PM₂.₅ (fine particulate matter with aerodynamics diameter <2.5 μm) is the most important component of air pollutants, and has a significant impact on the atmospheric environment and human health. Using satellite remote sensing aerosol optical depth (AOD) to explore the hourly ground PM₂.₅ distribution is very helpful for PM₂.₅ pollution control. In this study, Himawari-8 AOD, meteorological factors, geographic information, and a new deep forest model were used to construct an AOD-PM₂.₅ estimation model in China. Hourly cross-validation results indicated that estimated PM₂.₅ values were consistent with the site observation values, with an R² range of 0.82–0.91 and root mean square error (RMSE) of 8.79–14.72 μg/m³, among which the model performance reached the optimum value between 13:00 and 15:00 Beijing time (R² > 0.9). Analysis of the correlation coefficient between important features and PM₂.₅ showed that the model performance was related to AOD and affected by meteorological factors, particularly the boundary layer height. Deep forest can detect diurnal variations in pollutant concentrations, which were higher in the morning, peaked at 10:00–11:00, and then began to decline. High-resolution PM₂.₅ concentrations derived from the deep forest model revealed that some cities in China are seriously polluted, such as Xi ‘an, Wuhan, and Chengdu. Our model can also capture the direction of PM₂.₅, which conforms to the wind field. The results indicated that due to the combined effect of wind and mountains, some areas in China experience PM₂.₅ pollution accumulation during spring and winter. We need to be vigilant because these areas with high PM₂.₅ concentrations typically occur near cities.

Forest understory plant communities in the eastern United States are often diverse and are potentially sensitive to changes in climate and atmospheric inputs of nitrogen caused by air pollution. In recent years, empirical and processed-based mathematical models have been developed to investigate such changes in plant communities. In the study reported here, a robust set of understory vegetation response functions (expressed as version 2 of the Probability of Occurrence of Plant Species model for the United States [US-PROPS v2]) was developed based on observations of forest understory and grassland plant species presence/absence and associated abiotic characteristics derived from spatial datasets. Improvements to the US-PROPS model, relative to version 1, were mostly focused on inclusion of additional input data, development of custom species-level input datasets, and implementation of methods to address uncertainty. We investigated the application of US-PROPS v2 to evaluate the potential impacts of atmospheric nitrogen (N) and sulfur (S) deposition, and climate change on forest ecosystems at three forested sites located in New Hampshire, Virginia, and Tennessee in the eastern United States. Species-level N and S critical loads (CLs) were determined under ambient deposition at all three modeled sites. The lowest species-level CLs of N deposition at each site were between 2 and 11 kg N/ha/yr. Similarly, the lowest CLs of S deposition, based on the predicted soil pH response, were less than 2 kg S/ha/yr among the three sites. Critical load exceedance was found at all three model sites. The New Hampshire site included the largest percentage of species in exceedance. Simulated warming air temperature typically resulted in lower maximum occurrence probability, which contributed to lower CLs of N and S deposition. The US-PROPS v2 model, together with the PROPS-CLF model to derive CL functions, can be used to develop site-specific CLs for understory plants within broad regions of the United States. This study demonstrates that species-level CLs of N and S deposition are spatially variable according to the climate, light availability, and soil characteristics at a given location. Although the species niche models generally performed well in predicting occurrence probability, there remains uncertainty with respect to the accuracy of reported CLs. As such, the specific CLs reported here should be considered as preliminary estimates.

Quantifying methane (CH₄) emissions from paddy fields is essential for evaluating the environmental risks of the paddy rice production system, and improving the accuracy of CH₄ modeling is a key issue that needs to be addressed. Based on a database containing 835 field measurements, both single national and region-specific models were established to estimate CH₄ emissions from paddy fields considering different environmental factors and management patterns using 70% of the measurements. The remaining 30% of the measurements were then used for model evaluation. The performance of the region-specific model was better than that of the single national model. The region-specific model could simulate CH₄ emissions in an unbiased manner with R² values of 0.15–0.70 and efficiency values of 11–60%. The paddy rice type, water regime, organic amendment, latitude, and soil characteristics (pH and bulk density) were identified as the main drivers in the models. By inputting the high-resolution spatial data of these drivers into the established model, the CH₄ emissions from China’s paddy fields were estimated to be 4.75 Tg in 2015, with a 95% confidence interval of 4.19–5.61 Tg. The results indicated that establishing and driving a region-specific model with high-resolution data can improve the estimation accuracy of CH₄ emissions from paddy fields.

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).

Identifying factors that influence the land surface temperature (LST) of urban forests can help improve simulations and predictions of spatial patterns of urban cool islands. This requires a quantitative analytical method that combines spatial statistical analysis with multi-source observational data. The purpose of this study was to reveal how human activities and ecological factors jointly influence LST in clustering regions (hot or cool spots) of urban forests. Using Xiamen City, China from 1996 to 2006 as a case study, we explored the interactions between human activities and ecological factors, as well as their influences on urban forest LST. Population density was selected as a proxy for human activity. We integrated multi-source data (forest inventory, digital elevation models (DEM), population, and remote sensing imagery) to develop a database on a unified urban scale. The driving mechanism of urban forest LST was revealed through a combination of multi-source spatial data and spatial statistical analysis of clustering regions. The results showed that the main factors contributing to urban forest LST were dominant tree species and elevation. The interactions between human activity and specific ecological factors linearly or nonlinearly increased LST in urban forests. Strong interactions between elevation and dominant species were generally observed and were prevalent in either hot or cold spots areas in different years. In conclusion, quantitative studies based on spatial statistics and GeogDetector models should be conducted in urban areas to reveal interactions between human activities, ecological factors, and LST.

Underlying levels of atmospheric pollutants, assumed to be governed by smoothing mechanisms due to atmospheric dispersion, can be estimated from global emissions source databases on greenhouse gases and ozone-depleting compounds. However, spatial data may be contaminated with noise or even missing or zero-valued at many locations. Therefore, a problem that arises is how to extract the underlying smooth levels. This paper sets out a structural spatial model that assumes data evolve across a global grid constrained by second-order smoothing restrictions. The frequency-domain approach is particularly suitable for global datasets, reduces the computational burden associated with two-dimensional models and avoids cumbersome zero-inflated skewed distributions. Confidence intervals of the underlying levels are also obtained. An application to the estimation of global levels of atmospheric pollutants from anthropogenic emissions illustrates the technique which may also be useful in the analysis of other environmental datasets of similar characteristics.

Concern regarding the impacts of continued nitrogen and sulfur deposition on ecosystem health has prompted the development of critical acid load assessments for forest soils. A critical acid load is a quantitative estimate of exposure to one or more pollutants at or above which harmful acidification-related effects on sensitive elements of the environment occur. A pollutant load in excess of a critical acid load is termed exceedance. This study combined a simple mass balance equation with national-scale databases to estimate critical acid load and exceedance for forest soils at a 1-km2 spatial resolution across the conterminous US. This study estimated that about 15% of US forest soils are in exceedance of their critical acid load by more than 250 eq ha-1 yr-1, including much of New England and West Virginia. Very few areas of exceedance were predicted in the western US. This simple mass balance equation estimated that 17% of US forest soils exceed their critical acid load by more than 250 eq ha-1 yr-1, and these areas are predominantly located in the northeastern US.

Atmospheric nitrogen dioxide (NO₂) is an important reactive gas pollutant harmful to human health. The spatiotemporal coverage provided by traditional NO₂ monitoring methods is insufficient, especially in the suburban and rural areas of north China, which have a high population density and experience severe air pollution. In this study, we implemented a spatiotemporal neural network (STNN) model to estimate surface NO₂ from multiple sources of information, which included satellite and in situ measurements as well as meteorological and geographical data. The STNN predicted NO₂ with high accuracy, with a coefficient of determination (R²) of 0.89 and a root mean squared error of 5.8 μg/m³ for sample-based 10-fold cross-validation. Based on the surface NO₂ concentration determined by the STNN, we analyzed the spatial distribution and temporal trends of NO₂ pollution in north China. We found substantial drops in surface NO₂ concentrations ranging between 9.1% and 33.2% for large cities during the 2020 COVID-19 lockdown when compared to those in 2019. Moreover, we estimated the all-cause deaths attributed to NO₂ exposure at a high spatial resolution of about 1 km, with totals of 6082, 4200, and 18,210 for Beijing, Tianjin, and Hebei Provinces in 2020, respectively. We observed remarkable regional differences in the health impacts due to NO₂ among urban, suburban, and rural areas. Generally, the STNN model could incorporate spatiotemporal neighboring information and infer surface NO₂ concentration with full coverage and high accuracy. Compared with machine learning regression techniques, STNN can effectively avoid model overfitting and simultaneously consider both spatial and temporal correlations of input variables using deep convolutional networks with residual blocks. The use of the proposed STNN model, as well as the surface NO₂ dataset, can benefit air quality monitoring, forecasting, and health burden assessments.

Urban tree canopies are a significant sink for atmospheric elemental carbon (EC)––an air pollutant that is a powerful climate-forcing agent and threat to human health. Understanding what controls EC deposition to urban trees is therefore important for evaluating the potential role of vegetation in air pollution mitigation strategies. We estimated wet, dry, and throughfall EC deposition for oak trees at 53 sites in Denton, TX. Spatial data and airborne discrete-return LiDAR were used to compute predictors of EC deposition, including urban form characteristics, and meteorologic and topographic factors. Dry and throughfall EC deposition varied 14-fold across this urban ecosystem and exhibited significant variability from spring to fall. Generalized additive modeling and multiple linear regression analyses showed that urban form strongly influenced tree-scale variability in dry EC deposition: traffic count as well as road length and building height within 100–150 m of trees were positively related to leaf-scale dry deposition. Rainfall amount and extreme wind-driven rain from the direction of major pollution sources were significant drivers of throughfall EC. Our findings indicate that complex configurations of roads, buildings, and vegetation produce “urban edge trees” that contribute to heterogeneous EC deposition patterns across urban systems, with implications for greenspace planning.

Particulate pollution is closely related to public health. PM₁ (particles with an aerodynamic size not larger than 1 μm) is much more harmful than particles with larger sizes because it goes deeper into the body and hence arouses social concern. However, the sparse and unevenly distributed ground-based observations limit the understanding of spatio-temporal distributions of PM₁ in China. In this study, hourly PM₁ concentrations in central and eastern China were retrieved based on a random forest model using hourly aerosol optical depth (AOD) from Himawari-8, meteorological and geographic information as inputs. Here the spatiotemporal autocorrelation of PM₁ was also considered in the model. Experimental results indicate that although the performance of the proposed model shows diurnal, seasonal and spatial variations, it is relatively better than others, with a determination coefficient (R²) of 0.83 calculated based on the 10-fold cross validation method. Geographical map implies that PM₁ pollution level in Beijing-Tianjin-Hebei is much higher than in other regions, with the mean value of ∼55 μg/m³. Based on the exposure analysis, we found about 75% of the population lives in an environment with PM₁ higher than 35 μg/m³ in the whole study area. The retrieval dataset in this study is of great significance for further exploring the impact of PM₁ on public health.