Attribution of Air Quality Benefits to Clean Winter Heating Policies in China: Combining Machine Learning with Causal Inference

Abstract

Heating is a major source of air pollution. To improve air quality, a range of clean heating policies were implemented in China over the past decade. Here, we evaluated the impacts of winter heating and clean heating policies on air quality in China using a novel, observation-based causal inference approach. During 2015–2021, winter heating causally increased annual PM2.5, daily maximum 8-h average O3, and SO2 by 4.6, 2.5, and 2.3 μg m–3, respectively. From 2015 to 2021, the impacts of winter heating on PM2.5 in Beijing and surrounding cities (i.e., “2 + 26” cities) decreased by 5.9 μg m–3 (41.3%), whereas that in other northern cities only decreased by 1.2 μg m–3 (12.9%). This demonstrates the effectiveness of stricter clean heating policies on PM2.5 in “2 + 26” cities. Overall, clean heating policies caused the annual PM2.5 in mainland China to reduce by 1.9 μg m–3 from 2015 to 2021, potentially avoiding 23,556 premature deaths in 2021.

Measuring the Emission Changes and Meteorological Dependence of Source‐Specific BC Aerosol Using Factor Analysis Coupled With Machine Learning

“…Such activity usually begins in the middle of November and ends in the middle of next March. Quantifying the impacts of winter‐heating on air quality has received much concern because of the severe air pollution that occurs during the cold season (Song et al., 2023; Yun et al., 2021; J. Zhang et al., 2020). To reduce heating‐related air pollutant emissions, the central government enacted the “Clean Winter Heating Plan for Northern China (2017–2021)” in northern China in 2017, with the goal of increasing the share of clean heating to 50% by 2019 and to 70% by 2021 with respect to the 2016 level (NDRC, 2017).…”

“…Such activity usually begins in the middle of November and ends in the middle of next March. Quantifying the impacts of winter-heating on air quality has received much concern because of the severe air pollution that occurs during the cold season (Song et al, 2023;Yun et al, 2021;. 5b), respectively.…”

Toward Better and Healthier Air Quality: Global PM2.5 and O3 Pollution Status and Risk Assessment Based on the New WHO Air Quality Guidelines for 2021

“…Under specific meteorological conditions and in certain regions, the formation and concentrations levels of O 3 and PM 2.5 are determined by the ratio of emissions of NOx and VOCs. [76][77][78][79][80] For instance, in areas with VOC restrictions, reducing NOx emissions may result in increased O 3 concentrations and vice versa. [81,82] Additionally, Jiang et al research indicates that the atmospheric oxidizing capacity's strength impacts the production of PM 2.5 and O 3 in both northern and southern cities.…”

Differentiating Periodic Drivers of Air Quality Changes: A Two‐Step Decomposition Approach Integrating Machine Learning and Wavelet Analysis

“…Apart from atmospheric processes operating at different timescales, emission intensities from various sources also exhibit temporal changes with distinct periodic patterns. For instance, emissions from cooking and vehicles display prominent diurnal patterns (He et al., 2016; Li et al., 2019), while heating‐related emissions vary with the seasons (Song et al., 2023). To better discern the impacts of different drivers (sources and meteorological processes) on air quality, it is essential to effectively separate these periodic signals within air quality time series data.…”