The Application of Pumping Induced Ebullition Based Carbon Dioxide Measurements in Northern Wisconsin Lake Studies

Abstract

Few lake studies have used a method that directly measures dissolved carbon dioxide (CO2). Direct measurements are dissolved carbon dioxide estimates obtained with an analyzer or sensor, such as the headspace analysis method that uses gas collection syringes and an ion chromatography CO2 analyzer (Cole and Caraco, 1998). In lake studies summarized in Sobek et al. (2005) for the period 1968 to 2005, the vast majority of estimates of dissolved carbon dioxide were indirectly derived from measurements of pH and dissolved inorganic carbon (DIC) or pH and acid neutralizing capacity (ANC). Traditional direct measurements are difficult to obtain. However, indirect measurements can under or overestimate the carbon dioxide concentration due to uncertainties associated with pH, DIC and ANC. Recently, a new field based technology called pumping induced ebullition (PIE) was developed for the direct measurement of carbon dioxide (Browne, 2004a). This approach simplifies direct carbon dioxide measurements by using a process that mechanically induces ebullition of a dissolved gas in a water sample, traps the harvested gas within a collection chamber, and measures this harvested gas with an infrared CO2 gas analyzer (IRGA). This study explored the potential of this new technology for the collection of field based carbon dioxide data. The accuracy and precision of PIE based measurements of carbon dioxide were characterized using solutions of defined composition. PIE based carbon dioxide measurements were also made in lakes across three trophic classes (eutrophic, mesotrophic, oligotrophic). These results were compared to traditional direct (headspace) and indirect (pH, ANC; pH, DIC) carbon dioxide measurements. Carbon dioxide measurements made using PIE showed high precision and accuracy for defined carbon dioxide solutions from 0 to 100 μmol/liter DIC and 0 to 2610 μatm PCO2. Linear regressions between PIE based [H2CO3 *] and the sodium bicarbonate concentration added to the solutions showed a close relationship with r2 values of 1.00, RMSE ranging from 0.57 to 1.99 μmol L-1 [H2CO3*], and coefficients of variation ranging from 1.29 to 4.32 %. PIE’s trend with the sodium bicarbonate concentration was also close to the one-to-one line, with slope values ranging from 0.94 to 0.98. PIE based PCO2 also proved to be a precise, accurate, and reliable measurement for surface analyses of lake bodies ranging from zero to 2314 μatm PCO2 and for depth analyses of lake bodies ranging from 450 to 4250 μatm PCO2. The PIE method substantially increases the ability to measure carbon dioxide variations in time and space. In a continuous twenty-four hour diurnal pattern, PIE based PCO2 and DIC measurements showed the actual respiration and photosynthesis of the lake from day to night and back to day at temporal resolutions inaccessible via traditional indirect and other direct sampling methods. PIE based PCO2 and DIC whole lake maps also showed spatial variation of dynamic lake systems at rarely documented spatial resolution. PIE gives researchers direct measurement methods of known quality (accuracy and precision), spatial resolution, and temporal resolution for studying dissolved carbon dioxide, which were inaccessible via traditional sampling methods (indirect and other direct carbon dioxide measurement methods).