Prediction of Rice Growth in Chiba Prefecture from Weather Data by Cloud Computing

Abstract

To manage rice cropping effectively, managers need to predict crop growth. We developed a cloud-based model to predict heading of the main cultivars grown in Chiba Prefecture from AMeDAS data and to calculate crop growth stages and appropriate work periods. We developed the “Deruta” interface to enable viewing of this information on smartphones and tablets. Heading dates in 15 fields in Chiba had a root mean square error, an indicator of the accuracy of numerical prediction, of 2.99 days. Of those predicted by the model, 86.7% lay within ±3 days, the target value. Our results show that it is easy to generate and transmit crop information via information and communications technology. “Deruta” will serve as a model for the forthcoming development of models for predicting crop growth and pest outbreaks.