Urban form regulation for synergetic PM_2.5 and O₃ control: A multi-indicator constrained DNN simulation

Area of Science:

• Environmental Science
• Urban Planning
• Air Quality Management

Background:

• Optimizing urban air quality through urban form indicators (UFIs) is crucial for sustainable development.
• Synergistic multi-pollutant control considering UFIs is challenging due to complex pollutant interactions and spatiotemporal variations.

Purpose of the Study:

• To develop a novel Urban Form Regulation-Aided Air Quality Optimization Model (UFR-AQOM) for synergistic control of PM2.5 and O3.
• To establish a nonlinear mapping between PM2.5, O3 concentrations and UFIs using deep learning.
• To investigate the effectiveness of UFI regulation for air quality improvement in China.

Main Methods:

• Developed UFR-AQOM using a deep residual network based on autoencoders to map UFIs to PM2.5 and O3 concentrations.
• Employed an inequality constraint strategy to regulate UFIs for pollutant optimization.
• Conducted comparative experiments in mainland China focusing on PM2.5 attainment and collaborative PM2.5/O3 attainment.

Main Results:

• The model demonstrated high accuracy with R² values of 0.97 and low RMSE for both PM2.5 and O3.
• Collaborative optimization (Scenario 2) effectively reduced O3 while maintaining PM2.5 control, outperforming single-pollutant optimization.
• Identified varying regulation effectiveness across regions and a higher sensitivity of PM2.5 to UFIs compared to O3.

Conclusions:

• UFR-AQOM provides an accurate and effective approach for synergistic PM2.5 and O3 control in urban planning.
• Recommended enhancing regulation of green spaces and road density, and zoned regulation of land use proportions for optimal air quality.
• Findings support evidence-based urban planning for improved air quality and public health.