SOM models or agroecosystem models with specific treatment of SOC dynamics represent our conceptual and quantitative understanding of the mechanisms regulating C transformations in soil. Well-tested models have successfully been used to simulate SOC changes at various temporal (day to century) and spatial (site to national) scales. When well-initialized and calibrated (if necessary), simulation models have been capable of describing historical trends in SOM dynamics observed in long-term experiments worldwide (Paustian et al., 1992; Par-ton and Rasmussen, 1994; Grant et al., 2001). When driven with more generic soil and climate databases, these models have been able to provide insight on scaling issues (Izaurralde et al., 2001a) or regional trends in SOM evolution (Falloon et al., 1998). What level of model complexity is needed for project level applications? On the one hand, simple models, or simplified versions of complex models, can provide rapid assessment of the impact of management practices on SOC content under a given set of climate and soil conditions. On the other, models of increased complexity can yield an in-depth understanding of the processes regulating SOC dynamics in addition to providing ancillary information on the environmental impacts of carbon sequestration practices such as trace gases (Del Grosso et al., 2000) and erosion (Izaurralde et al., 2001c).

At the project level, however, there seems to be a scale missing in SOC prediction: the landscape or field scale. SOM models have been applied either at the site scale with specific information, or at the county or smaller spatial scales with generalized soil and climate databases. A key limitation in applying models at the landscape scale has been the input data needed to drive them. As discussed in previous sections, advances in methodologies to estimate soil parameters from ancillary, more easily acquired parameters (digital elevation maps), may facilitate the application of models in landscape, whole farm, mode. The APEX model (agricultural policy environmental extender) (Williams, 1999) is an example of one such model with this type of capability. APEX is a watershed-scale model capable of simulating plant growth, watershed hydrology, soil C dynamics (Izaurralde et al., 2001b), erosion, nutrient cycling and routing of water, nutrients, pesticides, and soil across connected fields under various land use and management practices. The APEX model contains all the functions of the EPIC model and can be applied to whole farms and small watersheds (<2500 km2) including multiple fields, soil types, and landscapes. Further, APEX simulates routing of water, sediment, eroded C, nutrients, and pesticides across complex landscapes and channel systems to the watershed outlet.

Model evaluations will also be needed for rapid assessment of the possible impacts of management practices on soil C sequestration. CSTORE is an example of a simple model being developed with such an objective (Paustian, 2004). In this case, the model is intended for use in field-level prediction of SCS, and also as part of a decision-support system, with minimum data requirements. The CSTORE model is based on the widely used CENTURY model with a simpler representation of the SOM pools.

0 0

Post a comment