Dataset Information

Can dispersion modeling of air pollution be improved by land-use regression? An example from Stockholm, Sweden.

ABSTRACT: Both dispersion modeling (DM) and land-use regression modeling (LUR) are often used for assessment of long-term air pollution exposure in epidemiological studies, but seldom in combination. We developed a hybrid DM-LUR model using 93 biweekly observations of NO_x at 31 sites in greater Stockholm (Sweden). The DM was based on spatially resolved topographic, physiographic and emission data, and hourly meteorological data from a diagnostic wind model. Other data were from land use, meteorology and routine monitoring of NO_x. We built a linear regression model for NO_x, using a stepwise forward selection of covariates. The resulting model predicted observed NO_x (R²=0.89) better than the DM without covariates (R²=0.68, P-interaction <0.001) and with minimal apparent bias. The model included (in descending order of importance) DM, traffic intensity on the nearest street, population (number of inhabitants) within 100?m radius, global radiation (direct sunlight plus diffuse or scattered light) and urban contribution to NO_x levels (routine urban NO_x, less routine rural NO_x). Our results indicate that there is a potential for improving estimates of air pollutant concentrations based on DM, by incorporating further spatial characteristics of the immediate surroundings, possibly accounting for imperfections in the emission data.

SUBMITTER: Korek M

PROVIDER: S-EPMC5658676 | biostudies-literature | 2017 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Can dispersion modeling of air pollution be improved by land-use regression? An example from Stockholm, Sweden.

Korek Michal M Johansson Christer C Svensson Nina N Lind Tomas T Beelen Rob R Hoek Gerard G Pershagen Göran G Bellander Tom T

Journal of exposure science & environmental epidemiology 20160803 6

Both dispersion modeling (DM) and land-use regression modeling (LUR) are often used for assessment of long-term air pollution exposure in epidemiological studies, but seldom in combination. We developed a hybrid DM-LUR model using 93 biweekly observations of NO<sub>x</sub> at 31 sites in greater Stockholm (Sweden). The DM was based on spatially resolved topographic, physiographic and emission data, and hourly meteorological data from a diagnostic wind model. Other data were from land use, meteor ...[more]

PMID: 27485990

Similar Datasets

Project description:Multiple land use regression models (LUR) were developed for different air pollutants to characterize exposure, in the Durban metropolitan area, South Africa. Based on the European Study of Cohorts for Air Pollution Effects (ESCAPE) methodology, concentrations of particulate matter (PM10 and PM2.5), sulphur dioxide (SO2), and nitrogen dioxide (NO2) were measured over a 1-year period, at 41 sites, with Ogawa Badges and 21 sites with PM Monitors. Sampling was undertaken in two regions of the city of Durban, South Africa, one with high levels of heavy industry as well as a harbor, and the other small-scale business activity. Air pollution concentrations showed a clear seasonal trend with higher concentrations being measured during winter (25.8, 4.2, 50.4, and 20.9 µg/m3 for NO2, SO2, PM10, and PM2.5, respectively) as compared to summer (10.5, 2.8, 20.5, and 8.5 µg/m3 for NO2, SO2, PM10, and PM2.5, respectively). Furthermore, higher levels of NO2 and SO2 were measured in south Durban as compared to north Durban as these are industrial related pollutants, while higher levels of PM were measured in north Durban as compared to south Durban and can be attributed to either traffic or domestic fuel burning. The LUR NO2 models for annual, summer, and winter explained 56%, 41%, and 63% of the variance with elevation, traffic, population, and Harbor being identified as important predictors. The SO2 models were less robust with lower R2 annual (37%), summer (46%), and winter (46%) with industrial and traffic variables being important predictors. The R2 for PM10 models ranged from 52% to 80% while for PM2.5 models this range was 61-76% with traffic, elevation, population, and urban land use type emerging as predictor variables. While these results demonstrate the influence of industrial and traffic emissions on air pollution concentrations, our study highlighted the importance of a Harbor variable, which may serve as a proxy for NO2 concentrations suggesting the presence of not only ship emissions, but also other sources such as heavy duty motor vehicles associated with the port activities.

Project description:BackgroundIndoor air pollution, including fine particulate matter (PM2.5) and carbon monoxide (CO), is a major risk factor for pneumonia and other respiratory diseases. Biomass-burning cookstoves are major contributors to PM2.5 and CO concentrations. However, high concentrations of PM2.5 (> 1000 μg/m3) have been observed in homes in Dhaka, Bangladesh that do not burn biomass. We described dispersion of PM2.5 and CO from biomass burning into nearby homes in a low-income urban area of Dhaka, Bangladesh.MethodsWe recruited 10 clusters of homes, each with one biomass-burning (index) home, and 3-4 neighboring homes that used cleaner fuels with no other major sources of PM2.5 or CO. We administered a questionnaire and recorded physical features of all homes. Over 24 h, we recorded PM2.5 and CO concentrations inside each home, near each stove, and outside one neighbor home per cluster. During 8 of these 24 h, we conducted observations for pollutant-generating activities such as cooking. For each monitor, we calculated geometric mean PM2.5 concentrations at 5-6 am (baseline), during biomass burning times, during non-cooking times, and over 24 h. We used linear regressions to describe associations between monitor location and PM2.5 and CO concentrations.ResultsWe recruited a total of 44 homes across the 10 clusters. Geometric mean PM2.5 and CO concentrations for all monitors were lowest at baseline and highest during biomass burning. During biomass burning, linear regression showed a decreasing trend of geometric mean PM2.5 and CO concentrations from the biomass stove (326.3 μg/m3, 12.3 ppm), to index home (322.7 μg/m3, 11.2 ppm), neighbor homes sharing a wall with the index home (278.4 μg/m3, 3.6 ppm), outdoors (154.2 μg/m3, 0.7 ppm), then neighbor homes that do not share a wall with the index home (83.1 μg/m3,0.2 ppm) (p = 0.03 for PM2.5, p = 0.006 for CO).ConclusionBiomass burning in one home can be a source of indoor air pollution for several homes. The impact of biomass burning on PM2.5 or CO is greatest in homes that share a wall with the biomass-burning home. Eliminating biomass burning in one home may improve air quality for several households in a community.

Dataset Information

Can dispersion modeling of air pollution be improved by land-use regression? An example from Stockholm, Sweden.

Publications

Can dispersion modeling of air pollution be improved by land-use regression? An example from Stockholm, Sweden.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets