Delineation of management zones in a grain production area

Abstract

The delineation of management zones (MZs) is an important strategy for implementing precision agriculture. However, it is a complex process that requires further study. The objective of this study is to delineate MZs and validate them with respect to soil characteristics as well as corn and soybean yield in a 97-ha land cultivated under a no-till farming system. Samples were collected for physio-chemical analysis of soils and apparent electrical conductivity (EC). Moreover, data on the altitude and yield of soybean and corn were obtained. The data were initially analyzed descriptively and using Pearson correlation. Data interpolation and the elaboration of spatial variability maps for each characteristic were then performed. Furthermore, MZ thematic maps were developed. The ideal number of MZs for each combination or strategy studied was determined by the lowest value of the fuzziness performance index and normalized classification entropy. For MZ validation, Fuzzy K-means algorithm and Kappa index were used. The delineation of MZ was possible with the use of the soil characteristics with a higher temporal stability, such as the combined use of EC, soil organic matter, and clay, enabling the validation of differences in corn and soybean yield using Fuzzy algorithm and Kappa index. Data collected in different sample density did not interfere in the definition and validation of the MZ. No correlation was found between the soil EC and other chemical characteristics and yield. Furthermore, the correspondence of soil chemical properties for each MZ was not be feasible in areas with built-up soil fertility.