India - Pilot Project for Watershed Development in Rainfed Areas

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 15107 PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (CREDIT 1424-IN) NOVEMBER 20, 1995 Agricultural and Water Operations Division Country Department II South Asia Regional Office This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (Credit 1424-IN) CURRENCY EQUIVALENTS At appraisal (September 1984) - US$1.00 = Rupees (Rs.) 9.75 At PCR (June 1994) - US$1.00 = Rs.30.85 FISCAL YEAR OF BORROWER GOI: April 1 to March 31 ABBREVIATIONS AED Agricultural Engineering Directorate AEO Agriculture Extension Officer APAU Andhra Pradesh Agriculture University CIAE Central Institute of Agricultural Engineering CS&WCR&TI Central Soil and Water Conservation Research and Training Institute DA Department of Agriculture DEA Department of Economic Affairs DPAP Drought Prone Areas Project DwDC District Watershed Development Committee FAO/CP Food and Agriculture Organization Cooperative Programme GOI Government of India ICAR Indian Council of Agricultural Research ICRISAT International Crops Research Institute for the Semi-arid Tropics IDA International Development Association IG&FRI Indian Grasslands Fodder Research Institute M&E Monitoring and Evaluation MOA Ministry of Agriculture NARP National Agricultural Research Project NBSS&LUP National Bureau of Soil Survey and Landuse Planning NLRCDC National Land Resources Conservation and Development Commission NWDPC National Watershed Development Policy Committee ORP Operating Research Programme PKV Punjabrao University (Maharashtra) SAU State Agriculture Universities SWDC State Watershed Development Cell SWFC Subswatershed Farmer Committee SWDPC State Watershed Development Policy Committee UAS University of Agriculture Science (Karnataka) WDC Watershed Development Council WDT Watershed Development Team  FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operations Evaluabon November20, 1995 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on India Pilot Project for Watershed Development in Rainfed Areas (Credit 1424-IN) The India Pilot Project for Watershed Development in Rainfed Areas, supported by Credit 1424- IN for US$3 1.0 million equivalent, was approved in December 1983. The Project Completion Report (PCR) was prepared by the South Asia Regional Office. No comments on the PCR were received from the Borrower. The project had two objectives. First, to develop technologies that would increase crop, forage, fuelwood, and timber production, and reduce the variation in crop yields, in selected rainfed areas of Andhra Pradesh, Karnataka, Madhya Pradesh, and Maharashtra, totalling about 250,000 ha. Second, to develop criteria, procedures, and guidelines for replication in other areas of successful technologies. The main components of the project were the development of arable and non-arable lands through soil and moisture conservation works; strengthening extension systems and farmer access to inputs; and the strengthening of training, research, and administrative institutions at the State and watershed levels. The project succeeded as a pilot action because it promoted innovation and learning, some of which is paying off on a broader scale in Bank-supported projects and government programs. But implementation was marked by several weaknesses. The project paid insufficient attention to land tenure and ownership, and beneficiary participation, required by the project's design, was minimal in practice. Monitoring and evaluation were underdeveloped and, as a result, data are insufficient to recalculate an economic rate of return. The decision to adopt vegetative techniques for soil and water conservation following the mid-term review was sound, but the exclusive use of vetiver grass limited the impact of these measures. The outcome of the project is rated as satisfactory because of the learning and that took place. Institutional development is rated as modest because arrangements for working with villagers on common and public lands were underdeveloped. Sustainability is rated as uncertain because beneficiary participation was not sought and farmers were reluctant to take responsibility for activities about which they had not been consulted. The main lessons of implementation are that pilot programs must be managed as experiments, with higher than normal standards of monitoring and evaluation; improving resource conservation on common lands requires equal attention to conservation techniques and to institutional arrangements for management; and, farmers will adopt soil and water conservation methods on their private lands if the incentives they face make it sufficiently attractive. No audit is planned. Attachment This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   FOR OFFICIAL USE ONLY PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (Credit 1424-IN) TABLE OF CONTENTS PREFACE ................................................. i EVALUATION SUMMARY ................................................ ii PART I. PROJECT REVIEW FROM BANK'S PERSPECTIVE ................... I 1. Project Identity .1 2. Background . 3. Project Objectives and Description .2 4. Project Design and Organization. 3 5. Project Implementation. 5 6. Project Results . I1 7. Project Sustainability ......................14...................... .. 1 4 8. Bank Performance .15 9. Borrower Performance.15 10. Lessons Learned .1 6 11. Adherence to Covenants .16 12. Consulting Services .17 13. Project Relationship. 1 7 14. Project Documentation and Data .17 PART II. PROJECT REVIEW FROM BORROWER'S PERSPECTIVE ................ 1 8 (not received) This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed wiLhout World Bank authorization.  PART III. STATISTICAL INFORMATION ............... 19 1. Related Bank Loans and Credits .19 2. Project Timetable .19 3. Credit Disbursement .20 4. Project Implementation ................................................................... 21 A. State-wise and Treatment-wise Targets and Achievements .21 B . Details of Targets and Achievements for Watershed Treatments .22 5. Project Cost and Financing ..25 A. Project Cost .25 B . Project Financing .27 6. Project Results ..27 A. Direct Benefits .27 B. Economic Impact .28 C. Financial Impact .28 7. Status of Covenants .29 8. Use of Bank Resources. 30 A. Staff Inputs. 30  i PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (Credit 1424-IN) PREFACE This is the Project Completion Report (PCR) for the Pilot Project for Watershed Development in Rainfed Areas in India for which Credit 1424-IN in the amount of SDR 29.4 million (US$ 31.0 million equivalent) was approved on Decemb r 8, 1983. SDR 5.2 million and SDR 3.0 million were cancelled on December 5, 1991 and May 1, 1993, respectively, as part of the funds redeployment exercise. The Credit was closed on December 31, 1993, two years behind schedule. With Government of India's (GOI) refund of the unused balance of the Special Account on June 15, 1994, the undisbursed balance of SDR 2.16 million was cancelled. Parts I and III of the PCR were prepared by a Food Agriculture Organization/World Bank Cooperative Program (FAO/CP) mission which visited India in May 1994. Borrower's Part ll and comments have been requested but have not been received. PCR preparation was based on a review of the Staff Appraisal Report (SAR) and legal documents, supervision reports, project files, and correspondence, field visits to project sites in all the four states where the project was implemented, and discussions with officers of the project implementing agencies concerned and Bank staff associated with the project.   iii PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (Credit 1424-IN) Evaluation Summary Objectives 1. The overall objectives of the pilot project were to: (a) develop, on a watershed basis, and through the use of appropriate organization and management systems, suitable technologies for increasing and stabilizing crop and forage yields, as well as the production of fuelwood and timber in selected rainfed farming areas; and (b) define technical criteria, administrative procedures and investment limits useful in replicating watershed development activities in other areas. Thus, the emphasis was not only on the development of technologies and the parameters for their implementation, but also on the establishment of an organizational and institutional structure for their sustainable implementation. 2 . The means to attain the above objectives covered a range of measures that included the implementation of improved soil and moisture management measures, better crop and grassland husbandry practices and reforestation to improve productivity of both crop and non-arable lands. The measures are directed so as to mitigate the effects of variation in the quantity and distribution of rainfalll by exploiting such opportunities as: (a) introducing more stable farming systems; (b) retaining more water in the soil profile so as to recharge the groundwater for expanding or increasing the intensity of irrigation; and (c) controlling erosion by providing safe disposal of runoff and drainage. 3. The project provided for the development of about 250,000 ha of land covering eight watersheds, two each in the states of Andhra Pradesh, Kamataka, Maharashtra, and Madhya Pradesh. The main components of the project were: (a) the development of crop land through conservation works; (b) stabilization or improvement in crop production through strengthening extension services and access to inputs; (c) the development of non-arable lands through conservation works and forest plantations; (d) provision of increased staff training, research and operational research programs; and (e) institutional strengthening of project administration at State and Watershed levels. Project support also included the strengthening of the Watershed Development Council of the Government of India (GOI), monitoring and evaluation of project progress, and the preparation of a possible follow-on project. Implementation Experience 4. Despite an initial delay in project start-up, and a subsequent change in project focus from a concentration on engineering structures to the use of vegetative approaches to soil and water conservation, the project's physical targets were largely achieved. Soil and moisture conservation on arable lands included a combination of measures, such as land smoothing, graded bunds and grass waterways, diversion channels/interceptor drains, waste weirs and check structures, and, particularly, the promotion of in-field vegetative barriers. These have been considered very effective by farmers. The promotion of only one type of in-field vegetative barriers, vetiver (khus), however, limited the prospect of an even wider impact of conservation measures on arable private lands. Afforestation on private lands was also very successful and the majority of farmers  iv have continued to protect and maintain their woodlots. Adaptive research covering in situ vegetative soil and moisture conservation measures in the project area was also effectively carried out by all the three research agencies in Andhra Pradesh, Karnataka and Maharashtra. Quantitative parameters are now available for a variety of conservation measures. Project implementation with its technological bias, however, overlooked land tenure or ownership and other general sociological issues. As a consequence, the same technical conservation measures did not succeed on the non-arable or common property lands. Results 5. The project was more successful in meeting its physical objectives than in attaining its institutional or organizational goals. It has been successful in demonstrating the benefits from a range of soil and moisture management measures as evidenced by the satisfactory adoption rates of the technologies promoted by the project. The introduction of vegetative techniques (vetiver bainiers) for soil and moisture conservation has been successful in parts of the areas, but the over- emphasis on one particular species to the exclusion of others has limited a wider possible impact. Research in alternative approaches started toward the end of the project and was unable to have any major impact. 6. Improved agronomic techniques (contour cultivation and vegetative barriers) and inputs (seeds, fertilizers and pesticides) along with project conservation measures, had some impact on agricultural production. However, a fortuitous change in relative prices provided the incentive for farmers in general to respond. As a consequence, farm incomes from non-project farmers have also increased almost at the same rate as that of participating farmers. It is, therefore, difficult to isolate and quantify the benefits and the associated costs from specific project supported investments. The objective of establishing financial parameters to evaluate cost effectiveness was not achieved. Also, the pilot nature of the project, which covers a wide range of management measures not all of which were successful, does not permit a rational evaluation of the economic rate of return (ERR) even for comparison purposes. Furthermore, the project, either through its monitoring mechanism or evaluation studies, had not compiled any data on the costs and benefits of interventions in any individual watersheds. For these reasons, it was not possible for the PCR mission to reestimate ERRs for individual watersheds, as was done at appraisal. Sustainability 7. While farmer adoption rates of the project-promoted soil and moisture conservation measures have been satisfactory, beneficiary involvement in project implementation was minimal, and the continued implementation of project components by project staff is not sustainable for the Government. The experience with conservation measures on private arable land, however, indicates acceptance and maintenance of structures where a distinct and clear linkage to increased productivity can be established. Conservation and afforestation measures on private non-arable land also appear sustainable as farmers have undertaken active maintenance measures. However, the sustainability issue--both in respect of the methods of implementation and maintenance works-- will remain a clear problem for watershed management on common property resources unless appropriate mechanisms for community participation and benefit-sharing are developed. Findings and Lessons Learned 8 . There are two important lessons to be learned from the project. The first relates to the requirement of flexibility. Project design should ensure flexibility to tune the prescriptions for watershed management to the specific characteristics of the watersheds. Flexibility should include,  v not only the adoption of measures to cover a wide range of micro-agro-ecological environmenits (with varying rainfall, slope and soil conditions), but also the prospect of a combination of both vegetative and limited engineering solutions. The second relates to the need for explicit farmer insvolvement as part of implementation strategy. Sustainability issues, both from the point of view of cost efficiency of the conservation measures. as well as for the maintenance of the stuLictures constuLicted, require an active beneficiary involvement in watershed management.   I PROJECT COMPLETION REPORT INDIA A PILOT PROJECT FOR WATERSHED DEVELOPMENT IN RAINFED AREAS (Credit No. 1424-IN) PART I. PROJECT REVIEW FROM BANK'S PERSPECTIVE 1. Project Identity Project Name : A Pilot Project for Watershed Development in Rainfed Areas Credit No. : 1424-IN RVP Unit : South Asia Country : India Sector : Agriculture Sub-sector Rainfed Area Development 2. Background 2. 1 About 70 percent of the net cropped area in India is based on rainfed agriculture. Even at full development of irrigation in the distant future, about 105 million hectares of crop lands would remain rainfed. National investment priorities in India have been however traditionally directed toward the development of irrigable lands, and, until the early 1980s, relatively little economically promising new technology was available to farmers in rainfed areas. The risk of crop failure due to the hazards of weather has been deterring farmers from investing in inputs and expensive crop care, resulting in generally low and stagnant productivity from rainfed areas. Compounding this overall environment in most rainfed farming areas, increasing population and livestock numbers have led to denudation of vegetation for fuel and feed, and increasing cultivation of fallow, often marginal and sloping lands leading to accelerated erosion. It was, therefore, considered urgent to develop, at least on a pilot basis, a program for the development of rainfed farming areas. Based on past experience, two aspects were recognized. First, the sustenance and income of the Indian rainfed farmer depended on the crop and livestock products derived by him from his arable land and from public land in his vicinity. Second, rainfed farming development has to address one key problem: that is to optimize, in the long-term, the utilization of rainfall for plant and animal production within a catchment area. 2.2 It was against the above background that the project in question was conceived to develop technologies and mechanisms for increasing agricultural production in areas receiving reasonably dependable rainfall by improved land and crop management, intercropping and sequence cropping, and development of fodder and fuel in non-arable areas.  2 3. Project Objectives and Description 3.1 Project Objectives. The overall objectives of the pilot project were: (a) to develop, on a watershed basis, and through the use of appropriate organization and management systems, suitable technologies for increasing and stabilizing crop and forage yields, as well as the production of fuelwood and timber in selected rainfed farming areas; and (b) to define technical criter-ia, administrative procedures, and investment limits useful in replicating watershed development activities in other areas. Thus, the emphasis was not solely on the development of technologies and the parameters for their implementation, but also on the establishment of an oreanizational and institutional structure for their sustainable implementation. 3.2 The mneans to attain the above objectives covered a range of measures that included the implementation of improved soil and moisture management interventions, better crop and grassland husbandry practices, and reforestation, to improve productivity of both crop and non-arable lands. These measures were directed so as to mitigate the effects of variation in the quantity and distribution of rainfall by exploiting such opportunities as: (a) introducing more stable farming systems; (b) retaininig more water in the soil profile so as to recharge the groundwater for expandinig or increasing the intensity of irrigation; and (c) controlling erosion by providing safe disposal of runoff and drainage. 3.3 Project Description. The project provided for the development of about 250,000 ha cover-inig eight watersheds: two each in the states of Andhra Pradesh, Karnataka, Maharashtra and Madhya Pradesh. Initially a single watershed (of about 25,000 ha) was to be selected for development in each state and this was to be followed up by the development of an additional watershed in each state. Selection of watersheds was to conform to an agreed development criteria an(d characterized by rainfall (above 750 mm annually), soil types, and the availability of basic agricultural sUpport services. Actual watershed development plans were necessarily to vary accordinlg to climate. land capability, population density, accessibility, and existing level of development. 3.4 The development of watersheds was to include: (a) crop production activities including provision of staff, equipment and training: (b) constructioni of soil conservationi works on crop lands to include about 300 kms of drains, drop structures, graded bunds/terraces and waterways over 140,00() ha, of privately owned farmland including land smoothing (60,000 ha), and land shaping (10,500 ha), and rehabilitation of existing soil conservationi structures on 30.000 ha. It also included construction of soil conservation structures to protect about 55,000 ha of public lands including 10,000 brushwood/stone check dams, 400 wire crate dams and 100 drop structures: (c) oversowing of 45,000 ha of grazing lands with fodder legumes, replanting with fuel and fodder trees of about 14,000 ha of forest reserves and reveniue lands, and distribution of seedlings for tree planting on about 8.000 ha of permnanent fallow and marginial waste land:  3 (d) development of rainwater utilization through limited project investment im farm ponds; (e) provision of increased staff training, research and operational research programs at four State Agricultural Universities (SAU) and three Indian Council for Agricultural Research (ICAR) institutions, and of 50 man-months of technical assistance and 50 man-months of overseas training; and. (f) establishment of project administration at State and watershed levels and strengthening of the Watershed Development Council of the Government of India (GOI) for appraisal of sub-projects and for monitoring and evaluation. 3.5 Project Costs and Financing. Total project costs (including contingencies) over a seven-year implementation period were estimated at US$45.5 million (Rs.443.3 million). These were to be financed by an IDA credit of SDR29.4 million (US$31.0 million), a grant of about US$0.8 million from the British Overseas Development Agency (ODA) and contributions from the Government of India and participating states of US$13.7 million. Project costs included about US$0.7 million in duties and taxes and US$1.5 million in foreign costs. 3.6 Mid-term Review. Given the innovative nature of the project. an agreemenit was reachied at negotiations that a Mid-Term Review of the project would be undertaken by a joint team from GOI, the participating States, and the Bank during the third year of the project (see para 4.5). It was noted then that the exercise could result in substantial modifications to the project and approach originally approved. 4. Project Design and Organization 4.1 Project preparation dated back to 1976 and culminated in a final round of preparation which started in March 1981 as a joint effort of GOI's Ministry of Agriculture (MOA). the participating State Governments and the Bank's Resident Mission. Formulation of project concepts, the choice of technologies, and the selection of participating states required considerable time to arrive at a consensus. In addition, there were also fluctuations in the priotity assigned to a project of this type, both by GOI and State Governments. Two Bank missions which visited India in September 1982 and January 1983 appraised the project and produced a Staff Appraisal Report dated November 17, 1983. 4.2 At the time of preparation. India had little substantive experience in the planning and implementation of systematic programs for the development of rainfed areas. The technological base, as well as the management and economic aspects relevant to such efforts, were not well established. The pilot approach adopted by the project was, therefore, well suited to the objective of establishing the parameters for watershed management. This implied large-scale verification and extension of technical packages which are applicable under different environmental and social conditions in the four selected states, including those already developed prior to the project and those introduced by the project. In addition, the efficiency of the organiztional structure for implementing essential works and the unit costs of such works and technical packages remained important criteria in determining the replicability of the development proposals.  4 4.3 Project design emphasized flexibility in its approach to project interventions and laid down technical and financial criteria for both crop land and non-arable land development. While some variation in the financial criteria among the different watersheds was permitted, an overall financial criterion was established to limit the investment needed for eventual large-scale replication. Project design also recognized the need for establishing cost effectiveness of the different known technical approaches and for any newer ones, by providing resources for research through the State Agricultural Universities and other research institutions, as well as for technical assistance and training. But perhaps the most significant element in project design was the considerable emphasis given to project organization and management, particularly with respect to farmer involvement. The project concept emphasized the need for selecting components which were acceptable to the local populations and also that the great majority of people living within the areas selected for rehabilitation should perceive themselves as direct beneficiaries of project activities and be involved in the planning and execution of the project activities. A three-tier structure for project implementation at the Center, at the State, and at the District/Watershed level was supported by a multi-disciplinary approach in the design of specific interventions and measures to ensure farmer participation, not only to promote project "ownership", but also to formulate the measures for cost recovery. 4.4 While the project design and organization were appropriate, in retrospect it became apparent that in its attempt to cover a wide range of technical variations the project overextended itself. The desire to provide a 'normalized' menu of solutions applicable to both private, communal and state lands in a wide range of (micro) agro-ecological environments (with varying rainfall, slope and soil conditions) was also over ambitious. As a consequence, despite the emphasis on flexibility and on the need for the application of measures to suit specific watershed requirements, it was inevitable that practical aspects of implementation resulted in the application of the same technical packages for each watershed, and this impaired their effectiveness significantly. Added to this, treatments were based purely on physical land capability criteria and, despite conceptual statements on the need for the involvement of people of the area, did not take into account the socio-economic and socio-political conditions prevailing in the communities living within the watersheds. Furthermore, project success criteria as described in the Staff Appraisal Report were based on physical targets rather than the development and adoption of processes. Although the project's evaluation studies were careful to assess the verification and adoption of technical packages introduced by the project, the project's own monitoring system, following the guidelines of the Staff Appraisal Report, concentrated on measuring the coverage of watershed treatments against strictly controlled project physical targets and annual budgets. The inclusion of process monitoring involving the application of efficiency parameters and the qualitative evaluation of watershed management measures in project design would have given the project a better chance of identifying and linking a variety of approaches to the specific technical specifications of the different areas. 4.5 As already noted (para 3.6), the project design included a Mid-Term Review which was extremely useful in changing the focus of the project. The overall concept of integrated watershed development was retained by the Mid-Term Review. However, affordable, simple and replicable technologies with greater reliance on contour cultivation, in-situ moisture conservation, vegetative (khus) contour hedges (porous barriers), became the main strategy of the project. The Mid-Term Review recommended a systems approach which integrated conservation and production in the three sub-systems identified by maximizing the utilization of vegetative measures through the super-imposition of appropriate mixes of productive crops, grasses, shrubs and trees.  5 It must be noted that this change of approach was immediate in that it did not involve any phasing in or out of technical packages. The technical approaches proposed by the Mid-Term Review were as follows: (a) Arable Land - Vegetative contour hedges/strips, contour cultivation, dead furrows at appropriate intervals, cropping systems - high-yielding drought resistant varieties, integrated nutrient and pest management with elements of organic farming, appropriate farm implements and machinery, contingency farming, etc., replacing the more expensive and land-consuming engineering- based earth bunds and grass waterways with drop structures. (b) Non-arable Land (rangelands. wastelands, degraded forest) - vegetative fencing/social fencing, stabilization through vegetative contour barriers, contour furrowing, seeding of grasses and legumes and planting of shrubs and trees. (c) Natural Drainage Courses - Diversion drains, gully control and bank stabilization through bio-engineering and vegetative measures, storage and management of run-off in natural depressions/partially dug out structures with low dams to promote the recharge of groundwater, and decongest blocked drainage lines. Farm ponds were not promoted in post-Mid Term Review stages because they occupied space on crop land, suffered from seepage losses, and did not suit marginal farmers with small and scattered holdings. 4.6 The Mid-Term Review also recommended that, in place of a single second watershed of about 25.000 ha, the States could select two to three watersheds of about 8,000 to 10,000 ha each located in different agro-climatological zones. 5. Project Implementation 5.1 General. The project became effective on June 29, 1984 and was closed on December 31. 1993, two years behind schedule. Project implementation was initially very slow and, by the original closing date, project disbursements totalled about 40 percent of the total credit. The major reason for this slow disbursement was a slow start-up of project actions. The rate of disbursements was also affected by the declining value of the Rupee vis-a-vis the US Dollar (at appraisal US$1=Rs.9.75; current value US$1=Rs 31.00) as well as the movement in the US$/SDR exchange rate which implied a substantial increase in local currency resources, and therefore, a slower drawdown of the IDA credit. As a result, SDR 5.2 million and SDR 3.0 million were cancelled in December 1991 and May 1993, respectively, as part of the funds redeployment exercise. Between 1990 and the extended closing date, GOI in agreement with the Bank, extended the scope of the project to cover additional watersheds in the participating States. This resulted in a substantial increase in disbursement over the last two years of the project reaching SDR 19.04 million as against the revised credit amount of SDR 21.2 million. With GOI's refund of the unused balance of the Special Account on June 15, 1994, the undisbursed balance of SDR 2.16 million was cancelled. A problem that persisted throughout the project implementation period was the delay by State Governments in seeking reimbursements from GOI for amounts already spent. This was pointed out by several Bank supervision missions.  6 5.2 Physical Progress. The project's physical targets were largely achieved. In addition to the four original watersheds, the project covered twenty three new ones. The land area covered is about 350,000 ha located in 27 watersheds (compared to 250,000 ha in 8 watersheds targeted at appraisal). Details of targets and achievements by each state are set out in Part III/Section 4. For the four states together, a ready comparison with the main targets set at appraisal reveals the following: Achievements as % of Revised Achieve- Revised Treatment SAR a/ Targets ments Targets Soil conservation on arable lands (ha) 170,000 185.304 154,745 84 Horticulture (ha) - 15,270 16,792 110 Afforestation (ha) 22.000 51,763 42,700 82 Silvipasture/pastures (ha) 45,000 16,196 12,326 76 Drainage line treatment structures (no.) 10,500 40,651 44,542 110 a! SAR targets refer to the original estimate of approximately eight watersheds. Both the number of watersheds and the nature and variety of treatments were changed as a consequence of the Mid-Term Review. It is useful to make comparisons for achieve- ments against the revised targets. 5.3 Soil and Moisture Conservation Measures on Arable Land. The physical achievemenits of the project's soil and moisture conservation program were impressive and reached 84 percent of the revised targets on 27 watersheds. Although watershed development packages remained the same for all the States. individual treatments (and their terminology) varied between the States and watersheds as evidenced by the achievement of physical targets under nine key indicators which are consistent for all watersheds (see Part lI/Section 4). The treatment of arable lands with land smoothing, graded bunds and grass waterways has significantly helped reduce the hazards of soil erosion and, together with other agronomic and financial packages, aided a revival in rainifed production. Farmers reported that earthen bunds had been 75 to 90 percent effective in checking run-off from their fields and 75 to 80 percent effective in preventing soil loss. It was also revealed that productivity of crops increased by 10 to 15 percent, other factors being equal. Farmners were frequently not satisfied with earth bunds constructed by project-managed bulldozers: compaction levels were low and longitudinal and cross sections were uneven, consequently bunds lost their shape and gaps appeared within one monsoon season. Because farmers were rarely involved in the design, siting and construction of the bunds and waterways, they tended not to maintain the channel and repair the breaches, even at the risk of increased erosion. Farmers much preferred the man- and bullock-made bunds; they participated more and quality was consistently better.  7 5.4 Diversion channels/interceptor drains constructed along the contour between non- arable and arable lands were very successful in protecting the crop land from unwanted run-off of upper catchment areas, which previously caused severe sheet and gully erosion in cultivated fields. and guiding it safely into farm ponds and/or natural drainage lines. Some farmers used the diversion structure as a V-ditch along which they planted lines of horticultural and forest-produce trees. As there was no community involvement in their design, siting and construction. farmers were not willing to maintain and repair the diversion channels - unless there was some personal gain attached. such as fodder trees and farm ponds. 5.5 Waste weirs and waterway check structures made of maintenance-free boulders and live hedges were also well built and effective in arresting silt and reducing the erosive velocity of run-off water collected by the earth bunds while guiding it safely into farm ponds and drainage lines. 5.6 The promotion of in-field vegetative barriers, for the interception of run-off and prevenition of soil erosion, was perhaps the single most important contribution of the Mid-Term Review. Although, in principle, a variety of vegetative barriers were envisaged for trials, project implementation appeared to have been largely involved with only one type of vegetative barrier -- khus, or vetiver -- to the neglect of other possibilities. This is because research establishments at the state agricultural universities started testing alternative vegetative and bio-engineering techniques, including several grass and shrub species. only recently (in 1990) and significant results were not available for dissemination in the field. 5.7 The experience with vetiver as the major element in vegetative conservation, however, has been variable. In the upland areas of Maharashtra and Karnataka, vetiver worked well as a conservation measure. Farmers were ready to accept and maintain vetiver barriers planted in association with their boundary bunds, especially in the lower sections of their fields where run-off water (and silt deposits) collected before overtopping and draining from the field. In Andhra Pradesh and Madhya Pradesh, however, the vetiver-based vegetative barrier was not successful. Farmers found the survival rate of the khus to be extremely poor with no visible impact on increased productivity and many of them, therefore, ploughed up the areas planted. 5.8 The disparate achievement of vetiver as a technology for soil and water conservation brings out a number of technical points. Since most of the vetiver was planted by the project, there was little or no commitment on the part of the farmers to manage it. This aspect had a significalnt impact on the eventual adoption rate. Farmers were unwilling to participate and wait for a three- year establishment period before any visible impact on productivity could be seen. There were also questions on the establishment of the vetiver hedges. Vetiver's hardiness as a crop that required little or no attention was oversold. In point of fact, losses of 25 to 50 percent within individual key lines under "good" conditions and failure in the majority of total barrier plantings were not uncommon. Furthermore, there were cases of die-back after the establishment period of three years. Success with vetiver seemed to be associated with rainfall in areas in excess of 1000 mm, while its impact in semi-arid and arid areas was poor. What was most significant, however, was that vetiver was over-prescribed as a soil and moisture conservation measure on the flat vertisols where the majority of the slopes were less than 2 percent. Vetiver barriers are not strictly necessaiy in many of the places where they have been planted -- contour cultivation and dead furrows would be more appropriate in dry areas, drainage channels where water logging is a risk. As a consequence, farmers ignored both earth bunds and vetiver barriers. Furthermore, for areas with  8 small farm holdings, the introduction of the vetiver barriers on farmers fields resulted in the dissection and loss of productive land which was already in short supply. Watershed management plans which should have focused on Land Classes IV to VI (according to Land Capability) instead invested a considerable amount of resources in soil conservation measures on Class I and II lands when all that was required was more appropriate tillage systems and improved cropping practices. 5.9 Afforestation. The project's afforestation program on private lands was successful, and awareness of the farm forestry program was high (80-90 percent) among farmers. More than 80 percent have continued to protect/maintain their woodlots through some form of fencing, watering and replacement of dead seedlings. Nearly all farmers requested that the project provide watering facilities (through farm ponds or wells) and fencing before the planting of tree seedlings. Some farmers also felt that the seedlings should have been planted at increased depths to avoid further moisture stress. Farmers generally ranked the benefits of farm forestry as follows: fuel, cash crop/monetary return, served as fencing for cropped fields, timber for the home and agiicultural implements, and higher milk yields due to improved supplies of fodder. Continuous contour 'V' ditches were applied on farm forests and block plantations and have now been widely accepted as the basis for land preparation for trees. Survival rates with 'V' ditches were mainly above 80 percent (if protection was provided) and improved growth rates were also evident. Afforestation on public lands was not relatively successful. With the exceptioll of Andhra Pradesh, which recruited forest guards to protect the project's forestry plantations, the survival rate of Block Plantations was very low (0 to 40 percent, and usually below 10 percent). A lack of genuine participation from the local communities in the planning and implementation of forestry works meant that there was little interest in managing the forest resources and, without the provision of forest guards, the plantations were left unprotected and soon became overgrazed and irreparably damaged. Towards the end of the project, particularly in the additional watersheds, the forest protection committees for managing the rehabilitated community forests were formed. This was generally after the project had completed the plantation works; the users were not directly involved in the selection of species, density and mix of planting, construction of water-harvesting ditches, etc. 5. 10 Perceived as part of the treatment for the management of non-arable land, dryland horticulture was an innovation introduced during the course of the project and has been relatively successful, particularly in the states of Karnataka and Andhra Pradesh. The production and provision of quality grafts and seedlings was the main project activity. The major species demanded by the farmers included mango, guava, pomegranate, jack fruit, custard apple, tamarind etc. While the survival rates of the trees at first year of planting varied between states, they were reasonably high. Farmers were conscious of the benefits, and demand for planting materials was high. 5.11 Drainage Line Structures. This covered a number of measures such as check dams, nala bunds, and other run-off management structures. In general, most of these structures were effectively implemented. Depending on the specific type, the measures had the effect of reducing the velocities of flood water flow, restricting the deepening and widening of gullies and inducing large quantities of silt to collect upstream which have helped recharge local groundwater. They also frequently became vegetated and added further strength to the structure and gully stabilization objectives. 5.12 Despite an abundance of potential sites and a pressing demand from farmers, farm ponds (stop dams and small tanks) were not constructed after the Mid-Term Review. The project  9 replaced farm ponds with more affordable run-off management structures which were constructed on public land and benefitted more farmers. However. Kamataka disagreed with this decision and continued to construct farm ponds after the Mid-Term Review. Wherever farm ponds were constructed, they generally performed well in providing much needed water for livestock and supplementary irrigation for tree seedlings and horticultural crops. Farmers did not see the siltation of ponds as a problem; they frequently cleaned out the silt and used it as additional fertilizer on their fields. 5.13 Pasture/silvipasture development constituted an additional measure to improve land management and to increase average productivity of the farmers' non-arable land. Pasture development was generally confined to sowing of stylo seeds on bunds and ridges of contour ditches. Typically all activities under the program -- ploughing, clearing of the area and sowing -- were undertaken by the project with no involvement of the farmers. Farmers were of the opinion that pasture development was not possible because it was impossible to stop open grazing without some form of fencing. Although no data on the productivity of the pastures was available, visits to the project area by evaluation teams have shown that the area degenerated to its original state under grazing pressure. 5.14 Conservation on Public/Common Property Lands. Soil and moisture conservation interventions on private lands as stated above were well implemented. Project implementation with its technological bias however overlooked land tenure or ownership or other general sociological issues. As a consequence, the same technical conservation measures did not succeed on the non- arable or common property lands. This was very apparent in the afforestation and drainage line structures implemented on common property areas. The project failed to take advantage of lessons leamt from social forestry projects (ILO, ODA, World Bank) which had previously identified the need for the formation user groups for the planning, implementation, and management of these resources. Towards the end of the project period, particularly in some of the additional watersheds, there was a more successful effort to involve the beneficiaries in the management of common resources. The design and management of conservation measures for common property resources need further development. 5.15 Research. Adaptive research covering in situ vegetative soil and moisture conservation measures in the project area was effectively executed by multi-disciplinary teams of the research agencies in the three states of Andhra Pradesh, Karnataka and Maharashtra. Quantitative parameters are now available for a variety of conservation measures. Research was also undertaken by the SAU at Bangalore on a more flexible bio-engineering approach to soil and water conservation on arable lands. This compared the efficiency of vetiver contour key lines and graded earth bunds on alfisols (sandy loams and sandy clay loams) with slopes of 2.5 percent. The research concluded that a combination of small graded earth bunds (cross section - 0.35 m), vetiver key lines and land smoothing produced the best results for minimizing nutrient, soil and water losses. To maximize the conservation of water, the SAU also included a series of waste weirs, grass waterways and small farm ponds in this package. The SAU further concluded that the maintenance costs to the farmer for large earth bunds, vetiver contour key lines, and a combination of small bunds and key lines, were the same, and consequently, the most appropriate approach to the problem would be to leave the choice of technology to the farmer who could seek the cooperation of his neighbors when necessary. All construction works were carried out by the SAU with maximum participation of the farmers. The involvement of farmers at all stages, including the choice of technology to be introduced contributed to a perception of "ownership" and the prospect of maintenance by the farmers later.  10 5.16 Project Organization and Management. The three-tier structure for operation and maintenance envisaged at appraisal was established but its effectiveness was variable between states. Overall responsibility for coordination and management was vested with the National Watershed Development Policy Committee (NWDPC) at the level of GOI which met once a year to appraise watershed masterplans submitted by participating states. Operational responsibility vested in the Watershed Development Council (WDC) was appropriately managed. The WDC, which was established as a special project unit within the Ministry of Agriculture for the coordination of World Bank assisted projects, was upgraded to an operational division responsible for rainfed agriculture within the Ministry, including coordination of all watershed development projects, irrespective of sources of funding. WDC was also responsible for monitoring and supervision of project implementation, and this was effectively carried out. Independent evaluation studies comprising of a bench-mark survey followed by a final evaluation study were commissioned by each of the states. 5.17 A State Watershed Development Policy Committee (SWDPC) was eventually established in each state and was responsible for the approval of masterplans for watersheds and project plans for sub-watersheds before their submission to the WDC for review and funding. In addition to the differing priorities provided by the states, the multi-disciplinary concept of watershed development requiring the expertise of agriculture, land use planning, soil conservation. forestry and socio-economics was difficult to achieve in the vertical administrative structure of the Indian Government. The States found it difficult to establish the SWDPCs, most takin, two to three years to achieve, and in the case of Madhya Pradesh, as much as five years. As a consequence, SWDPCs were never fully operational. District Watershed Development Committees (DWDCs) were intended to act as sub-committees to the existing Distlict Coordination Committee. In most states the DWDCs did not function effectively, the district authorities preferring to coordinate multi-disciplinary actions and sanction watershed development interventions through existing administrative organizations. Watershed Development Teams (WDTs) were the actual operating arm of watershed development responsible for detailed sub- project planning and implementation. As a rule, all WDTs were adequately staffed and functioned well to achieve project targets despite an agricultural or soil conservation bias within their staff. Like the SWDPCs, the WDTs also took two to three years to establish. Project components during the intervening period were implemented by the respective technical line departments. What is important, however, is the fact that until the start-up of the development of the 23 additional watersheds, all project funding were channelled through the individual line departments. Although watershed masterplans were designed in an integrated manner (linked through a 'fixed' land capability classification), the funding mechanism worked against a multi-disciplinary approach, in that. each team member prepared work plans and budgets for his own technical requiremenits. It was only during the last three years of the project when all sub-watershed funds were channelled through the WDTs that sub-project implementation began to be coordinated in a more integrated mnanner. 5.1 8 Even so, non-agricultural sections like forestry were not constituted as conceived, nor fully staffed, and therefore rarely integrated into WDT operations. There was also a tendency to overlook land use planning, and socio-economic and livestock inputs into the WDTs. WDTs did not develop a calendar of operations and a roster of inspections for their treatment programs to coordinate management of their working practices and ensure timely implementation and monitoring of their planned interventions. Likewise, there were few periodical meetings and seminars to: (a) discuss progress under the various components and their allied activities; (b)  II promote the success of tested technologies, and (c) analyze the causes of any failed interventions -- the continued plantation of unprotected government forest land was an example of the latter point. Furthermore. WDT staff were not fully trained in the skills required for integrated development and community participation which seriously affected their understanding of common property resource management and the need for communicating with the local people. 5. 19 Beneficiary Participation. The involvement of farmers at district and sub-watershed levels was incorporated in project design. Farmers' representation on DWDCs, discussions between the WDTs and residents from each village for deciding and locating development works, formation of village action groups, and preparation of action plans on agreed priorities were envisaged at appraisal. However, at the time of the Mid-Term Review in 1988, beneficiary involvement was limited to a subdued presence of two farmers' representatives in DWDC meetings only. There was very little evidence of farmer's involvement and community participation regarding the planning and implementation of project actions. A number of project interventions were considered not sustainable because village communities were not prepared to operate and maintain the structures and plantations handed over to them on completion of the micro-watershed development works. In a number of cases, the treatments soon fell into disrepair and were lost to the communities. 5.20 The formation of forest protection committees in all four states during the last two years of the project was based on a user group concept but cannot be considered as people's participation per se. The block plantations were usually planned and implemented by the forest departments who, towards the end of the project and for budgetary reasons, were required to withdraw their forest guards and turn over the protection and management of the forests to the local communities. Project forestry officers, with a limited background in social organization, helped users to elect a management committee, briefed them on the care of the tree species they had inherited, and handed over all responsibility to them. Training of the beneficiaries in the formation and management of user groups usually entailed two or three joint meetings between forest officers and community leaders. Whether this 'turnover' approach will work or not is too early to say but experience from Madhya Pradesh has shown that rural communities preferred to be involved from the vety beginning in the planning and implementation of common property resources to meet their local needs. Often this meant a preference for particular species for specific uses. A combination of forestry with improved fodder grasses ('silvipasture') which required a less dense planting of trees, and planting of trees and grasses on earth buinds ('agro-forestry'). were treatments frequently requested by fanners which. where implemented, proved quite successful. Under the project's more top-down approach, species, densities and mixes were pre-selected by the forest departments. Where these did not satisfy the needs of the intended users, the local communities were not prepared to provide protection and management of their common property assets when handed over to them. 6. Project Results 6.1 Project results have been more impressive on the physical front than on the institutional side. For a pilot project of this nature, project results are to be seen from the impact on the adoption rates of the technologies promoted. Information on these are available from the  12 bench-mark and final evaluation studies' undertaken by four independent agencies, as part of coiisultancy services provided in the four states. Adoption rates were assessed by the number of farmers adopting new cultural practices promoted by the project (e.g. contour cultivation and horticulture), maintenance of conservation works constructed (e.g. gap-filling of earth bunds and vegetative barriers) and protection of plantations. Taking 50 percent farmer adoption as a success criteria, evaluation studies indicate the following watershed management measures as having been successfully adopted by the farmers: contour cultivation, dead furrows and intercropping; diversioni channels, earth bunds, grass waterways, and land smoothing; horticulture; farm forestry; silvipasture and fodder farms; check dams and run-off management structures and farmponds. Although proven successful at project-supported research stations and in some localized micro- watersheds, a number of other watershed treatments do not appear to have been universally accepted and require further trials for their technical and social applications. These include: vegetative barriers and key lines; community forestry/government plantations; silvipasture on community/govemment land; nala bank stabilization/plantation or, more generally, the management of c&r)moon property resources. 6.2 Adoption rates for soil and water conservation measures have been generally higher in respect of the more gently sloping and marginal alfisols of Andhra Pradesh and Karnataka rather than on the flat and more fertile vertisols of Madhya Pradesh and Maharashtra. Watersheds in Madhya Pradesh and Maharashtra had more than 60 percent of their land belonging to Land Capability Class I to III with a greater percentage of large landowners. The adoption rates for earth bunding and vegetative barriers were low as farmers saw little benefits from these measures. On the contrary, watersheds in Andhra Pradesh and Karnataka had less than 30 percent of the land within the same classification. The adoption of watershed treatments was higher in these watersheds with a greater proportion of land in the Class IV-VI category (marginal and non- arable), with a greater slope and a higher density of small holdings. 6.3 Applied research carried out by the SAUs was productive and well organized. The impact of the results, however, was limited as this research concentrated primarily on only one vegetative aspect of conservation -- vetiver. Research in alternative approaches, which started toward the end of the project, was unable to have any major effect on project implementation. There was little attempt also at linking research to an organizational framework relevant for the treatment of common property resources. 6.4 On the institutional side, the project's most noteworthy achievement has been the establishment of the Watershed Development Council (WDC) as a permanent division of the Ministiy of Agriculture, responsible for the coordination of all watershed development projects, thereby increasing the prospects of learning from the experiences of watershed development proojects in other areas of the country. This has had an impact in that WDC has adopted a number of additional measures in the second generation of watershed development projects currently under implementation in India. Actual implementation, however, continues to be dominated by a centralized top-down strategy with a heavy reliance on district level watershed development teams (WDTs) for actual implementation. As a result, the emphasis given in watershed management has often reflected the technical background of the team leader -- usually from an agricultural extension or soil conservation background. Thus, for example, Maharashtra used extension workers to consult farmers for improving crop production while Madhya Pradesh showed an engineering bias l The final evaluation studies were not completed in two of the four states, viz., Karnataka and Maharashtra.  13 to achieve targets for soil conservation works without consulting the farmers. The multi- disciplinary approach for watershed management was adopted largely as an ad hoc measure and, excepting for Karnataka which has an independent division of watershed management within the Department of Agriculture, none of the other states have adopted the approach as standard. 6.5 Project achievements in terms of ensuring beneficiary involvement in planning and implementation have not been commensurate with the level of physical achievements because of an inadequate attention given to this aspect. A limited effort at farmers' involvement at the planning stage was attempted through the participation of two farmer representatives from each sub- watershed in DWDC meetings, but, by and large, the project was implemented by project staff and/or hired labor. Added to this, farmer representatives were usually large landowners nominated by the District Collectors and not elected by or typical of the sub-watershed communities they represented. Farmer involvement during project implementation and subsequent operation and maintenance was also minimal. Watershed management measures in the last three years of the project had attempted a formn of beneficiary involvement through the formation of micro-watershed user groups, e.g. forest protection committees in all states, sanghas (farmers' associations) in Karnataka and agricultural cooperatives in Maharashtra. During this short period, the farmers became responsible for most sub-project decision-making, and extension of watershed management treatments with the help of technical assistance from the WDTs. Similarly, the associations have been organizing paid labor for treatment works from within their own communities, generating much needed local employment. However, this marginal involvement was insufficient to provide the grounds for any cost recovery as envisaged at appraisal. 6.6 Project conservation measures were combined with a greater emphasis on extension which focused on the expansion of gross cropped areas, through increase of cropping intensities, largely as a consequence of the conservation measures. More importantly, it included changes in the cropping pattern (introduction of safflower, inter-cropping -- pigeon peas with sorghum, soya beans, groundnuts and finger millet; cotton with green gram), and the increased use of better technical approaches (contour ploughing) and inputs (better seeds, improved varieties, fertilizers and pesticides). The impact of these changes on improving farm incomes has been significant in the case of Karnataka vis-a-vis other states. Also, in Andhra Pradesh and Madhya Pradesh. fortuitous changes in relative prices in favor of oilseeds and beans provided a major incentive for farmers in general to respond. As a consequence, production and farm incomes for non-project farmers have also increased at almost the same rate as that of project farners. Net finanicial farm incomes for farmers, at current prices, before project, without project and with project situatiolns are as follows:  14 Current Current Farm Before Without With Farm Models Size Project Project a! Project b/ (ha) (Rs.) (Rs.) (Rs.) Madhya Pradesh (ParuLa Nala) 2 7.790 16.500 18.720 Andhra Pradesh (Maheswaram) 2 4,370 5,429 5,960 Karnataka (Kabbalanala) 1.50 2,590 4,600 10,300 Maharashtra (Manoli) 6,590 - 14,300 a/ For non-project farmers. b/ For project farmers 6.7 Benefits accruing to farmers as a result of the project, as stated earlier, were largely as a consequence of a change in cropping pattems (Andhra Pradesh and Madhya Pradesh) and only partly due to an increase in cropping intensity (Karnataka). The financial results in the table above (para 6.6) indicate the marginal nature of incremental benefits with respect to Andhra Pradesh and Madhya Pradesh, but significant with respect to Karnataka. Given this variation, the objective of establishing financial parameters to evaluate cost effectiveness was not achieved. There is. therefore, a great need to clearly establish the costs and benefits from the range of watershed management measures as such an assessment was not attempted during the evaluation studies supported by the project. 6.8 An economic rate of return (ERR) for the project as a whole was not estimated at appraisal, but ERRs were estimated for three individual watersheds initially to be covered in the states of Andhra Pradesh, Karnataka, and Maharashtra. Re-estimation of the economic impact of the project for the sake of meaningful comparison is seriously constrained because: (a) the project introduced a wide range of interventions and no data are available on the extent to which these different interventions were actually implemented in each single watershed so as to rationally determine the project-supported investment costs: and (b) it is also difficult to segregate and quantify the benefits due to the project supported interventions for the reasons expressed in para 6.6 above. Furthermore, the project. either through its monitoring mechanism or evaluation studies, had not compiled any data on the costs and benefits of interventions in any individual watersheds. For these reasons, it was not possible for the Project Completion mission to re- estimate ERRs for individual watersheds, as was done at appraisal. 7. Project Sustainability 7. 1 The lack of fainer involvement in the implementation of the project raises a number of questions regarding project sustainability. Public expenditure covering virtually the entire costs of watershed management measures for private gains (from private crop land) while attractive to the farmer is unlikely to be sustainable for the Government. The financial burden of the project- developed watershed management measures. if replicated for countrywide implementation, would be far too large for the government to absorb. However, now that the beneficial effect of  15 watershed management measures on private crop land has been amply demonstrated, farmers are quite willing to group together and participate in some measure of cost sharing. Consequently, the second generationi of watershed management projects under implementation in India lay significanit emphasis on farmer participation both in planning and implementation and maintenance of pro'ject components. 7.2 In the above conitext, project measures on private crop land can be viewed as sustainable provided the necessaiy adjustments to operational modalities are made. The packet of measures for the management of common property resources. however, cannot be consider-ed to be sustainable. There is a need for the development of technical and. more importantly, organizational modules2 before watershed managem-ent measures for common property reSOurCeS canl be considered sustainable. 8. Bank Performance 8. I As one of the first in a series of resource manlagemenit projects supported in India. the Banik has made a significanit contributioni to the development of technologies for wiatershed management. Lessons from the experience of this project have already been applied to some of the other (second generation) projects currently under implementation in India. 8.2 The main strength of the Bank has been its abliity to adapt and change its technical approach to the problem. Despite an initial emphasis on watershed management along traditionial en gineering lines implicit in the Staff Appraisal Report, the Bank was able to change its focus to a vegetative-based approach that is more cost effective and allows easier replication. In addition, the Bank's strength was reflected in regular and effective supervision of the project. Its decisionl to cancel a part of IDA credit as a consequence of savings in local currency alising from SDR:LUS$ and US$:Rupee exchange rate movements was most practical. The Bank was also tlexible in allowing expansion of areas to be covered by the project to permit the inclusion of additional watersheds and extension of the closling date of the pro ject to capitalize on a momentum that had takeen lon,ger to generate than oiiginally anticipated. A weakniess in the Bank's role however was the adoptioni of an almost doctrinal approach to one particular line of vegetative-based watershed manag,ement technology. Despite the significant input of supervision missions which repeatedly identified difficulties with a blanket vetiver-based approach. there was a persistence in applying the same approach across widely different ecological environments. This was however modified towards the end of the project period when a greater focus was placed on other vegetative measures. 9. Borrower Performance 9.1 Despite some variances between individual States, overall periformlance of the Borrower was positive. There were initial delays in the establishment of the institutional flairaewor-k for project implementation, but once this was established, project implementation proceeded at a reasoniable pace. Despite the difficulties of organizing multi-disciplinary teamis fromn staff belonging to distinct line agenicies, the Govemiment was reasonably quick in the fomnation of the teams. However, staff continuity remained a problem and project management was affected by 2 For example, appriopriate mechanisms for community participation and bellefit sharing.  16 a high rate of turnover in top level project management. Although an implicit objective was the institutionalization of the multi-disciplinary structure for watershed management, participating states did not continue with that structure. Only one state. which in any case, had adopted the multi-disciplinary approach before the project, continued its support after the project had ended. 21 Although Borrower's performance on beneficiary participation was inadequate under this project, by learning from its experience it has responded well in this respect under the second generation of projects for watershed management. The formation of farmer user groups is now being actively undertaken to ensure beneficiary involvement in both planning and implementation stages. 10. Lessons Learned 10. 1 The main lessons that can be leamed from the project are as follows: (a) Project design should ensure flexibility to tune the prescriptions for watershed management to the specific characteristics of the watersheds. Flexibility should include not only the adoption of measures to cover a wide range of micro-agro-ecological environment (with varying rainfall, slope, and soil conditions) but also the prospect of a combination of, where appropriate and possible, both vegetative and limited engineering solutions. For instance, the focus on the introduction of a single vegetative conservation technology (vetiver), which was successful only in relatively higher rainfall areas, should have been adjusted to include other varieties under this project. Project design should avoid the provision of a "normalized" menu of solutions applicable to both private, communal and state lands. More specifically, there is a need to establish the institutional mechanism and treatmnents for the management of common property resources in a manner different from that of individual resources. Project success resulted in large part because of the administrative and technical flexibility introduced at the Mid-Term Review and the strong practical orientation of the Bank's supervision effort. (b) Project impact has not been commensurate with the level of physical achievement because of inadequate beneficiary involvement. If the project is to be sustainable, and replication at a national level is to remain an objective, farmer involvement has to be an explicit part of implementation strategy. (c) Project supervision needs to focus on the overall objective of the project in addition to that of the simple quantitative elements. Where new technologies are being tried there should be an effort to attempt alternative approaches to develop a menu of technologies. 11. Adherence to Covenants 11.1 All major project covenants and requirements were complied with. However, there was a need to strengthen staffing of the agricultural extension services in the Master Watershed Plan for all states.  17 12. Consulting Services 12.1 Consultancy services were utilized for undertaking a bench-mark socio-economic survey followed by a project impact study at the end of the project. These studies were adequately undertaken by four independent agencies in the four states. 13. Project Relationship 13.1 Relationships between the Borrower and the Bank were cordial throughout the project implementation period mainly through frequent communication between the Implementing agencies and the staff of the Resident Mission in India. The decision to close the project appeared to have been somewhat abrupt since the Borrower was under the impression that the project would be extended one more year. 14. Project Documentation and Data 14.1 The Staff Appraisal Report provided a basic framework for the Bank and the Borrower during project implementation. The Mid-Tenn Review Report also provided a revised and firm operational basis for project implementation. Project supervision reports contain considerable data on the project largely focused on the physical and financial aspects of project implementationl. A good deal of relevant infonmation and data are also available from the reports of the base-line and end-of-project studies of the four external agencies involved in project impact evaluation. However, the evaluation studies could have managed to establish the costs and benefits from the range of watershed management measures supported by the project so as to facilitate the evaluation of the economic impact of the project supported investments. 14.2 Under the adaptive research component, a number of useful technical booklets and manuals were also published.  18 PART II. PROJECT REVIEW FROM BORROWER'S PERSPECTIVE (NOT RECEIVED)  19 PART II1. STATISTICAL INFORMATION 1. Related Bank Lcans and Credits Law/Ciit Tle Purpa Yewr of Stw Cm_ts Kandi Watenhed and Arta To tackle the degradation problem of 19S0 COsed 1988 PCR cites one benefit as the Development Project the Himalayan sub-mountainous zone in generation of a great number (Punjab) Punjab. of lessons relevant for similar (LA1897-IN) P Himalayan Watershed DEwlop and apply technologies to 1983 C. oe t PCR cites substantial succs in Management Project increa and stabilize rainfed crop, "VD Sept. 92 incrasing agricultural (CrG2295-IN) forage, fucwvood and timber production, productivity and farm incomes speciflly through impm-ed soil nd but a medio;re performunce in moisture and crop production methodL fosy and soil conservation measures. Rainfed Areas Watershed Pilot project to increase crop and forage 1983 Clo sXee Quality of implementation Development Project yieids and production of fuchvood and 31 Dcc. 199' vanes considerably between (Cr.1424-IN) timber. StateL Integrated Watershed Watershed development through 1990 On-going - Development (Plains) stabilization of selected watenshed. Project (Cr.2131-IN, Ln31974N) Integrated Watershed Remedial measur to slow and reversc 1991 On-Oing Supervision cites need to Development (Hills) Project degradation of the natural environment improve people's participation (Cr.2100-IN, LnU3175N) of the Himalayan Foothills and improve in projet implementation. agricultural production 2. Project Timetable Activity Planned Actual Identification) Preparation ) 1976 1976 to 1981 Pre-Appraisal Mission Aug-Sept. 1982 Aug-Sept. 1982 Appraisal Mission Jan-Feb. 1983 Jan-Feb. 1983 Loan Negotiations October 1983 October 1983 Board Approval December 8,1983 Loan Signing February 8, 1984 Loan Effectiveness 29 June 1984 Loan Closing1/ December 31, 1991 December 31, 1993 / Two extensions.  20 3. Credit Disbursement Credit Disbursements: Cumulative, Estimated and Actual (USS million) FY Appraisal Estimate Revised Estimatel/ Actual Total Actual as % of Appraisal Estimate (up to FY91) and Revised Estimate (after FY91) 1984 0.25 0 0 1985 2.10 no 0 1986 5.60 0.50 9 1987 10.70 1.50 14 1988 15.60 3.20 21 1989 21.30 5.00 23 1990 26.20 7.50 29 1991 31.00 11.80 38 1992 - 22.35 15.70 70 1993 22.35 20.8 94 1994 _ 22.35 25.76 115 i/Credit amount was reduced by SDR 8.2 million (equivalent to USS8.65 million).  4. Project leplmeentation A. State-vise ard Treett t-visa Targets wid Achdeveawnts S. Key Indicators Ihift SA Andhra Pradesh Karnataka Adhyea Pradesh Pdiewashtre Totals ho. Targets Tgt Aich Tgt Adch X tt Ach X Tgt Ach X Tot Ach X 1 Soil Conservation on Arable Lands: 1 1) earth bunds ha 3000 5316 177 11800 5900 50 13400 10602 79 7146 7556 106 35346 29374 83 & waterways 11) vegetative ha 42600 28339 67 51490 40788 79 30340 28468 94 25528 27776 109 149958 125371 84 barriers i11) total ha 170000 45600 33655 74 63290 46688 74 43740 39070 89 32674 35332 108 185304 154745 84 2 Land smoothingl ha 70500 0 0 0 11000 2851 26 239 239 100 450 28 6 11689 3118 27 shaping 3 Diversion km 300 113 55 49 447 411 92 223 174 78 132 113 86 915 753 82 Channels 4 Horticulture ha 8085 8196 101 6363 5977 94 94 1204 1281 728 1415 194 15270 16792 110 5 Afforestatf0n 1) private ha 8000 31400 23046 73, - 2015 2092 104 247 215 87 - - - 1i)government/ ha 14000 6892 6220 90 2551 1418 56 3626 2611 72 , rs community M) totat ha 22000 38292 29266 76 5032 7098 141 4566 3510 77 3873 2826 73 51763 42700 82 6 Silvipasture/ ha 45000 7180 4400 61 573 1300 227 2004 1677 84 6439 4949 77 16196 12326 76 Pastures 7 Wasteland ha 1550 1207 78 - 600 42 7 745 631 85 2895 1880 65 Development 8 Drafnage Line (ha) (55000) Treatment Structures no 10500 1759 4979 283 21110 24609 117 4113 5786 141 13669 9168 67 40651 44542 110 9 Form Ponds/Stop no Limited 70 207 296 145 543 374 2 2 100 8 39 488 225 791 352 Dam * Footnotes: I - Nission EstiMate; 2 - Stream Bank Protection; 3 - includes Sericulture; 4 - Agave Plantation 5 Targets are Based on Watershed Nasterplans (includes final (Post-NTR) revision of four original watersheds and final plans for 23 additional watersheds)  H. Details of Targets end Achieventt for Watershed Treat na Andhra Pradesh S.No. Key Irdicators Lknit Naheshuarm Watherahed Additional Watershedh (7) Totats of Uatersheds (8) (25,331 ha) (53,330 ha) (78.661 ha) Tgt Ach X Tgt Ach X Tgt Ach X I Solt Conservation on Arabia Lands: I) earth bunds & waterways ha 3000 5316 177 0 0 0 3000 5316 177 If) vegetative barriers ha 6000 7396 123 36600 20943 57 42600 28339 67 111) total ha 9000 12712 141 36600 20943 57 45600 33655 74 2 Land smoothing/shaping ha - 3 Diversion Channets km - - - 113 55 49 113 55 49 4 Horticulture ha 500 4085 817 7585 4111 54 8085 8196 101 5 Afforestation i) private ha 0 0 0 31400 23046 73 31400 23046 73 11) goverrmnent/coummity ha 2673 3819 143 4219 2401 5' 6892 6220 90 III) total ha 2673 3819 143 35619 25447 71 38292 29266 76 6 Sltvipasture/Pastures ha 2880 1858 65 4300 2542 59 7180 4400 61 7 Wasteland Development ha - - 1550 1207 78 1550 1207 78 8 Drainage Line Treatment Structures no 1020 1805 177 739 3174 430 1759 4979 283 9 Farm Ponds/Stop Dams no 70 207 296 70 207 296 * Footnotes: 1 - Mission Estimate; 2 - Stream Bank Protection; 3 - Includes Sericulture; 4 - Agave Plantation Karrutaka SNo. Key Indicators Unlt KbbsIinala Watershed Additionat Watersheds (5) Totaels of Wstersheds (6) (29,803 ha) (100r768 ha) (130,571 ha) Tgt Ach X Tgt Ach X Tgt Ach I 1 Solt Conservation on Arabie Lands: 1) earth bunds & waterways ha 11800 5900 50 0 0 0 11800 5900 50 11) vegetative barriers he 1750 538 334 49740 34950 70 51490 40788 79 ifi) totat ha 13550 11738 87 49740 34950 70 63290 46688 74 2 Land smoothing/shaping ha 11000 2851 26 0 0 0 11000 2851 26 3 Diversion Channels km 112 262 234 335 149 44 447 411 92 4 Morticulture ha 1288 2085 162 5075 3892 T7 6363 5977 94 5 Afforestation I) private ha - - - - - -  11) government/community he - - - 111) total he 1600 4031 252 3432 3085 90 5032 7098 141 6 Sflvipasture/Pestures ha 255 694 272 318 606 191 573 1300 227 7 Wasttlard Development ha - e - - - - - - 8 Drainage Line Treatment Structures no 3790 4416 117 17320 20193 117 21100 24609 117 9 Fare Ponds/Stop Dasn no 55 414 753 90 129 143 145 543 374 Footnotes: I - Mission Estimate. NRd"qe Pradesh S.1o. Key Irdicators unit Peru Uala Watershed Additional Watersheds (4) Totals oaf Wtersheds (5) (30.524 be) (42.100 he) (72.624 ha) Tgt Adc I Tgt Adc X Tgt Adc S I Solt Conservation on Arable Lards: I) earth bunds & waterways ha 13400 10602 79 0 0 0 13400 10602 79 11) vegetative barriers ha 4340 11437 264 26000 17031 66 30340 28468 94 111) total ha 17740 22039 124 26000 17031 66 43740 39070 89 2 Land smoothing/shaping ha 239 239 100 0 0 0 239 239 100 3 Diversion Charwels km 98 153 156 125 21 17 223 174 78 4 HorticuLture ha 300 683 228 670 521 78 970 1204 124 rJ 5 Afforestation I) private ha 315 494 157 1700 1598 94 2015 2092 104 11) goverrnent/comn.fity ha 1251 1002 80 1300 416 32 2551 1418 56 111) total ha 1566 1496 96 3000 2014 67 4566 3510 77 6 Sltvipasture/Pastures ha 954 1144 120 1050 533 51 2004 1677 84 7 Wasteland Development ha 0 0 0 600 42 7 600 42 7 8 Drainage Line Treatment Structures no 1625 3089 190 2488 2697 108 4113 5786 141 9 Farm Pords/Stop Dams no 2 2 100 0 0 0 2 2 100 Maharashtra S.1o. Key Indicators Lint Nanoli Watershed Additional Watersheds (7) Totals of Watersheds (a) (25,403 ha) (40.821 ha) (66.224 ha) Tgt Ach I Tgt Ach X Tgt Ach S I Soil Conservation on Arabta Lands: I) earth bunds & waterways ha 6915 7396 107 231 160 69 7146 7556 106 11) vegetative barriers ha 13910 16599 119 11618 11177 96 25528 27776 109 III) total ha 20825 23995 115 11849 11337 96 32674 35332 108 2 Land smoothing/shaping ha 450 28 6 0 0 0 450 28 6  3 Dlversion Chanels km 52 62 119 80 51 S 132 113 86 4 Horticulture ha 226 1078 477 502 337 67 728 1415 194 5 Afforestat ion I) private ha 247 215 87 - - 247 215 87 11) goverrinent/con0.nity ha 1436 770 54 2190 1841 84 3626 2611 72 Ill) totat ha 16U3 985 59 2190 1841 100 3873 2826 73 6 Sltvipasture/Pastures ha 1547 1222 79 4892 3727 76 6439 4949 77 7 Wasteland Development ha 385 421 109 360 210 58 745 631 85 8 Drainage Line Treatmnt Structures no 132 242 183 13537 8926 66 13669 9168 67 9 Farm Ponds/Stop Dow no 8 39 488 0 0 0 a 39 488  25 5. Project Cost and Finance A. Project Cost (Rs million) ~mgmy 134 19844 1984 19887 19874 19664 1964 1990. 19-2 1992-3 1993,4 Toda ANDHRAFDB P_AMFLod r E9 F- j d ad - Cld Works 0"07 .36 2.1131 31006 3.00S 7.323 9:m2 7.35 1986 2L.796 30.256 114.061 Vahclas