Management zone delineation in arable crop systems

Abstract

Management Zone Delineation (MZD) could be of increasing importance for its economic and environmental benefits through varying rates of crop inputs to meet site-specific demands across individual fields. However, research on MZD in arable cropping systems is limited in New Zealand. Previous work from Lincoln Agritech Ltd. presented the benefits of Precision Agriculture for large scale farmers and contractors using yield, soil and aerial images to adopt Variable Rate Application of seeds, fertilizer and agro-chemicals. Furthermore, improved irrigation efficiency in wheat and maize cropping systems and maize zone delineation using Active Light-Nitrogen sensor, NIR camera technology and SPAD chlorophyll meter, were studied. The objective of the presented project is to develop methods and tools to identify optimal N supply rate for maize production with the emphasis on management zones and site adapted plant population densities in New Zealand. Two rain-fed maize fields from Waikato and one irrigated field from Canterbury were selected for field experiments. Yield and elevation data in 2013 and 2014 collected with a John Deer 7050 Series Self-Propelled Forage Harvester. Data analysis and mapping of management zone were done in Esri, ArcGIS 10.2.2. software. VESPER 1.6 free version from Australian Centre for Precision Agriculture was used to interpolate the data. Three management zones with low, medium and high yield potential were derived in the final map. Satellite images from Google Earth Archive of previous years were also used to delineate field boundaries and management zones. A strip-plot experimental design with 3 replicates was allocated to each management zone at each field. Three treatment levels of N fertilizer: farmer’s best N-fertilization practice, +35% and -35% were applied using calculated N-fertilizer demand prescription maps with an 8-row VRA fertilizer spreader. We expect that soil electrical conductivity field survey data, compatibility of yield data and prescription maps between different software packages, and the inclusion of annual yield data would improve the discriminating power of the various approaches.