A Web-based System for Scheduling Irrigation in Sugarcane

Adoption of decision support services (DSS) based on crop growth models has been poor and yet the concept of transferring an increasing body of scientific knowledge via DSSs remains attractive. This study explores the use of the Internet and participatory action research (PAR) to develop an irrigation management service called WaterSense for sugarcane irrigators in Australia. WaterSense combines two earlier web-based DSSs, one for use of limited water and the other for scheduling abundant water. In both cases, a small number of dedicated growers and extension staff were involved in designing and conducting field experiments to test concepts that were later included in the DSSs. An optimisation procedure ('Caneoptimiser'), based on the APSIM-Sugarcane model was developed for applying limited irrigation when most needed. However, the optimisation took too long to operate as a web service. Growers with abundant water prompted the development of a simple water balance technique (called 'WaterBalance') for scheduling irrigation over the Internet. Growers with limited irrigation wanted the service offered by Caneoptimser with the speed and format of WaterBalance. A new service called WaterSense was developed using algorithms for canopy development and soil water hydrology from Caneoptimser and algorithms for reference evapotranspiration and crop factors from WaterBalance. Adoption by those familiar with the background research has been immediate. Indications are that the combined use of the Internet, PAR and concurrent research and demonstration will open the way for the adoption of research findings in the management of irrigation and possibly other aspects of sugarcane farming. However, widespread adoption of DSSs cannot depend on participation of the intensity required for their development. Processes of adoption beyond the case study need further research. Introduction Over the past 25 years, a plethora of software tools has been developed for farmers but use of these tools has not grown with computer ownership as expected (McCown, 2002). Crop simulation models which often support these software tools have nevertheless been used extensively in research and education. Researchers have continued to develop and improve their models. Some attempts to transfer the knowledge encapsulated in these models to growers have been reasonably successful but many have been disappointing (McCown, 2002). In Australia, remarkably few have