Long-range transport (LRT) of air pollutants from a range of sources can substantially enhance background pollution levels, especially in urbanized regions, which can exacerbate high-pollution episodes. In the Aburrá Valley (AV), Colombia, and other cities in northern South America, biomass burning (BB), dust, and volcanic degassing have been identified as sources of long-range aerosol transport. However, the impact of these sources on air quality and the characterization of these sources have yet to be thoroughly studied. This work investigates the influence of these sources on the chemical composition of PM_2.5 during annual and intra-annual high-load aerosol events in the AV. We identified, tracked, and meteorologically characterized LRT events and evaluated their influence on PM_2.5 concentration and chemical composition. We found that the LRT of aerosols from BB, dust, and volcanic degassing influenced approximately 13 %, 8 %, and 13 % of days in the year, respectively. We applied the positive-matrix-factorization (PMF) statistical model to quantify PM_2.5 concentrations and chemical compositions for the different LRT event types (e.g., BB). For BB events, we identified large contributions from organic carbon (OC1 and OC2), F⁻, and secondary aerosol tracers (${{\mathrm{SO}}_{\mathrm{4}}}^{\mathrm{2}-}$ and ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ ). For dust LRT events, crustal mineral components, along with Ti and Ca, were the primary contributors to aerosol composition, while ${{\mathrm{SO}}_{\mathrm{4}}}^{\mathrm{2}-}$ , Na, Al, and Ca were the primary contributors during volcanic events. The concentrations of some ions and toxic heavy metals (Cr, Mn, Cd, and Ni) were also elevated during BB and volcanic-degassing events. BB contributed the most to PM_2.5 levels during the LRT events (∼11 µg m⁻³), while contributions from aerosols arising from dust and volcanic events were also substantial (<7 µg m⁻³). Our study identifies the Orinoco and the Middle Magdalena Valley as sizable sources of BB aerosols and Nevado del Ruiz as a source of volcanic aerosols. Additionally, we found that African dust reached the Andean region via the Caribbean route. As a result, we identify the need for future chemical-transport modeling studies in the region and new support strategies to manage internal and external pollution sources that degrade air quality in the AV and the surrounding region.

Mesoscale convective systems (MCSs) are crucial in shaping large‐scale tropical circulation and the hydrological cycle, particularly in Northwestern South America (NwSA), a region marked by complex terrain and significant MCS activity. Understanding MCSs in NwSA is vital due to their impact on precipitation patterns and potential for severe weather events. To enhance this understanding, the ATRACKCS algorithm was developed for tracking convective systems, utilizing precipitation and brightness temperature data sets. This research focuses on documenting the spatiotemporal variability of MCS occurrence, life cycle, and movement. Notably, MCS hotspots were identified to the west of the major orographic features in the region, with maximum occurrences at night, contrasting with the region's typical afternoon peak in land convection. MCS movement is also heavily influenced by topography, with higher velocities on the eastern (windward) side of the Andes compared to velocities on the western (leeward) side. MCSs generally move westward, driven by easterly winds, but this pattern is not consistent throughout the year or region. Northward movement is predominant to the west of the Andes, while southward movement is observed to the east. These seasonal and regional movement variations are linked to factors such as the intertropical convergence zone position, moisture availability, topography, and low‐level jets. This research underscores the complexity of MCSs in NwSA and emphasizes the need for detailed studies on the atmospheric environment shaping these systems. Additionally, it provides a robust 21‐year MCS database for NwSA and an advanced tracking tool for research in various geographic contexts and impact areas.

Governments’ responses to the COVID-19 pandemic provide a unique opportunity to study the effects of restricted socioeconomic activity on air quality. Here, we study the changes in air pollution levels during the lockdown in Medellín and its metropolitan area, Colombia, for periods with and without enhanced regional fire activity, considering the effects of meteorology using random forest and multiple linear regression methods. The lockdown measures, which reduced mean traffic volume by 70% compared to 2016–2019, resulted in reductions for PM_2.5 (50–63%), PM₁₀ (59–64%), NO (75–76%), NO₂ (43–47%), and CO (40–47%), while O₃ concentration increased by 19–22%. In contrast, when fire activity was high, the effects of the lockdown on air quality were shadowed by the long-range transport of biomass burning emissions, increasing fine particulate matter and ozone. This study shows that healthier levels are achievable through significant efforts from decision-makers and society. The results highlight the need to develop integral measures that do not only consider reductions in the local emissions from transportation and industry, but also the role of fire activity in the region, as well as the difficulties of achieving reductions in ozone from measures that are effective at reducing primary pollutants.

We develop a framework to forecast 24 h averaged particulate matter (PM_2.5) concentrations 4 d in advance in ground-based stations over the metropolitan area of the Aburrá Valley, Colombia. The input variables are gathered from a highly diverse set of sources, including in situ real-time PM_2.5 observations, meteorological forecasts from the Global Forecasting System (GFS), aerosol optical depth (AOD) forecasts from the European Copernicus Atmosphere Monitoring Service (CAMS), and the Moderate Resolution Imaging Spectroradiometer (MODIS) active fire products. We compare the performance of two tree-based machine learning (ML) methods, random forests (RFs) and gradient boosting (GB), with linear regression as a baseline for error metrics. One of the disadvantages of tree-based models is their inability to make skillful predictions out of the domain in which the models were trained. To address that problem, we implement piecewise linear regression learners within the models. Additionally, to enhance the performance of the models, we use a customized loss function that considers the probability distribution of the target values. Tree-based models highly outperform the linear regression, with GB showing the best results in most of the 19 stations used in this study. We also test two approaches for the multi-step output problem, a direct multi-output (MO) scheme and a recursive (RC) scheme, with the GB–MO approach showing the best results. According to the performance analysis, the predictability is less for values away from the mean and decreases between 06:00 LT (local time) and the early afternoon, when the expansion of the boundary layer occurs. To contribute to understanding the sources of predictability and uncertainty of air quality in the city, we perform a feature importance analysis revealing that the relevance of the different independent variables is a function of the lead time. Particularly, apart from the past concentrations, the variables that most affect the predictability are the forecasted aerosol optical depth (AOD), the integrated fire radiative power over a forecasted back trajectory (BT-IFRP), and the predicted planetary boundary layer height (PBLH). In the testing period, the models showed the ability to forecast poor-air-quality events in the valley with more than 1 d of anticipation. This study serves as a framework for developing and evaluating the ML-based air quality forecasting models over the Andean region.

Convection-permitting modeling allows us to understand mechanisms that influence rainfall in specific regions. However, microphysics parameterization (MP) and planetary boundary layer (PBL) schemes remain an important source of uncertainty, affecting rainfall intensity, occurrence, duration, and propagation. Here, we study the sensitivity of rainfall to three MP [Weather Research and Forecasting (WRF) Single-Moment 6-class (WSM6), Thompson, and Morrison] and two PBL [the Yonsei University (YSU) and Mellor–Yamada Nakanishi Niino (MYNN)] schemes with a convection-permitting resolution (4 km) over northwestern South America (NWSA). Simulations were performed by using the WRF model and the results were evaluated against soundings, rain gauges, and satellite data, considering the spatio-temporal variability of rainfall over diverse regions prone to deep convection in NWSA. MP and PBL schemes largely influenced simulated rainfall, with better results for the less computationally expensive WSM6 MP and YSU PBL schemes. Regarding rain gauges and satellite estimates, simulations with Morrison MP overestimated rainfall, especially westward of the Andes, whereas the MYNN PBL underestimated precipitation in the Amazon–Savannas flatlands. We found that the uncertainty in the rainfall representation is highly dependent on the region, with a higher influence of MP in the Colombian Pacific and PBL in the Amazon–Savannas flatlands. When analyzing rainfall-related processes, the selection of both MP and PBL parameterizations exerted a large influence on the simulated lower tropospheric moisture flux and moisture convergence. PBL schemes significantly influenced the downward shortwave radiation, with MYNN simulating a greater amount of low clouds, which decreased the radiation income. Furthermore, latent heat fluxes were greater for YSU, favoring moist convection and rainfall. MP schemes had a marked impact on vertical velocity. Specifically, Morrison MP showed stronger convection and higher precipitation rates, which is associated with a greater latent heat release due to solid-phase hydrometeor formation. This study provides insights into assessing physical parameterizations in numerical models and suggests key processes for rainfall representation in NWSA.