A sub-neighborhood scale land use regression model for predicting NO(2).

Area of Science:

• Environmental Science
• Urban Planning
• Atmospheric Chemistry

Background:

• Nitrogen dioxide (NO2) is a key air pollutant with significant health impacts.
• Accurate, fine-scale NO2 monitoring is crucial for urban environmental health assessments.
• Land use regression (LUR) models offer a promising approach for predicting air pollutant concentrations.

Purpose of the Study:

• To develop and validate a sub-neighborhood scale land use regression model for NO2 prediction in Portland, Oregon.
• To assess the impact of various geographic and meteorological variables on NO2 concentrations.
• To evaluate the generalizability of LUR models in a North American context.

Main Methods:

• Utilized passive NO2 measurements at 77 locations across Portland.
• Incorporated variables such as road density, railroad density, traffic volume, and land use within defined buffer zones (50-750 m).
• Employed regression analysis, including wind direction, and assessed model stability through iterative site exclusion.

Main Results:

• The initial LUR model explained 66% of the variation in NO2 concentrations.
• Inclusion of wind direction improved the model's predictive power by 15%.
• Model calibration showed a minor 3% variation in predictive power with iterative site exclusion.

Conclusions:

• Land use regression is a validated and effective method for predicting fine-scale NO2 concentrations in North America.
• Key factors for accurate NO2 modeling include road/railway infrastructure, land use patterns, and meteorological conditions like wind direction.
• The developed model provides a valuable tool for urban air quality management and health studies in Portland.