Improved 1-km-resolution PM_2.5 estimates across China using the space-time extremely randomized trees

Abstract

<p><strong>Abstract.</strong> Fine particulate matter with aerodynamic diameters ≤ 2.5 μm (PM<sub>2.5</sub>) shows adverse effects on human health and atmospheric environment. Satellite-derived aerosol products have been intensively adopted in estimating surface PM<sub>2.5</sub> concentrations, but most previous studies failed to monitor air pollution over small-scale areas limited by coarse spatial-resolution (3–50 km) and low data-quality aerosol optical depth (AOD) products. Therefore, a new space-time extremely randomized trees (STET) model is developed that integrates spatiotemporal information to improve PM<sub>2.5</sub> estimates at both spatial resolution and overall accuracy across China. To this end, the newly released MODIS MAIAC AOD product, meteorological and other auxiliary data are inputs to the STET model. Daily 1-km PM<sub>2.5</sub> maps in 2018 across mainland China are produced. The STET model performs well with a high out-of-sample (out-of-station) cross-validation coefficient of 0.89 (0.88), a low root-mean-square error of 10.35 (10.97) μg/m<sup>3</sup>, a small mean absolute error of 6.71 (7.17) μg/m<sup>3</sup>, and a small mean relative error of 21.37 % (23.77 %), respectively. Particularly, it can well capture the PM<sub>2.5</sub> concentrations at both regional and individual site scales. In addition, it posed a strong predictive power (e.g., monthly-R<sup>2</sup> = 0.80) and can be used to predict the historical PM<sub>2.5</sub> records. The North China Plain, the Sichuan Basin, and Xinjiang Province always are featured with high PM<sub>2.5</sub> pollution, especially in winter. The STET model outperforms most models presented in previous related studies. More importantly, our study provides a new approach to obtain high-quality PM<sub>2.5</sub> estimates, which is important for air pollution studies over urban areas.</p>

Association of Ambient Particulate Matter Pollution of Different Sizes With In-Hospital Case Fatality Among Stroke Patients in China

“…We used the 10 × 10 km grid China High Air Pollutants (CHAP) daily data set as the data source for assessing exposure to air pollutants. [22][23][24][25] CHAP is a long-term, full-coverage, highresolution, high-accuracy, ground-level air pollutant data source. It uses a combination of advanced satellite remote sensing and space-time models, achieving a high cross-validation coefficient of determination of 0.89 and low root mean square error of 10.33 μg/m 3 .…”

“…To have a more comprehensive capture of comorbidities and their weights, we alternatively adjusted for Elixhauser comorbidity scores, [37][38][39] which were weighted summation of 31 comorbidities and showed high predictive performance for inhospital case fatality in Chinese samples. 15 Fifth, to test the robustness of our exposure resolution, we assessed the long-term exposure to the pollution for each case using a much finer resolution annual 1 × 1 km CHAP data set as the alternative data source for air pollutants, 22,24 and the main models and sensitivity analyses models were re-estimated using these high-resolution grids. The 1 × 1 km exposure data were not used in our primary results because daily exposure at this resolution is not currently available.…”

Acute and chronic health impacts of PM2.5 in China and the influence of interannual meteorological variability

“…Higher resolution distributions of surface PM 2.5 are now becoming available from satellite products (e.g. Wei et al, 2019) although these are derived from aerosol optical depth retrievals using downscaled meteorological variables and interpolation/scaling approaches, and thus the reliable information content is less than the resolution suggests. In contrast, our studies are based on sound understanding of the underlying physical processes, and have information content that is as high as these alternative products.…”

Mapping Seasonal High-Resolution PM2.5 Concentrations with Spatiotemporal Bagged-Tree Model across China

“…Surface PM 2.5 concentrations estimated by satellite-based AOD have been widely applied in recent years to monitor air quality [10,11]. The AOD products commonly used to estimate PM 2.5 concentrations include MODIS AOD [12,13], MERRA-2 AOD, and Himawari-8 AOD [14,15]. A new high-resolution (1 km) daily MCD19A2 AOD retrieved by a multiangle implementation of atmospheric correction (MAIAC) algorithm was released on 30 May 2018 [16].…”

“…In this study, the AOD products with high coverage ratio are generated by integrating MODIS MAIAC AOD [13] and VIIRS IP AOD [25]. Integrating the advantages of the two AOD products by considering similar pixels between the two products can improve the coverage of MAIAC AOD products.…”