The adverse health effects associated with the inhalation and ingestion of naturally occurring radon gas produced during the uranium decay chain mean that there is a need to identify high-risk areas. This study detected radon-prone areas using a geographic information system (GIS)-based probabilistic and machine learning methods, including the frequency ratio (FR) model and a convolutional neural network (CNN). Ten influencing factors, namely elevation, slope, the topographic wetness index (TWI), valley depth, fault density, lithology, and the average soil copper (Cu), calcium oxide (Cao), ferric oxide (Fe₂O₃), and lead (Pb) concentrations, were analyzed. In total, 27 rock samples with high activity concentration index values were divided randomly into training and validation datasets (70:30 ratio) to train the models. Areas were categorized as very high, high, moderate, low, and very low radon areas. According to the models, approximately 40% of the study area was classified as very high or high risk. Finally, the radon potential maps were validated using the area under the receiver operating characteristic curve (AUC) analysis. This showed that the CNN algorithm was superior to the FR method; for the former, AUC values of 0.844 and 0.840 were obtained using the training and validation datasets, respectively. However, both algorithms had high predictive power. Slope, lithology, and TWI were the best predictors of radon-affected areas. These results provide new information regarding the spatial distribution of radon, and could inform the development of new residential areas. Radon screening is important to reduce public exposure to high levels of naturally occurring radiation.

Nitrogen (N) is an essential nutrient but may become a pollution source in the environment when the N concentration exceeds a certain threshold for humans and nature. Nitrate is a major N species in river water with notable spatial and temporal variations under the influences of natural factors and anthropogenic N inputs. We analyzed the relationship between riverine N (focusing on nitrate) concentration and various factors (land use, climate, basin topography, atmospheric N deposition, agricultural N sources and human-derived N) in 104 rivers located throughout the Japanese Archipelago except small remote islands. We aimed to better understand processes and mechanisms to explain the spatial and temporal changes in riverine nitrate concentration. A publicly available river water quality database observed in the 1980s (1980–1989) and 2000s (2000–2009) was used. This study is the first to evaluate the long-term scale of 20 years in the latter half of Japan's economic growth period at the national level. A geographic information system (GIS) was employed to determine average values of each variable collected from multiple sources of statistical data. We then performed regression analysis and structural equation modeling (SEM) for each period. The forestland area influenced by the basin topography, climate (i.e., air temperature) and other land uses (i.e., farmland and urban area) played a major role in decreasing nitrate concentrations in both the 1980s and 2000s. Atmospheric N deposition (especially N oxides) and agricultural N sources (fertilizer and manure) were also significant variables regarding the spatial variations in riverine nitrate concentrations. The SEM results suggested that human-derived N (via food consumption) intensified by demographic shifts during the 2000s increased riverine nitrate concentrations over other variables within the context of spatial variation. These findings facilitate better decision making regarding land use, agricultural practices, pollution control and individual behaviors toward a sustainable society.

Shandong is the most populous and highly industrialized province in eastern China, and the resultant poor air quality is a cause for widespread concern. This study combines bottom–up and top–down approaches to develop a high-resolution anthropogenic emission inventory of air pollutants for 2017. The inventory was developed based on updated emission factors and detailed activity data. The emissions of sulfur dioxide (SO₂), nitrogen oxides (NOₓ), particulate matter with aerodynamic diameters smaller than 2.5 and 10 μm (PM₂.₅ and PM₁₀, respectively), carbon monoxide (CO), volatile organic compounds (VOCs), and ammonia (NH₃) were estimated to be 1387.8, 2488.6, 5281.7, 3193.0, 9250.7, 2254.7, and 1210.6 kt, respectively. Power plants were the largest contributors of SO₂ and NOₓ emissions accounting for 43.7% and 41.9% of the total emissions, respectively. CO emissions mainly originated from industrial processes (40.1%), mobile sources (24.8%), and fossil fuel burning (21.2%). The major sources of PM₁₀ and PM₂.₅ emissions were industrial processes and fugitive dust, contributing 83.0% and 86.9% of their total emissions, respectively. Industrial processes (60.0%) contributed the largest VOC emissions, followed by mobile sources (16.8%) and solvent use (14.5%). Livestock and N-fertilizers were major emitters of NH₃, accounting for 69.9% and 21.2% of the total emissions, respectively. Emissions were spatially allocated to grid cells with a resolution of 0.05 ° × 0.05 ° based on spatial surrogates, using Geographic Information System (GIS). Heavy pollutant emissions were mainly concentrated in the central and eastern areas of Shandong, while high NH₃–emissions occurred in the western region. Most pollutant emissions from industrial sectors occurred in June and July, while low emissions were recorded between January and February. Range uncertainties in emission inventory were quantified using Monte Carlo simulations. Our inventory provides effective information to understand local pollutant emission characteristics, perform air quality simulations, and formulate pollution control measures.

Ozone has become a major atmospheric pollutant in China as the pattern of urban energy usage has changed and the number of motor vehicles has grown rapidly. The Beijing-Tianjin-Hebei Urban Agglomeration, also known as the Jing-Jin-Ji Urban Agglomeration (hereafter, JJJUA), with a precarious balance between protecting the ecological environment and sustaining economic development, is challenged by high levels of ozone pollution. Based on ozone observation data from 13 cities in the JJJUA from 2014 to 2017, the spatio-temporal trends in the evolution of ozone pollution and its associated influencing factors were analyzed using Moran’s I Index, hot-spot analysis, and Geodetector using ArcGIS and SPSS software. Five key results were obtained. 1) There was an increase in the annual average ozone concentration, for the period 2014–2017. Comparing the 13 prefecture-level cities, ozone pollution in Chengde and Hengshui decreased, while it worsened in the remaining 11 cities. 2) Ozone pollution was worse in spring and summer than in autumn and winter; the peak ozone pollution season was from May to September; the average ozone concentration on workdays was higher than that on non-workdays, showing a counter-weekend effect. 3) Annual average concentrations were high in the central and southern parts of the study region but low in the north. 4) Prominent positive spatial correlations were observed in ozone concentration, with the best correlations shown in summer and autumn; concentrations were high in Baoding and Xingtai but low in Beijing and Chengde. 5) Concentrations of PM10, NO2, CO, SO2, and PM2.5, as well as average wind speed, sunshine duration, evaporation, precipitation, and temperature, all had significant effects on ozone pollution, and interactions between these influencing factors increased it.

Useful information on the potential radon risk in existing buildings can be obtained by combining data from sources such as potential risk maps, the ‘Sistema de Información sobre Ocupación del Suelo de España’ (SIOSE) [information system on land occupancy in Spain], cadastral data on built property and population surveys. The present study proposes a method for identifying urban land, premises and individuals potentially subject to radon risk. The procedure draws from geographic information systems (GIS) pooled at the municipal scale and data on buildings possibly affected. The method quantifies the magnitude of the problem in the form of indicators on the buildings, number of premises and gross floor area that may be affected in each risk category. The findings are classified by type of use: residential, educational or office. That information may guide health/prevention policies by targeting areas to be measured based on risk category, or protection policies geared to the construction industry by estimating the number of buildings in need of treatment or remediation. Application of the methodology to Greater Madrid showed that 47% of the municipalities have houses located in high radon risk areas. Using cadastral data to zoom in on those at highest risk yielded information on the floor area of the vulnerable (basement, ground and first storey) premises, which could then be compared to the total. In small towns, the area affected differed only scantly from the total, given the substantial proportion of low-rise buildings in such municipalities.

Pesticides move to surface water via various pathways including surface runoff, spray drift and subsurface flow. Little is known about the relative contributions of surface runoff and spray drift in agricultural watersheds. This study develops a modeling framework to address the contribution of spray drift to the total loadings of pesticides in receiving water bodies. The modeling framework consists of a GIS module for identifying drift potential, the AgDRIFT model for simulating spray drift, and the Soil and Water Assessment Tool (SWAT) for simulating various hydrological and landscape processes including surface runoff and transport of pesticides. The modeling framework was applied on the Orestimba Creek Watershed, California. Monitoring data collected from daily samples were used for model evaluation. Pesticide mass deposition on the Orestimba Creek ranged from 0.08 to 6.09% of applied mass. Monitoring data suggests that surface runoff was the major pathway for pesticide entering water bodies, accounting for 76% of the annual loading; the rest 24% from spray drift. The results from the modeling framework showed 81 and 19%, respectively, for runoff and spray drift. Spray drift contributed over half of the mass loading during summer months. The slightly lower spray drift contribution as predicted by the modeling framework was mainly due to SWAT's under-prediction of pesticide mass loading during summer and over-prediction of the loading during winter. Although model simulations were associated with various sources of uncertainties, the overall performance of the modeling framework was satisfactory as evaluated by multiple statistics: for simulation of daily flow, the Nash-Sutcliffe Efficiency Coefficient (NSE) ranged from 0.61 to 0.74 and the percent bias (PBIAS) < 28%; for daily pesticide loading, NSE = 0.18 and PBIAS = −1.6%. This modeling framework will be useful for assessing the relative exposure from pesticides related to spray drift and runoff in receiving waters and the design of management practices for mitigating pesticide exposure within a watershed.

Noise is associated with poor reproductive health. A number of animal studies have suggested the possible effects of exposure to high noise levels on fertility; to date, a little such research has been performed on humans.We examined an association between daytime and nocturnal noise exposures over four years (2002–2005) and subsequent male infertility.We used the National Health Insurance Service-National Sample Cohort (2002–2013), a population-wide health insurance claims dataset. A total of 206,492 males of reproductive age (20–59 years) with no history of congenital malformations were followed up for an 8-year period (2006–2013). Male infertility was defined as per ICD-10 code N46. Data on noise exposure was obtained from the National Noise Information System. Exposure levels of daytime and night time noise were extrapolated using geographic information systems and collated with the subjects' administrative district code, and individual exposure levels assigned.During the study period, 3293 (1.6%) had a diagnosis of infertility. Although there was no association of infertility with 1-dB increments in noise exposure, a non-linear dose-response relationship was observed between infertility and quartiles of daytime and night time noise after adjustment for confounding variables (i.e., age, income, residential area, exercise, smoking, alcohol drinking, blood sugar, body mass index, medical histories, and particulate pollution). Based on WHO criteria, adjusted odds for infertility were significantly increased (OR = 1.14; 95% CI, 1.05–1.23) in males exposed to night time noise ≥ 55 dB.We found a significant association between exposure to environmental noise for four years and the subsequent incidence of male infertility, suggesting long-term exposure to noise has a role in pathogenesis of male infertility.

Many regulatory screening level exposure assessments are based on simple large scale conceptual scenarios. However, exposure, and therefore risks associated with chemicals, are characterised by high spatial variability. The Scenario assembly tool (ScenAT) is a global screening level model to enable spatially resolved local predictions of environmental concentrations of home and personal care chemicals. It uses the European Union Technical Guidance Document (TGD) equation to predict local scale freshwater concentrations (predicted environmental concentrations - PECs) of chemicals discharged via wastewater. ScenAT uses Geographic Information System (GIS) layers for the underlying socio-economic (population) and environmental parameters (per capita water use, sewage treatment plant connectivity, dilution factor). Using a probabilistic approach, we incorporate sources of uncertainty in the input data (tonnage estimation, removal in sewage treatment plants and seasonal variability in dilution factors) for two case-study chemicals: the antimicrobial triclosan (TCS) and the anionic surfactant linear alkylbenzene sulphonate (LAS). We then compare model estimates of wastewater and freshwater concentrations of TCS and LAS to UK monitoring data. Comparison showed that modeled PECs were on average higher than mean measured data for TCS and LAS by a factor 1.8 and 1.4, respectively. Considering the uncertainty associated with both model and monitoring data, the use of a probabilistic approach using the ScenAT model for screening assessment is reasonable. The combination of modelled and monitoring data enables the contextualisation of monitoring data. Spatial PECs can be used to identify areas of elevated concentration for further refined assessment.A probabilistic approach for large scale screening assessments to contextualise monitoring data for risk assessment.

Previous studies investigating health conditions of individuals living near livestock farms generally assessed short time windows. We aimed to take time-specific differences into account and to compare the prevalence of various health conditions over seven consecutive years. The sample consisted of 156,690 individuals registered in 33 general practices in a (rural) area with a high livestock density and 101,015 patients from 23 practices in other (control) areas in the Netherlands. Prevalence of health conditions were assessed using 2007–2013 electronic health record (EHR) data. Two methods were employed to assess exposure: 1) Comparisons between the study and control areas in relation to health problems, 2) Use of individual estimates of livestock exposure (in the study area) based on Geographic Information System (GIS) data. A higher prevalence of chronic bronchitis/bronchiectasis, lower respiratory tract infections and vertiginous syndrome and lower prevalence of respiratory symptoms and emphysema/COPD was found in the study area compared with the control area. A shorter distance to the nearest farm was associated with a lower prevalence of upper respiratory tract infections, respiratory symptoms, asthma, COPD/emphysema, allergic rhinitis, depression, eczema, vertiginous syndrome, dizziness and gastrointestinal infections. Especially exposure to cattle was associated with less health conditions. Living within 500m of mink farms was associated with increased chronic enteritis/ulcerative colitis. Livestock-related exposures did not seem to be an environmental risk factor for the occurrence of health conditions. Nevertheless, lower respiratory tract infections, chronic bronchitis and vertiginous syndrome were more common in the area with a high livestock density. The association between exposure to minks and chronic enteritis/ulcerative colitis remains to be elucidated.

In this study, we used publicly available GIS layers and statistical tree-based modeling (CART and Random Forest) to predict pathogen indicator counts at a regional scale using 88 spatially explicit landscape predictors and 6657 samples from non-estuarine streams in the Oregon Coast Range. A total of 532 frequently sampled sites were parsed down to 93 pathogen sampling sites to control for spatial and temporal biases. This model's 56.5% explanation of variance, was comparable to other regional models, while still including a large number of variables. Analysis showed the most important predictors on bacteria counts to be: forest and natural riparian zones, cattle related activities, and urban land uses. This research confirmed linkages to anthropogenic activities, with the research prediction mapping showing increased bacteria counts in agricultural and urban land use areas and lower counts with more natural riparian conditions.