Zonal yield expectation maps for Precision Agriculture generated using a combined Fuzzy-Expert system

The development of procedures which facilitate the determination of realistic yield expectations for management units within agricultural fields represents a scientific challenge. For this an applicable and scenario-apt Fuzzy-Expert system was developed. The procedure for generation of sites-specific yield expectation maps is divided into three methodical steps – step 1: estimation of yield potential, step 2: influences of pre-crop and crop variety and step 3: influence of site-specific parameters. The regional approach takes into account physical parameters as well as meteorological input data or results of variety testing trials for prediction of regional average yields. This prediction is starting point for the joined site-specific fuzzy model. For yield estimation site-specific heterogeneity spatial information of parameter influencing yield formation are taken into account. Physical soil parameter (plant available water, potential capillary rise, ground water table, landform attributes) are used to generate the site specific yield expectation map. This map is generated using the Spatial Analysis and Modeling Tool (SAMT). A method for training of the fuzzy model is described. The procedure was tested on a 45 ha field cropped with winter wheat in a Chernozem area of Saxony-Anhalt in East-Germany for the years 2000 and 2005.