Rates of vertical mixing, gas exchange, and new production : estimates from seasonal gas cycles in the upper ocean near Bermuda

Abstract

Argon measurements, obtained from three years of monthly detailed vertical profiles near Bermuda (Station S, 32°N 64°W), show a maximum in argon supersaturation of about 4% in the seasonal thermocline in late summer. Since the argon supersaturation is 3-4 times smaller than that of oxygen, most of the oxygen supersaturation is not of physical origin and hence must result from biological production. In the winter mixed layer, air injection produces argon supersaturation despite high gas exchange rates. During spring and summer, radiative heating, air injection, and an upward argon flux create an even larger supersaturation in the mixed layer. In the seasonal thermocline, radiative heating creates argon supersaturations that persist in spite of vertical mixing. The observed seasonal cycles of temperature, argon, helium, and oxygen are simulated with an upper ocean model. I linearize the model's response to variations in vertical diffusivity, air injection, gas exchange rate, and new production and then use an inverse technique (singular value decomposition) to determine the values of these parameters that best fit the data. Results for the 1985-1987 average are as follows: A vertical turbulent diffusivity of 1.0 ± 0.1 x 10-4 m2 s-1 is consistent with both the thermal history and subsurface argon distribution. The rate of air injection, determined to ±15%, is similar to previous estimates. The seasonally-averaged gas exchange rate, determined to ± 11%, is consistent within errors with that predicted by Liss and Merlivat (1986). I estimate a lower limit to depth-integrated new production below the mixed layer of 5.0 ± 1.0 moles 0 2 m-2 yr-1 , and obtain an estimate of 6.2 ± 0.9 moles 0 2 m-2 yr-1 if new production in the mixed layer is fixed at zero. The period 1985-1987 appears to be typical of the climatological mean conditions at Station S and comparable to the 1960-1970 average period analyzed by Jenkins & Goldman (1985) and Musgrave et al. (1988). I propose that a mesoscale anticyclonic eddy is responsible for excess 3He and nitrate in the euphotic zone observed at a July, 1986 occupation of the Station S site. Hydrographic profiles are consistent with a type of eddy observed by Brundage & Dugan (1986), characterized by an unusually thick lens of subtropical mode (18°C) water. Analysis of the 35 year hydrographic record suggests that such eddies may arrive at Station S with an average frequency of 2-6 times per year, mostly during the summer and in years of vigorous 18°C water formation. Their timing and character suggest that they may be formed during winter convection events in the northeastern Sargasso Sea, advected southwestward by the gyre-scale circulation, and eventually absorbed by the Gulf Stream. Their magnitude and frequency indicate that they may supply a significant portion of the 3He and nutrient flux into the euphotic zone near Bermuda, and suggest a mechanism by which newly formed subtropical mode water is incorporated within the gyre interior. However, enhanced new production in such eddies could account for only a small portion of the new production integrated over the Sargasso Sea.