Air pollution assessment using land-use regression model

Abstract

Air pollution exposure assessment is the most important step in epidemiological studies for determining the relationship between air pollution and human health effects. One of new collaborative research project in Europe is the Human Early-Life Exposome (HELIX) and the most important part of this project is air pollution exposure assessment. In order to maximize the accuracy of exposure assessment the validated and accurate methods of this process need to be used. The land-use regression (LUR) model is widely applied to personal air pollution exposure studies. LUR model is based on multiple regression equations, which are used to describe the relationship between measured pollutant concentrations and potential predictor variables computed, using geographic information system (GIS). The aim of this study was to assess the air pollution of nitrogen dioxide (NO2), nitrogen oxides (NOx), and particulate matter (PM2.5, PM10) in Kaunas city using LUR model. The measurements of nitrogen oxides were carried out in 40 sites and particulate matter - in 20 sites in study area from the ESCAPE project. GIS predictors were used to develop LUR models for each air pollutant separately. Raster maps of the study area were created to assess the exposure of air pollution. LUR modelling results showed that annual average variation of nitrogen dioxide and nitrogen oxides concentrations ranged from 8.8 to 66.4 jg/m3 and from 10.9 to 102.0 jg/m3. The average annual concentration of particulate matter (PM2.5, PM10) varied from 15.6 to 36.2 jg/m3 and from 23.5 to 45,0 jg/m3 in Kaunas city. The average annual values of NO2 and NOx were 15.4 and 25.2 jg/m3and the concentrations of particulate matter (PM2.5, PM10) were 18.7 and 26.9 |jg/m3