Land use regression (LUR) models have been widely used to provide long-term air pollution exposure assessment in epidemiological studies. However, models have rarely offered variables that account for the dispersion environment close to the source (e.g., street canyons, position and dimensions of buildings, road width). This study used newly available data on building heights and geometry to enhance the representation of land use and the dispersion field in LUR. Models were developed for PM10, NO(X), and NO2 for 2008-2011 for London, U.K. A separate set of models using "traditional" land use and traffic indicators (e.g., distance from road, area of housing within circular buffers) were also developed and their performance was compared with "enhanced" models. Models were evaluated using leave-one-out (n - 1) (LOOCV) and grouped (n - 25%) cross-validation (GCV). LOOCV R(2) values were 0.71, 0.50, 0.66 and 0.73, 0.79, 0.78 for traditional and enhanced PM10, NOX, and NO2 models, respectively. GCV R(2) values were 0.71, 0.53, 0.64 and 0.68, 0.77, 0.77 for traditional and enhanced PM10, NO(X), and NO2 models, respectively. Data on building volume within the area common to a 20 m road buffer within a 25 m circular buffer substantially improved the performance (R(2) > 13%) of NO(X) and NO2 LUR models.