Description of the method for apportioning dioxin and furan groups to individual congeners.

Table S1. Surrogates for apportioning county-level area source emissions to census tracts.

Table S2. Background air concentrations determined from literature (MNRiskS) or from monitoring and emissions data (NATA).

Table S3. Summary statistics of the model monitor comparison. Modeled mean values are highlighted if the mean difference between model and monitor is not significantly different from zero (p < 0.05). Fractional bias values between -0.67 and +0.67 (within a factor of 2) are highlighted.

Table S4. Summary statistics for regression analyses relating models to monitor and relating models to one another.

Figure S1. Boxplots of MNRiskS-ISC cancer risk and noncancer hazard indices by source category across all modeled receptors for the resident and farmer scenarios.

Figure S2. Boxplots of MNRiskS-ISC cancer risk and noncancer hazard indices for high risk pollutants for the resident, farmer, and fisher scenarios.

Figure S3. Boxplots of MNRiskS-ISC cancer risk and noncancer hazard indices by scenario for the case including all sources and all pollutants.

Figure S4. Benzene emissions breakdown by mobile and area source subcategories

Figure S5. Frequency distributions of benzene inhalation cancer risks using a) MNRiskS-ISC and b) NATA (MNRiskS-AERMOD was similar to MNRiskS-ISC). The distribution is clearly bimodal, with peaks in the distribution at high values and at low values, corresponding to urban and rural locations, respectively. The NATA frequency distribution has a peak at high values corresponding to urban census tracts but not at low values due to the relatively lower number of model receptors in rural areas in NATA.

RISA_1640_sm_SuppMat.doc796KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.