Advances in coupling big data technique with nutrient site-specific management: scheme, methods and outlook

In this review, the methods and techniques of the field nutrient precision management were summarized, including soil testing and formulated fertilization technique, soil nutrient testing technique and crop nutrition diagnosis technique. The long term promotion of precise nutrient management work in our country has accumulated large amounts of soil types, soil fertility and other basic data, as well as some data related to field nutrient management, such as rural cadastral survey, agricultural census, characteristics of crop’s demand for nutrient and so on. However, some shortages remind to deal with. A lot of attention has been paid on theoretical researches but less on the application of the technical researches. The researchers are in state of over-divided and fragmentized, hardly connecting traditional field nutrient managements with modern precision ones. The research achievements are also need to be modified for better practical application for both technicians and farmers. Nowadays, the big data technology provides solutions for the explosive growing data integration, exchanging and sharing. In this review, a fertilization management scheme of " basic fertilization and precise adjustment” was proposed by using the big data technology. Firstly, the basic fertilization plan was created through a comprehensive analysis and data mining, combined with the principle of formulated fertilization technique. The basic plan included the time and quantity of basal and top-dressed fertilizers. At the same time, the spatial visualization of field management of fertilization scheme could be realized by the use of cloud GIS, which laid the foundation for pushing fertilization scheme on mobile terminal and realizing the spatial and precise management of field information. The type and degree of crop nutritive stress were recognized by the nondestructive diagnostic technology such as digital image processing and the nutritional diagnostic model by machine learning characteristic parameters of crops. Therefore, the near-real-time monitoring and diagnosing could be performed to generate a precise adjustment to the quantity of topdressing in basic fertilization plan. An operating mode of " Pushing basic fertilization scheme, nondestructive diagnosing, precise adjusting, receiving feedback and optimizing the scheme” by the system framework of " cloud technique, handheld terminal and digital image” was proposed to support the application and updating of the fertilization scheme.