A DYNAMIC PROGRAMMING OPTIMIZATION MODEL FOR FLEXIBLE IRRIGATION SCHEDULING AT FARM LEVEL

Abstract

The northern parts of Sudan face high water scarcity due to aridity and desertification but excess water is often applied to the main irrigated crops. This over-irrigation contributes to aggravate water scarcity problems. Improved water saving irrigation is therefore required, mainly through appropriate irrigation scheduling. Consequently, a two-step (simulation and dynamic programming) approach has been introduced in this study to solve the complex problem of optimal water scheduling for single crop at a farm level operating under flexible water delivery system. The complexity of a real-world situation is represented by incorporating in the optimization model the nature of irrigation water supply and rain fall and the dynamic nature of crop growth. A linear, dated, and additive crop water production function of Doorenbos and Kassam (1979) was incorporated with financial benefits as a useful tool to evaluate alternative irrigation management scenarios. Knowledge of the crop behavior to varying levels of water use through crop growth stages is vital for making rational decisions based on the most economical use of limited water. The developed irrigation scheduling model consists of a database management interfest, a water balance simulation base and dynamic programming. It performs irrigation scheduling under various management options for single crop in a single field. This single crop model tries to find the best allocation of the available water in time in order to maximize the total expected yield of a given crop. It uses several sub- models that interface with each other and with the dynamic programming (DP) to generate the optimal tradeoff between irrigation levels. xii The model is developed under Excel spreadsheet environment and employ mouse-driven approach. It is base on utilizes of a daily water balance approach and uses climatological, crop and soil data as input. The dynamic programming model (DP) allocates a given amount of water optimally over the different growth stages to maximize the yield per hectare for a given crop, accounting for the sensitivity of the crop growth stages to water stress the model was applied to schedule cotton crop in tayba block of Gezira The model was verified by testing it against field data and FAO CROPWAT model. The model-predict seasonal number of irrigation water use and total volume of applied water comparison the DP model and CROPWAT gave significant different. Due non consideration of variation of crop water need with various crop growth stage The developed irrigation scheduling simulation model was explored to simulate improved irrigation scheduling alternatives. Several irrigation strategies were analyzed, including a variety of climatic demand conditions (irrigation at wet and dry periods, at high medium and low crop evapotranspiration rate and at early mid and late sowing dates). The model was used to study the optimal irrigation scheduling of a cala cotton (Sunni Minor) is located in Tyba block Area in Gezira Scheme. Data collected included metrological data soil data, crop and agricultural Data and economic data. The results obtained by the developed model is all mostly typical to the value used in design and for standard practices of Gezira scheme. The developed model is employed as experimental tool to quantify the impacts of changing the most important model input parameters (crop xiii evapotranspiration, sowing date, and rainfall amounts) affecting irrigation scheduling. This is viewed by many investigators as a test of model sensitivity to variations of its inputs on resulting outputs. The model was manage to determine amount and timing of irrigation of cotton crop through the growing season and compared with CROPWAT models. The developed model was found to schedule on farm irrigation and predict cotton yield at different senarious and deficit irrigation water supply satisfactory. conclusion and recommendation plans were stated for policy making Model results concerning the timing of deficit irrigation and determination of watering calendar may be integrated in extension program to build capacity of users and for the future research the developed model can be adapted in the future for prediction of the yield of other crops especially sorghum in rain fed areas .