ABSTRACT
Numerous researchers have addressed the question of determining the optimal allocation of water to irrigated crops (Dudley et al., 1971; Burt and Stauber, 1971; Bernardo et al., 1987). In general, intraseasonal irrigation water allocation is both time and quantity dependent. Therefore, much of the research has focused on estimating optimal rules for scheduling irrigation (Bras and Cordova, 1981; Harris and Mapp, 1980; Yaron et al., 1987; Zaveleta et al., 1980; Dudley, 1988). However, this work usually implicitly assumes that the irrigation hardware was designed to meet such a rule, that the water supply can be manipulated sufficiently to provide water when required, and that management actually has the ability to operate the sophisticated hardware (Svendsen, Merrey, and Fitzgerald, 1983). These assumptions may be valid for newer, reservoir-serviced irrigation systems in more developed countries, but they rarely hold for the vast majority of the existing irrigation systems in Asia. In particular, they do not hold when the water supply is provided from a run-of-the-river type system that offers no means of effectively regulating the actual daily water supply. Such systems often result in extreme fluctuations in both daily and seasonal water supply (Hashimoto, Stedinger, and Loucks, 1982; Kuczera, 1987). Unfortunately for many countries in Asia, run-of-the-river water supply is the standard rather than the exception.
