Dynamics of energy, water, carbon and nitrogen in agricultural ecosystems: simulation and experimental validation

If soil-plant-atmosphere models are to perform reliably under the wide range of edaphic, climatic and management conditions encountered in ecosystem studies, then the use of explanatory vs. functional algorithms must be expanded in these models. In this article, the soil-plant-atmosphere model ecosys, based mainly upon explanatory algorithms, is described and tested against field data recorded during a 1990–1991 sugar beet-barley rotation on a sandy sediment in the Eisenbach catchment north of Braunschweig, Germany. Water movement below 0.50 m was underestimated by the model, likely because hydraulic conductivity was underestimated for the extremely sandy subsoil in this catchment. The period of time that N in extractable mineral form resided in the topsoil following fertilization was overestimated, as has been found elsewhere by this and other models. Ways by which these overestimations could be resolved are discussed. The model was able to reproduce general trends in the distribution of water and mineral N in the soil, and of C and N in the crops during the two-year field study without the use of site-specific values for any of its parameters. Examples of hourly model output for soil and crop water relations during a selected 3-day period in June 1991 are presented as examples of how the model could be more rigorously validated in studies carried out at higher temporal resolution. Because parameters used in explanatory algorithms have values that are independent of the model itself, models based on such algorithms are well adapted to perform reliably under a wide range of environmental conditions. Examples of such adaptations are given for the model used in this study.