Spatial Non-stationarity Research Articles

Efficient Large-Scale Nonstationary Spatial Covariance Function Estimation Using Convolutional Neural Networks

Spatial processes observed in various fields, such as climate and environmental science, often occur at large-scale and demonstrate spatial nonstationarity. However, fitting a Gaussian process with a nonstationary Matérn covariance is challenging, as it requires handling the complexity and computational demands associated with modeling the varying spatial dependencies over large and heterogeneous domains. Previous studies in the literature have tackled this challenge by employing spatial partitioning techniques to estimate the parameters that vary spatially in the covariance function. The selection of partitions is an important consideration, but it is often subjective and lacks a data-driven approach. To address this issue, in this study, we use the power of Convolutional Neural Networks (ConvNets) to derive subregions from the nonstationary data by employing a selection mechanism to identify subregions that exhibit similar behavior to stationary fields. We rely on the ExaGeoStat software for large-scale geospatial modeling to implement the nonstationary Matérn covariance for large scale exact computation of nonstationary Gaussian likelihood. We also assess the performance of the proposed method with synthetic and real datasets at large-scale. The results revealed enhanced accuracy in parameter estimations when relying on ConvNet-based partition compared to traditional user-defined approaches.

Structural fluctuations of the Arctic Oscillation tied to the Atlantic Multidecadal Oscillation

The Arctic Oscillation (AO) has been observed to undergo distinct decadal structural fluctuations that significantly influence regional weather and climate. Understanding the drivers and mechanisms behind the AO’s spatial nonstationarity is critical for improving climate predictions related to the AO. We present evidence that the Atlantic Multidecadal Oscillation (AMO) plays a pivotal role in modulating AO’s Pacific center in recent decades. The poleward amplified cooling associated with negative AMO enhances the north-south temperature gradient and results the strengthened westerly winds and stratospheric polar vortex (SPV) responses, which reflects more planetary waves from the North Pacific to the North Atlantic. This enhances the atmospheric coupling between these regions and leads to a more pronounced Pacific center within the AO pattern. Numerical simulations from ECHAM5 and 35 CMIP6 models further corroborate the essential role of the AMO. These findings advance our understanding of the mechanisms driving the variability of the AO pattern.

Spatial non-stationarity effects of the driving factors on landscape ecological risk: A case of the Xiamen-Zhangzhou-Quanzhou Urban Agglomeration, China

Increasing construction land demands and population growth within urban agglomerations have led to increasingly prominent landscape ecological risk (LER) issues. However, the multi-scale nature and spatial non-stationarity characteristics of the driving patterns have not been fully investigated. In this study, taking the Xiamen-Zhangzhou-Quanzhou Urban Agglomeration (XZQUA) as a case, we analyzed the spatio-temporal evolution characteristics of LER utilizing the long time series land use data from 1980 to 2020; then, the Exploratory Regression Analysis (ERA) was employed to investigate the optimal combination of key influential elements affecting LER. Finally, the Multiscale Geographically Weighted regression (MGWR) model was applied to explore the spatial non-stationarity effects of the key driving elements. The results reveal that: (1) The LER of the XZQUA has shown a declining trend from 1980 to 2020, characterized with significantly and increasingly spatial autocorrelation as well. (2) The optimal variable combination is the temperature, distance from roads, elevation, and construction land area. (3) The MGWR model results reveal that the impact of various factors on LER demonstrates significant spatial non-stationarity. The mechanisms and intensity of these influences change depending on geographical location. The negative proportion of regression coefficients for construction land area (−0.203 to −0.002) and elevation (−0.046 to −0.001) all exceeded 80%, indicating a high negative effect on LER; The positive proportion of the regression coefficient for distance from roads (−0.026 to 0.042) exceeded 60%, indicating a predominantly positive effect on LER. Our findings provide guidance for regional policymakers to adopt targeted ecological risk intervention measures, thus reducing LER in the urban agglomeration.

Hybrid Precoding Design for Near-Field Wideband THz Systems With Spatial Non-Stationarity

Hybrid Precoding Design for Near-Field Wideband THz Systems With Spatial Non-Stationarity

Spatial Nonstationarity in Phenological Responses of Nearctic Birds to Climate Variability.

Climate change is shifting the phenology of migratory animals earlier; yet an understanding of how climate change leads to variable shifts across populations, species and communities remains hampered by limited spatial and taxonomic sampling. In this study, we used a hierarchical Bayesian model to analyse 88,965 site-specific arrival dates from 222 bird species over 21 years to investigate the role of temperature, snowpack, precipitation, the El-Niño/Southern Oscillation and the North Atlantic Oscillation on the spring arrival timing of Nearctic birds. Interannual variation in bird arrival on breeding grounds was most strongly explained by temperature and snowpack, and less strongly by precipitation and climate oscillations. Sensitivity of arrival timing to climatic variation exhibited spatial nonstationarity, being highly variable within and across species. A high degree of heterogeneity in phenological sensitivity suggests diverging responses to ongoing climatic changes at the population, species and community scale, with potentially negative demographic and ecological consequences.

Analysis of multi-scale effects and spatial heterogeneity of environmental factors influencing purse seine tuna fishing activities in the Western and Central Pacific Ocean

Understanding the spatial fishing activity distribution characteristics is important for the sustainable development of fisheries. Spatial nonstationarity is always present, especially in marine ecosystems. To explore how marine environmental factors affect the fishing effort of tuna purse seine vessels, data from 2015 to 2020 on the fishing activities of these fleets and environmental variables in the Western and Central Pacific Ocean (WCPO) were analyzed. A Generalized Additive Model (GAM), Geographically Weighted Regression model (GWR), and Multi-Scale Geographically Weighted Regression (MGWR) model were applied to explore the drivers of fishing activity and the impacts of environmental factors on spatial heterogeneity. The results indicate that: (1) The MGWR models has the highest prediction accuracy and effectively reflects the spatial heterogeneity and multi-scale effects of environmental factors in a year. (2) Environmental factors exhibit significant multi-scale effects and spatial heterogeneity on the fishing activities of purse seine tuna vessels. Sea floor depth, salinity at 200 m depth and sea surface temperature show the greatest spatial heterogeneity in their impact on fishing activities. (3) Sea surface temperature, distance to port, and primary productivity and salinity at 200 m depth are key variables influencing the fishing activities of purse seine tuna vessels. These findings are expected to provide scientific and effective guidance for fishery management and sustainable development by assessing the spatial variations in fishing activities at multiple scales.

Spatial nonstationarity and the role of environmental metal exposures on COVID-19 mortality in New Mexico

Worldwide, the COVID-19 pandemic has been influenced by a combination of environmental and sociodemographic drivers. To date, population studies have overwhelmingly focused on the impact of societal factors. In New Mexico, the rate of COVID-19 infection and mortality varied significantly among the state's geographically dispersed, and racially and ethnically diverse populations who are exposed to unique environmental contaminants related to resource extraction industries (e.g. fracking, mining, oil and gas exploration). By looking at local patterns of COVID-19 disease severity, we sought to uncover the spatially varying factors underlying the pandemic. We further explored the compounding role of potential long-term exposures to various environmental contaminants on COVID-19 mortality prior to widespread applications of vaccinations. To illustrate the spatial heterogeneity of these complex associations, we leveraged multiple modeling approaches to account for spatial non-stationarity in model terms. Multiscale geographically weighted regression (MGWR) results indicate that increased potential exposure to fugitive mine waste is significantly associated with COVID-19 mortality in areas of the state where socioeconomically disadvantaged populations were among the hardest hit in the early months of the pandemic. This relationship is paradoxically reversed in global models, which fail to account for spatial relationships between variables. This work contributes both to environmental health sciences and the growing body of literature exploring the implications of spatial nonstationarity in health research.

Evaluation of thermal effects on urban road spatial structure: A case study of Xuzhou, China

Urban heat islands (UHI) are important environmental issue in cities where urban spatial structure has been proven to play an important role in alleviating UHI effects. The relationship between land surface temperature and urban spatial structures has been explored, providing strong support for their cooling effects. Urban roads are the skeleton of urban spatial structures, with obvious spatial structure characteristics; however, research on the relationship between roads and the thermal environment has been mostly focused at the micro and meso level, lacking exploration at the macro spatial structure scale. Xuzhou—a typical average-sized city in China—was selected as the research object and the road system as the carrier. The thermal environmental effects of road elements such as their structural attributes, geometric attributes and unique construction attributes were quantitatively analyzed using geographically weighted regression analysis. The results revealed that 1) the contribution of roads in the study area to the UHI effect is relatively stable; therefore, this area should become an important cooling space to decompose UHI patch connectivity and thus decrease the UHI effect. 2) the self-organizing structural characteristics of urban roads affect their thermal environments where in the straightness of the road structure and road thermal environment showed a clear overall negative correlation And 3) the length and width of the road segments had negative and positive effects on the thermal environment, respectively. The green coverage of the roads has a global negative effect on the thermal environment, but shows obvious spatial non-stationarity. Therefore, green measures must be implemented in different regions. The results here provide a quantitative basis for urban road system planning and urban form management and control that incorporates thermal environment improvements, as well as a reference for the study of urban thermal environments under different spatial forms and planning control systems in other countries and regions.

Exploring Spatial Non-Stationarity and Scale Effects of Natural and Anthropogenic Factors on Net Primary Productivity of Vegetation in the Yellow River Basin

Investigating the spatiotemporal dynamics of vegetation net primary productivity (NPP) and its influencing factors are crucial for green and low-carbon development and facilitate human well-being in the Yellow River Basin (YRB). Although the research on NPP has advanced rapidly, in view of the regional particularity of the YRB, the persistence of its NPP change trend needs to be further discussed and more comprehensive impact factors need to be included in the analysis. Meanwhile, the spatial non-stationarity and scale effects of the impact on NPP when multiple factors are involved remain uncertain. Here, we selected a total of twelve natural and anthropogenic factors and used multi-scale geographically weighted regression (MGWR) to disentangle the spatial non-stationary relationship between vegetation NPP and related factors and identify the impact scale difference in the YRB. Additionally, we analyze the spatiotemporal variation trend and persistence of NPP during 2000–2020. The results revealed the following: (1) The annual NPP showed a fluctuating increasing trend, and the vegetation NPP in most regions will exhibit a future trend of increasing to decreasing. (2) The effects of different factors show significant spatial non-stationarity. Among them, the intensity of the impact of most natural factors shows a clear strip-shaped distribution in the east-west direction. It is closely related to the spatial distribution characteristics of natural factors in the YRB. In contrast, the regularity of anthropogenic influences is less obvious. (3) The impact scales of different factors on vegetation NPP were significantly different, and this scale changed with time. The factors with small impact scales could better explain the change in vegetation NPP. Interestingly, the impact size and scale of relative humidity on NPP in the YRB are both larger. This may be due to the arid and semi-arid characteristics of the YRB. Our findings could provide policy makers with specific and quantitative insights for protecting the ecological environment in the YRB.

Exploring spatial non-stationarity of child labour and its related factors: A multiscale geographically weighted regression study of India

The relationships between various factors and child labour have been explored in the literature but, despite findings that suggest the predictive factors of child labour can vary according to context, there has been little research that has used spatial methods of analysis or attempted to estimate local relationships between covariates and the prevalence of child labour. This paper seeks to address this knowledge gap by using geographically weighted regression (GWR) and multiscale geographically weighted regression (MGWR) models. Using India 2011 Census Data as a case study, the findings show that GWR and MGWR models both perform significantly better than a global regression model across the whole of India. The study also finds significant spatial non-stationarity in the relationships between child labour and its covariates, with the association between district-level child labour rates and both the Muslim population and the child sex ratio found to have opposite directions in different parts of the country. Using MGWR, it was also possible to demonstrate that different covariates interact with child labour at various spatial scales, suggesting that interventions aiming to address varying aspects of the child labour problem may need to be deployed at different administrative levels to maximise their efficacy.

Spatiotemporal geographically weighted regression analysis for runoff variations in the Weihe River Basin

In order to investigate the effects of vegetation changes on runoff and to obtain recommendations for improving runoff in the Weihe River Basin (. In this study, a spatiotemporal geographic autocorrelation weighted regression analysis (SGAWRA) approach was newly developed based on previous studies. This approach investigates spatial non-stationarity of the dynamic response from vegetation variations to climatic change and human activity. Implications of spatial non-stationarity related to runoff variability were also discussed, which in turn yield the effect that vegetation changes have on runoff. The method systematically analysed the spatial non-stationarity of vegetation variations and its associated effects on runoff. Therefore, more closely related results with less error were produced at each step, and results with more accuracy were obtained. These results indicated that the average trend rates of NDVI in the annual average, each season, and the growing season (Growing season refers to April to September) exceeded 0. Areas where NDVI show a growing trend cover more than 50%, which is greater than the area with a decreasing trend. The GWR regression parameters of precipitation, average temperature, and NDVI are all greater than 0. The GWR regression parameters of human activities and NDVI also have more than 50% of the area greater than 0. Based on the visual analysis of the calculation results, it can be seen that there are obvious spatial trends in the data, and the spatial data are significantly different between different regions. Therefore, WRB can be regarded as spatio-temporally non-stationary. In the WRB, the underlying surface change with vegetation change as the prominent feature is the leading cause (about 60%) of the runoff attenuation. The results showed that WRB has spatial and temporal non-stationarity. The spatial non-stationarity of vegetation has a greater effect on runoff changes. The results of this study support recommendations for improving runoff in the WRB.

Spatial Big Data Analytics of Inequity in Access to Live Music Among Disadvantaged Populations

In recent years, music cities around the world have developed various supportive policy frameworks ensuring live music is accessible to large populations. The question of whether live music is equitably distributed by socioeconomic status remains unanswered, however. This study investigated the accessibility of music events in socially and economically disadvantaged communities. The level of accessible concerts varies substantially across U.S. counties, revealing significant inequity in access to concerts. Almost 80 percent of concerts are located in 5 percent of counties. Ordinary least squares (OLS) regression, geographically weighted regression (GWR), and multiscale geographically weighted regression (MGWR) methods were then employed to explore associations between accessible concerts and the socioeconomic characteristics of communities where these concerts are located. The regression models revealed that areas with a higher proportion of residents without a high school degree, higher unemployment rate, and higher poverty rate were less likely to have access to concerts, yet positive associations were found between accessible concerts and mean household income. Local R 2 values further confirmed the robustness of regression results of both GWR and MGWR and showed significant spatial nonstationarity of model performance. GWR and MGWR methods improved model performance significantly over the OLS method and proved how local conditions influence the OLS regression. The results from this study can guide decision makers and stakeholders in making appropriate policies to achieve equitable access to music events, especially for disadvantaged communities.

Spatial Downscaling of Nighttime Land Surface Temperature Based on Geographically Neural Network Weighted Regression Kriging

Land surface temperature (LST) has a wide application in Earth Science-related fields, and spatial downscaling is an important method to retrieve high-resolution LST data. However, existing LST downscaling methods have difficulties in simultaneously constructing and expressing spatial non-stationarity, spatial autocorrelation, and complex non-linearity during the LST downscaling process, which limits the performance of the models. Moreover, there is a lack of research on high-resolution nighttime land surface temperature (NLST) reconstruction based on spatial downscaling, which does not meet the data needs for urban-scale nighttime urban heat island (UHI) studies. Therefore, this study combined Geographically Neural Network Weighted Regression (GNNWR) with Area-to-Point Kriging interpolation (ATPK) to propose a Geographically Neural Network Weighted Regression Kriging (GNNWRK) model for NLST downscaling. To verify the model’s generality and robustness, this study selected four study areas with different landform and climate type for NLST spatial downscaling experiments. The GNNWRK was compared with four benchmark downscaling methods, including TsHARP, Random Forest, Geographically Weighted Regression, and GNNWR. The results show that compared to these four benchmark methods, the GNNWRK method has higher accuracy in NLST downscaling, with a maximum Pearson’s Correlation Coefficient (Pcc) of 0.930 and a minimum Root Mean Square Error (RMSE) of 0.886 K. Moreover, the validation based on MODIS NLST data and ground-measured NLST data also indicates that the GNNWRK model can obtain more accurate, high-resolution NLST with richer and more detailed texture. This enhances the potential of NLST in studying the effects of urban nighttime heat islands at a finer scale.

Optimizing cultivated land under non-grain expansion in rural China: unveiling the spatial dynamics of trade-offs and synergies among ecosystem services

ABSTRACT Non-grain expansion in rural China has affected the stability of cultivated land structure and the sustainability of ecosystem services (ESs). Delving into the spatio-temporal heterogeneity of ES interactions is essential for refined and differentiated cultivated land management practices. Taking Tongxiang City as an example, this study mapped the variations of four ESs of cultivated land from 2000–2020, identified their socio-ecological drivers, explored the spatial non-stationarity characteristics of trade-off/synergy among different ESs, and proposed a zoning control framework for cultivated land. The results indicated that (1) food supply (FS) and habitat quality (HQ) decreased by 40.40% and 1.27%, respectively, while carbon sequestration (CS) and soil conservation (SC) increased by 8.5% and 26.21%, respectively. Areas of decreased ES were mainly concentrated in the southwest, while areas of increased ES supply were dispersed. (2) CS-SC and CS-FS ES pairs exhibited obvious trade-offs, with concave trade-offs dominating between 2000 and 2010, and convex trade-offs prevailing from 2010 to 2020. These ES trade-offs were mainly distributed in the southeast and northern regions. HQ-CS and FS-SC presented significant synergies, predominantly characterized by linear synergy and situated in the northern and central areas. (3) The spatial differentiation of ESs was mostly explained by landscape configuration drivers. (4) Different cultivated land zones were delineated based on the spatial non-stationarity of trade-offs/synergies and ES advantage areas. This study contributes to our understanding of the changes in ESs resulting from non-grain expansion and offers valuable guidance for implementing differentiated measures for cultivated land zoning and management.

Reinforced deep learning approach for analyzing spaceborne-derived crop phenology

Vegetation Index (VI) curves capture the crop-specific phenological events from multi-temporal Earth observation images. Although advanced machine learning classifiers are the existing state-of-the-art, most of them assume the samples to be independent and identically distributed, which is not true for VI curve classification. This research proposed the consideration of spatial non-stationarity for VI curve-based crop classification. The proposed approach does not assume the samples to be independent and identically distributed. Different state-of-the-art approaches attempted to solve this problem by introducing architectural variations and considering the neighborhood patches. However, the inherent nature of the problem demanded an online approach. In the current research, this problem is formulated in a reinforcement learning framework where the generalizability of the predictor is improved using a feedback system. Different approaches for feedback were investigated in this study. The predictor was composed of a graph convolutional framework where the VI curves were transformed into a graph representation, incorporating the spatial, spectral, and temporal context. Graph convolutional operations were then used to learn the embedded representations and assign labels to the unlabeled points based on the labeled ones in the proximity. Further, a semi-supervised strategy was proposed where some initial feedback was employed to improve the generalizability of the model. Additionally, neighborhood-based feedback was also considered. The experiments using the VI curves collected over three farms in Israel covering wheat, potato, and barley crops illustrated the need for accounting the spatial non-stationarity in VI curve classification. The proposed dynamic feedback system and the graph-based strategy to consider spatial heterogeneity significantly improved the accuracy of the proposed framework. Additionally, the proposed reinforcement learning strategy ensures the dynamic refinement of model weights to facilitate the real-time applications even when the target data varies significantly. Although the proposed approach is generic, the end-to-end framework, as adopted in this research, is suitable for aerial or agricultural datasets. Hence, further fine tuning and remodeling may be required for non-agriculture or non-spatial datasets. However, the dynamic weight update may be adopted for any applications irrespective of the datasets.

The impact of socioeconomic factors on vegetation restoration in karst regions: A perspective beyond climate and ecological engineering

Socioeconomic factors are important parameters that affect vegetation changes in karst areas. Previous studies primarily focused on ecological engineering when analyzing the impact of human activities on vegetation restoration, whereas the impact of socioeconomic factors has been less studied. The study employs structural equation modeling and geographically weighted regression to quantitatively assess the impact of socioeconomic factors on vegetation dynamics within karst regions, with a focus on the county level. The results showed a significant upward trend in both the Enhanced Vegetation Index and Net Primary Productivity across China's humid karst regions from 2000 to 2020. Among the socioeconomic factors, non-farm economy and rural economy had a positive effect on vegetation change, with maximum path coefficients of the structural equation model of 0.79 and 0.64, respectively; whereas population pressure had a negative impact with a minimum path coefficient of −0.80. Over time, the positive impact of rural economy on vegetation restoration showed an increasing trend, and the path coefficient increased from-0.92 to 0.64; in turn, non-farm economy and population pressure showed a decreasing trend. Moreover, because they were affected by the heterogeneity of the karst mountain environment, the impact of various socioeconomic factors on vegetation restoration had obvious spatial non-stationarity. The results of this study will promote our understanding of the mechanism underlying vegetation change in humid karst areas and provide scientific reference for ensuring the sustainability of the ecological effects in these areas.

A Novel Flexible Geographically Weighted Neural Network for High-Precision PM2.5 Mapping across the Contiguous United States

Air quality degradation has triggered a large-scale public health crisis globally. Existing machine learning techniques have been used to attempt the remote sensing estimates of PM2.5. However, many machine learning models ignore the spatial non-stationarity of predictive variables. To address this issue, this study introduces a Flexible Geographically Weighted Neural Network (FGWNN) to estimate PM2.5 based on multi-source remote sensing data. FGWNN incorporates the Flexible Geographical Neuron (FGN) and Geographical Activation Function (GWAF) within the framework of Artificial Neural Network (ANN) to capture the intricate spatial non-stationary relationships among predictive variables. A robust air quality remote sensing estimation model was constructed using remote sensing data of Aerosol Optical Depth (AOD), Normalized Difference Vegetation Index (NDVI), Temperature (TMP), Specific Humidity (SPFH), Wind Speed (WIND), and Terrain Elevation (HGT) as inputs, and Ground-Based PM2.5 as the observation. The results indicated that FGWNN successfully generates PM2.5 remote sensing data with a 2.5 km spatial resolution for the contiguous United States (CONUS) in 2022. It exhibits higher regression accuracy compared to traditional ANN and Geographically Weighted Regression (GWR) models. FGWNN holds the potential for applications in high-precision and high-resolution remote sensing scenarios.

Geographically weighted regression model-calibration for finite population parameter estimation under two stage sampling design

In many agricultural, forestry, environmental, and ecological surveys, data are often spatial in nature and exhibit spatial nonstationarity. A well-known method for addressing spatial nonstationarity and capturing the spatially varying relationships between different variables is Geographic Weighted Regression (GWR). The calibration approach is one of the most widely used techniques in sample surveys for incorporating the known population characteristics of auxiliary variables by changing the original sampling design weights. The model-calibration approach is an improvement on the conventional calibration approach that can handle a variety of assisting working models. Two-stage sampling is one of the most frequently used sampling strategies in large-scale sample surveys. In the present study, a couple of GWR model-calibration estimators were proposed under two-stage sampling, assuming the availability of population-level complete auxiliary information. Under a set of regularity assumptions, the asymptotic properties of the developed estimators have been evaluated such as design unbiasedness, model unbiasedness, approximate variance, and estimators of variances. The performance of the developed estimators has been compared with the existing estimators through a spatial simulation study and a design-based simulation based on real data. The performance of the proposed estimator was found to be more precise than the existing estimators under two-stage sampling.

Spatial non-stationarity test of regression relationships in the multiscale geographically weighted regression model

By allowing covariate-specific bandwidths for estimating spatially varying coefficients, the multiscale geographically weighted regression (MGWR) model can simultaneously explore spatial non-stationarity and multiple operational scales of the corresponding geographical processes. Treating the constant coefficients as an extreme situation which corresponds to the global scale and infinite covariate bandwidth, the traditional linear regression, GWR and mixed GWR models are special cases of the MGWR model. An appropriately-specified GWR-based model would be beneficial to the understanding of the general underlying processes, especially for their operational scales. To specify an appropriate model, the key issue is to determine how many MGWR coefficient(s) should be constant. Along the traditional statistical line of thought, we propose a residual-based bootstrap method to test spatial non-stationarity of the MGWR coefficients, which can underpin our understanding of the characteristics of regression relationships in statistics. The simulation experiment validates the proposed test, and demonstrates that it is of valid Type I error and satisfactory power, and is robust to different types of model error distributions. The applicability of the proposed test is demonstrated in a real-world case study on the Shanghai housing prices.

Association of neighborhood-level superfund sites with atmospheric toxic exposures and increasing cancer risk in the United States: A proof-of-concept geospatial study.

10577 Background: Decades of underinvestment have disproportionately exposed several neighborhoods in the US to hazardous conditions. Superfund is the name given to the environmental program established by the US Congress in 1980, to address abandoned hazardous waste sites. The extent to which Superfund sites have contributed to differences in cancer risk remains unknown. We investigated atmospheric toxic exposures increasing cancer risk by proximity to Superfund sites at the neighborhood level. Methods: In this cross-sectional study of US neighborhoods, census-tract level air toxic cancer risk data from the Environmental Protection Agency for 2019 were linked with census-tract level geolocation data of ~1300 Superfund sites from the Superfund Enterprise Management System database. Linear regression models were utilized to explore association between cancer risk and proximity to Superfunds. Geographically Weighted Regression models explored spatial non-stationarity by modeling spatial relationships that vary across the study area. Bivariate Local Indicators of Spatial Association (LISA) cluster maps were created for superfund site proximity and cancer risk (high-high, low-low, low-high, high-low). Kruskal-Wallis testing compared the distributions of different variables across the four LISA cluster types. Results: Study explored 82,770 census tracts nationwide. For every 10 percentile increase in Superfund proximity, there was an increase of 4% in cancer risk (p < 0.001). This association was strongest in the Intermountain West, South-Southeast, and Northeast regions. High-high clusters (high proximity to Superfund, high cancer risk) had higher percentage of non-White, low-income, unemployed and lower education attainment populations compared to low-low clusters (all p < 0.001). Within high-high clusters, non-White population had the strongest correlation with proximity to Superfunds. Cancer risk was much higher for the top 10% than the bottom 10% of proximity to Superfunds (Table). Conclusions: Our study highlights higher clustering of cancer risk around neighborhoods with existing Superfund sites. Further, we demonstrate the closer a neighborhood is to a Superfund site, the greater the cancer risk is with associated sociodemographic changes. Using a novel approach to geospatial analytics, the results demonstrate the first clear evidence of a dangerous association between Superfunds and healthcare, and encourage the urgent prioritization of Congressional policies that optimize investment in Superfund site clean up for public health. [Table: see text]