A site-optimised multi-scale GIS based land use regression model for simulating local scale patterns in air...

Area of Science:

• Environmental Science
• Geographic Information Systems (GIS)
• Air Quality Modeling

Background:

• Standard Land Use Regression (LUR) models use a single equation, failing to capture spatial variations in pollutant dispersion across diverse urban landscapes.
• This limitation hinders model transferability and accuracy, especially when land use types are unevenly sampled.
• Heterogeneous urban environments require localized modeling approaches for effective air pollution assessment.

Purpose of the Study:

• To develop and evaluate a site-optimized, multi-scale GIS-based LUR modeling approach for simulating nitrogen dioxide (NO2) concentrations.
• To compare the performance of the multi-scale LUR model against standard LUR, Universal Kriging (UK), Ordinary Kriging (OK), and Inverse Distance Weighting (IDW) models.
• To assess the model's effectiveness in distinct urban scales: central business district (CBD), urban, and suburban areas in Auckland.

Main Methods:

• Developed a site-optimized, multi-scale GIS-based LUR model to simulate NO2 concentrations at CBD, urban, and suburban scales.
• Employed leave-one-out cross-validation (LOOCV) to assess model performance at each scale.
• Validated the multi-scale LUR model against external observations from eight fixed regulatory monitoring stations.

Main Results:

• The multi-scale LUR model demonstrated strong performance with R² values of 0.62 (CBD), 0.86 (urban), and 0.73 (suburban) during LOOCV.
• External validation showed a high R² of 0.85 and the lowest RMSE (8.48 μg·m⁻³) compared to other models.
• Predictor variables varied significantly across scales, indicating scale-specific influences on NO2 dispersion.

Conclusions:

• The site-optimized multi-scale LUR model provides a robust and accurate method for mapping NO2 concentrations in complex urban environments.
• This approach overcomes the limitations of standard LUR models, offering improved spatial representation of air pollution.
• The model's superior performance highlights the importance of scale-specific analysis in air quality modeling for distinct urban configurations.