Optimal water application decisions with deficit irrigation

Abstract

A linear programming based optimization model was developed for tubewell irrigation system in the high Barind tract area of Bangladesh to maximize profit for wheat and Boro rice from the available land and water supply. The study area comprises four thanas, viz. Tanore of Rajshahi district, Nachole and Gomstapur thanas of Chapai Nawabgonj district and Niamatpur thana ofNaogaon district. The area is within a deep tubewell irrigation project and the available land area for irrigation is about 90660 ha. There are 1463 deep tubewells in the study area and the design capacity of each tubewell is about 56 litre per second. Available water for irrigation from 1463 tubewells in the area ranges from 13355 to 14786 ha.m at design discharge during the irrigation season from November to May assuming 16 operating hours a day. At 80% and 60% of design discharge, the water availability varies from 10684 to 11829 ha.m and 8013 to 8871 ha.m, respectively over the irrigation season. Irrigation equalling full crop water requirement and 10, 20, 30, and 40 percent deficit crop water requirement along with 20, 50 and 80 percent dependable rainfall were considered in the model. Deficit irrigation was applied at vegetative and yield formation stages of Boro rice and wheat. For estimating yields of Boro rice and wheat under different crop water requirements, yield response factors at vegetative and yield formation stages of the crops were determined by field experiments. The values for Boro rice were 1.53 and 0.60 at vegetative stage and 0.29 and 0.28 at yield formation stage, respectively in the first and second years offield experiments. The corresponding values for wheat were 0.21 and 0.18 at vegetative stage and 0.47 and 0.46 at yield formation stage. The model was first solved without any constraint on land area under Boro rice and wheat using both experimental yields and current farmer's yields. In the solution obtained with experimental yields, all area was covered with wheat whereas with farmer's yields about 98% ofthe area was covered by wheat, only 2% area being under xvi i • Boro rice at full discharge level of tubewells. At 80% and 60% of design discharge of .tubewells, all area was found to be under wheat. In this situation, the seasonal profit using experimental yields were found to be 140 to 155% higher than that obtained with farmer's yields. Next the model was solved with constraints on the maximum and minimum areas under each crop using both experimental and farmer's yields. For Boro rice, the specified maximum and minimum areas were 50000 and 31449 ha, respectively and those for wheat were 60000 and 3613 ha, respectively. At full discharge level, after satisfying the requirement of minimum area under Boro rice, the remaining area was found to be under wheat. At 80% design discharge, the minimum area specified under Boro rice was found to split under 20% and 30% deficit irrigation, the remaining area being under wheat .mostly at 40% deficit irrigation. In this case, the seasonal profit using experimental yields were found to be 27 to 71% higher than that obtained with farmer's yields. Comparing the incremental profit with and without any constraint on area under crops, the profit under unrestricted condition was found to be higher. However, all area under wheat, as found in the case of unrestricted situation, may not be acceptable to the farmers who are mostly rice growers. It seems judicious to consider 80% of the design discharge of tubewells in irrigation planning as the pump efficiency gradually decreases with time thereby reducing the amount of pumping water. Under restricted condition, deficit irrigation appears in solution and keeping Boro area close to the present practice, remaining land area is left for wheat thus encouraging crop diversification. Practicing deficit irrigation, not only the existing farmers of the project will be benefited but also additional farmers will be benefited from BMDA deep tubewells.