The scale of the problem is hard to overstate. Recent estimates attribute between 4.2 and 4.6 million premature deaths in 2021 to outdoor air pollution, and a significant share of the mortality linked to fine particulate matter is specifically tied to the combustion of fossil fuels. The relationship between exposure and risk is non-linear, with harmful effects observed even at low concentrations, and there is no safe threshold below which adverse health effects disappear for vulnerable groups such as children, the elderly, and people with pre-existing disease. Because these emissions are largely regulable, they represent one of the most actionable frontiers in public health policy.

Latin America’s intermediate cities sit at a particularly acute intersection of pressures. Argentina is one of the most urbanized nations on Earth, with 92 percent of its population living in urban areas and a projected 21 percent urban growth between 2020 and 2050. Sustained growth of the vehicle fleet combined with fragmented urban planning has made road traffic one of the main sources of environmental degradation in these cities, while the short-term adoption of zero-emission technologies such as electric mobility remains out of reach. Yet unlike large metropolises, these urban areas frequently lack baseline environmental information, and the scarcity of high-resolution diagnostics limits environmental governance and hinders effective sustainable mobility policy.

The technical challenge is deceptively specific. Emissions models such as the International Vehicle Emissions (IVE) model, developed by the University of California at Riverside and partner institutions, can estimate total daily emissions from a registered fleet with reasonable accuracy, drawing on chassis dynamometer tests, real-world driving measurements from portable emissions measurement systems, and local inputs like fuel composition, ambient temperature, humidity, and altitude. But the IVE model produces no spatial distribution of the emissions it accounts for. Allocating that inventory across an entire road network normally requires exhaustive vehicle flow models or continuous traffic counts for every street section, information that growing intermediate cities rarely possess once the initial inventory stage is complete.

The research team, led by Beatriz Soledad Sosa of the Center for Environmental Research and Studies at the National University of the Center of the Province of Buenos Aires, linked three sequential computational models to bridge that gap. The first stage produced a bottom-up emissions inventory for Tandil’s 54,305 private vehicles using the IVE model, fed by primary data collected on the ground: vehicle counting campaigns with an intelligent traffic camera across arterial, highway, and residential hierarchies in three socioeconomic strata, driving pattern surveys recorded with GPSLogger smartphone applications on seven circuits, and a survey of daily vehicle starts. Meteorological data from the Aero Tandil station completed the local inputs. The second stage characterized the road network through Geographic Information Systems, calculating commercial density within linear buffers, redistributing census tract population proportionally along road segments, and counting educational establishments within 500-meter buffers, all standardized per 100 meters of segment.

The third stage is where the innovation becomes most visible. The team adapted DROVE, an open-source algorithm written in R that disaggregates on-road vehicle emission inventories spatially and temporally, originally developed by researchers working in the Colombian city of Manizales. DROVE offers three disaggregation options of increasing informational demand, and the researchers worked with Option 2, which weights emissions by road type using a factor for each vehicle category. Their intervention redefined the road type variable inside the algorithm’s data loading matrix so that it could accept as many road hierarchies as the user requires, and externalized the weighting factors into a simple text file, making the tool dramatically more flexible for other cities with different urban structures.

The classification step proved to be the methodological heart of the project. Applying k-means non-hierarchical clustering to the three quantitative urban densities, validated with the elbow method and Pseudo F-statistic, the researchers identified three behavioral groups: 4,141 low-intensity segments dominated by quiet residential streets, 236 segments of highest intensity and specialization forming commercial and educational corridors, and 2,007 intermediate or mixed segments serving as transition zones. Crossing these clusters with the five functional road hierarchies yielded 10 significant road classes, from high-density avenues to route collectors. Population density showed an R-squared of 0.70 across the partition and educational establishment density 0.63, indicating that the typology captures the population and educational variation of the city more strongly than the commercial one.

The results for particulate matter under 10 micrometers (PM10) demonstrate why the extra effort matters. When emissions were disaggregated using the standard three road categories, the model treated quiet residential streets and dense commercial corridors as equivalent. Moving to the 10 behavioral categories increased the standard deviation of cell-level emissions by up to 49 percent and more than doubled the maximum values at every grid resolution tested, proving that the model now captures the functional heterogeneity of the city. At the finest 100-meter resolution, the mean cell value was 2.05 grams per day with a maximum of 589.94 grams, while the coarser 500-meter grid concentrated mass into fewer cells and generated what the authors describe as a false alarm of saturation in isolated cells. The 100-meter grid, by contrast, precisely identified the continuity of critical corridors, which is essential for designing effective environmental zoning plans.

Perhaps the most striking finding is a matter of efficiency: 49.7 percent of total PM10 emissions occur in only 5.63 percent of the analyzed road segments. The high-intensity corridors and the intermediate or mixed areas, which together account for over half of daily emissions while representing less than a third of segments, emerge as clear priorities for intervention. The temporal analysis added a further surprise. Comparing the morning off-peak period with the evening peak, emissions rose by a constant 17 percent, far less than the 61.2 percent increase in distances traveled between the two blocks. Because travel distances in intermediate cities are shorter, the conventional notion of peak hours typical of megacities must be reconsidered: the pollutant load remains structural and sustained throughout the day, so management should focus on the spatial location of schools and commercial centers rather than sweeping temporal restrictions.

The implications reach well beyond one Argentine city. The structural axes identified in Tandil concentrate infrastructure that attracts vulnerable populations, including schools and healthcare centers, exposing children and the elderly to fine particulate loads, and the high-resolution diagnostic enables precise siting of air quality monitoring equipment, more realistic dispersion modeling, and more reliable health exposure analysis. The authors frame the contribution within Sustainable Development Goal 11, Sustainable Cities and Communities, arguing that smart transport management must encompass small and intermediate cities because they represent a significant share of the global population. With the adapted DROVE algorithm dispensing with the massive traffic data that constrains resource-limited contexts, cities can now prioritize infrastructure improvements, optimize traffic flows, plan pedestrian zones and bike lanes, or schedule time windows for street pedestrianization in a strategic and localized manner, turning a lack of baseline information from an excuse into a solvable technical problem.

Subject of Research: Environmental Science

Article Title: Computational model integration for high-resolution spatial emission simulation: a tool for urban transport management in resilient cities

Article References: Sosa, B. S., Cisneros Basualdo, N. E., Naveyra, J., Linares, S., & Porta, A. (2026). Computational model integration for high-resolution spatial emission simulation: a tool for urban transport management in resilient cities. Frontiers in Future Transportation, 7, Article 1873329. https://doi.org/10.3389/ffutr.2026.1873329

Image Credits: AI Generated

DOI: 10.3389/ffutr.2026.1873329

Keywords: air pollution, vehicle emissions, urban mobility, PM10, GIS, DROVE algorithm, intermediate cities, sustainable transport, Argentina, SDG 11, Computational, model