Post-processing methods to eliminate erroneous grain yield measurements: review and directions for future development

Yield mapping is increasingly used in agricultural management. The distributions produced from the majority of these datasets are non-normal and can be misleading if used in the decision making process. Numerous studies over the last 25 years, published in various formats, have highlighted the sources of errors that contribute to this non-normality and have proposed a variety of post-processing methods to reduce their effect. A comprehensive cataloging of the types of errors present and methods used to remove them as well as the approaches to and effects of post-processing error removal is needed. This review identifies four types of yield mapping measurement errors: issues associated with harvesting dynamics of the combine harvester, the continuous measurement of moisture and yield, the accuracy of positional data and errors caused by the harvester operator. Methods to remove errors range from simple thresholds to complex routines that incorporate harvest position and local yield variation. The benefits of applying filters have focused on the removal of erroneous yield variation based on simple descriptive statistics, the creation of yield distributions that show greater normality and the decrease of the nugget-variance relationship and prediction variance estimated in yield map interpolation. Publication of these parameters should accompany the interpolated yield map for both unprocessed and post-processed datasets to provide an insight into the reliability of the collected measurements and the effectiveness of the routines implemented. Examples from commercial growers in Western Australia are used to reinforce the review’s findings and highlight the potential for extensions to current methods to remove errors associated with harvester speed, narrow finishes and harvester turns and overlaps. This paper suggests extending the current routine implementation of error removal towards an automated post-processing system.

G. Lyle, B. A. Bryan, B. Ostendorf