ATMOSPHERIC OXYGEN-EXCHANGE IN THE HUDSON RIVER - DOME MEASUREMENTS AND COMPARISON WITH OTHER NATURAL-WATERS

The measurement of metabolism using diel free-water oxygen techniques requires the estimation of atmospheric oxygen exchanges. We measured s uch exchange on nine different occasions in the freshwater, tidally-in fluenced Hudson River estuary using a floating dome technique. We also analyzed previously published data on the exchange of a variety of ga ses measured in lakes, estuaries, and open ocean waters using a wide v ariety of techniques. Data were expressed as a ''transfer velocity'' a nd normalized to an exchange of oxygen at 20 degrees C. Considered tog ether, these data indicate a significant predictive relationship when the natural log of transfer velocity is regressed with measured wind s peed (r(2) = 0.55; p = 0.0001). The influence of wind was particularly pronounced in estuaries and in lakes. Data from open ocean waters sho wed much less influence of wind, probably because surface turbulence i n these deeper waters can be temporally and spatially decoupled from w ind. Our Hudson data agreed well with data collected in other systems. In general, data from estuaries-including the hudson-indicated slight ly higher transfer velocities at any given wind speed than do data fro m lakes (although this difference was less pronounced for our Hudson d ata than for other estuaries). The difference may result from some int eraction of wind and tidal currents, or it may reflect a bias in the d ome method of measurements; all of the estuarine data were collected u sing the dome approach, while the majority of the lake data were deter mined using an added tracer. If the dome method actually gives a biase d, high estimate of oxygen flux, this is in contradiction to previous criticisms of this method that domes may underestimate fluxes by block ing wind at the water surface. We have used the regression of the natu ral log of transfer velocity versus wind speed developed here to estim ate respiration in the Hudson estuary from diel changes in dissolved o xygen. To allow for possible biases in technique and for measurement e rror, we estimated 95% confidence limits around the regression. Estima tes of respiration in the Hudson determined using the upper and lower 95% confidence limits are 30% higher and 12% lower than that determine d when using the best-fit regression. An independently-constrained car bon budget for the tidally-influenced, freshwater Hudson River estuary indicates that respiration rates cannot be much higher than our mean estimate as calculated using the linear regression of the gas transfer and wind data to correct for air-water oxygen exchange. Gas transfer in natural systems is difficult to measure and is controlled by many i nterrelated physical factors. In the absence of extensive, system spec ific held studies, the regression presented here should be useful in e stimating atmospheric oxygen exchange in other estuarine or riverine e cosystems which are relatively deep and wide.