Soil CO2 flux in a tallgrass prairie

Soils are an important component of the global carbon budget due to their large C storage capacity and the ability to replenish atmospheric C via soil surface CO2 flux. Our objectives were to quantify year-round soil CO2-C fluxes in a tallgrass prairie and to develop an equation to predict flux using soil temperature and soil water content. Soil CO2-flux, soil temperature and soil water were measured on selected days throughout the year from 1993 through 1998, n = 216, at the Blackland Research Center, Temple, Texas, USA. On any date, there was little difference in average daily soil temperatures among years, but there were large differences in soil water content among years that often were related to differences in precipitation totals. Soil CO2-C flux had a seasonal pattern that was more similar to soil temperature than soil water (minimum in the winter and maximum in the early summer). Average annual soil CO2-C fluxes, which were 1.6, 1.3, 1.2, 1.0, 2.1 and 1.5 kg CO2-C m-2 yr-1 in 1993 through 1998, respectively, increased with annual precipitation. Regressed separately, the exponential relationship between soil CO2-C flux and soil temperature accounted for approximately 46% of flux variability while a quadratic relationship between flux and soil water content accounted for 26% of the variability. Both terms were combined into one equation that explained about 52% of the flux variance. Predicted and measured fluxes showed similar patterns throughout the year, there was little bias between predicted and measured fluxes, averages were essentially equal and the root mean square error between measured and predicted fluxes was about 38% of the average flux. The equation accounted for 76% of flux variability of an independent data set from the Konza Prairie in Kansas. The relationship between flux, soil temperature and soil water content should provide accurate predictions of soil CO2 flux in tallgrass prairies in the midwestern US.