We investigated carbon cycling and ecosystem characteristics among two prairie restoration treatments established in 1987 and adjacent cropland, all part of the Conservation Reserve Program in southwestern Wisconsin, USA. We hypothesized that different plant functional groups (cool-season C3 vs. warm-season C4 grasses) between the two prairie restoration treatments would lead to differences in soil and vegetation characteristics and amount of sequestered carbon, compared to the crop system. We found significant (P < 0.05) differences between the two prairie restoration treatments in soil CO2 respiration and above- and belowground productivity, but no significant differences in long-term (approximately 16-year) carbon sequestration. We used a biometric approach aggregating short-term observations of above- and belowground productivity and CO2 respiration to estimate total net primary production (NPP) and net ecosystem production (NEP) using varied methods suggested in the literature. Net ecosystem production is important because it represents the ecosystem carbon sequestration, which is of interest to land managers and policymakers seeking or regulating credits for ecosystem carbon storage. Such a biometric approach would be attractive because it might offer the ability to rapidly assess the carbon source/sink status of an ecosystem. We concluded that large uncertainties in (1) estimating aboveground NPP, (2) determining belowground NPP, and (3) partitioning soil respiration into microbial and plant components strongly affect the magnitude, and even the sign, of NEP estimates made from aggregating its components. A comparison of these estimates across treatments could not distinguish differences in NEP, nor the absolute sign of the overall carbon balance. Longer-term quantification of carbon stocks in the soil, periodically linked to measurements of individual processes, may offer a more reliable measure of the carbon balance in grassland systems, suitable for assigning credits.