The estimation of pollution fields, especially in densely populated areas, is an important application in the field of environmental science due to the significant effects of air pollution on public health. In this paper, we investigate the spatial distribution of three air pollutants in Tehran’s atmosphere: carbon monoxide (CO), nitrogen dioxide (NO2), and atmospheric particulate matters less than 10 μm in diameter (PM10μm). To do this, we use four geostatistical interpolation methods: Ordinary Kriging, Universal Kriging, Simple Kriging, and Ordinary Cokriging with Gaussian semivariogram, to estimate the spatial distribution surface for three mentioned air pollutants in Tehran’s atmosphere. The data were collected from 21 air quality monitoring stations located in different districts of Tehran during 2012 and 2013 for 00UTC. Finally, we evaluate the Kriging estimated surfaces using three statistical validation indexes: mean absolute error (MAE), root mean square error (RMSE) that can be divided into systematic and unsystematic errors (RMSES, RMSEU), and D-Willmot. Estimated standard errors surface or uncertainty band of each estimated pollutant surface was also developed. The results indicated that using two auxiliary variables that have significant correlation with CO, the ordinary Cokriginga scheme for CO consistently outperforms all interpolation methods for estimating this pollutant and simple Kriging is the best model for estimation of NO2 and PM10. According to optimal model, the highest concentrations of PM10 are observed in the marginal areas of Tehran while the highest concentrations of NO2 and CO are observed in the central and northern district of Tehran.

Finding an appropriate satellite image as simultaneous as possible with the sampling time campaigns is challenging. Fusion can be considered as a method of integrating images and obtaining more pixels with higher spatial, spectral and temporal resolutions. This paper investigated the impact of Landsat 8-OLI and Sentinel-2A data fusion on prediction of several toxic elements at a mine waste dump. The 30 m spatial resolution Landsat 8-OLI bands were fused with the 10 m Sentinel-2A bands using various fusion techniques namely hue-saturation-value (HSV), Brovey, principal component analysis (PCA), Gram-Schmidt (GS), wavelet, and area-to-point regression kriging (ATPRK). ATPRK was the best method preserving both spectral and spatial features of Landsat 8-OLI and Sentinel-2A after fusion. Furthermore, the partial least squares regression (PLSR) model developed on genetic algorithm (GA)-selected laboratory visible-near infrared-shortwave infrared (VNIR–SWIR) spectra yielded more accurate prediction results compared to the PLSR model calibrated on the entire spectra. It was hence, applied to both individual sensors and their ATPRK-fused image. In case of the individual sensors, except for As, Sentinel-2A provided more robust prediction models than Landsat 8-OLI. However, the best performances were obtained using the fused images, highlighting the potential of data fusion to enhance the toxic elements’ prediction models.

Different prediction models have important effects on the accuracy of spatial distribution simulations of heavy metals in soil. This study proposes a model (RFOK) combining a random forest (RF) with ordinary kriging (OK), multi-source environmental data such as terrain elements, site environmental elements, and remote sensing data were incorporated to predict the spatial distribution of heavy arsenic (As) in soil of a certain large arsenic slag site. The predictions results of RFOK were compared with those obtained using the RF, OK, inverse distance weighted (IDW), and stepwise regression (STEPREG) models for assessment of prediction accuracy. The results showed that arsenic pollution was widely distributed and the center of the site, including arsenic slag stacking area and production area were seriously polluted. The overall spatial distribution of arsenic pollution simulated by the five models was similar, but the IDW, RF, OK, and STEPREG showed less spatial variation of soil pollution, while RFOK simulation can better express the characteristics of details in change. The cross-validation results showed that RFOK had the lowest root-mean-square error (RMSE), mean absolute error (MAE), and mean relative error (MRE) relative to the other four models, followed by RF, OK, IDW, and STEPREG. The RMSE, MAE and MRE of RFOK decreased by 62.2%, 64.3% and 68.7%, respectively, relative to the RF model with the second highest accuracy. Compared with the traditional spatial distribution prediction model, the RFOK model proposed in this study has excellent spatial distribution prediction ability for soil heavy metal pollution with large spatial variation characteristics, which can fully explain the nonlinear relationship between pollutant content and its environmental impact elements.

The traditional incremental lifetime cancer risk (ILCR) model of urban soil polycyclic aromatic hydrocarbon (PAH) health risk assessment has a large spatial scale and commonly calculates relevant statistics by regarding the whole area as a geographic unit but fails to consider the high heterogeneity of the PAH distribution and differences in population susceptibility and density in an area. Therefore, the risk assessment spatial performance is insufficient and does not reflect the characteristics of cities, which are centered on human activities and serve the needs of humans, thus making it difficult to effectively support PAH prevention and treatment measures in cities. Here, the random forest model combined with the kriging residual model (RFerr-K) is used to estimate high-precision PAH distributions, separately considering the exposure characteristics of children and adults with different susceptibilities, and kindergarten point-of-interest (POI) and population density index (PDI) data were used to estimate the distributions of the kindergarten children and adults in the study area. Through the refined expression of these three dimensions, a new spatially explicit model of the incremental lifetime cancer-causing population distribution (MapPILCR) was constructed, and the risk threshold range delineation method was proposed to accurately identify regional risk levels. The results showed that the RFerr-K model significantly improves the accuracy of PAH prediction. The susceptibility index (SI) of children is 45% higher than that of adults, and POI and PDI data can be used effectively in population distribution estimation. The MapPILCR model provides a useful method for the spatially explicit assessment of the cancer risk of urban populations to inspire urban pollution grid management.

Nitrogen is one of the most significant pollutants in the Yangtze River estuary (YRE), China. Reliable estimation of nitrogen concentration in the water is crucial for assessment of the water quality of the estuary. Because ocean fronts exist in the YRE, which divide water masses into different regions, it is necessary to account for the heterogeneity of the water surface when predicting nitrogen concentrations. A new geostatistical method, called spatiotemporal point mean of surface with non-homogeneity (ST-PMSN), is proposed to model the non-stationary spatiotemporal random process of nitrogen concentrations between 2004 and 2013 in the YRE. The method considers the spatiotemporal correlation of surface water nitrogen and uses information from both sides of a boundary for heterogeneous water masses. Comparing with several other interpolating methods, including spatial ordinary kriging (OK), stratified ordinary kriging (SOK), point mean of surface with non-homogeneity (P-MSN), spatiotemporal ordinary kriging (STK), and stratified spatiotemporal ordinary kriging (SSTK), the cross-validation results show that ST-PMSN has the highest accuracy, followed by SSTK, STK, P-MSN, SOK, and OK in descending order. ST-PMSN is therefore demonstrated to be effective in estimating the nitrogen pollutant concentrations in a stratified estuary. According to interpolated nitrogen concentrations in the YRE, water quality has generally deteriorated—with fluctuations—from 2004 to 2013. The average annual reduction in area of water quality of Grades I and II from 2004 to 2013 was 1.10%. At the same time, the average annual increase in area of water quality of Grades III and IV was 0.89% and that of Grade V was 0.21%. The results of this study provide a new and more accurate interpolating method for assessing the pollutant concentration in the marine and offers guidance for more precise classification of water quality in the YRE.

Heavy metal soil contamination is associated with potential toxicity to humans or ecotoxicity. Scholars have increasingly used a combination of geographical information science (GIS) with geostatistical and multivariate statistical analysis techniques to examine the spatial distribution of heavy metals in soils at a regional scale. A review of such studies showed that most soil sampling programs were based on grid patterns and composite sampling methodologies. Many programs intended to characterize various soil types and land use types. The most often used sampling depth intervals were 0–0.10 m, or 0–0.20 m, below surface; and the sampling densities used ranged from 0.0004 to 6.1 samples per km², with a median of 0.4 samples per km². The most widely used spatial interpolators were inverse distance weighted interpolation and ordinary kriging; and the most often used multivariate statistical analysis techniques were principal component analysis and cluster analysis. The review also identified several determining and correlating factors in heavy metal distribution in soils, including soil type, soil pH, soil organic matter, land use type, Fe, Al, and heavy metal concentrations. The major natural and anthropogenic sources of heavy metals were found to derive from lithogenic origin, roadway and transportation, atmospheric deposition, wastewater and runoff from industrial and mining facilities, fertilizer application, livestock manure, and sewage sludge. This review argues that the full potential of integrated GIS and multivariate statistical analysis for assessing heavy metal distribution in soils on a regional scale has not yet been fully realized. It is proposed that future research be conducted to map multivariate results in GIS to pinpoint specific anthropogenic sources, to analyze temporal trends in addition to spatial patterns, to optimize modeling parameters, and to expand the use of different multivariate analysis tools beyond principal component analysis (PCA) and cluster analysis (CA).

Accurate mapping of air pollutants is essential for epidemiological studies and environmental risk assessments. Concentrations measured by air quality monitoring stations (AQMS) have primarily been used to assess the exposure of PM₂.₅. However, the low coverage and amount of monitoring stations affect the errors of spatial interpolation or geostatistical estimates. In contrast to other integrated approaches developed for improved air pollution estimates, this study utilizes data from low-cost microsensors densely deployed in Taiwan to improve the popular spatial interpolation approach called inverse distance weighting (IDW). A large dataset from thousands of low-cost sensors could improve spatial interpolation by describing the distribution of PM₂.₅ in detail. Therefore, this study presents a clustering-based method to assess the distribution of PM₂.₅. Then, a smarter IDW is performed based on correlated observations from the selected air quality stations. The publicly available data chosen for this investigation pertained to Taiwan, which has deployed 74 monitoring stations and more than 11,000 low-cost sensors since December 2020. The results of leave-one-out cross-validation indicate that there are fewer PM₂.₅ estimation errors in the developed approach than in estimations that use kriging across almost all of the months and sampled dates of 2019 and 2020, particularly those with higher PM₂.₅ spatial heterogeneities. Spatial heterogeneities could result in more significant estimation errors in mainstream approaches. The root mean square error of the monthly average estimate for PM₂.₅ ranged from 1.17 to 3.86 μg/m³. We also found that the clustering of one month characterizing the pattern of PM₂.₅ distribution could perform well in spatial interpolations based on historical data from monitoring stations. According to the information on the openaq platform, low-cost sensors are in demand in cities and areas. This trend might pave the way for the application of the proposed approach in other areas for superior exposure assessments.

Due to rapid urbanization in China, lead (Pb) continues to accumulate in urban topsoil, resulting in soil degradation and increased public exposure. Mapping Pb concentrations in urban topsoil is therefore vital for the evaluation and control of this exposure risk. This study developed spatial models to map Pb concentrations in urban topsoil using proximal and remote sensing data. Proximal sensing reflectance spectra (350–2500 nm) of soils were pre-processed and used to calculate the principal components as landscape factors to represent the soil properties. Other landscape factors, including vegetation and land-use factors, were extracted from time-sequential Landsat images. Two hybrid statistical approaches, regression kriging (RK) and geographically weighted regression (GWR), were adopted to establish prediction models using the landscape factors. The results indicated that the use of landscape factors derived from combined remote and proximal sensing data improved the prediction of Pb concentrations compared with useing these data individually. GWR obtained better results than RK for predicting soil Pb concentration. Thus, joint proximal and remote sensing provides timely, easily accessible, and suitable data for extracting landscape factors.

Mapping soil contamination enables the delineation of areas where protection measures are needed. Traditional soil sampling on a grid pattern followed by chemical analysis and geostatistical interpolation methods (GIMs), such as Kriging interpolation, can be costly, slow and not well-suited to highly heterogeneous soil environments. Here we propose a novel method to map soil contamination by combining high-resolution aerial imaging (HRAI) with machine learning algorithms. To support model establishment and validation, 1068 soil samples were collected from an arsenic (As) contaminated area in Zhongxiang, Hubei province, China. The average arsenic concentration was 39.88 mg/kg (SD = 213.70 mg/kg), with individual sample points determined as low risk (66.9%), medium risk (29.4%), or high risk (3.7%), respectively. Then, identified features were extracted from a HRAI image of the study area. Four machine learning algorithms were developed to predict As risk levels, including (i) support vector machine (SVM), (ii) multi-layer perceptron (MLP), (iii) random forest (RF), and (iii) extreme random forest (ERF). Among these, we found that the ERF algorithm performed best overall and that its prediction performance was generally better than that of traditional Kriging interpolation. The accuracy of ERF in test area 1 reached 0.87, performing better than RF (0.81), MLP (0.78) and SVM (0.77). The F1-score of ERF for discerning high-risk points in test area 1 was as high as 0.8. The complexity of the distribution of points with different risk levels was a decisive factor in model prediction ability. Identified features in the study area associated with fertilizer factories had the most important contribution to the ERF model. This study demonstrates that HRAI combined with machine learning has good potential to predict As soil risk levels.

Exposure to ambient air pollution is associated with maternal and child health. Some air pollutants exhibit similar behavior in the atmosphere, and some interact with each other; thus, comprehensive assessments of individual air pollutants are required. In this study, we developed national-scale monthly models for six air pollutants (NO, NO₂, SO₂, O₃, PM₂.₅, and suspended particulate matter (SPM)) to obtain accurate estimates of pollutant concentrations at 1 km × 1 km resolution from 2010 through 2015 for application to the Japan Environment and Children's Study (JECS), which is a large-scale birth cohort study. We developed our models in the land use regression framework using random forests in conjunction with kriging. We evaluated the model performance via 5-fold location-based cross-validation. We successfully predicted monthly NO (r² = 0.65), NO₂ (r² = 0.84), O₃ (r² = 0.86), PM₂.₅ (r² = 0.79), and SPM (r² = 0.64) concentrations. For SO₂, a satisfactory model could not be developed (r² = 0.45) because of the low SO₂ concentrations in Japan. The performance of our models is comparable to those reported in previous studies at similar temporal and spatial scales. The model predictions in conjunction with the JECS will reveal the critical windows of prenatal and infancy exposure to ambient air pollutants, thus contributing to the development of environmental policies on air pollution.