Modelling soil organic carbon (SOC) stock changes in agroecosystems can be performed with different approaches depending on objectives and available data. Our objective in this paper is to describe a scheme for developing a dynamic SOC algorithm for calculating net greenhouse gas emissions from Canadian farms as a function of management and local conditions. Our approach is flexible and emphasizes ease of use and the integration of available knowledge. Using this approach, we assessed the performance of several SOC models having two or more compartments for some common agroecosystems in Canada. Analysis of long-term data for conventional management practices at different sites (n = 36) in Canada, including recent model applications in the literature on some of those data, indicated that the results obtained with two-compartment models, such as the Introductory Carbon Balance Model (ICBM) and Modified Woodruff Model (MWM), yielded results comparable to those of a multi-compartment model (CENTURY). The analysis also showed that a model such as ICBM need stuning to be applied to management and conditions across Canada. Two-compartment models programmable in a simple spreadsheet format, though they may not supplant more complex models in allapplications, offer advantages of simplicity and transparency in whole-farm analyses of greenhouse gas emissions. Key words: Virtual Farm, soil organic carbon, soil disturbance, C inputs, Introductory Carbon Balance Model (ICBM), CENTURY, Modified Woodruff Model (MWM).
