A Multiscale Land Use Regression Approach for Estimating Intraurban Spatial Variability of PM_2.5 Concentration by...

Area of Science:

• Environmental Science
• Urban Planning
• Atmospheric Chemistry

Background:

• Poor air quality is a significant global urban environmental challenge, especially in Asia.
• High-density cities exhibit complex pollution dispersal, leading to spatial variability in exposure levels.
• Particulate matter (PM2.5) is a key air pollutant of concern.

Purpose of the Study:

• To investigate and map the fine-scale spatial variability of PM2.5 exposure in urban environments.
• To address the challenges of analyzing multiscale pollution complexity.
• To develop an improved Land Use Regression (LUR) model for PM2.5 estimation.

Main Methods:

• Utilized Land Use Regression (LUR) modeling integrating air quality, satellite, meteorological, and spatial data.
• Conducted independent modeling at city and neighborhood scales, treating predictor variables separately.
• Incorporated building morphological indices and road network centrality metrics for neighborhood-scale analysis.

Main Results:

• Achieved improved prediction performance in the Aerosol Optical Depth (AOD)-PM2.5 relationship at the city scale (R² from 0.72 to 0.80).
• Identified building morphology and road network metrics as effective indicators for neighborhood-scale PM2.5 estimation.
• Generated a combined city- and neighborhood-scale PM2.5 map detailing intraurban variability.

Conclusions:

• The study successfully developed a multiscale LUR modeling approach for PM2.5 estimation.
• The combined model provides a valuable geospatial tool for understanding intraurban PM2.5 variations.
• Findings contribute to better urban air quality management and public health strategies.