We did not adjust for
BC in the two-pollutant models because the time period of BC measurement overlapping
with other pollutants was less than 50%. Consequently, simultaneously
entering BC into the models might introduce large uncertainty when the pollutants
involved represent a different time period of data.