OPTIMAL UTILIZATION OF WATER RESOURCES IN TRANSBOUNDARY KRISHNA RIVER BASIN

Abstract

In planning optimal development of the water resources of a river basin screening models based on optimization techniques are very much useful for obtaining reasonable estimates of system yields for dependable release reliability targets. The Krishna river basin in India consisting of 12 sub-basins having 126 projects, i.e., 59 major and 67 medium projects is considered for study. Apart from these projects, water demands of9 projects are also included. In this study an attempt has been made to develop an implicit stochastic screening model based on linear programming (LP) that would assess the optimal integrated annual system yield from a large river basin such as the Krishna, consisting ofreservoirs and barrages. The developed model is termed as the modified integrated yield model (MIYM). The Krishna which is a transboundary river passes through the states of Maharashtra, Karnataka and Andhra Pradesh. The water utilizations by these co-basin riparian states are subjected to the allocations defined by the Krishna Water Disputes Tribunal (KWDT) award. Hence, multiobjective criterion is now introduced by identifying relevant development planning objectives and defining the relative importance of each of these objectives. For this purpose the MIYM is further improved and is termed as the multiobjective modified integratedyield model(MOIYM). The KWDT allocated the 75 percent water year dependable flow of 2060 TMC (58332.6 MCM) of Krishna river among the riparian states of Maharashtra, Karnataka and Andhra Pradesh in the ratios of0.2718:0.3398:0.3883 as per its final orders in 1976. As a post-award development, by a separate agreement, each state has agreed to spare 141.67 MCM of each of their shares in Krishna waters to Tamil Nadu towards the drinking water supply ofChennai city through the Telugu Ganga canal. The earlier existing reservoir yield models have been suitably modified to overcome certain limitations and eliminate possible causes of infeasibility conditions arising in the model applications. Four types of water uses are considered for a multipurpose project, viz, municipal demand, industrial demand, irrigation and hydropower generation. The model also determines import and export ofwater required in within-year time periods. For this within-year storage continuity constraint, IV continuity of annual yields and energy constraints have been modified by introducing excess firm and secondary releases in within-year time period. Further additional constraints pertaining to the KWDT award also have been included. The models also simultaneously optimize the cropping patterns at each project. The model consists of over-year and within-year time period constraints. The model estimates multiple annual and within-year time period yields, over-year and within-year time period storages and their capacities, crop plans, hydropower, exports and imports. The first objective considered in the models is represented by the maximization of annual system yield. The second objective considered is represented by. the minimization of the total absolute deviations of water used by each state from the water allocated to each state as per the tribunal decisions. The third objective considered is represented by the maximization ofannual gross benefit. The model assessed either the optimal integrated annual system yield (firm and secondary) from the basin or maximum annual gross benefit of the system subject to minimal deviations. The inflow series of28 years are considered for computation purposes. For the purpose of this study 28 over-year time periods, and 12 within-year time periods for the critical year only are considered with the water year starting from the month of June and ending in May. Annual reliabilities for the firm and secondary yields considered are 97 percent and 76 percent, respectively. The net inflow series at each project were calculated by the basin water balance method from the discharge data available at nearby river gauging site. Failure years at each project were identified from the respective net inflow series. The inflow fractions in within-year time periods are calculated for each reservoir considering inflow of the driest year. Storage area curves (linearized over dead storage) are used for computation of evaporation parameters. The within-year and over-year storages are obtained through model runs for projects having irrigation as a main purpose, and then average head corresponding to the storages available in each within-year time period is calculated by using storage-elevation relationship developed for the project. In case ofhydropower projects having number of power plants at different levels, weighted average head and combined plant and discharge capacities are considered in the analysis. The gross irrigation water requirements at each within-year time period of the proposed crop plan under each project was estimated by FAO methodology. Population ofthe basin in year 2001 is calculated from the district census data and then projected for year 2050. Population of a sub-basin is distributed proportionately among all the projects in proportion oftheir respective culturable command area (CCA). Municipal and industrial water demand at each project is calculated for projected population. The projected annual water demands in the basin by the year 2050 are estimated to be 67387.68 MCM. Site-specific values of allowable percentage yield (failure fraction) for satisfying the project specific demands as far as possible in successful years have been considered in the study. Protein and calorie requirements ofthe total as well as ofthe agricultural population have been computed. The LINDO software is used for solving the LP problem. A separate computer program is used to generate the model constraints and objective functions as input data required in LINDO format. Assessment of the water resources potential of the Krishna basin is presented at the Krishna basin level, at the sub-basin levels, and at the project levels wherever required. Detailed discussion is made on the system yield, irrigation and hydropower, benefit, nutritional requirements, and, exports and imports. The trade-off analysis is carried out for two options in detail; these are (i) the option ofzero absolute deviations from the KWDT award allocations and (ii) the best compromise solution. The study is limited to the surface water resources and also for major and medium projects only. The problem is solved in three stages. In Stage-A, single objective analysis is made using MIYM for three cases, i.e., maximizing annual system yield, maximizing annual gross benefits and minimizing total absolute deviations from the KWDT award allocations. Multiobjective analysis is done in Stage-B for three cases using MOIYM. Constraint method of solving multi-objective linear programming problem has been adopted for the solution. In MOIYM the first case was solved for only one trial with the objectives, i.e., maximizing annual system yield subject to minimum absolute deviations objective as a constraint. The second case was solved for maximizing annual gross benefit subject to minimum absolute deviation objective as a constraint, and for this case 13 options have been tried. The third case is solved for maximizing annual gross benefit subject to minimum deviations and optimal annual system yield objectives as the constraints. The Stage-C is solved for multiobjective analysis with inter-basin water transfers. The trade-off analysis has been done for the second case in Stage-B.. The thirteen options, which were tried, are for the deviations ranging between zero and the maximum. On the basis ofthe trade-offs and marginal curves analysis, the best compromise solution is selected. VI Without any conflicts in the river water shares the optimal annual system yield achieved would be 75332.47 MCM. But with the KWDT award considerations, i.e., with zero absolute deviations this yield is likely to reduce by 11.57 percent. In this case the Musi sub-basin would face the maximum deficit of 60.32 percent in meeting its annual water demands, whereas the middle Krishna would have the maximum surplus by 45.94 percent over its annual water demands. Out of the 126 major and medium projects in the river system about 14 of them would be most successful and would not face any failure years, whereas 53 projects would fail in meeting their respective target annual water demands. About 31.5 percent CCA of the basin would be annually irrigated with an enhanced irrigation intensity of 1.18. About 37 projects need crop diversifications. The minimum and maximum irrigation intensities achieved would be 0.15 and 1.68 at Visapur in the Upper Bhima and Vanivilas Sagar in Vedavathi subbasins, respectively. The Krishna basin is likely to meet the nutritional requirements completely of the agricultural population only. The system would provide 11768.11 MU ofhydropower annually. The annual gross returns from all the water uses are likely to be 206175.9 million rupees. From the trade-off analysis for the best compromise solution achieved the basin shows an increase, in the annual system yield of 3.09 percent, annual gross benefit of 9.63 percent and annual intensity of 2.59 percent than the respective values at zero deviation. However the water shares achieved would be inthe ratios of027&i 0.3356: 0.3BQfl. The use ofMOIYM could reduce the size of the problem significantly to 24729 variables and 25425 constraints for the system. Due to the model's overall flexibility in application and computational efficiency; it can be applied to any other transboundary river basin by including in the model, additional constraints pertaining to the specific issues related to the river water share.