S U M M A R Y1. Two pH electrodes and a thermistor were used to record conditions in the surface of Esthwaite Water every 15 min over a 12-month period. Combined with approximately weekly measurements of alkalinity they allowed inorganic carbon speciation to be calculated. 2. Large changes in pH from 7.1 to nearly 10.3, and hence in concentrations of inorganic carbon species, were measured over a year. Carbon speciation and pH varied on a diel, episodic and seasonal basis. Diel variation of up to pH 1.8 was recorded, although typical daily variation was between 0.03 and 1.06 (5 and 95 percentiles). Daily change in concentration of inorganic carbon varied between 4 and 63 mmol m-3 (5 and 95 percentiles). 3. During lake stratification, episodes of high pH, typically of 1-2 weeks' duration were interspersed with episodes of lower pH. These changes appeared to relate to the weather: e.g. low wind velocity, high pressure, low rainfall and high sunshine hours correlated with periods of high pH. 4. Seasonal progression of carbon depletion generally followed stratification and the development of high phytoplankton biomass. When the lake was isothermal, the phytoplankton biomass caused relatively small amounts of carbon depletion. 5. During autumn, winter and spring, the lake had concentrations of COz* (free C02) up to 0.12 mol m 3 which is nearly seven times the calculated atmospheric equilibrium concentration so the lake will accordingly be losing carbon to the atmosphere. In contrast, during periods of elevated pH the concentration of C02* was reduced close to zero and the lake will take up atmospheric CO2. The rates of transfer between water and the atmosphere were estimated using a chemical equilibrium model with three boundary layer thicknesses. The calculations show that over a year the lake loses COz to the atmosphere with the current mean atmospheric level of 360 pnol mol-', at between 0.28 and 2.80 mol m-2 y r l . During elevated pH, rates of CO&dlux increased up to nearly tenfold as a result of chemical-enhancement by parallel flux of H C Q Input of C02* to the lake from the catchment is suggested to be the main source of the carbon lost to the atmosphere. 6. The turnover time for COz between the air and water was calculated to be 1 year for the gross influx and 3.3 years for the net flux. These values are less than the average water residence time of 0.25 years, which indicates that over a year inflow from streams is a more important source of inorganic carbon than the atmosphere. 7. Influx of C 0 2 from the atmosphere was calculated to be roughly equivalent to between 1 and 4% of the rates of production in the water during mid-summer indicating that this source of inorganic carbon is not a major one in this lake. 0 1996 Blackwell Science Ltd 579 580 S. C. Maberly
8.Influx of CO2 from the hypolimnion was estimated on one occasion to be 6.9 10-9 mol m-2 s-l using transfer values based on mass eddy-diffusion. These rates are equivalent to 23% of the rate of influx of CO2 from the atmosphere on this occasion whic...