Izaurralde, R. Cesar (Joint Global Change Research Institute, 8400 Baltimore Ave., Suite 201, College Park, MD, 20740; Phone: 301-314-6751; Fax: 301-314 –6760;

Email: cesar.izaurralde@pnl.gov)

Modeling Long-Term Soil Organic Carbon Dynamics as Affected by Management and Water Erosion

R.C. Izaurralde *, W.M. Post, J.R. Williams, P. Puget, A.M. Thomson, W.B. McGill, R.Lal, L.B.Owens

The soil C balance is determined by the difference between inputs (e.g. litter, crop residues, decaying roots, organic amendments, depositional C) and outputs (e.g. soil respiration, dissolved organic C leaching and eroded C). Two competing hypotheses suggest erosion may either increase or decrease output. One hypothesis states that C from eroded fields becomes “sequestered” in depressional areas and thus is rendered unavailable for decomposition. An alternative hypothesis argues that due to aggregate breakdown during erosion events, physically-protected C becomes accessible, thereby increasing oxidation of C and emission of CO₂. The objectives of this paper are to (a) review literature to determine the extent to which empirical evidence supports either the sequestration or increased accessibility hypothesis in managed ecosystems and (b) present modeling results of two 60-year experiments documenting changes in soil C as affected by management and water erosion. The treatments were established in 1939 on two ~1 ha watersheds (W128 and W188) at the USDA North Appalachian Experimental Watershed facility north of Coshocton, OH. Soils in these watersheds were developed from loess-like deposits over residual bedrock and are silt loam in texture. Soil and crop management changed over time in both watersheds. Until the early 1970s both watersheds were under a corn-wheat-meadow-meadow rotation. Watershed 128 has been under continuous plow till corn since 1984. Watershed 188 has been under continuous no till corn since 1970. The EPIC model (v. 3060) was used for the simulations. It is capable of simulating wind and water erosion and contains subroutines to simulate C and N dynamics. Simulated C stocks were lower than observed values but reproduced the trend of higher soil C content under no till management. Eroded C under no till (W188) was about one fourth that under plow till (W128). The annual simulated rates of C erosion under plow till was 0.333 Mg C ha^-1 y^-1 while under no till it was 0.084 Mg Cha^-1y^-1. These rates were higher than those estimated with ¹³⁷Cs (0.041 Mg Cha^-1 y^-1) and measured with sediment collectors (0.026 Mg C ha^-1 y^-1) in a nearby watershed. It is possible that some of the discrepancy between predicted erosion and sediment collected may be explained by the occurrence of soil deposition within the watershed. These results clearly demonstrate the significance of C erosion on the C balance. However, they do not answer directly either of the two hypotheses. Further research using landscape models will be conducted to elucidate the effects of depositional C on the C balance.