Calibrations are necessary for most chemical sensors because the response is not consistent between sensors nor stable over time. If chemical sensors could be designed to have identical behavior from sensor to sensor and no drift, the need for sensor calibrations would be reduced. In the present paper, the feasibility of calibration-free optical chemical sensors is explored. An indicator-based pCO 2 (partial pressure of CO 2 ) sensor is designed that has excellent sensor-to-sensor reproducibility and measurement stability. This superior level of performance is achieved by using the following strategy:(1) renewing the sensing solution, (2) allowing the sensing solution to reach equilibrium with the analyte, (3) calculating the response from a ratio of the indicator solution absorbances, and (4) through careful solution preparation, wavelength calibration, and stray light rejection. Three pCO 2 sensors are calibrated, and the response curves are essentially identical within the uncertainty of the calibration. Long-term laboratory and field studies are presented that show the response has no drift over extended periods (months). The theoretical response, determined from thermodynamic characterization of the indicator solution, also predicts the observed calibrationfree performance. Other absorbance-based sensors, such as optrodes, can be designed and operated in a similar fashion, making calibration-free optical chemical sensors available for a wide range of biomedical, industrial, and environmental applications.

We present and compare direct and indirect pCO 2 observations taken in Lake Superior in the last decade and use them to understand temporal and spatial variability in lake carbon cycle processes. In situ observations from 2001 and biannual survey data for 1996-2006 indicate that Lake Superior was, on average, supersaturated (annual mean 5 46.7 6 17.3 Pa [461 6 171 matm]) with respect to atmospheric pCO 2 (mean 5 38.3 6 0.6 Pa) in April and close to equilibrium (mean 5 37.5 6 6.7 Pa) with respect to atmospheric pCO 2 (mean 5 36.4 6 0.7 Pa) in August. Both data sets indicate that temporal variability in surface lake pCO 2 from weekly to interannual timescales was predominantly controlled by changing dissolved inorganic carbon and associated changes in pH. An unstratified water column appears to have limited pCO 2 fluctuations in spring. Through summer and into early fall, pCO 2 variability on a daily timescale at 12 m increased with time to a maximum amplitude of 19 Pa, likely as a result of internal waves on the thermocline. Year-to-year changes in mean surface lake pCO 2 and temperature were of the same sign and approximate magnitude at all observed points, consistent with the lake's small size relative to the synoptic-scale meteorological systems that force it. Variability in pCO 2 was not correlated with major climate indices. While these data provide a first large-scale overview of Lake Superior's pCO 2 and its temporal variability, their time-space resolution and accuracy are not sufficient to further refine previously imbalanced lake-wide carbon budgets.

Sensors for the partial pressure of CO 2 (pCO 2 ) and dissolved O 2 (DO) were deployed near the surface and bottom of a freshwater lake (Placid Lake, Montana) during the period from ice cover to seasonal stratification. Sources of variability were examined using one-dimensional physical and biogeochemical models. Model predictions for pCO 2 and DO were compared to further constrain model parameters. A number of transient processes were documented that have not been well characterized in previous studies. The models made it possible to link these short-term events to specific forcings. We found that (1) 11 d of the 13-d turnover period occurred under ice through lightdriven convective mixing, (2) phytoplankton biomass increased to its highest seasonal level under ice, (3) weak stratification set up immediately after ice-out, causing bottom water pCO 2 and DO to diverge from surface levels, (4) subsequent diel convective mixing brought bottom pCO 2 and DO back toward surface levels, and (5) before stable stratification, vertical entrainment of CO 2 -rich water, net production, and air-water exchange drove 100-200 atm daily changes in pCO 2 , but, because of their counterbalancing effects, surface pCO 2 remained Ͼ1,000 atm for nearly 1 month after ice-out. Upon stable stratification, net production and air-water exchange overcame pCO 2 gains from mixing and heating and reduced pCO 2 to near atmospheric levels within 20 d. Net production and gas exchange accounted for ϳ75% and 25%, respectively, of the decrease in surface pCO 2 observed after ice-out. Diel convection was the dominant mixing process both under ice and after ice-out and may be an important underrepresented mechanism for CO 2 and DO exchange between surface and bottom water.Spring thaw heralds not only a new growing season but also an important transitional period for ice-covered lakes and reservoirs. Physical forcings change dramatically between ice cover and open water, creating large and rapid changes in lake biology and geochemistry. As ice and snow cover diminishes, light penetration can stimulate phytoplankton growth (e.g., Wright 1964) and convective mixing (Bengtsson 1996). Surface water temperatures Ͻ4ЊC continue to warm and sink or, after ice-out, to be mixed by winds to greater depths, which results in an isothermal water column. Nutrients accumulated over winter combined with increased insolation produce the characteristic spring phytoplankton bloom. With further solar heating, the water column becomes thermally stratified, isolating the bottom water from the atmosphere until turnover occurs again, which is usually in the fall for dimictic lakes. The duration and magnitude of ice cover, turnover, and stratification can dramatically influence the concentrations of biogeochemical species both seasonally and interannually (e.g., Cornett and

On average, the water column of Lake Superior is undersaturated with respect to dissolved oxygen and supersaturated with respect to carbon dioxide during the summer-stratified period. On the basis of temporal changes in water column dissolved oxygen, we calculate rates of oxygen consumption that range from 0.19 to 0.75 mmol m. These rates are a factor of 5-10 times larger than can be supported by the particulate carbon settling rates and benthic oxygen consumption rates. In addition, on the basis of the limited information available, dissolved allochthonous carbon inputs are insufficient to account for the calculated rates of carbon oxidation. Rates of nitrate and total CO 2 (⌺CO 2 ) production are 0.019 Ϯ 0.012 and 0.13 Ϯ 0.06 mmol m Ϫ3 d Ϫ1, respectively, and are consistent with the oxidation of a dissolved organic component that is similar in composition (C : N ratio) to the settling particulate material. Previously published estimates of total primary production were smaller but similar in magnitude to our integrated water column respiration rates. We interpret the observed imbalance between particulate carbon delivered to the deep lake and the calculated rate of carbon oxidation to be the result of the decomposition of dissolved organic carbon that appears to have both an autochthonous and an allochthonous component.A simplistic view of aquatic ecosystem carbon cycling is that photosynthetic production, limited by the availability of one or more major nutrients, generates a large pool of fixed (autochthonous) carbon in the euphotic zone. Most of that organic material is ''recycled'' within the euphotic zone with some, typically small, fraction being ''exported'' out of the euphotic zone. A portion of this export production is then remineralized within the deep water column or within the sediments, and the balance is permanently buried (e.g., Dymond et al. 1996 and references therein). The record of this buried ''residual'' production serves as the basis for paleoclimate studies. For these studies, variations in the accumulation rate of biogenic material in lacustrine systems are 1 Corresponding author (mcmanus@coas.oregonstate.edu). Present address: College of Oceanic and Atmospheric Sciences, 104 Ocean Admin. Bldg., Corvallis, Oregon 97331-5503. AcknowledgmentsSpecial thanks are extended to the Co-PIs on the project, Elise Ralph and Robert Sterner. Elise Ralph kindly processed the conductivity-temperature-depth data and provided an initial draft of the map. Chris Moser helped with much of the hard work on deck; his dedication and perseverance are greatly appreciated. Jason Agnich, Angela Cates, Bronwen Cumberland, Phillip Hommerding, Keely Johnson, Melissa Jones, Julie Klejeski, Kimberly Kolbeck, and Christina Willie all helped in the laboratory or in the field. Mike DeGrandpre (Univ. MT) provided the instrumentation for the comparative pCO 2 data. Simone Alin and Tom Johnson provided insightful comments on an early draft of this manuscript. Conversations with Jon Cole and Jim Cotner helped clarify...