Rice irrigation scheduling incorporating stochastic rainfall

Abstract

An irrigation-scheduling program has been developed to increase irrigation efficiency in a large-scale rice irrigation project in Malaysia. The study focused on predicting daily irrigation deliveries for the main season and off-season and its performance was analyzed. A water balance approach in which rainfall was considered as a stochastic variable was used. Markov chain was used to describe the random occurrence of daily rainfall, and a skewed normal distribution was applied to generate the amount of rainfall on rainy days. Predicted rainfall and evapotranspiration values were later used to estimate weekly irrigation deliveries through the water balance equation. Comparison of the observed and predicted irrigation deliveries values for the 2000/2001 season showed that the observed values were higher than the predicted value indicating excess water supply in the field. It was observed that by modifying the existing schedules, it could save a considerable amount of irrigation water during both the main and off seasons. The periodic results obtained could be used to monitor the allocation of irrigation deliveries for improving the water management system. The present irrigation system was evaluated using an adequacy indicator. The adequacy indicator was computed by comparing the crop water requirement with the water release data. This computation provides information on the uniformity of water distribution and any shortfall or excess.