[Variation Analysis of Daily PM_2.5 Concentrations Based on Boosted Regression Tree: A Case Study in Changzhou]

Area of Science:

• Environmental Science
• Atmospheric Chemistry
• Air Quality Monitoring

Context:

• Analysis of hourly concentration data from six air quality monitoring stations in Changzhou urban area (December 2014 - November 2015).
• Investigation into seasonal variations and spatial distribution of PM2.5 pollution.
• Quantification of contributing factors to daily PM2.5 changes.

Purpose:

• To analyze the seasonal variation of PM2.5 pollution in Changzhou.
• To quantify the contributions of PM10, gaseous pollutants, and meteorological factors to daily PM2.5 changes using Boosted Regression Tree (BRT).
• To identify key drivers of PM2.5 pollution for targeted mitigation strategies.

Summary:

• PM2.5 pollution exhibits significant seasonal differences, with severe pollution and longer duration in winter and lighter pollution in summer.
• The Boosted Regression Tree (BRT) model demonstrated a good fit for simulating daily average PM2.5 concentrations, with a cross-validation correlation coefficient of 0.957.
• Meteorological conditions (36.2%) were the most significant factor influencing daily PM2.5 changes, followed by gaseous pollutants (28%), PM10 (23.4%), and regional transport (12.6%).

Impact:

• Meteorological factors, secondary formation, primary emissions, and regional transport are key drivers of PM2.5 pollution.
• Positive associations were found between daily PM2.5 and PM10, relative humidity, CO, and O3.
• Negative correlations were observed with temperature, SO2, and mixed layer height, indicating complex relationships with atmospheric pressure and NO2.