VOCs matter for public health for several reasons. Some members of the family, such as benzene, formaldehyde, and 1,3-butadiene, are recognized carcinogens or respiratory irritants, while others participate in atmospheric chemistry that generates ground-level ozone and secondary organic aerosol, both of which are linked to cardiovascular and respiratory harm. Because VOCs are emitted by many different kinds of sources, from gasoline stations and diesel trucks to paint shops, refineries, and chemical storage, the air in an industrial-adjacent neighborhood is a blended mixture in which no single concentration measurement can reveal responsibility. Source apportionment addresses this problem by using the relative pattern of many co-measured compounds as a diagnostic signature of each emission type.

The THRIVEair campaign grew out of longstanding community concern about air quality in neighborhoods close to Philadelphia’s industrial corridor. Fenceline communities, a term used for residential areas directly bordering large industrial operations, often experience elevated and highly variable pollutant concentrations depending on wind direction, facility operations, and time of day. Residents in such areas have historically lacked the dense, locally relevant monitoring data needed to demonstrate which sources dominate their exposure, a gap that can leave environmental agency decisions based on sparse regional averages rather than block-by-block reality. The study’s authors positioned THRIVEair as an effort to close that gap with sustained, neighborhood-scale measurement.

Methodologically, the research relied on time-resolved measurements of a wide suite of VOC species collected over an extended monitoring period at locations within the fenceline community. Analytical instruments captured compounds characteristic of different emission categories: aromatic hydrocarbons such as benzene, toluene, ethylbenzene, and xylenes, which trace gasoline combustion and solvent use; light alkanes and alkenes associated with natural gas, petrochemical operations, and vehicle exhaust; and chlorinated species that often indicate industrial solvent release or historical contamination. High-frequency sampling allowed the researchers to resolve short-term plumes and diurnal cycles that would be invisible to 24-hour integrated canister sampling alone.

The core of the analysis was receptor-based source apportionment, most commonly implemented through positive matrix factorization, or PMF, a statistical technique that takes the time series of many measured species and decomposes it into a small number of factors, each representing a distinct source profile with its own chemical fingerprint and temporal behavior. Rather than requiring an emissions inventory in advance, PMF lets the data themselves reveal how many source types are present and how much each contributes to the measured concentrations at the receptor location. The stability and interpretability of the resolved factors depend on the number and quality of the measured species, the frequency of sampling, and careful uncertainty estimation, all of which the study addressed in its design.

Interpreting the resolved factors typically involves cross-checking their chemical profiles and temporal patterns against known local activity. A traffic factor, for example, tends to peak during morning and evening rush hours and to be enriched in benzene and lighter aromatics, while an industrial or petrochemical factor may show a different compound ratio pattern and correlate with winds arriving from the direction of specific facilities. Meteorological data, including wind speed and direction, are usually incorporated to test whether factor contributions align with plausible source locations. This triangulation of chemistry, timing, and wind direction is what transforms a statistical factor into a defensible attribution of pollution to a source category.

The study’s findings carry significance both locally and methodologically. Locally, quantifying the share of VOC exposure attributable to industrial sources versus mobile sources versus regional background gives community members, public health officials, and regulators a factual basis for prioritizing interventions. If a substantial fraction of carcinogenic VOC exposure traces to a small number of industrial source categories, then targeted emission controls, fenceline monitoring requirements, or operational changes at specific facilities become evidence-backed priorities. Conversely, if traffic dominates, the intervention levers shift toward transportation policy, fleet electrification, and street-level exposure management. The apportionment results therefore function as a decision map rather than a mere description.

Methodologically, the work adds to a growing body of literature demonstrating that community-scale monitoring paired with receptor modeling can resolve source contributions that regional networks average away. Traditional regulatory monitoring in the United States relies on a limited number of sites, often sited to represent broad urban backgrounds, which systematically underestimates the exposure of people living immediately adjacent to emission sources. Studies like THRIVEair illustrate how denser, community-led or community-partnered measurement can capture the plume dynamics, wind-driven variability, and compound-specific signatures that define fenceline exposure. This approach aligns with a broader movement in environmental health toward citizen-science-informed monitoring and environmental justice screening tools that identify communities bearing disproportionate pollution burdens.

The environmental justice dimension is central to the study’s framing. Communities of color and lower-income neighborhoods in many American cities are disproportionately located near industrial zoning, freight corridors, and port facilities, and Philadelphia is no exception. Documenting elevated or source-attributable VOC concentrations in such neighborhoods provides quantitative support for the lived experience of residents who have long reported odors, health symptoms, and industrial incidents that went unmeasured by official networks. Source apportionment strengthens this documentation because it links measured exposure to identifiable emission categories, making it harder for the contribution of specific activities to be dismissed as background noise.

For the broader scientific community, the THRIVEair results contribute to the ongoing refinement of VOC source profiles in a modern urban environment. Emission compositions change over time as vehicle fleets evolve, fuel formulations shift, natural gas infrastructure ages, and industrial processes modernize, meaning that source profiles derived from studies conducted a decade or more ago may no longer represent current conditions. Fresh, locally derived apportionment results help update the emission inventories and chemical transport model inputs that underpin air quality forecasting, health risk assessment, and regulatory modeling. They also provide benchmarks against which future measurements can be compared to evaluate whether interventions are actually reducing the targeted source contributions.

The study also highlights practical considerations for communities elsewhere that want to understand their own air quality. Effective fenceline apportionment requires sustained funding for instruments and analysis, careful site selection to capture both source-influenced and background-influenced air, quality assurance protocols that withstand scientific and legal scrutiny, and genuine partnership with residents so that monitoring reflects local priorities and knowledge. The THRIVEair model, in which measurement campaigns are designed around community questions and results are translated into actionable findings, offers a template that other fenceline communities near refineries, chemical plants, ports, and freight hubs could adapt.

Ultimately, the research transforms an abstract complaint about industrial air into a quantified, compound-by-compound accounting of who contributes what to the air a fenceline community breathes. By resolving the mixture of volatile organic compounds into its constituent sources, the THRIVEair study gives Philadelphia residents, health officials, and regulators a shared factual foundation, and it demonstrates that modern exposure science can deliver the neighborhood-scale evidence that environmental justice demands. As cities nationwide grapple with legacy industrial zoning and expanding freight activity, the study stands as an example of how targeted monitoring and rigorous source apportionment can turn ambient air data into leverage for public health protection.

Subject of Research: Source apportionment of volatile organic compounds in a Philadelphia fenceline community using the THRIVEair monitoring campaign

Article Title: Source apportionment of volatile organic compounds in a Philadelphia fenceline community: results from THRIVEair

Article References: Frueh, L., Moore, K., Tiegs, G., Wahl, K., Johnston, L., Clougherty, J. E., Johnston, N. A. C., & Tripathy, S. (2026). Source apportionment of volatile organic compounds in a Philadelphia fenceline community: results from THRIVEair. Journal of Exposure Science & Environmental Epidemiology. https://doi.org/10.1038/s41370-026-00976-2

Image Credits: AI Generated

DOI: 10.1038/s41370-026-00976-2

Keywords: volatile organic compounds, source apportionment, fenceline community, Philadelphia, air quality monitoring, environmental justice, positive matrix factorization, exposure science, industrial emissions, community health, THRIVEair, benzene

Lake Victoria sustains approximately 42 million people through fisheries, drinking water and agriculture, making the ecological trajectory of the basin a matter of profound regional consequence. Yet the lake has undergone substantial ecological change over the last century under the combined pressures of fishing intensity, land-use transformation and catchment degradation. The Winam Gulf, which receives discharge from five major river systems, has long been identified as a hotspot of land degradation. As early as 2006, the World Agroforestry Centre highlighted severe erosion and sediment delivery from surrounding catchments, warning that urgent action was needed to avert flooding and ecological harm. Despite community engagement programmes and policy initiatives in the intervening years, the new sediment record suggests those warnings went largely unheeded, with the most rapid degradation occurring in the past two decades.

The technical heart of the study is its dating framework. Sediment cores were collected at the mouths of the Nyando, Sondu-Miriu, Awach, Luanda and Kisat rivers, along a transect across the Nyando sediment plume, and from a reference site in the centre of the gulf. Chronologies were built using unsupported lead-210, calculated by subtracting supported lead-210, inferred from lead-214 activity under the assumption of secular equilibrium in the uranium-238 decay series. Ages and dry mass sedimentation rates followed the Constant Rate of Supply model using cumulative unsupported lead-210 inventories. Caesium-137, often used as an independent chronological marker of peak atmospheric fallout in 1963, proved unusable in these equatorial sediments because activities fell below detection limits, a common limitation in East African records owing to low fallout deposition and radioactive decay since peak weapons testing.

The resulting chronologies reveal a consistent inflection point in the 1960s, a period the authors associate with the transition from colonial to independent governance and accelerating land clearance for agriculture. In the Nyando system, sedimentation rose from roughly 0.1 grams per square centimetre per year in the 1960s to a peak of 0.717 grams in 2018, standing at 0.638 grams in 2021, an approximately 700 percent increase. The Sondu-Miriu climbed from 0.122 to 0.382 grams per square centimetre per year, peaking at 0.682 in 2007, coinciding with the commissioning of the first Sondu-Miriu hydroelectric power station. The smaller Luanda and Kisat rivers, which drain areas around the city of Kisumu, recorded 400 to 500 percent increases over the same decades, a trajectory the researchers link in part to the expansion of Kisumu’s urban footprint from 19 to 103 square kilometres between 1969 and 2019.

The geochemical record adds a second dimension to the story. Concentrations of phosphorus, sulphur, calcium and organic matter, estimated by loss-on-ignition at 450 degrees Celsius and measured by triple quadrupole ICP-MS after mixed-acid digestion, remained stable until the 1960s, dipped by 20 to 30 percent, and then surged after 1990, with phosphorus and sulphur nearly doubling within a decade in the Nyando core. Calcium showed a distinct step change around 2005, rising about 25 percent within a single year. Across nearly all cores, rare earth element distributions remained consistent, suggesting a largely stable mineral source, while the shifting chemistry of surface-reactive elements points to increasing mobilisation of agriculturally influenced topsoil. In practical terms, the lake is not simply receiving more dirt; it is receiving more of the nutrient-rich, carbon-bearing surface material on which both farm productivity and aquatic ecosystems depend.

To determine where this sediment was coming from, the team deployed a Frequentist sediment fingerprinting approach using the open-source FingerPro R package, version 2.0. A total of 318 composite riverbed sediment samples, each aggregated from eight to ten subsamples across the channel width, were collected from the Nyando, Sondu-Miriu and Awach catchments in a nested design. Conservative geochemical tracers were selected using a combination of conservativeness index, consensus ranking and consistent tracer selection criteria, and source contributions were quantified with linear variability propagation, a method designed to avoid the biases that non-linear mixing functions introduce under high source variability. The result is a time-resolved provenance reconstruction stretching from 1960 to 2020, effectively turning each dated core layer into a snapshot of catchment sediment supply.

The apportionment results expose how sharply sediment sources can shift in response to land-use change. In the Nyando catchment, contributions from the Nyando-Kipchorian sub-catchment rose rapidly between 2000 and 2005, while the historically erosion-prone Awach Kano and Nyaidho sub-catchment, previously the dominant source, declined in relative terms. Across the plume transect cores, the pattern appeared with a lag that lengthened with distance from the river mouth, tracing the progressive dispersal of material through the gulf. The timing aligns closely with satellite-derived land-cover data showing a 75 percent loss of woodland across the Nyando between 1985 and 2014, most of it between 1996 and 2000, alongside a 1022 percent expansion of urban area. The Tinderet Forest alone lost 26.6 square kilometres of tree cover between 2000 and 2020, a ten percent reduction concentrated within the very sub-catchments identified as rising sediment sources.

In the Sondu-Miriu and Awach systems, similar source shifts tell equally pointed stories. From 2005 onward, the Sondu-Miriu core recorded a growing proportional contribution from one sub-catchment, with the Yurith source trending toward 80 to 90 percent of delivered sediment by 2017, a pattern consistent with deforestation that has removed 32 percent of forest cover over six decades, most notably since 2000. In the Awach, the earlier 2000s sedimentation peak of 0.3 to 0.5 grams per square centimetre per year tracked rising inputs from the lower sub-catchments, where roads, settlements and farmland expanded, while the later 2010 to 2020 peak, reaching 0.9 grams, corresponded to intensifying pressure in the upper catchment, where steep slopes, higher rainfall, timber harvesting and agricultural conversion compound erosion risk. Independent modelling by other researchers reports sediment yield increases of 17 to 33 percent in Awach sub-catchments between 2018 and 2023, corroborating the core-based trends.

The implications reach well beyond academic curiosity. Accelerated sediment transfer represents the mobilisation of nutrient-rich topsoil, organic carbon and associated contaminants from productive landscapes into aquatic systems, threatening soil fertility upstream even as it drives eutrophication, altered nutrient cycling and declining fisheries habitat downstream. Previous work has estimated the cost of non-intervention against soil erosion at 390 million US dollars per year to Kenya’s economy, and the study’s stakeholder consultations identified fragmented governance, monitoring and environmental data as persistent barriers to effective management. The framework’s central promise is precision: rather than spreading scarce mitigation resources uniformly across entire basins, terracing, agroforestry, riparian buffer restoration and cover cropping can be concentrated in the specific sub-catchments shown by the sediment record to be disproportionate contributors. Combined with Earth observation, seasonal monitoring and erosion modelling, this integration of geochronology and fingerprinting offers a replicable template for adaptive catchment management, not only across the Lake Victoria Basin but in the many rapidly changing tropical catchments worldwide where the health of soil, water and food systems remains inseparably linked.

Subject of Research: Pb-210 geochronology and geochemical sediment source apportionment used to reconstruct historical land-to-lake sediment transfers in the Lake Victoria Basin

Article Title: Linking geochronology and source apportionment of sediments to understand land–lake transfers in the Lake Victoria Basin

Article References: Watts, M. J., Humphrey, O. S., Tuffield, L., Gowing, C., Marriott, A. L., Ongore, C. O., Isaboke, J., Osano, O., Blake, W. H., & Aura, C. M. (2026). Linking geochronology and source apportionment of sediments to understand land–lake transfers in the Lake Victoria Basin. Environmental Geochemistry and Health, 48(15), Article 597. https://doi.org/10.1007/s10653-026-03451-x

Image Credits: AI Generated

DOI: 10.1007/s10653-026-03451-x

Keywords: Lake Victoria, soil erosion, sediment cores, Pb-210 geochronology, sediment fingerprinting, source apportionment, Winam Gulf, Nyando catchment, deforestation, eutrophication, land-use change, catchment management