USING INDUSTRY DATA TO IMPROVE PRODUCTIVITY THROUGH BETTER VARIETY RECOMMENDATIONS

A LARGE amount of information is collected annually by mills at the rake, block, farm and mill area levels. In some areas, additional information is available, relating blocks to soil types, sub-districts, crop class and/or CCS by harvest time. QCANESelect™, the web-based decision support tool used to assist with variety selection at the block and farm level, makes recommendations for varieties based on soil type, diseases of concern and some management options. However, much of the information relating to variety performance on different soil types is not readily available and is not routinely analysed to assist in providing QCANESelect™ with the background data used to make variety recommendations. Currently QCANESelectTM recommendations are based on the collective knowledge of a group of field-based experts who provide a subjective assessment of the relative productivity of varieties over a full crop cycle. Preliminary analyses of data from two mill areas over five years (Bundaberg 2009–2013; Herbert 2008–2012) suggested that the recommendations could be greatly improved by using the information on variety × soil type × crop class. This paper will show how, within the limitations of mill data, variety recommendations can be improved by combining analyses of mill data and expert knowledge.