Land Use Regression Model and Spatial Variation of Oxidative Potential in the Netherlands

Abstract

Land Use Regression Model and Spatial Variation of Oxidative Potential in the NetherlandsAbstract Number:2166 Aileen Yang*, Meng Wang, Daan Leseman, Bert Brunekreeft, Flemming R Cassee, Nicole AH Janssen, and Gerard Hoek Aileen Yang* National Institute for Public Health and the Environment (RIVM), Netherlands, E-mail Address: [email protected] Search for more papers by this author , Meng Wang Institute for Risk Assessment Sciences, Division Environmental Epidemiology, Utrecht University , Netherlands, E-mail Address: [email protected] Search for more papers by this author , Daan Leseman National Institute for Public Health and the Environment (RIVM), Netherlands, E-mail Address: [email protected] Search for more papers by this author , Bert Brunekreeft Institute for Risk Assessment Sciences, Division Environmental Epidemiology, Utrecht University, Netherlands, E-mail Address: [email protected] Search for more papers by this author , Flemming R Cassee National Institute for Public Health and the Environment (RIVM), Netherlands, E-mail Address: [email protected] Search for more papers by this author , Nicole AH Janssen National Institute for Public Health and the Environment (RIVM), Netherlands, E-mail Address: [email protected] Search for more papers by this author , and Gerard Hoek Institute for Risk Assessment Sciences, Division Environmental Epidemiology, Utrecht University, Netherlands, E-mail Address: [email protected] Search for more papers by this author AbstractBackground Recently, oxidative potential (OP) of PM has been proposed as a more health-relevant indicator than PM mass concentrations as it better incorporates the biologically relevant properties of PM. Land use regression (LUR) models are increasingly used to estimate long-term exposure to air pollution in epidemiological studies, but to date few have been developed for OP.Methods Three two-week PM2.5 samples were collected at 40 sites spread over the Netherlands/Belgium in a one-year period. The sites included 10 regional background, 18 urban background and 12 street sites. Two a-cellular OP methods electron spin resonance (OPESR) and dithiothreitol (OPDTT) were applied. LUR models were developed on annual average concentrations and a range of land-use and traffic related GIS variables.Results Mean [min, max] annual average concentrations for OPDTT and OPESR were 1.31[0.68, 2.61] nmol DTT min-1/m3 and 1150 [496, 2228] A.U./m3 respectively. OP was highest at the street sites and lowest at the regional background sites. Street/urban site ratio was 1.18 and 1.38 for OPDTT and OPESR respectively, while regional/urban background ratio were 0.82 for both. OPESR was moderately correlated with OPDTT (Pearson’s R=0.59), but showed higher correlations with Fe, Cu, PM2.5 mass concentrations and PM2.5 absorbance (R=0.70-0.80). LUR models could be developed for both OP methods, with an R2 of 0.64 and 0.67 for OPDTT and OPESR respectively and cross validation R2 >0.50 for both models. Models included 3 to 5 predictors, with regional estimates as the common predictor. OPESR included traffic load and population density, while OPDTT had nearest major road, traffic intensity on the major road, road length and natural land as additional predictors.Conclusions High spatial contrasts were found for both OPDTT and OPESR. LUR models for OP explained more than 60% of the spatial variations within our study area.