Classification of management zones for precision farming in saline soil based on multi-data sources to characterize spatial variability of soil properties

Recent research in precision agriculture has focused on use of management zone as a method to more efficiently apply crop inputs and precise soil management. In this paper, the variables of NDVI data, soil salinity data and cotton yield were selected as clustering variables and fuzzy c-means clustering algorithm was used to define management zone in an about 15 hm² field in a coastal saline land, fuzzy performance index and normalized classification entropy were used to determine the optimal number of clusters. The results revealed that the optimal number of management zones for the study area was 3. To assess whether the defined management zones can be used to characterize spatial variability in soil chemical properties, 224 georeferenced soil sampling points were examined by using One-way variance analysis. It was found that there exist significantly statistical differences between the chemical properties of soil samples in each defined management, and management zone 3 presented the highest nutrient level and potential crop productivity, whereas management zone 1 worst. The results reveal that fuzzy c-means clustering algorithm can be used to delineate management zones by using the given three variables. The defined management zones can not only be useful for the sampling design, but provide an effective decision-making support for variable input in precision agriculture.