Land use regression models for estimating individual NOx and NO₂ exposures in a metropolis with a high density of traffic roads and population

Sci Total Environ. 2014 Feb 15;472:1163-71. doi: 10.1016/j.scitotenv.2013.11.064. Epub 2013 Dec 28.


This study is conducted to characterize the intra-urban distribution of NOx and NO2; develop land use regression (LUR) models to assess outdoor NOx and NO2 concentrations, using the ESCAPE modeling approach with locally specific land use data; and compare NOx and NO2 exposures for children in the Taipei Metropolis by the LUR models, the nearest monitoring station, and kriging methods based on data collected at the measurement sites. NOx and NO2 were measured for 2 weeks during 3 seasons at 40 sampling sites by Ogawa passive badges to represent their concentrations at urban backgrounds and streets from October 2009 to September 2010. Land use data and traffic-related information in different buffer zones were combined with measured concentrations to derive LUR models using supervised forward stepwise multiple regressions. The annual average concentrations of NOx and NO2 in Taipei were 72.4 ± 22.5 and 48.9 ± 12.2 μg/m(3), respectively, which were at the high end of all 36 European areas in the ESCAPE project. Spatial contrasts in Taipei were lower than those of the European areas in the ESCAPE project. The NOx LUR model included 6 land use variables, which were lengths of major roads within 25 m, 25-50 m, and 50-500 m, urban green areas within 300 m and 300-5,000 m, and semi-natural and forested areas within 500 m, with R(2)=0.81. The NO2 LUR model included 4 land use variables, which were lengths of major roads within 25 m, urban green areas within 100 m, semi-natural and forested areas within 500 m, and low-density residential area within 500 m, with R(2)=0.74. The LUR models gave a wider variation in estimating NOx and NO2 exposures than either the ordinary kriging method or the nearest measurement site did for the children of Taiwan Birth Cohort Study (TBCS) in Taipei.

Keywords: GIS; LUR; Land use regression; Nitrogen dioxide; Nitrogen oxides; RMSE; Traffic pollution; geographic information systems; land use regression; root mean square error.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Air Pollutants / analysis*
  • Air Pollution / statistics & numerical data*
  • Environmental Monitoring / methods*
  • Nitrogen Dioxide / analysis
  • Nitrogen Oxides / analysis*
  • Regression Analysis
  • Spatial Analysis
  • Taiwan
  • Transportation / statistics & numerical data


  • Air Pollutants
  • Nitrogen Oxides
  • Nitrogen Dioxide