Analysis of ecosystem controls on soil carbon source-sink relationships in the northwest Great Plains
Article first published online: 22 NOV 2006
Copyright 2006 by the American Geophysical Union.
Global Biogeochemical Cycles
Volume 20, Issue 4, December 2006
How to Cite
2006), Analysis of ecosystem controls on soil carbon source-sink relationships in the northwest Great Plains, Global Biogeochem. Cycles, 20, GB4012, doi:10.1029/2005GB002610., , , , , and (
- Issue published online: 22 NOV 2006
- Article first published online: 22 NOV 2006
- Manuscript Accepted: 14 JUL 2006
- Manuscript Revised: 11 APR 2006
- Manuscript Received: 29 AUG 2005
- carbon sources and sinks;
- Ecosystem control;
- General Ensemble Biogeomechanical Modeling System (GEMS)
 Our ability to forecast the role of ecosystem processes in mitigating global greenhouse effects relies on understanding the driving forces on terrestrial C dynamics. This study evaluated the controls on soil organic C (SOC) changes from 1973 to 2000 in the northwest Great Plains. SOC source-sink relationships were quantified using the General Ensemble Biogeochemical Modeling System (GEMS) based on 40 randomly located 10 × 10 km2 sample blocks. These sample blocks were aggregated into cropland, grassland, and forestland groups based on land cover composition within each sample block. Canonical correlation analysis indicated that SOC source-sink relationship from 1973 to 2000 was significantly related to the land cover type while the change rates mainly depended on the baseline SOC level and annual precipitation. Of all selected driving factors, the baseline SOC and nitrogen levels controlled the SOC change rates for the forestland and cropland groups, while annual precipitation determined the C source-sink relationship for the grassland group in which noticeable SOC sink strength was attributed to the conversion from cropped area to grass cover. Canonical correlation analysis also showed that grassland ecosystems are more complicated than others in the ecoregion, which may be difficult to identify on a field scale. Current model simulations need further adjustments to the model input variables for the grass cover-dominated ecosystems in the ecoregion.