Combining Eddy Covariance Fluxes, High-Precision Trace Gas Measurements, Chemical Transport Modeling, and Inverse Modeling to Estimate Regional CO2 Fluxes in the Southern Great Plains, USA

We use the radon tracer method to estimate monthly average net ecosystem exchange (NEE) of carbon dioxide in the Southern Great Plains of the USA for the year 2007. These estimates are compared with optimized flux estimates of NEE from NOAA CarbonTracker, sampled with a Lagrangian particle dispersion model to identify the upwind area influencing each measurement. The radon-tracer equation is very simple: F(CO2) = F(Rn) x delta(CO2)/delta(Rn), where F(CO2) is NEE, F(Rn) is the flux of radon out of the soil, and delta(X) is the discrepancy between a measured concentration and "background" levels, caused by near-field terrestrial fluxes. The F(Rn) term presents a challenge in applying the radon-tracer method, since neither its mean value nor seasonality are well known. We present two lines of evidence to help constrain the seasonal cycle of radon flux in the Southern Great Plains during 2007 (a year with record high early-summer soil moisture), increasing our confidence in applying the radon-tracer method. The first line of evidence comes from a comparison between the observed seasonal cycle of radon at ARM-CART SGP and simulations made by 14 global transport models as part of the "Transcom4" experiment which assumed constant radon emissions. The second line of evidence uses the radon-tracer approach in reverse, using measured eddy covariance CO2 fluxes and nighttime accumulation of CO2 and radon to calculate radon fluxes.