Spatial Driving Factor Analysis for Specialty Crops

A spatial decision support system (SDSS) has been developed that combines data from trial sites with expert knowledge and biophysical data to predict where different crops may be successful. The tool is proving useful for the analysis of highly specialised crops, such as specialty coffee, where little trial data is available and variables influencing crop response are crop-specific. The SDSS provides maps showing the probability of success and the certainty level of the prediction in each location. Researchers require not only maps showing where niche crops are likely to succeed, but also the reasons behind predicted success and failure. Driving factors are interpreted as the most important variables in a model, and analysing these can provide this information. Simple driving factor analysis has been included in the SDSS, but methods are needed to determine combinations of driving factors within a model, which may be spatially variable throughout a study area. This paper discusses the present status of the SDSS, and explores ways of developing and implementing site-specific driving factor analysis.