The performance of the Danish simulation model DAISY in prediction of N_min at spring

Abstract

In Denmark the Danish Agricultural Advisory Centre has for some years used the soil content of mineral nitrogen in spring in fertilizer recommendations. Since 1986 these recommendations have been based on soil samples carried out at all intersections of a nationwide 7 km square grid in Denmark. It was hoped that it may be possible to replace soil measurements with values of soil mineral-N calculated with a model. The Danish simulation model DAISY, which among other things simulates changes in the inorganic N content of the soil, was evaluated with respect to the N_min content in the early spring under bare soil and under winter cereal. For both situations the precrop was cereals. The performance of the model was evaluated in farming systems receiving mineral fertilizer and in some instances organic manures. The results were analysed according to type of subsoil: sandy or loamy. Predictions were 11 kg N ha⁻¹ less than the measured values as a mean and the differences between simulated and measured values were high for fields receiving organic manures. Predictions were less than ± 10 kg N ha⁻¹ of measured values in 25–58% of cases for the different types of crop cover at the time of soil sampling, type of subsoils and fertilizer strategies, respectively. Predictions were less than ± 20 kg N ha⁻¹ of measured values in 48–89% of cases for the different situations. The best predictions were obtained for sandy subsoils covered by winter cereal and supplied with mineral fertilizer only. It is concluded that the quality of the data used as input in the model has to be increased and that further developments of parts of the DAISY model are needed before modelling can be a useful tool in N fertilizer recommendations.