Water allocation in the Krishna River Basin to improve water productivity in agriculture

After the initial administrative procedures were completed in October 2004, the project was launched through an inception workshop and advisory panel meeting held in Hyderabad on Nov 19th, 2004. In attendance were officials and civil society representatives from the three basin states of Maharashtra, Andhra Pradesh, and Karnataka, in addition to staff of the International Water Management Institute (IWMI), University of Melbourne (UM), and Jawaharlal Nehru Technological University (JNTU). The project goal and objectives, as well as methods, outputs, and dissemination were discussed, clarified, and agreed upon. The advisory panel recommended a range of additional aspects to be included in the project, which have been accepted by the team.

In conjunction with the UM visit to Hyderabad for the advisory panel meeting, the project team visited the Nagarjuna Sagar Left Bank Canal irrigation command area and strengthened contacts with the A.P. Irrigation Department. From this process specific irrigation command areas (Nagarjuna Sagar Left Bank Canal & Musi Medium Irrigation Project), including the major canals of the Musi sub-basin, were identified and prioritised for more in-depth study on water deliveries, return flows, agricultural production, and water productivity trends. The project team also held meetings with JNTU in which students interested in working with the project met with IWMI and UM researchers and identified a list of 6 student research topics.

During the first year, the project team has made significant progress with data assembly including Census data, hydrologic data (stream discharge, rainfall, and meteorological data), water release data (canal release and reservoir release) and remote sensing data (MODIS, LANDSAT, ASTER, AVHRR, IRS, SRTM etc.).

Analyses are underway on the hydrologic and land use trends in the basin. The analysis is progressing to establish a rainfall runoff relationship using a simple conceptual rainfall-runoff model, SYMHYD and a more sophisticated distributed hydrologic routing model SWAT. A modeling study was carried out for Hyderabad city water balance using future water supply scenarios and the results showed that additional water supply to meet the future water demand in 2030 from Nagarjuna Sagar project (NJSP) on the Krishna river to Hyderabad is likely to be a relatively small component of total reservoir releases to irrigated agriculture (5.3-10.2% by 2030). Moreover wastewater irrigation compensates for more than half of the traditional irrigated area lost by bringing water from NJSP.

The first component of the study analysis of the water availability and water use is essential to the remaining two, i.e., water resources management as well as economic and water productivity assessments. The hydrology component has generated three publications submitted for peer review (details in the main section of the annual report); as a result, this component will largely be completed in year 2.

Water balances and resource management assessments of the Musi, Malaprabha, and Upper Bhima sub-basins have been initiated and will be carried forward in year 2 and will form the basis to complete the water resources component of the entire Krishna Basin scale in year 3. This process has allowed the project team to develop a draft methodology of the water allocation modelling framework and to identify the modelling tools and begin a process of familiarisation with them before starting their implementation.

Finally, collection of economic data on urban water use and demand has been initiated in year 1. Work on agricultural water productivity is progressing and data collection is in process. A first approximation of the water-sharing arrangements between the three states in the Krishna Basin has been made; additional information is being collected to improve this. An economic method of evaluating these shares and how future allocations could be evaluated is currently being undertaken. Finally, a start has been made on quantifying urban water use in Hyderabad, including a time series of water sources to the city from 1950-2004. Water productivity assessments and scenario analyses (component 3) are scheduled for years 2 and 3. Year 4 will be for synthesis and dissemination.

Research activities were focused on three sub-basins: Musi, Malaprabha, and Upper Bhima from the three riparian states, Andhra Pradesh, Karnataka and Maharashtra, respectively. The Musi represents a wide spectrum of water use and demand patterns: urban demand in the Hyderabad mega-city area, medium-sized irrigation projects in mid-basin, and the lower catchment dominated by Nagarjuna Sagar, a major irrigation project, and increasing industrial demand in various places. The Malaprabha sub-basin is located in a semi-arid area. The Malaprabha major irrigation project and several groundwater and lift irrigation schemes are located in the sub-basin. The upper Bhima has Pune as the major city in Maharashtra and has water use and water quality components similar to those of the Musi. There is also hydropower generation in the Bhima sub-basin.

The tasks set out in the work plan for the year were largely accomplished. For the WEAP model the basic input file configuration was completed and Krishna basin data loaded. In addition, WEAP and REALM nodes for the basin were identified and data loading is in progress. A preliminary REALM model for the Musi sub-basin is ready with six supply and seven demand nodes (the data gaps to be filled). A concept paper on "water allocation modelling" was presented in an IWMI-TATA conference. The paper outlines the proposed methodology not only for assessing the social costs and benefits of competing demands for water but also for encompassing a number of issues such as urbanization, extensive spatial and temporal variability, public investment in infrastructure and changing water-quality parameters.

In the course of the year, several research papers, journals articles and conference papers were published. A research report on water balance and changes in the hydrology of the Krishna basin demonstrates that the largest water transfers takes place from the Upper to the Lower Krishna river. The same study report shows that the development of irrigation projects and rainwater harvesting schemes have dramatically decreased outflow to the ocean from 28 percent of rainfall in the first 60 years of the twentieth century to almost 5 percent over the remaining 40 years and thereby effectively "closing" the basin. Another article presents the mapping of diverse irrigated areas in the Krishna river basin using an irrigated fraction approach and MODIS images (year 2000). The mapping indicated that out of the total basin area, 24 percent was irrigated in 2000. The other articles focus on assessment of water resources and water balance modelling in the Musi sub-basin. This sub-basin has a very large area irrigated by different sources such as canals, tanks, groundwater and wastewater. The upper part of this sub-basin is dry compared to the lower part and the runoff to rainfall coefficient varies from 0.07 to 0.24. Altogether, 854 million m3 (MCM) of water are available in the Musi sub-basin, with 170 MCM and 110 MCM imported from Godavari and Nagarjuna Sagar, respectively, for the Hyderabad city. A water balance model was developed for the Hyderabad city to study the changes in the local water cycle and evaluate the potential water sources within the city. The net groundwater recharge from the city was estimated as 39 MCM which is much less than the groundwater extraction for urban use (60 MCM). The total outflow from Hyderabad city in the form of runoff and sewerage is 298 MCM, where the runoff is the major component with 56 percent contribution, which is the result of the urbanization. It is anticipated that 30 percent of the total water will be met by the Krishna river, which will place some stress on agricultural water supply in the Nagarjuna Sagar reservoir during dry years. The spatial variation in land use of the Nagarjuna Sagar command was assessed for a dry year and it was found that head-enders suffered most due to lack of information on reservoir operation and no alternative source of irrigation.

The economics component of the project is well under way. The initial focus is on the Musi sub-basin. The economic model has been constructed with the objective of linking it to the hydrological model. Thus, the economic implications of changes in the allocation of water to different nodes can be assessed on a monthly basis. The economic model is deterministic, but one that is capable of being changed into a stochastic one, if required. The methodology used in the Musi will be applied to the other sub-basins selected for study.

A study in the Lower Krishna river basin looks at the evolution of formal and informal institutional arrangements and agrarian changes over time in response to growing water scarcity. This study will come up with an effective governance system that would allow water allocations, which satisfy human needs and environment requirements and potentially minimize water-related conflicts in the basin.

A meeting of the Project Advisory Panel was held on November 15, 2005. Representatives from the three riparian states attended the meeting. The panel supported the overall direction of the project, in particular, the two-tier approach to study water allocation at the basin and sub-basin levels. The panel recommended that groundwater assessments should be included in the project.

During the mid term review held in Nov 2006, it was agreed that inter-sectoral water allocation modeling would focus on the three sub-basins: Musi, Malaprabha, and Upper Bhima from the three riparian state of the Krishna basin namely Andhra Pradesh, Karnataka and Maharashtra respectively.

During the period under review (June 2006 - June 2007), the work on the project centered on the following activities:
Groundwater modeling in Musi sub-basin to estimate ground water recharge.
The development and completion of a water allocation model for the Musi that includes a groundwater component.
The simulation of water allocation scenarios in the Musi sub-basin.
The development of an approach that can be applied to assess the economic components of various water allocation scenarios.
The completion of a Social Benefit Cost analysis of water allocations in the Musi sub-basin.
An assessment of the relative values of water in the Musi sub-basin
Hydrological modeling for Malaprabha.
Setting up the Soil and Water Assessment Tool (SWAT) model for Upper Bhima.

Musi Sub-basin (Andhra Pradesh)

The Musi sub-basin represents a wide spectrum of water use and demand patterns: urban demand for domestic and industrial use in the greater Hyderabad City area. Agricultural demands include a medium-sized irrigation project in the middle of the basin and irrigated area in lower reaches as a part of major irrigation project, Nagarjuna Sagar, in the lower catchment and ground water irrigation in the entire sub-basin. Groundwater overexploitation of the Musi sub-basin has been quantified using historical piezometric time-series analysis. Based on records for 1989-2005, it was found that 2.5 m of saturated depth has been depleted with a depletion rate of 0.18 m per year. A physical groundwater model, MODFLOW was used to estimate the groundwater recharge. It was estimated that the average recharge during monsoon was 17% of the mean total annual rainfall (1,180 Mm3). The sustainable utilizable water resource was estimated at 1,220 Mm3. It was also found that the artificial recharge structures could enhance natural recharge by 11%, which warrants further investigation.

REALM was used to build the Musi water allocation model, and simulate the water supply system. In the model, the Musi sub-basin is integrated with the Nagarjuna Sagar project, which lies outside the sub-basin but is one of the water sources for Hyderabad city and irrigation. The calculation of the Musi water balance, completed in Year 2 of the project, was improved with the addition of a groundwater modeling component. The model now consists of 11 supply nodes and 14 demand centers.

A scenario analysis was used to investigate possible changes to water allocations in the Musi sub-basin. For each scenario, levels of assured supply were estimated for each demand center. Results from this analysis indicate that competition for water is very high. It was found that transfers of water from agriculture to urban use may well grow over years and the irrigation allocation is going to worsen over time. To improve the productivity of available irrigation water, improvement of traditional irrigation efficiency and crop diversification is essential. Additional scenarios analyzed include satisfying future urban demand from the Nagarjuna Sagar project alone and the impact of reduced stream flow and crop diversification.

In assessing the relative values of water in the Musi sub-basin, it was necessary to adapt an approach (Social Benefit Cost analysis) discussed earlier in the project into a model of the sub-basin. This was done in a two stage procedure. First a model of the private benefits and costs was developed of the surface flows of water for one year. In this model no account was made for the influence of time. Its purpose was to evaluate whether the approach would work and was undertaken because of data limitations that existed at the time. In the second stage, as data became available, the model was expanded to include groundwater, a separate node for the wastewater region and time. This model was then linked to the water allocation model of the sub-basin. The modeling results indicate that the average net values placed on water used for agricultural purposes ranged from Rs.23.68/m3 in the wastewater region to Rs.55.44/m3 in the region upstream from Hyderabad.

Malaprabha Sub-basin (Karnataka)

The Malaprabha sub-basin is located in a semiarid area of the Krishna basin. The Malaprabha Major Irrigation Project and several groundwater and lift irrigation schemes are located in the basin. The current drinking and industrial demand are not significant, but are expected to grow to 20% of total water available in coming years. There is only one major reservoir, Malaprabha, which has a live storage of 866 Mm3; yet with a 75% dependable inflow of 608 Mm3, the reservoir fills to 763 Mm3. There is a provision to import 198 Mm3 of water to secure assured supply in the command and also to export 156 Mm3 from Ramthal lift irrigation scheme to the Middle Krishna sub-basin. The SYMHYD model was set up for Malaprabha using reservoir inflow data. The simulation period was from 1993 to 2003 (however, as no data existed for 2002 it was excluded from the analysis). From the model, runoff was predicted to be 40% of rainfall and 60% went to either groundwater or evapotranspiration (ET). Based on district-wise groundwater statistics, the groundwater potential was estimated to be approximately 20% of the rainfall. A REALM model is being set up for two supply node, the Malaprabha reservoir and groundwater. The demand nodes include the lift irrigation schemes, left and right canals, the Kochi weir and urban (in Hubli-Dharwar) and rural drinking water demand. Data population of the model is currently in progress. As with the analysis of the Musi sub-basin, an economic component will also need to be incorporated into the model. The preliminary scenarios to be developed include assessing the effects of projected urban demand, the impacts of export and import into the basin.

Upper Bhima Sub-basin (Maharashtra)

The hydrology and degree of water resource development in the Upper Bhima sub-basin is a microcosm of the whole Krishna basin. There is a gross storage of approximately 8,194 Mm3 in eight major and more than 30 medium reservoirs in the sub-basin. During year 2004-05, out of total water available (8,700 Mm3), 1,274 (15%) was diverted westward for hydropower generation from Tata Hydel Power projects. The water uses for non-irrigation and irrigation were accounted for 25% (2,207 Mm3) and 35% (3,022 Mm3) of total water available, respectively while 25% (2,167 Mm3) resulted in excess runoff. During the drought in 2002-04, the downstream projects suffered drastically when total water available reduced to half and dead storages were used to meet the drinking water needs without any supply for irrigation. There is a provision to import 2,238 Mm3 from Upper Krishna sub-basin to augment water supply in water scarce regions. Agricultural water use in the Upper Bhima sub-basin was assessed by integrating ET data derived from remote sensing techniques with a process based hydrological model, SWAT. The calibrated model was used to derive a monthly basin water balance for the irrigation year 2004-2005 and the agricultural water use was quantified. The ET in the sub-basin was estimated to be 38,172 Mm3 (835 mm). Of the total evaporative depletion, 42% was estimated to be non-beneficial. Thus, different water uses for this non-beneficial use can be explored. The model will be recalibrated with finer land use resolution data and simultaneously REALM nodes will be established.

The aim in this project is to investigate water availability and allocation issues in the Krishna Basin of India. This Basin is nearly a closed basin (i.e. all the water available is fully allocated to some purpose for a large part of the year), densely populated with some 73 million people, crosses jurisdictional boundaries, experiences both spatial and sectoral demand pressures for more water and suffers from a number of environmental pressures. This basin as a whole is suffering from severe water scarcity and the problem is only expected to get worse as populations grow, economic development occurs and readily available low cost supplies of water become exploited. The complexity of the water system in the basin requires the development of specialised analytical tools to inform policy makers of the impacts different water allocation strategies may have. The information provided to policy makers and stakeholders with these tools is derived from analyses that link total water availability in a water allocation model, with an economic framework that has the capability to measure the regional and inter-sectoral impacts of alternative allocations of water.
During 2007-08 much of the initial thinking and work undertaken in this project was brought together into a set of models that explain the allocation of water in three sub-catchments within the basin. The most complex sub-basin modelled (the Musi) was the most reflective of the concerns addressed in the study. The water allocation model drew on the hydrology of surface and ground water supplies in the Musi basin and aligned them with the agricultural, industrial household and power generation demands for the resource. These physical quantities were valued and incorporated within a Social Cost-Benefit analysis. Simulations were then conducted to ascertain the impacts of changes in climate, inter-sectorial demand changes and different cropping patterns over the next two decades. The model was used to simulate scenarios for the Musi catchment that included supplying all future urban water requirements from the Nagarjuna Sagar reservoir, implementing water harvesting structures, various climatic effects and changing cropping patterns. It was found, unsurprisingly, that while a majority of the water was used in agriculture, it only yielded a return of between Rs.1 and 2/m3. While only a limited amount of water was used in Hyderabad City, it returned over Rs.50 and 10/m3 to household and industrial users, respectively. Supplying more water from Nagarjuna Sagar to Hyderabad City reduced the overall benefits of the system (as the costs of providing water outweighed any benefits to users), yet had little impact on agricultural producers. Conversely, changes in both climate and the cropping patterns were found to have little impact on the overall worth of the system, yet had some impact on agricultural water users.
Models of the Malaprabha and Upper Bhima catchments are also close to completion. These two models were somewhat simpler, as ground water was not (to date) incorporated in the final versions of the model and inter-sectorial demand pressures are great concern to stakeholders and policy makers in these two sub catchments.
During 2007-08 great strides were made to interact with various stakeholders and policy makers in the basin. Project stakeholder workshops were organised in Hyderabad, Bangalore and Mumbai to coincide with the final review of the project. The aims in these workshops are not only to inform and gain valuable insights for the evaluation process, but also to discuss the possible alternate allocation scenarios and possible benefits. The workshops were attended by officials from State and Central Government agencies and the collaborating scientists in India and from Australia. The workshops provided a vehicle for project scientists to present concepts and project results. The deliberations during the workshops provided a good feed back and fostered closer collaboration between the stakeholders and the researchers. The general exchanges and attitudes of the stakeholders in these meetings to basin water management, in particular, and to the project in general, has changed considerably since 2004. The three riparian states are now more aware of the inter-sectoral issues and the productivity of water use across the whole of the basin. There is also more openness to and awareness of integrated basin water management amongst the stakeholders. At these meetings a number of competing and opposing exchanges between State officials occurred, highlighting the sensitivity of the sectorial and State interest over the whole-of-basin productivity.
In April, 2008, high level officials from all three states participated in a tour of the Murray-Darling basin, Australia. The aim of the tour was to provide the participants with an insight into the challenges and opportunities of aspects of water management in the Murray-Darling Basin. During the tour participants gained an understanding of the strategies adopted by the Murray - Darling basin authorities to cope with the problems of water scarcity. The perspectives gained by participants have important implications for the even more acutely water stressed Krishna Basin.
In addition, it is likely that the stakeholder meetings and associated activities, including the tour, have started harmonising the conflicting positions held by individual states and their policy makers, as they provide a forum for exchanging ideas in a neutral and non-threatening environment.
A final review of the project was held in February 2008 and a one year extension was sought and granted. The main project activities during the extension period are to:
1: Finalize the existing hydro-economic sub-basin and State scale models
2: Apply the models to a suite of end-user defined water resource allocation and
3: Gain acceptance of the modelling outputs amongst State and Federal policy makers