DEVELOPMENT AND EVALUATION OF A CERES-TYPE MODEL FOR WINTER OILSEED RAPE

Because of its large N fertiliser requirements and long growth cycle, winter oilseed rape (Brassica napus L.) is considered to expose its en vironment to substantial risks of N losses. Soil-crop models provide u nique tools to analyse such impacts, with an accuracy that primarily r elies on the simulation of crop C and N budgets. Here, we describe a m odel simulating the growth and development of oilseed rape that was ad apted from CERES-N Maize and a previously existing rape model. In addi tion to its soil components, the model, called CERES-Rape, has modules for crop phenology, net photosynthesis, leaf area development and gra in filling, as influenced by crop N status. A new feature compared to previous rape models is the ability to predict the crop's C and N budg ets throughout its growth cycle, including losses from leaves by senes cence. It also contains a mechanistic description of N translocation f rom vegetative parts to pods and grains after the onset of flowering. The model has been calibrated on a one-year experiment with three fert iliser N levels conducted in France, and subsequently tested on a simi lar experiment from Denmark for which no parameters were adjusted. In the vegetative phase, the time course of biomass and N accumulations i n the various plant compartments was well simulated, with predicted va lues falling within one or two standard deviations from the mean in th e measurements, except for the low-N treatments for which the high rat es of leaf senescence could not be mimicked. After the onset of flower ing, some bias appeared in the simulation of crop N uptake which impai red the predictions of final grain N yields. Simulated grain dry matte r yields matched observations within +/-15% for the calibration data s et, but were over-estimated by a factor of 2 for the other data set. D espite the above shortcomings, the simulation of fertiliser effects on the dynamics of crop N uptake and dry matter was judged sufficiently satisfactory to allow an investigation of N losses from rapeseed-cropp ed soils with the CERES-Rape model. (C) 1998 Elsevier Science B.V.