Table 1. Description of the Dutch LUR models.

Pollutanta	LUR modelb
Predictors with subscripted radii (m) are variables corresponding to the accumulated attribute value within the given circular buffer centred at the cell under consideration. Each model is calculated for each raster cell in the Netherlands.
aConcentration of PM2.5absorbance is given in 10−5 m−1, the residual pollutant concentrations are given in μg m−3. bRegional background concentration estimates in μg m−3 for NO2 (BEO), NOx (BEX), and PM2.5 (BEP), and in 10−5 m−1 for PM2.5absorbance (BEA); traffic load (vehicles·day−1·m) on all roads (TL) and major roads (TML); heavy traffic load (vehicles·day−1·m) all roads (HTL); road length (m) of all roads (RL) and major roads (MRL); inverse distance (m−1) to all roads (IDC) and major roads (IDM); population density (inhabitants·m−2, POP); the surface area (m2) of industrial area (IND), harbour area (HAR), and residential area (RES).
NO2	−7.8 + 1.18·BEO + 2.3e−05·POP5000 + 2.47e−06·TL50 + 1.06e−4·RL1000 + 9.84e−05·HTL25 + 12.19·IDC + 4.47e−07·HTL25−500
NO2background	3.21 + 0.74·BEO + 2.29e−05·POP5000 + 6.4e−07·IND5000 + 4.72e−07·HAR5000
NOx	3.25 + 0.74·BEX + 4.22e−06·TL50 + 6.36e−04·POP1000 + 2.39e−06·HTL500 + 71.65·IDM + 0.21·MRL25
PM2.5	9.46 + 0.42·BEP + 0.014·MRL50 + 2.28e−09·TML1000
PM2.5absorbance	0.07 + 2.95e−09·TL500 + 0.0029·MRL50 + 0.85·BEA + 7.90e−09·RES5000 + 1.72e−06·HTL50
PM10	23.71 + 2.16e−08·TML500 + 6.68e−06·POP5000 + 0.015·MRL50