Comparison of microenvironmental CO concentrations in two cities for human exposure modeling

The microenvironmental components of the CO concentration in two cities are compared by subtracting the ambient background concentration from personal exposures measured in Denver, Colorado, and Washington, DC. Two surrogate measures for the ambient background concentration are tested. Both improve the similarity of the means in the two cities, but the Denver standard deviations are higher than those in Washington, DC. Microenvironments containing the internal combustion engine have both higher means and standard deviations in Denver compared with Washington, DC. The Washington, DC, mean concentration for automobiles, for example, was 59% of the Denver mean (2.0 ppm versus 4.9 ppm). Washington, DC, had approximately 57% of the Denver emissions, and the difference in mean CO concentrations is roughly consistent with the lower emissions in Washington, DC, due to lower elevation. A surprising finding is that mean CO exposure levels caused by cooking with gas stoves in Washington, DC, were only 58% of the levels in Denver (1.9 ppm and 3.3 ppm, respectively). This result suggests that elevation may exert an influence on gas stove emissions that is similar to its influence on internal combustion engines. Using an averaging time model, analysis of the autocorrelation of sleeping and office microenvironments suggests that considerable serial dependency exists. The microenvironmental data and findings in this paper have important implications for constructing human exposure-activity pattern models. For future human exposure field studies, the findings emphasize the importance of measuring background values in a location that is extremely close to each microenvironment studied.