P.C. Struik

Evaluation of a model for sugar beet production by comparing field measurements with computer predictions

Abstract PIEteR is a field specific bio-economic production model for sugar beet, developed for decision support at field and farm level. The model simulates the effects of plant density, N-availability and harvest date on root and sugar yield, internal and external quality, financial return and remaining N in crop residue. The model was tested for a series of 183 commercial and experimental fields in The Netherlands. Average prediction errors for root and sugar yield and financial returns were about 12%, and the variances accounted for about 50%. The data of the commercial fields were also totalled up per farm for each year and compared with the reported ones. The respective predition errors of total root and sugar yields and financial returns were about 13% and the variances accounted for about 93%. A sensitivity analysis was carried out on the decisions on plant density, N fertilization and harvest date and their effects on the output (yield, quality and financial returns). Changes in harvest date had greater effects than changes in N-fertilization and plant density for a field studied. PIEteR needs to be improved for simulation of sugar content and local yield. For application in a decision support system, giving advice to growers, relative rather than absolute results might be presented. A detailed recording of growing conditions could improve the performance of the model.

Introduction to a bio-economic production model for sugar beet growing.

Yield prediction is a basic tool for practical purposes in sugar beet growing. Connected with an economic module, a crop model can serve as a basis for decision support at field level. PIEteR as such is a bio-economic model, mainly focusing on nitrogen fertilization, plant density and harvest date. Its main component is a crop growth model, that simulates crop responses to weather, soil factors and growers’ decisions. Its input and (future) output make PIEteR a useful basis for decision support. Its potential for accurate predictions will be tested and compared with other models. If PIEteR will prove to be the most accurate and useful model, it will be developed further, thus enabling improvements in the quality of the growers’ decisions.