The Yellow River Basin serves as a critical ecological barrier in China, with soil and water conservation being the primary focus of ecological protection. Despite significant achievements from comprehensive measures by the Chinese government, future runoff (sediment) trend and meteorological droughts remain uncertain due to complex environmental conditions, which are key issues for regional water resource management. Using climate and land use change data with the SWAT model, 9 scenario combinations were constructed to project spatial-temporal runoff and sediment evolution, meteorological drought characteristics, and the contributions of environmental factors in the Fen River Basin. Future projections indicate a warmer and more humid climate with annual precipitation increase of 108 mm, and temperature rise of 2–6°C. The frequency of dry in the basin from 2061 to 2100 is significantly lower than that from 2020 to 2060, while the wet frequency presents the opposite. Under SSP1-2.6, SSP2-4.5, SSP5-8.5 scenarios, runoff and sediment increase progressively with higher emissions. The ecological protection scenario shows the highest runoff and sediment. In the farmland protection scenario, runoff is lower than in the natural development scenario, while sediment is higher. Climate change dominates upstream hydrological processes, with the contribution more than 78.00

Lead (Pb) and chromium (Cr) are priority soil hazards: Pb is a potent neurotoxicant and Cr (VI) is a recognized carcinogen, both persisting for decades and threatening groundwater and human health through ingestion and leaching. At large industrial sites, the delineated three-dimensional extent of these metals directly governs remediation volume, cost, and residual exposure risk. To resolve hazard distributions at decision-relevant resolution, we developed D-MSN, a deep-learning-enhanced extension of the Mean Surface with Stratified Nonhomogeneity (MSN) model. D-MSN integrates masked spatial attention, gated residual networks, and meta-polynomial trend learning to jointly capture in-stratum correlation, between-stratum heterogeneity, and cross-stratum dependency. Applied to Pb and Cr at an abandoned integrated iron-and-steel complex in China (2.81 km², 12,900 samples spanning coking, sintering, ironmaking, steelmaking, rolling, and captive power generation) under stratified spatial-block cross-validation, D-MSN outperformed inverse distance weighting, 3D ordinary kriging, 3D-MSN, and DKNN, reducing MAE and RMSE by ∼35% and ∼38% relative to ordinary kriging. The model-derived hotspots spatially coincide with known process units, revealing diffuse Cr signatures near slag-handling and material-storage areas and compact Pb plumes near sintering, blast-furnace, and captive-power-plant operations. Unlike conventional smoothing-based interpolation methods, D-MSN preserves localized hotspot structures and supports source-oriented interpretation. Coupling D-MSN with Monte Carlo dropout further yields probabilistic exceedance maps under GB 36600-2018, identifying decision-uncertain volumes that deterministic interpolators systematically conceal. The practical implications of stratified-heterogeneity-aware modeling for risk-based contaminated site management are highlighted by the fact that the interpolation method alone changes the delineated remediation volume by about 36% in comparison to standard kriging, with cost implications on the order of 108 CNY.

IntroductionGross ecosystem product (GEP) provides a monetary indicator of ecosystem contributions to human well-being, yet its spatiotemporal heterogeneity and associated factors remain insufficiently understood in geomorphologically complex karst regions.MethodsUsing Guizhou Province, Southwest China, as a global karst hotspot, we quantified GEP in 2000, 2010, and 2020 and examined its spatial patterns and associated drivers across typical karst (TK), sub-karst (SK), and non-karst (NK) zones. A task-oriented multi-scale framework was employed, integrating global Moran’s I for spatial dependence analysis and GeoDetector for factor and interaction detection.ResultsThe GEP increased from CNY 1.68 × 1012 in 2000 to CNY 2.10 × 1012 in 2020, following a decline in 2000–2010 and a strong rebound in 2010–2020. High GEP areas were concentrated in southwestern and southern Guizhou and expanded northward and eastward. GEP exhibited significant positive spatial autocorrelation in all three years. NDVI was the dominant factor associated with GEP spatial differentiation, while elevation and slope acted as persistent background constraints. Interaction effects were mainly characterized by nonlinear enhancement, and natural—anthropogenic interactions became more important over time. Distinct association regimes of GEP drivers were identified across the three geomorphological zones. The multi-scale framework further provided complementary evidence on continuous spatial heterogeneity, clustering patterns, and governance-relevant territorial differences.DiscussionGEP-oriented ecological management in karst regions should be both geomorphology specific and scale-aware.

Study region: Qiyi Glacier in the Qilian Mountains, China. Study focus: Glacier flow is one of the important processes in the glacier development and evolution. It provides a scientific basis for assessing glacial disaster risks, and is of great significance for formulating adaptive strategies in response to glacial environmental variations. In this study, a two-dimensional higher-order flow-band glacier flow model (PoLIM) was constructed to analyze the spatio-temporal patterns of surface flow velocity on Qiyi Glacier, and the future dynamics and dominant influencing factors of flow velocity under various climate scenarios were investigated. New hydrological insights for the region: Impacted by glacier (ice) volume and thickness, the annual variation in surface flow velocity of Qiyi Glacier exhibits a decreasing trend, declining from 16 m a-1 in 1958-5.97 m a-1 in 2021, with significantly higher flow velocity during the ablation season compared to the non-ablation season. The mean glacier surface flow velocity along the main flowline was 6.92 f 0.13 m a-1 from 2017 to 2021. Under three future scenarios, this velocity is projected to decrease to 0.71 f 0.13 m a-1, 0.73 f 0.12 m a-1 and 0.47 f 0.09 m a-1 by 2050, respectively. The spatio-temporal patterns of glacier flow velocity are primarily related to glacier scale (ice thickness) and its variation, with climate warming-induced basal sliding serving as a principal driver of velocity changes in certain part of the glacier.

Risk factors at different stages of COVID-19 may interact with each other, forming a risk network. Identifying the key risk factors within this network and their interrelationships is crucial for reducing the overall risk of COVID-19. We constructed three Bayesian Belief Network (BBN) models by combining data-driven approaches with expert validation. Using the Tree-Augmented Naive Bayes (TAN) algorithm, we developed the INFORM COVID-19 Risk BBN model and the COVID-19 Regional Safety Assessment BBN model. The joint BBN model was established using the Greedy Thick Thinning (GTT) algorithm. Parameter learning was performed through maximum likelihood estimation. Expert validation, 10-fold cross-validation, and model performance metrics were employed to comprehensively assess the overall performance of the models. Additionally, mutual information analysis and sensitivity analysis were used to explore the importance of risk factors at each stage and their interdependencies. "INFORM Vulnerability" and "INFORM Lack of Coping Capacity" were identified as the two key risk factors influencing the risk of early outbreak. In the mid-to-late stages of the pandemic, "Emergency Preparedness" and "Monitoring and Detection" had the greatest impact on regional safety and control measures. Furthermore, the joint BBN model indicated that the most important risk factors affecting the overall COVID-19 risk were "Lack of Coping Capacity," "Government Risk Management Efficiency," and "Regional Resiliency," while the influence of other variables was relatively minor. The main contribution of this study lies in identifying the key risk factors at different stages of the pandemic and their interdependencies, providing policymakers with valuable insights for the rational allocation of limited health resources and the formulation of appropriate and effective prevention and control policies.

Recent global influenza resurgences, escalating to pandemics, emphasize the urgency for effective vaccinations. Despite their efficacy, vaccines offer limited protection against A/H3N2 variants. Thus, elucidating the spatial patterns and underlying drivers of A/H3N2 seasonality is critical for its management. However, the mechanisms governing this seasonality are not fully understood. The study conducted a collaborative and interdisciplinary analysis of influenza A/H3N2 epidemiology in China from 2012 to 2018, utilizing national influenza surveillance data, viral gene sequence data, and meteorological information. We initially examined the spatiotemporal distribution of influenza A/H3N2 across different temperate zones in China. Subsequently, we employed Bayesian "SkyGrid" reconstruction analysis to gain insights into the population dynamics of the influenza A/H3N2 virus within China's temperature zones. Additionally, we utilized generalized additive models (GAM) to assess the influence of meteorological factors on the seasonal prevalence of influenza A/H3N2. Our analysis of China's national influenza data revealed distinct seasonal patterns for A/H3N2: winter epidemics prevailed in temperate zones, while summer and autumn outbreaks occurred in subtropical and tropical areas. The seasonality of influenza A/H3N2 across China's diverse climatic zones is shaped by the interplay of virus migration and meteorological factors. Virus migration introduced new variant populations during seasonal epidemics of influenza A/H3N2 to different temperature zones in China, thereby seeding subsequent seasonal outbreaks. Our findings also indicate that meteorological elements trigger influenza A/H3N2 activity following virus migration. Moreover, the spatial variations in influenza A/H3N2 seasonality in China can be attributed to specific temperature thresholds, approximately 1 °C and 24 °C. These thresholds could serve as potential indicators for A/H3N2 prevalence. This insight is invaluable for tailoring region-specific prevention and control strategies in China and other regions with similar environmental conditions.

The heterogeneous and complex nature of prediabetes presents a major challenge in identifying individuals predisposed to developing incident diabetes and related complications. We aimed to identify phenotypic subgroups of prediabetes at risk and to explore their distinct associations with cardiometabolic outcomes. This study included 79,000 individuals with prediabetes from the three large-scale prospective cohorts in China. Phenotypic heterogeneity was identified using a soft-clustering algorithm based on the proximity network derived from uniform manifold approximation and projection (UMAP), combined with graph-clustering and Gaussian mixture models. Associations between phenotype probabilities and the incidence of type 2 diabetes (T2D), cardiovascular disease (CVD), and kidney events were assessed to evaluate risk differences across the identified profiles. Six phenotypic profiles were identified, including five with distinct metabolic features (representing 70

Respiratory syncytial virus (RSV), a major cause of acute respiratory infections (ARIs) globally, poses a significant threat, especially to vulnerable populations. However, the spatial transmission dynamics of RSV strains, including the influence of environmental and socioeconomic factors, remain inadequately understood. This study applied genetic sequences and phylogenetic methods to quantify evolutionary and spatial dispersal dynamics of RSV subgroup A (RSVA) across China from 2011 to 2019. We assessed viral population trends, mapped interprovincial transmission patterns, and evaluated the influence of meteorological and socioeconomic factors on viral spread. Our results revealed cyclical fluctuations in effective population size every 3-5 years, and a predominant southward spread driven by interprovincial transmission networks. We found that higher winter relative humidity (RH), urbanization rate, and human mobility promoted viral spread, while higher winter temperature and elevated urban population density appeared to inhibit it. These findings provide crucial insights into RSVA dispersal in China, underscoring the importance of regional surveillance networks and targeted interventions to curb cross-regional spread, and offer a valuable framework to inform RSV vaccine rollout strategies and guide resource allocation in high-risk areas.

Identifying causal relationships is essential for understanding the mechanisms through which natural and anthropogenic factors interact within Earth systems. However, in spatial cross-sectional data, the absence of temporal ordering poses significant challenges to traditional causal inference methods. This study proposes a novel Geographical Pattern Causality (GPC) model to detect positive, negative, dark causality and its strength between variables in spatial data. Grounded in dynamical systems theory and generalized embedding principles, the method transforms spatial neighbourhoods into lagged sequences, reconstructs the phase space, and compares symbolic trajectories to assess predictability and consistency in pattern changes-thereby inferring both the direction and type of causality. Case studies demonstrated that, compared to correlation analysis and Linear Non-Gaussian Acyclic Model (LiNGAM), the GPC model could reveal latent causal relationships among weakly correlated variables in geographical systems and capture diverse causal patterns. Despite limitations, such as sensitivity to noise and potential biases from proxy variables, the GPC model provides a novel framework for causal inference based on spatial observations, and it advances both the methodological and theoretical development of causality analysis in complex geographical systems.

Spatial count data are a prevalent data type in natural and social sciences. As the data present complicated spatial autocorrelation and heterogeneity inherent in geographical analysis, the current methods lack a theoretical approach to model and predict counts, especially with limited spatial samples. To address the gap, this study develops a new method named Poisson Means of Stratified Nonhomogeneity (PoiMSN). The method considers both autocorrelation and heterogeneity but not covariates. Moreover, it incorporates local samples and out-stratum neighbors that traditional methods neglect to model and predict the latent process for data in a Poisson distribution. This study compares PoiMSN with Poisson geostatistics and traditional MSN and designs simulations to validate the model. PoiMSN outperforms the other models as it has the lowest mean absolute error and root-mean-squared error, and furthermore, at least 5% improvement in accuracy for autocorrelated and stratified Poisson data. The case study with hand, foot, and mouth disease data shows PoiMSN can precisely map the disease risks with lower uncertainty. PoiMSN has the ability to accommodate spatially non-stationary count data from autocorrelated and heterogeneous populations and leverage extensive sample information.

Life expectancy (LE) is one of crucial metrics of human evolution. However, the evolutionary trajectories of LE in different regions of China and the regional inequalities expected in 2030 are still unclear yet. This study collected provincial LE data and relevant explanatory variables for the years of 2000, 2010, 2020 in China. The Geotree method was employed to reconstruct the evolution trajectories of LE, while a multilevel model was used to predict LEs at the provincial levels in the country for the year 2030. The LE in China exhibited significant geographical pattern, decreasing from the east to the west of the country. LE increases with socio-economic development but is constrained by the natural environment. The physical limitation to LE is significant in western China but are being alleviated with the development of socio-economic conditions. LE will increase in all provinces by 2030, with the overall LE in China reaching 80.05 years (95

Ecological corridor is conducive to promoting biological migration and flow, improving ecosystem services, and better maintaining ecological balance and human well-being. Using multi-source data (meteorology, remote sensing, and socio-economy), an indicator and model system covering four categories key ecosystem services (ESs) was built to analyze the spatial pattern in the Fenhe River Basin. The trade-offs/synergies among ESs were explored by Spatial Overlay, and the driver's contributions for ESs were estimated by Random Forest model. On this basis, the current and planned eco-corridors were identified by Minimum Cumulative Resistance (MCR) model, and providing the suggestions for eco-corridor optimization. The results showed that forests in the Lvliang and Taihang Mountains provided high-quality regulation (water conservation, soil retention, water purification, sand fixation) and support services (environmental purification, biodiversity), which were 70.79 % and 50.28 % higher than the regional average. Compared to the regional average, supply service in the cultivated land of river valley exceeded by 62.46 %, while cultural service in certain unused land increased by 1.75 times. There were strong trade-offs among four ESs, with the area accounting for 37.23 %. Natural factors played a decisive role in the key ecosystem services, and the contribution exceeded 64 %, especially terrain and vegetation, with the contributions of 31.62 %-65.38 %. The explanatory power of social-economic factors (<27 %) is weak, but the role of population and GDP cannot be ignored. For the eco-corridors construction, taking the main stream of the Fenhe River as the axis, the Taihang and Lvliang Mountains as ecological barriers, a new corridor system connecting the east and west of the basin needed to be built. The findings provided scientific guidance for the formulation and implementation of ecological sustainability policies in China, and reference for the improvement of biodiversity and ecological security in ecologically sensitive regions worldwide.

Soil pollution threatens human health and food security, particularly in industrial legacy sites. Accurate three-dimensional distribution modeling of soil contamination is crucial for understanding pollutant migration and guiding targeted remediation. Yet, the strong heterogeneity of contaminants limits the performance of traditional methods. We proposed a GeoAI-based approach, the three-dimensional deep kriging neural network (3D-DKNN), which combines deep learning with geostatistical principles to enhance interpolation accuracy in heterogeneous environments. Applied to polycyclic aromatic hydrocarbons at a typical industrial site, 3D-DKNN was benchmarked against traditional three-dimensional ordinary kriging (3D-OK) and inverse distance weighting (IDW). Cross-validation shows that 3D-DKNN achieved the lowest RMSE and MAE, and the highest correlation coefficient. Relative to IDW, it reduced RMSE by 36 %-80 %, MAE by 40 %-58 %, and increased correlation by over 19 %. Based on risk thresholds, contamination hotspots were identified in the northern and northwestern areas, particularly manufacturing and warehouse areas, which were recognized as key risk and remediation areas. This study demonstrates the potential of GeoAI for modeling complex pollutants and improving soil risk assessment.

What is already known about this topic?:Disability-free life expectancy (DFLE) is a vital measure of older adults' quality of life. Although its temporal trends and determinants have been examined in previous studies, spatial heterogeneity has often been underestimated, because most analyses were conducted at aggregated national or provincial levels, masking local variations due to limited data availability. What is added by this report?:Compared with 2018, DFLE increased in 2023, and the urban-rural gap narrowed overall, with slower improvement in the western region. The determinants of DFLE varied by region, age, and urban-rural context. What are the implications for public health practice?:Public health strategies should be tailored to urban and rural contexts. Priority should be given to strengthening healthcare access, social security systems, and climate-adaptive infrastructure, particularly in rural areas of western China.

BACKGROUND:High temperature beyond the comfort threshold is the main hazard to cause heat-related mortality. However, existing methods of defining the heat thresholds are usually based on case studies in data-rich regions and rarely considers the acclimatization. METHODS:Based on the temperature-mortality relationship observed in 36 locations covering all six major climate zones in China, we found that the relative risk (RR) of heat-related mortality and the annual frequency of temperature (AFT) have a power function relationship (adjusted R2 = 0.74)), and the association is independent to the variation of the temperature across the territory. Furthermore, the association is slightly changed when the GDP/capita, proportion of elderly population and latitude are adjusted. According to this association, we proposed a new method to choose the heat threshold at finer resolution using only AFT. As the temperature frequency is easy to calculate, this method can be promoted to any geographical location without mortality data. RESULTS:According to the relationship between AFT and RR, using the daily time series of temperature at 2405 observation stations in China, we estimated and mapped the distribution of heat thresholds at the county level across China. We find that when the AFT is just 1 day per year, the corresponding RR is approximately 1.4 (95% CI, 1.2-1.8). As the AFT increases to 5 days per year, the RR decreases to about 1.2 (95% CI, 1.1-1.3). When the AFT reached 10 days per year, the RR further decreased to about 1.05 (95% CI, 1.0-1.1). CONCLUSIONS:This study advances the understanding on the driver of human beings' adaptation to high temperature. It also contributes significantly to the research on heat-related mortality in the context of global climate change.

The scientific new urbanization in mineral resource abundant regions (MRARs) in China is not only a guarantee for energy and resource security but also of great significance to realize strong sustainability. However, the new urbanization of MRARs in China is facing a series of difficulties due to the particularity of mineral resources. Therefore, to identify the policy foci of MRARs suitable for different new urbanization paths, this study conducts an evolutionary game analysis to find the urbanization evolution trends of MRARs. First, we established an asymmetric evolutionary game model to study the process of strategy selection between rural and urban residents in MRARs. Second, MATLAB was used to conduct numerical simulation to find the urbanization evolution trends in different types of MRARs. The main findings we obtained are as follows. (1) Environmental carrying capacity has a restrictive effect on the new urbanization in MRARs. (2) Socioeconomic foundations have a fundamental impact on the new urbanization in MRARs. (3) The policy intensity of “people-oriented” has an important influence on the new urbanization in MRARs. (4) Different types of MRARs greatly vary with new urbanization trends. Finally, we proposed some policy recommendations to effectively promote the new urbanization in MRARs.

BackgroundWhile stationary links between childhood hand, foot and mouth disease (HFMD) and air pollution are known, a comprehensive study on their heterogeneous relationships (nonstationarity), jointly considering numerical, temporal and spatial dimensions, has not been reported.MethodsMonthly HFMD incidence and air pollution data were collected at the county level from Sichuan-Chongqing, China (2009-2011), alongside meteorological and social environmental covariates. Key influential factors were identified using random forest (RF) under the stationary assumption. Factors' numerically, temporally, and spatially heterogeneous relationships with HFMD were assessed using generalized additive model (GAM) and geographically and temporally weighted regression (GTWR).ResultsOur findings highlighted the relatively higher stationary contributions of fine particulate matter (PM2.5) and ozone (O3) to HFMD incidence across Sichuan-Chongqing counties. We further uncovered heterogeneous impacts of PM2.5 and O3 from three nonstationary perspectives. Numerically, PM2.5 showed an inverse 'V'-shaped relationship with HFMD incidence, while O3 exhibited a complex pattern, with increased HFMD incidence at low PM2.5 and moderate O3 concentrations. Temporally, PM2.5's impact peaked in autumn and was weakest in spring, whereas O3's effect was strongest in summer. Spatially, hotspot mapping revealed high-risk clusters for PM2.5 impact across all seasons, with notable geographical variations, and for O3 in spring, summer, and autumn, concentrated in specific regions of Sichuan-Chongqing.ConclusionsThis study underscores the nuanced and three-perspective heterogeneous influences of air pollution on HFMD in small areas, emphasizing the need for differentiated, localized, and time-sensitive prevention and control strategies to enhance the precision of dynamic early warnings and predictive models for HFMD and other infectious diseases, particularly in the fields of environmental and spatial epidemiology.

The phenomena with within-strata characteristics that are more similar than between-strata characteristics are ubiquitous (e.g., land-use types and image classifications). It can be summarized as spatial stratified heterogeneity (SSH), which is measured and attributed using the geographical detector (Geodetector) q-statistic. SSH is typically calibrated by stratification and hundreds of algorithms have been developed. Little is discussed about the conditions of the methods. In this work, a novel stratification method based on head/tail breaks is introduced for the purpose of better capturing the SSH of geographical variables with a heavy-tailed distribution. Compared to conventional sample-based stratifications, the presented approach is a population-based optimized stratification that indicates an underlying scaling property in geographical spaces. It requires no prior knowledge or auxiliary variables and supports a naturally determined number of strata instead of being subjectively preset. In addition, our approach reveals the inherent hierarchical structure of geographical variables, characterizes its dominant components across all scales, and provides the potential to make the stratification meaningful and interpretable. The advantages were illustrated by several case studies in natural and social sciences. The proposed approach is versatile and flexible so that it can be applied for the stratification of both geographical and nongeographical variables and is conducive to advancing SSH-related studies as well. This study provides a new way of thinking for advocating spatial heterogeneity or scaling law and advances our understanding of geographical phenomena.