Soil Carbon Dynamics in Corn-Based Agroecosystems: Results from Carbon-13 Natural Abundance H. P. Collins, R. L. Blevins, L. G. Bundy, D. R. Christenson, W. A. Dick, D. R. Muggins, and E. A. Paul* ABSTRACT We used natural "C abundance in soils to calculate the fate of Cf-C inputs in fields cropped to continuous corn (Zea mays L.). Soil samples were collected from eight cultivated and six adjacent, noncultivated sites of the Corn Belt region of the central USA. The amount of organic C in cultivated soils declined an average of 68%, compared with adjacent, noncultivated sites. The 8 13 C of cultivated soil profiles that had been under continuous corn for 8 to 35 yr increased in all depth increments above that of the noncultivated profiles. The percentage of soil organic C (SOC) derived from corn residues and roots ranged from 22 to 40% of the total C. The propor- tion of corn-derived C, as determined by this technique, decreased with soil depth and was minimal in the 50- to 100-cm depth increments of fine-textured soils. The mean residence time of the non-corn C (C 3 ) ranged from 36 to 108 yr at the surface, and up to 769 yr at the subsoil depth. The longer turnover times were associated with soils high in clay. Prairie-derived soils have a higher potential to sequester C than those derived from forests. The significant loss of total C at all sites and the slow turnover times of the incorporated C lead us to conclude that there is a substantial potential for soils to serve as a C sink and as a significant nutrient reserve in sustainable agriculture. S OIL ORGANIC CARBON is a major controller of soil tilth and fertility that reflects and influences ecosystem processes. Agriculture reduces SOC levels and contrib- utes to increases in atmospheric CO 2 concentrations (Houghton et al., 1983). Altered management practices such as reduced tillage, decreased bare fallow, and in- creased residue inputs not only mitigate the loss of C, but in some cases can also increase soil C levels (Paus- tian et al., 1997; Flach et al., 1997). The significance of soil C pools and fluxes in global change scenarios and sustainable agriculture require that data for specific sites be related to landscape and regional scales. Atmospheric CO 2 contains both radioactive ( 14 C) and stable ( 13 C) isotopes suitable for tracer studies. These are incorporated into plants by photosynthesis and eventually into soils through decomposition and humifi- H.P. Collins and D.R. Christenson, Dep. of Crop and Soil Sciences, Michigan State Univ., East Lansing, MI 48824-1325; H.P. Collins, present address: 7535 Mesplay Avenue SE, Lacey, WA 98503; R.L. Blevins, Agronomy Dep., Univ. of Kentucky, Lexington, KY 40546- 0091; L.G. Bundy, Dep. of Soil Science, Univ. of Wisconsin, Madison, WI 53706-1299; W.A. Dick, School of Natural Resources, The Ohio State Univ., Wooster, OH 44691; D.R. Muggins, Southwest Experi- ment Station, Univ. of Minnesota, Lamberton, MN 56152; E. A. Paul, Dep. of Crop and Soil Sciences, Michigan State Univ., East Lansing, MI 48824-1325. Received 20 June 1997. Corresponding author (paulea@pilot.msu.edu). Published in Soil Sci. Soc. Am. J. 63:584-591 (1999). cation. Studies involving 13 C abundance have been used to measure vegetation shifts (Galimov, 1985; Cerri and Andreux, 1990), community composition (Dzurec et al., 1985), soil chronosequence analyses (Stout et al., 1981; Guillet et al., 1988), root growth and respiration (Qian and Doran, 1996; Gregorich et al., 1996), SOC turnover (Balesdent et al., 1988), and microbial biomass turnover (Ryan and Aravena, 1994). The isotope signal in the plant is retained essentially unchanged in the SOC (Wedin et al., 1995; Boutton, 1996). Plant C isotope ratios vary somewhat during the grow- ing season (Tieszen and Steuter, 1991) and are affected by moisture (van Kessel et al., 1994). Martel and Paul (1974) found that the soils of low, moist areas (gleysols) can be 2%o more negative than associated upland molli- sols of the catena. Production of wheat (Triticum aestlvum L.) for 100 yr on a C 3 -C 4 prairie of the Sanborn plots in Missouri, with an original 8 13 C of -18.6%o, resulted in replace- ment of 50% of the SOC by the residue from the agricul- tural crop (Balesdent et al., 1988). A Colorado site culti- vated to wheat for 84 yr had a 13 C content of -19.3%o compared with -16.1%o for adjacent native vegetation (Follett et al., 1997). Thus, 30% of its SOC was derived from wheat residues; the remaining 70% was from SOC present before cultivation. Calculation of the above- and belowground C inputs from wheat and associated weeds showed a residue incorporation efficiency of 5% into SOC during the period of 1909 to 1993 period. A nearby Nebraska site, cultivated with wheat for 22 yr, had 17% of its soil organic matter (SOM) derived from wheat. This represented an incorporation efficiency of 10.5% (Follett et al., 1997). Analysis of soil that had grown Q corn for 25 yr on an original C 3 forested site (-28.2%o) in Ontario, Canada allowed Gregorich et al. (1995) to determine that 30% of the SOC in the 0- to 30-cm layer came from corn. This represented a mean residence time of 35 yr for the non-corn-derived C. Monreal (personal communica- tion, 1998) combined 14 C and 13 C analyses to determine that the organic C of macroaggregates turned over in 61 yr, while that in microaggregates required 275 yr. Extrapolation of site-specific SOC turnover data to other soils and modeling of regional input parameters can only be accomplished using multiple sites with a known history of management and crop residue inputs. A series of long-term agricultural sites from Kentucky Abbreviations: KBS, Kellogg Biological Station; MRT, mean resi- dence time; SOC, soil organic C; SOM, soil organic matter.