Abstract:Ambient exposure to fine particulate matter (PM2.5) is known to harm public health in China. Satellite remote sensing measurements of aerosol optical depth (AOD) were statistically associated with in-situ observations after 2013 to predict PM2.5 concentrations nationwide, while the lack of surface monitoring data before 2013 have created difficulties in historical PM2.5 exposure estimates. Hindcast approaches using statistical models or chemical transport models (CTMs) were developed to overcome this limitation, while those approaches still suffer from incomplete daily coverage due to missing AOD data or limited accuracy due to uncertainties of CTMs. Here we developed a new machine learning (ML) model with high-dimensional expansion (HD-expansion) of numerous predictors (including AOD and other satellite covariates, meteorological variables and CTM simulations). Through comprehensive characterization of the nonlinear effects of, and interactions among different predictors, the HD-expansion parameterized the association between PM2.5 and AOD as a nonlinear function of space and time covariates (e.g., planetary boundary layer height and relative humidity). In this way, the PM2.5-AOD association can vary spatiotemporally. We trained the model with data from 2013 to 2016 and evaluated its performance using annually-iterated cross-validation, which iteratively held out the in-situ observations for a whole calendar year (as testing data) to examine the predictions from a model trained by the rest of the observations. Our estimates were found to be in good agreement with in-situ observations, with correlation coefficients (R2) of 0.61, 0.68, and 0.75 for daily, monthly and annual averages, respectively. To interpolate the missing predictions due to incomplete AOD data, we incorporated a generalized additive model into the ML model. The two-stage estimates of PM2.5 sacrificed the prediction accuracy on a daily timescale (R2 = 0.55), but achieved complete spatiotemporal coverage and improved the accuracy of monthly (R2 = 0.71) and annual (R2 = 0.77) averages. The model was then used to predict daily PM2.5 concentrations during 2000-2016 across China and estimate long-term trends in PM2.5 for the period. We found that population-weighted concentrations of PM2.5 significantly increased, by 2.10 (95% confidence interval (CI): 1.74, 2.46) μg/m3/year during 2000-2007, and rapidly decreased by 4.51 (3.12, 5.90) μg/m3/year during 2013-2016. In this study, we produced AOD-based estimates of historical PM2.5 with complete spatiotemporal coverage, which were evidenced as accurate, particularly in middle and long term. The products could support large-scale epidemiological studies and risk assessments of ambient PM2.5 in China and can be accessed via the website (http://www.meicmodel.org/dataset-phd.html).

Estimation of PM2.5 Concentration across China Based on Multi-Source Remote Sensing Data and Machine Learning Methods

National Scale Spatiotemporal Land-Use Regression Model for PM2.5, PM10 and NO2 Concentration in China

Estimation of PM<sub>2.5</sub> concentration in China using linear hybrid machine learning model

Estimating Ground-Level PM2.5 in China Using Satellite Remote Sensing.

Remote Sensing Fine Estimation Model of PM2.5 Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China

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.

Predicting monthly high-resolution PM2.5 concentrations with random forest model in the North China Plain

High-resolution Prediction of the Spatial Distribution of PM2.5 Concentrations in China Using a Long Short-Term Memory Model

Machine learning driven by environmental covariates to estimate high-resolution PM2.5 in data-poor regions

Estimating PM2.5 concentrations via random forest method using satellite, auxiliary, and ground-level station dataset at multiple temporal scales across China in 2017

Prediction of Fine Particulate Matter Concentration near the Ground in North China from Multivariable Remote Sensing Data Based on MIV-BP Neural Network

Retrieval of Fine-Grained PM2.5 Spatiotemporal Resolution Based on Multiple Machine Learning Models

Spatiotemporal dynamics and exposure analysis of daily PM2.5 using a remote sensing-based machine learning model and multi-time meteorological parameters

Estimation of hourly full-coverage PM2.5 concentrations at 1-km resolution in China using a two-stage random forest model

Improving the Estimation of PM2.5 Concentration in the North China Area by Introducing an Attention Mechanism into Random Forest

Regional Representativeness Analysis of Ground-Monitoring PM2.5 Concentration Based on Satellite Remote Sensing Imagery and Machine Learning Techniques

Prediction of PM2.5 Concentration Using Spatiotemporal Data with Machine Learning Models

Point-surface fusion of station measurements and satellite observations for mapping PM2.5 distribution in China: methods and assessment

Full Coverage Estimation of the PM Concentration Across China Based on an Adaptive Spatiotemporal Approach

Estimating Spatiotemporal Distribution of PM1 Concentrations in China with Satellite Remote Sensing, Meteorology, and Land Use Information

Estimation of PM2.5 concentrations in North China with high spatiotemporal resolution using the ERA5 dataset and machine learning models