| Long-term exposure to ambient pollutant could cause adverse health effect. Significant scientific evidences demonstrated that the increasing mortality and morbidity of respiratory and cardiovascular disease are associated with the severe ambient air pollution. In addition, the evidence of epidemiology studies showed that contrast of exposure to air pollution within-city is larger than the contrast between-cities. Therefore, health effects related to long-term exposure within city scale plays an important role among all the environmental epidemiology studies associated with air pollution in the world.Currently, land-use regression (LUR) model is one of the most popular utilized models. Obtaining high quality concentrations of pollutants is the principle basis for the model establishment. Hence, selecting an efficient, low-cost and multi-pollutants feasible monitoring approach is the key step. Although passive samplers have been used in most of studies, this approach features a number of limits, such as it cannot sample particle matters (PM), and it requires large amount of sample time for the chemical analysis. Comparatively, online monitoring is an alternative way to solve the limitation of multiple-pollutant and time-free, but it is very expense in case of lots of instruments are necessarily used in a small-scale area study. Under this circumstance, achieving a suitable monitoring, which is simple, realizable, also could represent the long-term exposure concentration level, is the essence for the intra-urban exposure assessment. Moreover, studies on the health effect of air pollution exposure assessment in China stared in recent. Only a few cities have been reported with one or two pollutants using a single approach. It is meaningful to further explore much more methodologies associated with the exposure assessment among cities in China.This study focuses on the assessment of long-term exposure to ambient pollutant within city area as well as the corresponding health effects. Firstly, the ambient pollutant data monitored using two different methodologies in four United States (US) cities are adopted to build the LUR model together with the combined model of LUR and Universal Kriging (UK). Predictions of the participant locations in the MESA cohort from the models will be applied for the health evaluation. Monitoring campaign of each city was conducted in the summer and the winter of the studied year, respectively. Two types of ozone (O3) and nitrogen oxide (NOx) concentration were obtained to represented daily average level and the peak hour average level, respectively, considering adopting the both passive sampling and mobile monitoring. The models were established using each type of pollutant concentration to demonstrated the characteristics of the pollutant spatial distribution at various time with the corresponding performance in the model. Secondly, the association between PM2.5 and the light scattering coefficient in Shanghai had been investigated, and the O3 concentration characteristics in three typical districts of Shanghai were observed as well. The exposure assessment model related to PM2.5, O3 and nitrogen dioxide (NO2) with geographic variation of Shanghai are build based on the identical methodology.The main results obtained from this study are given as follows:(1) In four US cities, passive badge data can be used successfully to build the well-performing LUR+UK exposure prediction models for NO2 and NOx, but cannot be used to build robust exposure model for O3. Mobile monitoring data can be used successfully to build prediction model for NOx and O3, R2 of cross-validation of successful model for both NOx and O3 were within the range of 0.5-0.9. The daily average total oxidants (NO2+O3) concentration cannot be successfully build models as well.(2) Using six well-built model, including two types of monitoring methods for NOx only mobile monitoring data for O3 in two seasons, we predicted the exposure concentration of the participant locations. With controlling confounders (age, gender, race, site, smoke status, education levels, scan voxel), only the ozone concentration during peak hours in the summer time have significant association with the increase of the emphysema percentage. On the contrary, NOx in the winter had significant negative association with the percent emphysema. The remaining pollutant concentrations of other models showed no significant association with the percent emphysema.(3) PM2.5 and light scattering coefficients had strong association for the Shanghai study. It showed that the values of the light scattering coefficients were higher in winter but lower in summer. It was obtained from the calculation of multiple regressions, EC, NO3-, NH4+ were primarily components, impacting on extinction coefficient significantly. According to the IMPROVE formulation, we separated OC into hygroscopicity and non-hygroscopicity parts, and considered ion concentration of sea salt. Consequently, the estimation of bext values was much closer to the monitoring values. OC, EC and ammonium sulfate salt were the main functional components for the estimation of the scattering coefficients.(4) Annual PM2.5 and O3 concentrations of at regular monitoring sites of Shanghai cannot be used to establish the well-performed LUR +UK model. However, NO2 can be used successfully to build the LUR+UK model. Cross validation R2 of LUR model was 0.33 and sharply improved to 0.85 by the UK model. The failure of PM2.5 was mainly due to the source of PM2.5 in Shanghai is complex, in which, fossil fuel and biomass burn account for a large portion of the sources. The traffic related geographic information is not adequate to explain PM2.5. |
PDF Full Text Request