The impact of wildfires on air pollution and health across land use categories in Brazil over a 16-year period.

Forest fires cause many environmental impacts, including air pollution. Brazil is a very fire-prone region where few studies have investigated the impact of wildfires on air quality and health. We proposed to test two hypotheses in this study: i) the wildfires in Brazil have increased the levels of air pollution and posed a health hazard in 2003-2018, and ii) the magnitude of this phenomenon depends on the type of land use and land cover (e.g., forest area, agricultural area, etc.). Satellite and ensemble models derived data were used as input in our analyses. Wildfire events were retrieved from Fire Information for Resource Management System (FIRMS), provided by NASA; air pollution data from the Copernicus Atmosphere Monitoring Service (CAMS); meteorological variables from the ERA-Interim model; and land use/cover data were derived from pixel-based classification of Landsat satellite images by MapBiomas. We used a framework that infers the "wildfire penalty" by accounting for differences in linear pollutant annual trends (ß) between two models to test these hypotheses. The first model was adjusted for Wildfire-related Land Use activities (WLU), considered as an adjusted model. In the second model, defined as an unadjusted model, we removed the wildfire variable (WLU). Both models were controlled by meteorological variables. We used a generalized additive approach to fit these two models. To estimate mortality associated with wildfire penalties, we applied health impact function. Our findings suggest that wildfire events between 2003 and 2018 have increased the levels of air pollution and posed a significant health hazard in Brazil, supporting our first hypothesis. For example, in the Pampa biome, we estimated an annual wildfire penalty of 0.005 µg/m3 (95%CI: 0.001; 0.009) on PM2.5. Our results also confirm the second hypothesis. We observed that the greatest impact of wildfires on PM2.5 concentrations occurred in soybean areas in the Amazon biome. During the 16 years of the study period, wildfires originating from soybean areas in the Amazon biome were associated with a total penalty of 0.64 µg/m3 (95%CI: 0.32; 0.96) on PM2.5, causing an estimated 3872 (95%CI: 2560; 5168) excess deaths. Sugarcane crops were also a driver of deforestation-related wildfires in Brazil, mainly in Cerrado and Atlantic Forest biomes. Our findings suggest that between 2003 and 2018, fires originating from sugarcane crops were associated with a total penalty of 0.134 µg/m3 (95%CI: 0.037; 0.232) on PM2.5 in Atlantic Forest biome, resulting in an estimated 7600 (95%CI: 4400; 10,800) excess deaths during the study period, and 0.096 µg/m3 (95%CI: 0.048; 0.144) on PM2.5 in Cerrado biome, resulting in an estimated 1632 (95%CI: 1152; 2112) excess deaths during the study period. Considering that the wildfire penalties observed during our study period may continue to be a challenge in the future, this study should be of interest to policymakers to prepare future strategies related to forest protection, land use management, agricultural activities, environmental health, climate change, and sources of air pollution.

Trace metal dynamics in an industrialized Brazilian river: A combined application of Zn isotopes, geochemical partitioning, and multivariate statistics.

The Paraiba do Sul (PSR) and Guandu Rivers (GR) water diversion system (120 km long) is located in the main industrial pole of Brazil and supplies drinking water for 9.4 million people in the metropolitan region of Rio de Janeiro. This study aims to discern the trace metals dynamics in this complex aquatic system. We used a combined approach of geochemical tools such as geochemical partitioning, Zn isotopes signatures, and multivariate statistics. Zinc and Pb concentrations in Suspended Particulate Matter (SPM) and sediments were considerably higher in some sites. The sediment partition of As, Cr, and Cu revealed the residual fraction (F4) as the main fraction for these elements, indicating low mobility. Zinc and Pb were mostly associated with the exchangeable/carbonate (F1) and the reducible (F2) fractions, respectively, implying a higher susceptibility of these elements to being released from sediments. Zinc isotopic compositions of sediments and SPM fell in a binary mixing source process between lithogenic (δ66/64ZnJMC ≈ + 0.30‰) and anthropogenic (δ66/64ZnJMC ≈ + 0.15‰) end members. The lighter δ66/64ZnJMC values accompanied by high Zn concentrations in exchangeable/carbonate fraction (ZnF1) enable the tracking of Zn anthropogenic sources in the studied rivers. Overall, the results indicated that Hg, Pb, and Zn had a dominant anthropogenic origin linked to the industrial activities, while As, Cr, and Cu were mainly associated with lithogenic sources. This work demonstrates how integrating geochemical tools is valuable for assessing geochemical processes and mixing source effects in anthropized river watersheds.

Mapping distance-decay of cardiorespiratory disease risk related to neighborhood environments.

Neighborhood characteristics affect an individual's quality of life. Although several studies have examined the relationship between neighborhood environments and human health, we are unaware of studies that have examined the distance-decay of this effect and then presented the risk results spatially. Our study is unique in that is explores the health effects in a less developed country compared to most studies that have focused on developed countries. The objective of our study is to quantify the distance-decay cardiorespiratory diseases risk related to 28 neighborhood aspects in the Federal District, Brazil and present this information spatially through risk maps of the region. Toward this end, we used a quantile regression model to estimate risk and GIS modeling techniques to create risk maps. Our analysis produced the following findings: i) a 2500 m increase in highway length was associated with a 46% increase in cardiorespiratory diseases; ii) 46,000 light vehicles in circulation (considering a buffer of ≤500 m from residences) was associated with 6 hospital admissions (95% CI: 2.6, 14.6) per cardiorespiratory diseases; iii) 74,000 m2 of commercial areas (buffer ≤1700 m) was associated with 12 hospital admissions (95% CI: 2.2, 20.8); iv) 1km2 increase in green areas intra urban was associated with less two hospital admissions, and; vi) those who live ≤500 m from the nearest point of wildfire are more likely to have cardiorespiratory diseases that those living >500 m. Our findings suggest that the approach used in this study can be an option to improve the public health policies.

Association between vehicular emissions and cardiorespiratory disease risk in Brazil and its variation by spatial clustering of socio-economic factors.

Many studies have suggested that socio-economic factors are strong modifiers of human vulnerability to air pollution effects. Most of these studies were performed in developed countries, specifically in the US and Europe. Only a few studies have been performed in developing countries, and analyzed small regions (city level) with no spatial disaggregation. The aim of this study was to assess the association between vehicle emissions and cardiorespiratory disease risk in Brazil and its modification by spatial clustering of socio-economic conditions. We used a quantile regression model to estimate the risk and a geostatistical approach (K means) to execute spatial cluster analysis. We performed the risk analysis in three stages. First, we analyzed the entire study area (primary analysis), and then we conducted a spatial cluster analysis based on various municipal-level socio-economic factors, followed by a sensitivity analysis. We studied 5444 municipalities in Brazil between 2008 and 2012. Our findings showed a significant association between cardiorespiratory disease risk and vehicular emissions. We found that a 15% increase in air pollution is associated with a 6% increase in hospital admissions rates. The results from the spatial cluster analysis revealed two groups of municipalities with distinct sets of socio-economic factors and risk levels of cardiorespiratory disease related to exposure to vehicular emissions. For example, for vehicle emissions of PM in 2008, we found a relative risk of 4.18 (95% CI: 3.66, 4.93) in the primary analysis; in Group 1, the risk was 0.98 (95% CI: 0.10, 2.05) while in Group 2, the risk was 5.56 (95% CI: 4.46, 6.25). The risk in Group 2 was 480% higher than the risk in Group 1, and 35% higher than the risk in the primary analysis. Group 1 had higher values (3rd quartile) for urbanization rate, highway density, and GDP; very high values (≥3rd quartile) for population density; median values for distance from the capital; and lower values (1st quartile) for rural population density. Group 2 had lower values (1st quartile) urbanization rate; median values for highway density, GDP, and population density; between median and third quartile values for distance from the capital; and higher values (3rd quartile) for rural population density. Our findings suggest that socio-economic factors are important modifiers of the human risk of cardiorespiratory disease due to exposure to vehicle emissions in Brazil. Our study provides support for creating effective public policies related to environmental health that are targeted to high-risk populations.

Increased preterm birth following maternal wildfire smoke exposure in Brazil.

Preterm birth (PTB) complications are the leading cause of death among children under 5 years of age, responsible for approximately 1 million deaths in 2015, according to the World Health Organization. Those infants born prematurely who survived the first 5 years, studies suggest that these infants are more likely to experience a range of poor health outcomes during childhood and even adulthood. Wildfire smoke has been suggested as a type of air pollution source with high toxicity for reproductive health. In this study, we estimated the association between preterm birth and wildfire periods in Brazil, a country included in the list of the 10 nations with the greatest number of preterm birth and also considered as a very fire-prone region. We applied a time-stratified case-crossover study design using conditional logistic regression models to estimate the odds ratio for preterm birth associated with wildfire-related prenatal PM2.5, during different windows of exposure, including trimesters 1-3. After adjusting for several confounders (other air pollutants, demographics, meteorological variables, and spatiotemporal terms), we found that wildfire smoke exposure during pregnancy may be associated with preterm birth in Brazil. Southeast was the region with the highest increase in the odds of PTB (OR:1.41 (95%CI: 1.31-1.51) when the exposure occurred in the first trimester. In the North, exposure to PM2.5 during wildfire periods in the second trimester of pregnancy was associated with increased odds of PTB (OR:1.05 (95%CI: 1.01-1.09) in preterm birth when the exposure occurred in the second trimester. This study suggests that wildfire smoke exposure during pregnancy may increase the risk for preterm birth in Brazil. This should be of great concern to the public health authorities, obstetricians, and policymakers.

Schools exposure to air pollution sources in Brazil: A nationwide assessment of more than 180 thousand schools.

A growing body of evidence demonstrates that children at schools who are exposed to increased concentrations of air pollutants may have a higher risk for several health problems, including cognitive deficits. In this paper we estimate the exposure to air pollution sources at 186,080 schools in Brazil. Specifically, we accounted for the exposure to three proxies of air pollution source emissions, including distance to roadways, the extent of roadways within a buffer around each school, and the number of wildfire occurrences within a buffer around each school. About 25% of the Brazilian schools evaluated in our study are located within a distance ≤250 m of a major roadway, have ≥2 km of roadway within a buffer of 1 km, and have ≥7 wildfires records within a buffer of 10 km. Our results indicate significant prevalence ratio of these schools exposed to air pollution sources when we stratified the analyses by socioeconomic factors, including geographic (public schools had an increased likelihood of being exposed), economic (low-income areas had an increased likelihood of being exposed), health (overall, areas with low public health status had an increased likelihood of being exposed), and educational conditions (overall, areas with low educational indicator had an increased likelihood of being exposed). For example, we estimated that private schools were 15% (95% CI: 13-17%) less likely to be located within 250 m of a major roadway compared with public schools; schools in areas with low child mortality were 35% (95% CI: 34-37%) less likely to be within 250 m of a major roadway; and schools in regions with low expected years of schooling were 25% (95% CI: 22-28%) more likely to be located within 250 m of a major roadway. The analysis of the spatial patterns shows that a substantial number of schools (36-54%, depending on the air pollution source) has a positive autocorrelation, suggesting that exposure level at these schools are similar to their neighbors. Estimating children's exposure to air pollutants at school is crucial for future public policies to develop effective environmental, transportation, educational, and urban planning interventions that may protect students from exposure to environmental hazards and improve their safety, health, and learning performance.