Modeling Irrigation Schedules for Lowland Rice with Stochastic Rainfall

Abstract

A procedure is presented to estimate probabilistic irrigation requirements for lowland rice cultivation. The procedure uses a water‐balance equation in which rainfall and evapotranspiration are considered as stochastic variables. The Leaky law, total probability theorem, and SMEMAX and power transformations were tried to estimate weekly rainfall and normal distribution for estimating weekly evapotranspiration. Because of the zero values in the weekly rainfall data series, the Leaky law was found to be most appropriate to describe the rainfall amounts and normal distribution for evapotranspiration data series. Rainfall and evapotranspiration values were later used to estimate weekly irrigation requirements through the water‐balance equation. When applied to an irrigation system in Thailand, this model gave satisfactory results, and showed that a significant amount of water can be saved during the rainy season even when the system is operated at relatively high probability (or reliability) levels.