A new approach for variable rate fertilization based on direct read of soil map image

Abstract

Aim of study: To develop a methodology for variable rate (VR) fertilization with less complexity in practice for variable rate fertilization.
 Area of study: Northwest of Iran.
 Materials and methods: A software was developed to read a soil map image pixel-by-pixel to provide the required information to tailor the fertilizer rate, regardless of which software was used for map generation. A total of 78 soil samples were collected and analyzed for soil potassium, and the results were used to generate an actual map including zones ranging from 70 to 190 kg/ha. The application rates were evaluated based on 50 deposition pans and compared with those calculated from the actual map. Based on the lag time in fertilization, three applied maps were also generated.
 Main results: The correlation coefficients found between the application rates computed based on the original soil samples and posted the locations of the sample points on the applied maps were 0.95, 0.95, and 0.94, over the ravel speeds of 6, 7, and 8 km/h, respectively. The results showed there is a correlation coefficient of 0.96 with an RMSE of 1.88 kg/ha, where the application rates computed from deposition pans compared with the corresponding location on the actual map. All applied maps were identical to the actual map. The results showed that the VR fertilization based on a direct read of a map image operated as expected.
 Research highlights: Fertilizer application was based on the direct read of map image. This study highlights also the need of new approaches in programing for simplicity of precision agriculture.