Abstract:

The purpose of this study was to explore ways to solve the problems in the traditional fertilizer recommendation such as lack of spatial attribute information and experiential knowledge can not be processed. Maize as an example, based on GIS intelligent decision model on fertilizer recommendation was built by the deficiency experiments in multiple years and fields, nutrient value of plow layer soil grid sampling and fertilizer recommendation knowledge base. It was observed that this model not only can quantitative simulation structural elements on fertilization, but also qualitative analysis unstructured elements on fertilization. So the fertilizer recommendation was intelligent decision supported relative comprehensively. There were some field experiments on maize for the model verification in the 20 Li-pu village, Xinzhou city, Shanxi province. These experiments confirmed that maize yield was significantly increased under the guidance of the model. It is concluded that this model has some practical value.