Evaluate the Validity of Land Use Regression Models for NO2 Based on Routine and Campaign Monitoring Data from an Urban Area of Japan

Abstract

Background: Generally, land use regression (LUR) models have been built using campaign monitoring data rather than routine monitoring data. However, the routine monitoring data have advantages such as low cost and long-term coverage. The objective of this study was to compare the predictability of LUR models for NO2 based on routine and campaign monitoring data obtained from an urban area. This evaluation considered that LUR models that could represent regional differences in air pollution exposure and regional road structure were optimal.Method: We selected the city of Suita in Osaka (Japan). We built a model based on routine monitoring data obtained from all sites (routine-LUR), and a model based on campaign monitoring data (campaign-LUR) within the city. The routine LUR models were based on monitoring networks across two prefectures (i.e., Osaka and Hyogo prefectures). Next, we selected 30 evaluation sites and developed LUR models excluding these evaluations sites. We calculated the predictability of each model, and compared the predicted NO2 concentrations from each model with measured annual average NO2 concentrations from evaluation sites.Results: The predictability for NO2 concentrations of routine-LUR model was better than that of campaign-LUR model: adjusted R2 = 0.68 and 0.59, respectively, and root mean square error = 3.4 ppb and 3.9 ppb, respectively. The routine-LUR model were highly correlated with the measured NO2 concentrations at evaluation sites [ρ = 0.88] comparing with the campaign-LUR model [ρ = 0.54]. Although the predicted NO2 concentrations from each model were correlated, the LUR models based on routine networks provided better visual representations of the local road conditions in the city.Conclusion: The present study demonstrated that LUR models based on routine data could estimate local traffic-related air pollution in an urban area. The importance and usefulness of data from routine monitoring networks should be acknowledged.