The Latin American and Caribbean (LAC) region faces persistent challenges of inequality, climate change vulnerability, and deteriorating air quality. The Aburrá Valley, where Medellín is located, is a narrow tropical valley with complex topography, strong thermal inversions, and unstable atmospheric conditions, all of which exacerbate the accumulation of pollutants. In Medellín, concentrations have remained nearly unchanged over the past eight years, consistently approaching critical thresholds, despite the implementation of air quality control strategies. These persistent high concentrations are closely linked to the variability of the atmospheric boundary layer (ABL) and are often intensified by prolonged dry periods. This study focuses on a representative street canyon in Medellín that has undergone recent urban interventions, including the construction of new public spaces and pedestrian areas, without explicitly considering their impact on dispersion. Using Computational Fluid Dynamics (CFD) simulations, this work evaluates the influence of urban morphology on accumulation. The results reveal that areas with high Aspect Ratios (AR > 0.65) and dense vegetation exhibit reduced wind speeds at the pedestrian level—up to 40% lower compared to open zones—and higher concentrations, with maximum simulated values exceeding 50 μg/m3. This study demonstrates that the design of pedestrian corridors in complex urban environments like Medellín can unintentionally create pollutant accumulation zones, underscoring the importance of integrating air quality considerations into urban planning. The findings provide actionable insights for policymakers, emphasizing the need for comprehensive modeling and field validation to ensure healthier urban spaces in cities affected by persistent air quality issues.
Urban bus stops represent critical micro-environments where pedestrians may be exposed to elevated concentrations of traffic-related air pollutants, particularly under low-wind conditions. This study presents a computational fluid dynamics (CFD) analysis of airflow patterns, nitrogen dioxide (NO2) dispersion, and exposure mitigation strategies within a representative urban bus stop in Medellín, Colombia. A three-dimensional model was developed in COMSOL Multiphysics v5.6, solving the Reynolds-Averaged Navier–Stokes equations coupled with the SST k– turbulence model and a passive scalar transport formulation for NO2. Simulated breathing-zone concentrations ranged from mol/m3 under low-wind conditions (0.5 m/s) to mol/m3 at higher ventilation (4.0 m/s), highlighting the strong dependence on atmospheric dilution. Recirculation zones formed beneath the shelter roof and near confined regions, promoting pollutant accumulation within the waiting area. Optimized ventilation configurations achieved exposure reduction efficiencies (ERE) between 6.8% and 15.6%, with the highest mitigation observed under low-ventilation scenarios. The results indicate that system performance is primarily governed by the interaction between induced airflow and ambient wind conditions, with recirculation and advection patterns driving pollutant dispersion beyond local removal. These findings provide quantitative guidance for the aerodynamic design of mitigation systems in open urban transport infrastructure.
Automated water quality monitoring systems increasingly combine low-cost sensors with artificial intelligence to improve accessibility and real-time analysis. However, many existing solutions rely on expensive instrumentation, pre-trained models, or large structured datasets, limiting their applicability in underserved regions. This project presents an end-to-end, low-cost monitoring platform tailored for resource-constrained environments. The system integrates custom-designed sensors that measure water parameters such as pH, turbidity, conductivity, temperature, and color, feeding real-time data into a locally trained artificial neural network (ANN). A total of 92 water samples were collected and labeled by expert evaluation into three classes: potable, non-potable, and regular. The ANN achieved 91% overall accuracy, with high precision and recall for potable and non-potable classes, and moderate performance for regular water. The system is deployed through a web interface that allows real-time data visualization, supporting continuous dataset expansion and future model retraining. This work demonstrates a novel and scalable approach for deploying IoT and AI-based water diagnostics in settings with limited infrastructure.