The theoretical basis of the simulation model CANDY (CArbon-Nitrogen-DYnamics) is briefly described along with results of its application to a test data set. The model contains modules for calculating soil temperature, moisture content and the processes of the soil carbon-nitrogen cycle. The model inputs are weather data (air temperature, precipitation and global radiation on a daily basis), plant development characteristics (seeding, harvesting, crop height and root depth), soil texture data and agricultural management (irrigation, fertilization, manuring, tillage). The test data material from two catchments in north Germany, one with loam soil (Neuenkirchen) and one with sand (Nienwohlde) included data from several nitrogen fertilizer treatments. The results show that the model outputs for temperature, moisture and nitrogen fit the observations quite well despite some deviations. Summarizing for all soil layers the square root of the mean quadratic differences of simulated and observed data are: for soil temperature less than or equal to 2.3 K for the loamy soil over a five-year period and less than or equal to 2.5 K for the sandy soil over a five-year period; for soil moisture less than or equal to 4.1 Vol.% for the loam and 14.4 Vol.% for the sand both over a three-year period. The results of soil nitrogen in 0-90 cm depth depend on the particular site under consideration. In most cases 2/3 of the whole data was contained within the deviation class less than or equal to 40 kg ha(-1). The differences between observed and simulated data are within the usual range for applications on farm fields.
