The basis for successful and sustainable agriculture is the utilization of adequate decision rules. When it comes to precision farming, these rules have to be applied to each sub-field, where they determine the actions to be taken.

There are many possibilities for achieving site-specific information for a field (e.g. measuring the electrical conductivity of soil or yield mapping). But which rules should be used to link this information with profit maximization treatment recommendations?

For site-specific nitrogen fertilization, a method is presented to generate decision rules automatically based on yield-predicting artificial neuronal networks. The database for the artificial neural net consists of information gained from sensors measuring apparent electrical conductivity, historical yield, historical fertilizer applications and in-season measurements such as the REIP index. It must be emphasized that all the necessary data are collected automatically, on the go and at very low cost.

One set of decision rules was generated for each of the three nitrogen applications within one crop season. The rules are visualized by means of decision trees. Thus, the importance of the various heterogeneity indicators, which the nitrogen recommendation is based on, can be analyzed.

The method was implemented on winter wheat fields between 2005 and 2010 on two farms in eastern Germany. The monetary result was compared with three other nitrogen fertilization strategies: uniform treatment, an online (sensor) approach and an offline (mapping) approach). All treatments were carried out with usual farm technology. From an economic point of view, the automatically generated decision rules were by far the best of all the fertilization strategies.

The automated generation of decision rules and visualization by means of decision trees will be the focus of the paper. Furthermore, the implementation of the field experiments and all economic calculations used for verifying the results will be shown in detail.