Forecasting Wet Land Rice Production for Food Security

Abstract

Rice is one of the major crops feeding the world population and is one of the most substantial ingredients in the food security chain. Therefore, a reliable forecast of rice production would have a predominant impact on assessing the world food security. In this paper we develop models to forecast the wet land rice production in two provinces of Indonesia. The four-monthly data are used to construct and develop the forecasting models. To forecast the rice production, we first forecast the harvested area and the yield. We then use a mathematical model to estimate the rice production in terms of the harvested area and yield. The proposed models are used to forecast the recorded data. The error of the forecasted data are analysed to assess the efficacy of the models. The analysis of the errors shows that ARIMA(p, q, d)) and Bayesian models are the best models for forecasting harvested area and yield. However, the results clearly indicate that the optimal model for one province it not necessarily the best model for the other province.