Surface exchange and transport processes governing atmospheric PCB levels over lake superior

Measurements of gas phase PCB concentrations obtained at a site on the shore of Lake Superior are analyzed to determine the importance of regional air/surface exchange as a determinant of PCB concentration over the lake. During periods of onshore flow, results obtained in congener-specific regression analyses of log concentration against reciprocal temperature are strongly dependent upon time of year. During months when the lake is colder than the overlying air (resulting in a stable atmosphere over the lake), regressions for onshore flow explain 48-63% of the observed variance in log C, and the regression slope and intercept are consistent, with Henry's law equilibrium between the lower atmosphere over the lake and aqueous concentrations. Results consistent with Henry's law are obtained only when water temperature is assumed to be similar to air temperature measured at the site during flow off the lake; poor regression results (r(2) = 0.00-0.06) are obtained when lake water temperature measured by data buoys is used. Based on consideration of the spatial variability of lake skin temperature during onshore flow periods, it is inferred that, under stable conditions, air/water equilibrium is reached rapidly and that atmospheric concentrations at the measurement sire are primarily affected by air/water exchange within 10-30 km offshore. These results are consistent with consideration of potential limitations of regression analyses of log C versus reciprocal tempera ture. Furthermore, they indicate that the net exchange of PCBs during the stable season over large lakes (April-August or September for Lake Superior) may be significantly less than previously estimated. Regression analyses based on samples obtained during onshore flow in the unstable season and during over-land flow exhibit increased scatter and are inconsistent, with the hypothesis of upwind equilibrium controlled by either Henry's law- or vapor pressure-mediated exchange at local temperatures. This result is consistent with rate limitations to atmosphere/surface exchange over the lake and land surface and with the effect of atmospheric mixing processes. Implications of these results to the interpretation of concentration measurements al. lakeside sites and the use of such measurements in calculations of the air/water exchange flux of PCBs over large lakes are discussed.