Large global variations in measured airborne metal concentrations driven by anthropogenic sources

Many regions globally exceed World Health Organization (WHO) air quality guidelines for PM2.5, significantly impacting human health. The Global Burden of Disease attributed 3 million deaths and 80 million years of lost healthy life to outdoor PM2.5 exposure in 2017. Ground-based monitoring of PM2.5 mass concentration and chemical composition is inadequate. Global observations of PM2.5 mass and composition are crucial for aerosol model development, exposure assessment, source understanding, and policy prioritization to reduce health impacts. The relationship between PM2.5 and human health is increasingly understood, linking it to cardiovascular and respiratory diseases, cancer, and type 2 diabetes. However, more analysis is needed to understand the effects of specific components, particularly trace metals, globally. The oxidative potential of PM2.5 is linked to its metal content, with increased redox-active elements inducing oxidative stress. Elevated concentrations of certain metals are associated with increased cardiovascular disease and mortality risks. Ground-based elemental composition data informs PM2.5 source identification. For example, potassium (K) is associated with wood burning, zinc (Zn) with traffic, and vanadium (V) with heavy fuel oil combustion. Coal and non-ferrous metal production are sources of multiple harmful elements. Vehicle traffic contributes a mix of elements, including heavy metals and crustal components. To the researchers' knowledge, no other global network has measured trace metal concentrations in PM2.5 at this scale. These observations are crucial for understanding particulate matter sources, assessing emission inventories, evaluating chemical transport model predictions, and understanding local and regional impacts of emission sources. The Surface PARTiculate matter Network (SPARTAN) provides new data to evaluate PM composition. This study investigates the trace metal composition of PM2.5, supplemented by PM10-2.5 and PM10 data, from SPARTAN sites worldwide, focusing on PM2.5 due to its significance for health.

The introduction extensively reviews existing literature on the health impacts of PM2.5, the limitations of current monitoring efforts, and the sources of various trace metals in atmospheric particulate matter. The review highlights the need for a globally consistent dataset on PM2.5 metal composition to improve exposure assessments, refine air quality models, and inform effective mitigation strategies. Specific studies cited include those on the global burden of disease attributable to air pollution, the relationship between PM2.5 and various health outcomes, the oxidative potential of PM2.5 and its metal content, and the sources of various trace metals.

The study utilized PM2.5 filter samples (N=800) collected from 19 globally distributed sites within the SPARTAN network between January 2013 and April 2019. The SPARTAN network prioritizes under-sampled locations, including megacities and developing regions, to address data gaps in exposure assessment. Site selection criteria emphasized representative environments avoiding anomalous sources, using low rooftops in urban areas. PM10 and PM2.5 filter masses were collected using two-stage stacked filter units within AirPhoton automated air samplers over 9-day periods. Starting in late 2017, the sampling stations were upgraded to AirPhoton SS5 models, using cyclone inlets to separate particles. The new setup used cartridges with eight Teflon filters for PM2.5 and PM10 sampling, including travelling blanks. Cartridges were pre-assembled and shipped to sites for installation by operators. Sampling protocols varied slightly depending on the site and project (e.g., MAPLE project in Canada). International sites sampled each PM2.5 filter for rotating 3-hour spans over 9 days (24h total), aiming for long-term average capture. The PM10 filter was sampled for a 30-minute period after each PM2.5 sample. MAPLE sites sampled for 48h total. After sampling, cartridges were returned to the Dalhousie University laboratory for analysis. Filters were analyzed for PM2.5/PM10 mass (gravimetrically), water-soluble ions (ion chromatography), black carbon (absorbance), and trace metals (inductively coupled plasma mass spectrometry – ICP-MS). For ICP-MS analysis, filters were extracted with 5% nitric acid. Field blank filter measurements were subtracted from sample measurements to account for trace metal baselines. Consistent ICP-MS analysis in the Dalhousie lab ensured consistency. Extraction efficiency was noted as low for some crustal elements (e.g., Fe), highlighting areas for improvement in future sampling protocols. Crustal enrichment factors (EFs) were calculated using two methods, one normalizing by Fe and the other by estimated coarse PM (PMc), to differentiate between natural and anthropogenic sources. Relative abundances (RAs) of trace metals were calculated by comparing concentrations at each site to those at Mammoth Cave National Park, a site with low trace metal concentrations. The study also compared PM2.5 and PM10 data to assess size-fraction contributions. A joint sampling campaign with the IMPROVE network was conducted in the US for independent data comparison.

The study's key findings include: 1. **High PM2.5 concentrations:** Kanpur had the highest mean PM2.5 concentration (102.8 µg/m³), followed by Beijing, Dhaka, and Hanoi. Sites in urban areas outside North America generally had higher PM2.5 levels, except for Mammoth Cave (due to biogenic sources). Canadian sites had the lowest levels. 2. **Elevated trace metal concentrations:** Significant spatial variation in trace metal concentrations was observed, with the highest concentrations in urban areas. Many elements (Pb, As, Cr, Zn) showed substantial enrichment compared to crustal concentrations (factors of 100-3000). Kanpur, Beijing, Dhaka, and Hanoi exhibited particularly high enrichments. 3. **Exceedances of health guidelines:** Lead (Pb) levels in Dhaka and Kanpur exceeded the US NAAQS 3-month guideline (150 ng/m³). Arsenic (As) concentrations in Kanpur, Hanoi, Beijing, and Dhaka approached or exceeded the WHO's 1:100,000 excess lifetime cancer risk level (6.6 ng/m³). 4. **Anthropogenic influence:** Crustal enrichment factors (EFs) demonstrated a strong anthropogenic contribution to trace metal concentrations, especially Pb, As, and Zn. EFs varied significantly depending on the element and location. Similar results using both Fe and PMc normalization validated the findings. 5. **Site-specific enrichments:** Relative abundance (RA) analysis relative to Mammoth Cave highlighted specific regional industries contributing to elevated trace metal levels. Bandung's high Pb was linked to lead smelting. Ilorin's high Cr was associated with the tanning industry. Singapore's high V was linked to shipping fuel and refineries. Beijing showed elevated Se due to coal combustion. Manila and Rehovot had relatively low trace metal concentrations. 6. **Correlation with population density:** A strong positive correlation was observed between population density and elevated trace metal levels, emphasizing the influence of anthropogenic activities.

The findings demonstrate the significant contribution of anthropogenic sources to globally varying airborne metal concentrations in PM2.5. The high levels of potentially toxic metals in densely populated areas, particularly in developing countries, raise serious public health concerns. The consistent sampling and analytical protocols employed by the SPARTAN network allow for a robust comparison of trace metal concentrations across diverse geographical locations and environments. The study's findings highlight the need for comprehensive monitoring efforts and effective air pollution control measures to minimize the adverse health impacts associated with exposure to these metals. The use of EFs and RAs provided valuable insights into the relative contributions of natural and anthropogenic sources. While some limitations in extraction efficiency were acknowledged, the overall conclusions remained robust due to the magnitude of observed variability. The study's focus on PM2.5 effectively addresses the most significant health concern related to particulate matter.

This study provides a globally consistent assessment of trace metal concentrations in PM2.5, highlighting significant anthropogenic influence and exceeding health guidelines in several megacities. The findings underscore the importance of ongoing monitoring and the need for effective mitigation strategies to protect public health. Future research should focus on refining sampling and analytical techniques, expanding the SPARTAN network geographically, and investigating specific source-receptor relationships to further refine our understanding of airborne metal pollution.

The study acknowledges some limitations in the analysis, specifically regarding the extraction efficiency of some crustal elements (like Fe) using nitric acid. Although this could underestimate the actual metal concentrations, the overall conclusions remain valid given the large differences observed in metal concentrations between sites. The study also acknowledges that the SPARTAN network does not yet have complete global coverage, with some regions remaining under-sampled. The temporal resolution is also limited by the 9-day sampling duration, which does not capture extreme short-term events. Future studies should explore the potential impacts of these limitations.