Contribution of ecological restoration projects to long-term changes in PM2.5

Fine particulate matter (PM2.5) concentration, a crucial indicator reflecting changes in air quality, has frequently been used in previous studies. However, the effects of large-scale ecological restoration (ER) projects on PM2.5 concentrations are often overlooked. Therefore, we used net primary productivity (NPP) as an indicator of ER engineering benefits, used ensemble empirical modal decomposition (EEMD) to reveal the trend of linear and nonlinear relationships of PM2.5 driven by different types and amounts of ER projects, and utilized the Extreme Gradient Boosting (XGBoost) and Shapley's Additive Interpretation (SHAP) values to quantify the impact of each factor on long-term PM2.5 concentrations. The results suggest that: (1) The nonlinear trend better describes the trend of PM2.5 change, with a shift from increasing to decreasing in areas covering 74.15% of China's area, especially in the four major ER zones; (2) PM2.5 concentration exhibits regional effects, L-L and H-H aggregation areas account for a larger proportion of the area, and are distributed in the areas with low and high PM2.5 concentrations, respectively. (3) The year 1990 marks a turning point for 36.8% of the regions. Across four ER regions emerge: a dramatic increase in PM2.5 concentration before ER implementation, followed by a slowing growth trend in the initial stages of the project, and ultimately a gradual decrease. (4) While the contribution of ER to the decrease of PM2.5 concentration is lower than the contribution of the growth caused by human activities and climate change, plant and newly plant areas are the main regions in which the ER project suppresses its growth. These findings highlight the influence of ER projects on PM2.5 concentration in addition to climate and human activities and provide a theoretical basis and scientific and technological support for quantifying the suppression of air pollution by ER projects.