preserving the spatial locality and structural continuity of the features. Third, an autoregressive moving average (ARMA) model is constructed to characterize the spatial dependencies of the feature sequences, and its state parameters are mapped onto a symmetric positive definite (SPD) matrix manifold, enabling the deep semantic representations of remote sensing scenes in a non-Euclidean geometric space. Finally, a Riemannian geometric barycenter network is designed to learn the Riemannian barycenter of each category on the SPD manifold, where a joint loss function is introduced to simultaneously optimize intra-class compactness and inter-class separability. Comprehensive experiments are conducted on three public remote sensing scene datasets: NWPU-RESISC45, UC Merced Land-Use, and WHU-RS19. Experimental results demonstrate that the proposed method consistently outperforms several representative state-of-the-art approaches under both 5-way 1-shot and 5-way 5-shot settings. Furthermore, ablation studies verify the effectiveness of each component within the proposed framework.

With the increasing diversity of remote sensing image acquisition conditions and imaging scenarios, building extraction models often experience significant performance degradation in cross-dataset applications due to variations in sensors and scene characteristics. Improving their cross-domain generalization ability has therefore become a critical research problem. To address the challenges of appearance style discrepancy and feature distribution shift in cross-domain building extraction, this paper proposes WCMNet, a wavelet-guided and CNN–Mamba hybrid network for unsupervised domain adaptation in building extraction. Specifically, a Mamba Wavelet Alignment (MWA) module is designed to align low-frequency style information in the wavelet domain while preserving directional high-frequency edge structures, thereby mitigating cross-domain appearance discrepancies and reducing structural degradation during domain translation. In addition, a Global–Local Mamba Block (GLMB) is developed to jointly model local textures and global semantic dependencies. In GLMB, the CNN branch captures fine-grained local details and boundary cues, while the Mamba branch models long-range contextual information; an adaptive gated fusion mechanism further integrates the two types of features. Experimental results on six cross-domain transfer tasks across the WHU, Massachusetts, and Potsdam datasets demonstrate that WCMNet consistently outperforms existing state-of-the-art domain adaptation methods. In particular, WCMNet achieves an average IoU of 65.13% and an average BIoU of 74.80% across all transfer settings, with improvements of up to 27.35 percentage points in IoU and 38.32 percentage points in BIoU compared with the strongest competing methods. These results demonstrate that the proposed MWA and GLMB effectively improve building completeness, boundary delineation accuracy, and cross-domain robustness.

Indicators of fluid migration and pressure-temperature changes in glaciated settings provide constraints on past ice-thickness, subglacial hydrology, and subsurface fluid behaviour. The East Irish Sea is a glaciated shelf area hosting hydrocarbon fields, wind farms, potential sites for Carbon-Capture and Storage and Hydrogen storage, and nuclear-waste sequestration. Two cross-cutting reflections (CCR) have been identified 80 m and 170 m below sea-level on seismic data acquired in 20 m of water in the EIS. The CCRs are imaged where they cross-cut normal seismic-stratigraphic reflections. An exploratory borehole provides petro-physical wireline data. The CCRs resemble gas-hydrate or diagenetic Bottom Simulating Reflectors (BSRs) representing palaeo-hydrate BSRs, formed during the last glacial maximum when conditions for a Gas-Hydrate Stability Zone (GHSZ) existed beneath thick ice accumulations. However an oxidation weathering front, in the presence of freshwater infiltration at sea-level low-stands during glacial periods, represents an equally plausible model. In the palaeo-BSR scenario, depth of the CCRs below basal Quaternary could be used to calculate ice-thickness required for establishment of a GHSZ. The weathering scenario would provide constraints on fresh-water flushing during glaciation. Both scenarios inform possible future conditions at the site, which is close to present and planned energy and waste facilities.

To identify the key determinants of traffic injury risk and clarify the relative roles of built environment factors and crash-context factors, this study develops a traffic injury risk identification framework that integrates 5D built environment variables with non-5D factors, using traffic crash data collected in Changsha, Hunan Province, China, from 2017 to 2019. After data cleaning, screening, and spatial matching, a total of 9,743 valid samples were obtained, and injury occurrence in a crash was defined as a binary dependent variable. On this basis, three feature sets were constructed, including a 5D feature set, a non-5D feature set, and a combined 5D + non-5D feature set. Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Categorical Boosting (CatBoost) models were then developed and compared, and the best-performing model was further interpreted using the Shapley Additive Explanations (SHAP) method. The results showed that the combined 5D + non-5D feature set consistently outperformed the models using either the 5D or non-5D feature set alone, indicating that traffic injury risk arises from the joint influence of the built environment and immediate crash-context conditions. Among the three models, CatBoost achieved the best performance under the combined feature set and produced the highest receiver operating characteristic-area under the curve (ROC-AUC) value, demonstrating superior overall discriminative ability. The SHAP results further revealed that, within the combined CatBoost model, 5D variables accounted for a larger share of the model-based contribution than non-5D variables. However, given the relatively small improvement in accuracy after adding 5D variables, this finding should be interpreted as evidence of complementary explanatory information rather than as a dominant source of predictive performance. Distance to the nearest metro station, lighting condition, road network density, point of interest (POI) mix, and distance to the nearest bus stop were identified as the most influential factors. Further dependence analysis showed that a greater distance to the nearest metro station generally increased traffic injury risk, whereas higher road network density and greater POI mix were generally associated with lower injury risk. From the perspective of the interaction between 5D built environment characteristics and non-5D factors, this study reveals the multidimensional pathways through which traffic injury risk is shaped. The findings also confirm the effectiveness of the CatBoost-SHAP framework for traffic injury risk identification and interpretation, and provide empirical support for urban traffic safety risk assessment, road environment optimization, and refined governance strategies.