Abstract:BACKGROUND: Short-term exposure to ambient PM2·5 is a leading contributor to the global burden of diseases and mortality. However, few studies have provided the global spatiotemporal variations of daily PM2·5 concentrations over recent decades.METHODS: In this modelling study, we implemented deep ensemble machine learning (DEML) to estimate global daily ambient PM2·5 concentrations at 0·1° × 0·1° spatial resolution between Jan 1, 2000, and Dec 31, 2019. In the DEML framework, ground-based PM2·5 measurements from 5446 monitoring stations in 65 countries worldwide were combined with GEOS-Chem chemical transport model simulations of PM2·5 concentration, meteorological data, and geographical features. At the global and regional levels, we investigated annual population-weighted PM2·5 concentrations and annual population-weighted exposed days to PM2·5 concentrations higher than 15 μg/m3 (2021 WHO daily limit) to assess spatiotemporal exposure in 2000, 2010, and 2019. Land area and population exposures to PM2·5 above 5 μg/m3 (2021 WHO annual limit) were also assessed for the year 2019. PM2·5 concentrations for each calendar month were averaged across the 20-year period to investigate global seasonal patterns.FINDINGS: Our DEML model showed good performance in capturing the global variability in ground-measured daily PM2·5, with a cross-validation R2 of 0·91 and root mean square error of 7·86 μg/m3. Globally, across 175 countries, the mean annual population-weighted PM2·5 concentration for the period 2000-19 was estimated at 32·8 μg/m3 (SD 0·6). During the two decades, population-weighted PM2·5 concentration and annual population-weighted exposed days (PM2·5 >15 μg/m3) decreased in Europe and northern America, whereas exposures increased in southern Asia, Australia and New Zealand, and Latin America and the Caribbean. In 2019, only 0·18% of the global land area and 0·001% of the global population had an annual exposure to PM2·5 at concentrations lower than 5 μg/m3, with more than 70% of days having daily PM2·5 concentrations higher than 15 μg/m3. Distinct seasonal patterns were indicated in many regions of the world.INTERPRETATION: The high-resolution estimates of daily PM2·5 provide the first global view of the unequal spatiotemporal distribution of PM2·5 exposure for a recent 20-year period, which is of value for assessing short-term and long-term health effects of PM2·5, especially for areas where monitoring station data are not available.FUNDING: Australian Research Council, Australian Medical Research Future Fund, and the Australian National Health and Medical Research Council.

Hyperlocal US PM2.5 Trace Elements Super-learned

Hyperlocal super-learned PM2.5 components across the contiguous US

Super-learning and Ensemble Weighted Averaging Models to Predict Hyperlocal Long-Term Exposure to Fine Particulate Matter Components in the United States

Monitoring and Source Apportionment of Trace Elements in PM2.5: Implications for Local Air Quality Management.

Characteristics, Sources, and Health Risks of PM2.5-bound Trace Metals in Northern Zhejiang Province: the Effects of Meteorological Variables Based on Machine Learning

An ensemble-based model of PM2.5 concentration across the contiguous United States with high spatiotemporal resolution

Space-time Trends of PM2.5 Constituents in the Conterminous United States Estimated by a Machine Learning Approach, 2005–2015

Elemental Characterization of Ambient Particulate Matter for a Globally Distributed Monitoring Network: Methodology and Implications

Dynamic landscape of multi-elements in PM2.5 revealed by real-time analysis

Quantifying the pollution changes and meteorological dependence of airborne trace elements coupling source apportionment and machine learning

PM10-bound trace elements in pan-European urban atmosphere

Pollution characteristics and ecological risks of trace elements in PM2.5 over three COALESCE network sites - Bhopal, Mesra, and Mysuru, India

Simulation of Airborne Trace Metals in Fine Particulate Matter over North America.

PM2.5 Prediction Based on Random Forest, XGBoost, and Deep Learning Using Multisource Remote Sensing Data

Trace elements in particulate matter from metropolitan regions of Northern China: Sources, concentrations and size distributions

Significant but Spatiotemporal-Heterogeneous Health Risks Caused by Airborne Exposure to Multiple Toxic Trace Elements in China

Sources and uncertainties of health risks for PM2.5-bound heavy metals based on synchronous online and offline filter-based measurements in a Chinese megacity

Global estimates of daily ambient fine particulate matter concentrations and unequal spatiotemporal distribution of population exposure: a machine learning modelling study

Spatiotemporal Continuous Estimates of PM2.5 Concentrations in China, 2000-2016: A Machine Learning Method with Inputs from Satellites, Chemical Transport Model, and Ground Observations.

Large global variations in measured airborne metal concentrations driven by anthropogenic sources

High spatio-temporal resolution predictions of PM 2.5 using low-cost sensor data