Prediction of daily mean and one-hour maximum PM2.5 concentrations and applications in Central Mexico using satellite-based machine-learning models.

BACKGROUND: Machine-learning algorithms are becoming popular techniques to predict ambient air PM2.5 concentrations at high spatial resolutions (1 × 1 km) using satellite-based aerosol optical depth (AOD). Most machine-learning models have aimed to predict 24 h-averaged PM2.5 concentrations (mean PM2.5) in high-income regions. Over Mexico, none have been developed to predict subdaily peak levels, such as the maximum daily 1-h concentration (max PM2.5). OBJECTIVE: Our goal was to develop a machine-learning model to predict mean PM2.5 and max PM2.5 concentrations in the Mexico City Metropolitan Area from 2004 through 2019. METHODS: We present a new modeling approach based on extreme gradient boosting (XGBoost) and inverse-distance weighting that uses AOD, meteorology, and land-use variables. We also investigated applications of our mean PM2.5 predictions that can aid local authorities in air-quality management and public-health surveillance, such as the co-occurrence of high PM2.5 and heat, compliance with local air-quality standards, and the relationship of PM2.5 exposure with social marginalization. RESULTS: Our models for mean and max PM2.5 exhibited good performance, with overall cross-validated mean absolute errors (MAE) of 3.68 and 9.20 µg/m3, respectively, compared to mean absolute deviations from the median (MAD) of 8.55 and 15.64 µg/m3. In 2010, everybody in the study region was exposed to unhealthy levels of PM2.5. Hotter days had greater PM2.5 concentrations. Finally, we found similar exposure to PM2.5 across levels of social marginalization. SIGNIFICANCE: Machine learning algorithms can be used to predict highly spatiotemporally resolved PM2.5 concentrations even in regions with sparse monitoring. IMPACT: Our PM2.5 predictions can aid local authorities in air-quality management and public-health surveillance, and they can advance epidemiological research in Central Mexico with state-of-the-art exposure assessment methods.

Characterization of aerosol particles during a high pollution episode over Mexico City.

More than 7 thousand wildfires were recorded over Mexico in 2019, affecting almost 640 thousand hectares. Most of these fires occurred during the spring season generating dense smoke plumes, impacting urban areas in the central part of the Mexican plateau. From May 10 to 17, 2019, biomass burning (BB) plumes affected Mexico City (MC) and diffused across the basin, producing PM2.5 levels ~ 2 times higher than the nation's air quality standards. Average PM2.5 concentrations increased sharply from 29.4 ± 7.2 µg m-3 to 65.1 ± 13.6 µg m-3 when the dense smoke plumes were detected. The higher particle concentration impacted the aerosol optical depth (AOD) as values ~ 3 times greater than the annual mean (0.32 ± 0.12) were measured, which resulted in a 17% loss of global horizontal irradiation (GHI). Under these severe pollution conditions, the visibility (Va) was reduced by ~ 80%. The high incidence of strong absorbent particles, such as soot and tarballs was revealed through electron microscopy and X-ray fluorescence (XRF) analysis. These techniques show chemical similarities between MC aerosols and those from the high-altitude (~ 4010 m. a. g. l.) Altzomoni Atmospheric Observatory, evidencing a strong influence of the BB emissions, suggesting a regional transport of these pollutants.

Impact of the COVID-19 Lockdown on Air Quality and Resulting Public Health Benefits in the Mexico City Metropolitan Area.

Meteorology and long-term trends in air pollutant concentrations may obscure the results from short-term policies implemented to improve air quality. This study presents changes in CO, NO2, O3, SO2, PM10, and PM2.5 based on their anomalies during the COVID-19 partial (Phase 2) and total (Phase 3) lockdowns in Mexico City (MCMA). To minimise the impact of the air pollutant long-term trends, pollutant anomalies were calculated using as baseline truncated Fourier series, fitted with data from 2016 to 2019, and then compared with those from the lockdown. Additionally, days with stagnant conditions and heavy rain were excluded to reduce the impact of extreme weather changes. Satellite observations for NO2 and CO were used to contrast the ground-based derived results. During the lockdown Phase 2, only NO2 exhibited significant decreases (p < 0.05) of between 10 and 23% due to reductions in motor vehicle emissions. By contrast, O3 increased (p < 0.05) between 16 and 40% at the same sites where NO2 decreased. During Phase 3, significant decreases (p < 0.05) were observed for NO2 (43%), PM10 (20%), and PM2.5 (32%) in response to the total lockdown. Although O3 concentrations were lower in Phase 3 than during Phase 2, those did not decrease (p < 0.05) from the baseline at any site despite the total lockdown. SO2 decreased only during Phase 3 in a near-road environment. Satellite observations confirmed that NO2 decreased and CO stabilised during the total lockdown. Air pollutant changes during the lockdown could be overestimated between 2 and 10-fold without accounting for the influences of meteorology and long-term trends in pollutant concentrations. Air quality improved significantly during the lockdown driven by reduced NO2 and PM2.5 emissions despite increases in O3, resulting in health benefits for the MCMA population. A health assessment conducted suggested that around 588 deaths related to air pollution exposure were averted during the lockdown. Our results show that to reduce O3 within the MCMA, policies must focus on reducing VOCs emissions from non-mobile sources. The measures implemented during the COVID-19 lockdowns provide valuable information to reduce air pollution through a range of abatement strategies for emissions other than from motor vehicles.

Characterizing the aging of biomass burning organic aerosol by use of mixing ratios: a meta-analysis of four regions.

Characteristic organic aerosol (OA) emission ratios (ERs) and normalized excess mixing ratios (NEMRs) for biomass burning (BB) events have been calculated from ambient measurements recorded during four field campaigns. Normalized OA mass concentrations measured using Aerodyne Research Inc. quadrupole aerosol mass spectrometers (Q-AMS) reveal a systematic variation in average values between different geographical regions. For each region, a consistent, characteristic ratio is seemingly established when measurements are collated from plumes of all ages and origins. However, there is evidence of strong regional and local-scale variability between separate measurement periods throughout the tropical, subtropical, and boreal environments studied. ERs close to source typically exceed NEMRs in the far-field, despite apparent compositional change and increasing oxidation with age. The absence of any significant downwind mass enhancement suggests no regional net source of secondary organic aerosol (SOA) from atmospheric aging of BB sources, in contrast with the substantial levels of net SOA formation associated with urban sources. A consistent trend of moderately reduced ΔOA/ΔCO ratios with aging indicates a small net loss of OA, likely as a result of the evaporation of organic material from initial fire emissions. Variability in ERs close to source is shown to substantially exceed the magnitude of any changes between fresh and aged OA, emphasizing the importance of fuel and combustion conditions in determining OA loadings from biomass burning.

A vehicle emissions system using a car simulator and a geographical information system: Part 1--System description and testing.

A methodology for estimating vehicular emissions comprising a car simulator, a basic traffic model, and a geographical information system is capable of estimating vehicle emissions with high time and space resolution. Because of the extent of the work conducted, this article comprises two sections: In Part 1 of this work, we describe the system and its components and use examples for testing it. In Part 2 we will study in more detail the emissions of the sample fleet using the system and will make comparisons with another emission model. The experimental data for the car simulator is obtained using on-board measuring equipment and laboratory Fourier transform IR (FTIR) measurements with a dynamometer following typical driving cycles. The car simulator uses this data to generate emission factors every second. These emission factors, together with information on car activity and velocity profiles of highways and residential and arterial roads in Mexico City in conjunction with a basic traffic model, provide emissions per second of a sample fleet. A geographical information system is used to localize these road emissions.

Higher excited electronic transitions of polyacetylene cations HC2nH+ n = 2-7 in neon matrixes.

The polyacetylene HC2nH+ n = 2-7 cations were produced from a mixture of diacetylene with helium in a hot cathode-discharge source. After a mass-selective deposition, their absorption spectra were studied in 6 K neon matrixes. Besides the known A2Pi <-- X2Pi system, several new transitions to higher excited 2Pi electronic states of these cations have been observed. In the case of HC4H+ and HC6H+, only one new weak absorption system has been detected with the onset at 336.1 and 417.2 nm, respectively. These C2Pi <-- X2Pi transitions form a series that extends to HC10H+. Two further electronic transitions are observed for HC8H+ through to HC14H+; a weaker B2Piu <-- X2Pig and a strong E2Piu <-- X2Pig in the UV. The integrated intensity of the UV system of the polyacetylene cations exceeds that of the A2Pi <-- X2Pi transition by an order of magnitude.