Spatial Decision Support for Tropical Agriculture: CaNaSTA Probability Modelling

When identifying suitable niches for specific crops, or identifying the probability of a phenomenon occurring at a specified location, the focus is often on quantifying spatial extent. Whilst this is a fundamental aspect of modelling, it is also important to understand why these locations are selected. This is specially so for agricultural products for which quality matters, i.e. most higher valued, income generating production alternatives. Furthermore, combinations of spatial and quality modelling will become increasingly important to support denomination of origins for successful product differentiation and product marketing. Another key issue is deciding how much confidence can be had in model results. This is particularly important when data may be sparse or spatially biased, as is often the case with agricultural trial data. Spatial decision support can facilitate the decision process by making available relevant data and knowledge. A Spatial Decision Support System has been developed called CaNaSTA (Crop Niche Selection in Tropical Agriculture). The engine of the tool is Bayesian probability modelling. CaNaSTA combines data from trial sites with expert knowledge and environmental data to provide maps, tables and graphs showing where selected species can be expected to thrive, or where a given phenomenon is expected to be present. CaNaSTA was initially developed as a tool to suggest niche forage species to smallholder farmers in the tropics. Recently, it has also been applied to coffee quality analysis, cowpea performance in tropical hillsides and carbon concentration in soils. The tool is proving useful for analysis of highly specialised crops, where little trial data may be available and variables influencing crop response are often cropspecific. Further research and development aims to provide further analysis options for specialised crop response research, as well as the application of CaNaSTA to other diverse spatial research problems.