Accounting for spatial effects in land use regression for urban air pollution modeling

Abstract

In order to accurately assess air pollution risks, health studies require spatially resolved pollution concentrations. Land-use regression (LUR) models estimate ambient concentrations at a fine spatial scale. However, spatial effects such as spatial non-stationarity and spatial autocorrelation can reduce the accuracy of LUR estimates by increasing regression errors and uncertainty; and statistical methods for resolving these effects--e.g., spatially autoregressive (SAR) and geographically weighted regression (GWR) models--may be difficult to apply simultaneously. We used an alternate approach to address spatial non-stationarity and spatial autocorrelation in LUR models for nitrogen dioxide. Traditional models were re-specified to include a variable capturing wind speed and direction, and re-fit as GWR models. Mean R(2) values for the resulting GWR-wind models (summer: 0.86, winter: 0.73) showed a 10-20% improvement over traditional LUR models. GWR-wind models effectively addressed both spatial effects and produced meaningful predictive models. These results suggest a useful method for improving spatially explicit models.

Comparison of predicted vehicular pollution concentration with air quality standards for different time periods

“…This increasing transportation load is making it difficult to manage the health and safety of society (Pawar and Patil 2015). Long-term exposure to air pollution leads to a variety of adverse health effects including respiratory, cardiovascular, developmental, reproductive, gastrointestinal, and neurological health outcomes (Bertazzon et al 2015). Vehicular sources have severe effect on human health and environment as these are ground level source (Amundsen et al 2008;Cheng et al 2013;Fan et al 2012;Fenger 2009;Kumar et al 2016a;Nagendra and Khare 2002;Neema and Jahan 2014;Province et al 2013;Sharma et al 2004;Sivacoumar and Thanasekaran 1999).…”

Mining Public Datasets for Modeling Intra-City PM2.5 Concentrations at a Fine Spatial Resolution

“…In Hoek et al [26], their RMSE ranged from 1.6 to 9.8 (g/m 3 ) for various types of air pollutants in study areas across the globe, and their temporal resolutions are commonly low (e.g., seasons). Moreover, LUR models rarely deal with spatial effects (e.g., spatial non-stationarity) except a more recent study that built a wind model to improve traditional LUR and had a 10-20% improvement on the prediction [1]. …”

“…These air monitoring stations also exist in many other countries. Scientists and government agencies use measurement data from these stations to build and validate air quality models (AQMs) to explain and predict the past and future air pollution levels for unmonitored locations (e.g., [1, 3, 13, 15, 16, 19, 20, 24, 25, 28, 29]). Predictions from these models can then be used to study the associations between long-term air pollution exposure and health impact at finer spatial scales (than simply using the monitored data) [26, 27].…”

Land Use Regression Models Using Satellite Aerosol Optical Depth Observations and 3D Building Data from the Central Cities of Liaoning Province, China

“…In this study, 1-km MODIS monthly vegetation indexes products (MOD13A3) from 2013 that were obtained from the LAADS website (http://ladsweb.nascom.nasa.gov). (5) Distance to Air Pollutant Point Source Emissions…”

“…The most common predictor variables in previous studies that are of interest to various traffi c descriptions (e.g., intensity, congestion, road length, and distance) are population density, housing, land use, physical geography (e.g., elevation and distance to the sea), and meteorological data [2][3][4]. Further developments of LUR modeling have included greater focus on transferable models and predictor variables such as wind speed, wind direction, and emission data; further development of models were considered [5,6].…”