Evolution of CO2 and soil carbon dynamics in biologically managed, row-crop agroecosystems

Field CO2 production was related to soil carbon pools and ̄uxes determined by laboratory incubation of soils from agroecosystems designed to test the possibility of substituting biological for chemical inputs. Treatments included: conventional and organic-based row crops, woody and herbaceous perennial crops and historically tilled and never tilled successional ®elds. The CO2 ef ̄ux in corn and soybeans was affected by crop residues from previous years and growing season temperatures but not soil moisture. Overwinter cover crops and perennials such as alfalfa and poplar, resulted in fairly uniform ̄uxes of approximately 20 kg CO2±C ha ÿ1 dayÿ1 throughout the non-frozen period. Highest ̄uxes occurred in alfalfa, historically tilled successional and never tilled, grassland successional treatments, although, highest aboveground productivity occurred in the corn and poplar. Summed, ®eld CO2 ̄uxes were similar to residue-C inputs. Measurement of CO2 mineralized in extended incubations in the laboratory made it possible to use soil enzyme activity to determine the size and dynamics of soil C pools. The residue of acid hydrolysis de®ned the size of the resistant pool Cr. Carbon dating determined its mean residence time (MRT). Curve analyses of CO2 evolution plotted on a per unit time basis gave the active (Ca) and slow (Cs) pool sizes and decomposition rate constants ka and ks. Temperature correction factors provided ®eld MRTs. The active pool of this coarse textured soil represents 2% of the soil C with a MRT of 30±66 days. The slow pool represents 40±45% of the SOC with ®eld MRTs of 9±13 years. The poplar soil has the greatest MRT for both the active and slow pools. The system approach to land use sustainability (SALUS) model, which predicts CO2 evolution from decomposition in the ®eld as part of a plant growth ± soil process model, was tested using the decomposition parameters determined by incubation and C dating. The model satisfactorily predicted the intra and inter year differences in ®eld CO2 but over predicted ̄uxes from residues in the fall. It does not yet adequately consider a lag period during which the residues lose their hydrophobicity, are comminuted and colonized. # 1999 Elsevier Science B.V.