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

Abstract

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.

Association between prenatal and childhood PM2.5 exposure and preadolescent anxiety and depressive symptoms

“…Daily average PM 2.5 exposure at each participant’s residence was estimated using a previously developed hybrid satellite land use regression model at a 1 × 1 km spatial resolution. 51 This model uses Extreme Gradient Boosting with inverse-distance weighted surfaces and several different meteorological and land use variables. Briefly, predictors include longitude and latitude, date, the density of roadways from OpenStreetMap, the aerosol optical depth from NASA’s Terra and Aqua satellites, daily mean predicted PM 2.5 from the NASA MERRA-2 GMI atmospheric simulation, daily mean of the height of the atmospheric mixing layer from the European Center for Medium-Range Weather Forecasts global climate dataset, temperature, dewpoint temperature, and the daily sum of precipitation.…”

“…The R 2 varied year to year, ranging from 0.64 to 0.86. 51 Participants were matched to the centroid of the nearest grid cell based on GPS coordinates collected at their residential address by study personnel. Residential addresses were updated throughout the study period, from birth to the most recent study visit.…”

Daily exposure to PM2.5and 1.5 million deaths: A time-stratified case-crossover analysis in the Mexico City Metropolitan Area

“…We utilized daily PM 2.5 predictions with spatial resolution of 1×1 km from our recently developed model based on extreme gradient boosting (XGBoost), and inverse-distance weighting (IDW) that uses aerosol optical depth data, meteorology, and land-use variables to predict daily mean PM 2.5 for the Mexico City Metropolitan Area (Gutiérrez-Avila et al, 2022). Daily mean air temperature with the same 1x1 km resolution was obtained from our satellite-based land surface temperature model for Central Mexico (Gutiérrez-Avila et al, 2021).…”