UBC Theses and Dissertations
Optimal long-term development and operation of irrigation systems with storage under hydrological uncertainty Igwe, Okay Cyril
The scarcity of water resources is increasing and marginal quantities of water are becoming more important. The need for sound and conservative management of water supply is imperative. Intensified agricultural production, to feed the ever-increasing world population, is requiring more irrigation, which is the heaviest consumptive use of water. It is desirable, therefore, to seek new techniques that can modify the irrigation water supply regime. Such a modification implies the urgent need for the development of an irrigation water supply regime dictated by seasonal hydrologic considerations and agricultural production that is technologically fully controlled on the basis of long-range stochastic considerations. The fact that most observed historical hydrologic data are usually short and may constitute poor representation of the possibilities for long-term planning in irrigation systems management, reinforces the postulation that any meaningful approach to the optimal development and operation of irrigation systems must take full cognizance of hydrological uncertainty. To achieve optimum competence in irrigation systems management under the predominating constraint of hydrological uncertainty, a methodology that first considers the systems operational policies as well as several levels of water consumption is necessary. To be realistic with operating rules one has to consider the stochastic variability in irrigation planning and thus has to consider uncertainty and risk relating to the major decision input information-(hydrologic information); and other considerations other than strictly maximizing expected economic monetary value must be brought into the model formulation. Any meaningful planning in agricultural water utilization has to be man-centered in approach and must provide objective analysis of subjective considerations. The above rationale led to the development of a stochastic Bayesian Decision Theory optimization model, which specifies expected utility as the criterial objective function to maximize, and which could be realistically employed to identify the best decision-criterion and adequate policies for optimal long-term planning, development and operation of irrigation systems with storage under hydrological uncertainty. The model, which is behavioral in approach, is applied to the Nicola Valley Irrigation District located in the dry, semi-arid interior of Brutish Columbia. Two irrigation operational procedures, two decision criteria, and different crop response function are employed in the analysis to identify the best planning policy for astute irrigation systems management in the region. The results obtained from the model indicate that the optimal areas to irrigate under hydrological uncertainty are dependent on the degree of hydrological uncertainty, the systems operating procedure, the crops irrigated and their responses to water, and on the decision criterion and utility function employed. Post-optimal analyses indicate that optimal policies obtained are very sensitive to discretized probability distribution of the uncertain states of nature, crop response function, utility function and decision criterion, and system operating procedure employed. For Nicola Valley Irrigation District the model shows that the practice of irrigating more alfalfa hectarage at a water consumption level that is below the designated maximum water requirement of alfalfa, - Procedure II, is superior to the practice of irrigating less hectarage to maximum consumptive use of crop and maximum water holding capacity of the soil, - Procedure I. It is also shown that the criterion of maximizing total expected utility, EU, is superior to the criterion of maximizing total expected monetary value, EMV, under uncertainty and risk. The model also shows that it is desirable to have some hydrological forecasting device. In the Nicola region for improved output from the model. Thus, the model has considerable promise as a valid tool for optimal long-term irrigation systems management decision-making under hydrological uncertainty.
Item Citations and Data