Table 6. Overall performance of exposure assessment methods by urbanization.
Correlation coefficient between PM2.5 concentration predictions and observed PM2.5 concentration in μg/m3 stratified by urbanization. The methods considered are: raw CMAQ output, ordinary least squares (“OLS”), inverse distance weighting (“IDW”), universal Kriging(“UK”), downscaler, random forests (“RF”), support vector regression (“SVR”) and Neural networks (“NN”). Methods use either CMAQ and/or other geographic covariates (“Covs”). The last line reports the mean and standard deviation, in parenthesis, for the observed PM2.5 concentration at urban versus nonurban monitoring sites.
| Method | Urban | Non-Urban |
|---|---|---|
| CMAQ | 0.54 | 0.50 |
| OLS (CMAQ) | 0.66 | 0.60 |
| OLS (Covs) | 0.75 | 0.69 |
| OLS (CMAQ+Covs) | 0.68 | 0.63 |
| IDW | 0.87 | 0.77 |
| UK (CMAQ) | 0.88 | 0.80 |
| UK (Covs) | 0.87 | 0.80 |
| UK (CMAQ+Covs) | 0.86 | 0.79 |
| Downscaler (CMAQ) | 0.88 | 0.80 |
| RF (CMAQ + Covs) | 0.74 | 0.67 |
| SVM (CMAQ + Covs) | 0.80 | 0.74 |
| NN (CMAQ + Covs) | 0.79 | 0.73 |
| PM2.5 | 10.16 (6.21) | 9.03 (6.01) |