The role of yield maps in Precision Farming

Research into the development of methods to turn yield map data into useful management information of known accuracy in the context of Precision Farming research and practice.

Year of Publication2003

The aim of this research was to develop methods to turn yield map data into useful management information of known accuracy in the context of Precision Farming research and practice. The global context was identified in terms of developing the principles of Precision Farming that apply to every country and every crop. The managerial context was established by identifying the management framework in which yield maps and spatially variable farm practices were used. The main research context was defined within a major Precision Farming project in the UK to develop new management guidelines.Yield data were recorded and analysed over ten years, resulting in a rich data set for both spatial and temporal trend analysis and management information. Most of the yield data included both systemic and systematic errors that were identified, classified and mostly removed. A new method was developed to automatically identify and remove known yield data errors by the use of an ‘expert filter’ program. Further routines were written to produce yield maps in a recommended format. Two main errors were identified; the dynamic time lag between detachment and sensing of the grain, and the unknown crop width entering the header. A patented method to measure the crop width, which also indicated the start of harvest, was developed.The yield data were further analysed to extract both spatial and temporal trends. A simple averaging of the data at each grid point over time was used to produce the spatial trend map. The development of the temporal trends went through a number of stages. Firstly the temporal increase and decrease was identified and discarded, then the Coefficient of Variation was used at each point. This was also discarded in favour of splitting the temporal effects into two: the inter-year offset that quantified the gross production change from year to year and the temporal variance that showed how each part of the field reacted relative to the mean. In these data sets, it was found that most spatial variability cancelled out over time and could not be used to predict the yield pattern in the following years but yield maps could be used to determine the nutrient off-take of Phosphate and Potassium through removing the crop biomass, which could then be replenished. Combining the spatial trends and the temporal variance resulted in a single management map showing homogenous management zones.

This item is categorised as follows

Additional keywords/tags

Error: The remote server returned an unexpected response: (302) Found.

What Next...?

This is a brief summary of an item in the OpenFields Library. This free online library contains items of interest to practitioners and researchers in the agricultural and landbased industries.