Concentrations of CO and NOy (total reactive oxidized nitrogen species) were measured simultaneously with many other atmospheric parameters, including SO2 concentration and wind direction, during three field studies (two rural and one suburban). The CO concentrations at the two rural sites can be described approximately by lognormal distributions, with the Scotia, Pennsylvania, site exhibiting higher levels (geometric mean, 214 parts per billion by volume (ppbv); standard deviation of the logarithms, 0.143) than Niwot Ridge, Colorado (geometric mean, 121 ppbv; standard deviation of the logarithms, 0.175). As expected, measurements made at Niwot Ridge when the wind was from the east, the direction of the principal nearby urban area, exhibited the highest CO levels. At Scotia, power-plant-dominated air parcels could be identified by larger 2> to y> ratios and were characterized by smaller to y> ratios. Simultaneous measurements of and y> allowed the identification of air parcels that arrived at Niwot Ridge and Boulder (the suburban site) without significant recent anthropogenic input.

The CO levels in these parcels were consistent with the tropospheric background concentration of CO at 40¿N latitude: approximately 84 ppbv in late summer and 127 ppbv in early spring. The to y> ratios observed in the Boulder suburban study provide a test for, and are consistent with, available emission inventories for these pollutants. At Niwot Ridge in east winds the observed ratios are consistent with the ratios expected from emission inventories, modified by photochemical transformation on time scales that are reasonable for transport from the location of emission to the measurement site. However, the rural data sets were not able to provide stringent tests of emission inventories. ¿1991 American Geophysical Union