Document Type


Publication Date

Fall 8-18-2003


Carbon dioxide is the most significant greenhouse gas other than water vapor and because of its contribution to global warming there has been an intense effort over the past decade to understand the global carbon cycle. The terrestrial biosphere provides large sources and sinks for CO2 and thereby strongly affects the atmospheric burden and rate of change of CO2 mixing ratio. However attempts to account for the distribution of CO2 in the atmosphere have uncovered a “missing” sink, most likely located in mid-latitude forests. Quantifying this sink has been hindered by a lack of flux measurements over land and uncertainties in the methods available. In this work, a new method for measuring vertical profiles of CO2 within and above the atmospheric boundary layer was developed. It involves sampling into Tedlar® bags and analysis by non-dispersive infrared (NDIR) spectroscopy. This method was demonstrated to have a precision of 0.06% and accuracy of 0.04%. The new method was used to measured CO2 profiles at two different sites; a forest near Park Falls, WI and a grassland site near Ponca City, OK. Both sites are located near CO2 flux towers, allowing intercomparison of the two methods. Both seasonal and regional variations in CO2 concentration were observed, with an average C02 difference between spring and fall of 4.84±1.86 ppm in Ponca City and 6.23±1.52 ppm in Park Falls. Data collected during the field studies was used to calculate CO2 surface fluxes by a budget method. Calculated average daily surface fluxes during the summer were -0.38±0.18 ppm m/s and -0.17±0.08 ppm m/s in Park Falls and Ponca City respectively and 0.06±0.19 ppm m/s and -0.03±0.09 ppm m/s during fall. These results follow a general trend of increasing downward flux with increasing photosynthesis. Our surface fluxes were compared to the surface fluxes measured by eddy covariance at the nearby flux towers and the difference between the two is generally less than 0.3 ppm m/s.