Background

While major public projects, notably those dependent on diversion from major river systems, make up the largest component of irrigation in some regions, elsewhere small schemes provide the greater part of canal irrigation. The area served by individual small schemes range from less than 50 ha to as much 5000 ha, but are generally a few hundreds of hectares. Where a scheme is based on direct diversion from a stream, without storage, construction and operation, in the past, has generally been entirely by the participating farmers. On the other hand where, a small reservoir is involved, as in the "tank" common in India, Sri Lanka and elsewhere, construction is commonly by state agency, with the state also being involved in some cases in operation, although generally to a limited degree. Included in the direct diversion systems are the notable "hill" schemes of Nepal. Although these were originally entirely farmer-constructed, government agencies are now providing assistance in rehabilitation, and are involved in construction of new projects. The manner in which this assistance should be provided and its impact on the previously autonomous character of the hill schemes are much debated.

In view of the attention currently being paid to the ecological costs of major storage projects, international development agencies have increasingly turned to small schemes, particularly those which enlist cultivators in their construction and management. It is noted, however, that although such schemes are conceptually very appropriate from the development point of view, their implementation poses a number of problems, some of which are referred to in the following discussion.

Farmer-constructed diversion systems

These commonly consist of a low weir, often requiring renewal after each high-flow season, a conveyance canal leading to the area to be irrigated, and distribution channels within that area. Much ingenuity and a great deal of labor have gone into construction of these works, and considerable labor is required in the annual task of reconstructing or repairing the diversion weir. The weir may be made of brush-wood and cobbles, or many rows of wooden stakes driven into the river-bed where it is of soft material, and infilled with mud. More recently wire mesh baskets filled with cobbles (gabions) are being used for weir construction, where stones of suitable size are available.

Such systems are exposed to a number of hazards. The diversion weir may be washed out during the irrigation season by a late flood, or the conveyance canal may be put out of service by slides (particularly in the very steep terrain of the hill schemes). In a dry year the stream-flow at the diversion may be excessively small. Many such schemes are on streams which normally cease flowing entirely early in the dry season (i.e. are non-perennial), in which case irrigation is confined to supplemental watering in the wet season and supply to a limited portion of the service area early in the dry. Others are on perennial streams, but usually with dry-season flow highly variable from year to year. In these circumstances, the extent of the area which can be irrigated in any particular year is uncertain, and farmers at the tailend of the distribution systems are likely to have very precarious supply. However, experience over the years indicates how far the distribution system can usefully be extended, granted that returns from the lower end of the system may be marginal.

Conventional economic analyses did not, of course, enter into the design of the existing farmer constructed schemes. They were constructed at a time when there was little avenue for other employment (low "opportunity cost" labor). In a subsistence-level situation even uncertain irrigation supply was judged to be better than none. This is obviously not an adequate basis for analyses of such a scheme from the viewpoint of an international development agency. The entry of such agencies into the field of the small farmer system poses a number of questions. These include the scope of the assistance, whether improvement of existing schemes or construction of new schemes, whether there are, in fact, many perennial streams suitable for development which are not already preempted by existing schemes, and what economic criteria should be used in determining the viability of new schemes (undeveloped sites are often undeveloped for the reason that they are problem sites and thus costly). In terms of the criteria usually employed in evaluation of irrigation projects many schemes on small unregulated nonperennial streams would be judged non-viable. However, in the development of such small schemes there can be important socio-economic factors which lie outside the compass of conventional economic evaluation of larger irrigation projects (such as, in the case of hill projects, the alternate cost of transport of food into remote areas without road access, or the social cost of migration from such areas). In any case, it would be impractical to analyze in detail each scheme as small as two or three hundred hectares, and simplified economic criteria have to be devised. The nature of these criteria strongly influences the scope for participation of international financing agencies in village-level irrigation development (Martin 1987, Sundar 1990).

In addition to the economic criteria, the irrigation system design approach needs to be tailored to the particular situation of the village scheme. The development in most cases is likely to be improvement or extension to existing schemes, and the concept of farmer ownership must be preserved at all costs. The improvements should be primarily those requested by the farmers, i.e. works aimed at remedying problems perceived by the farmers rather than those conceived by the lending agency as being desirable. This approach is doubly necessary if the farmers are to be asked to pay for or contribute to the work. However, it is reasonable to expect that the lending agency would wish to see some degree of up-grading as the result of its participation. In this regard, there are modern technologies which could indeed be introduced with considerable benefit, provided that farmers were persuaded of their value. Such technologies must, however, be adapted to the particular circumstances of such schemes, notably maximum use of local materials, minimum transportation (particularly in remote hill projects), and minimum cost to the farmers concerned. These requirements will often rule out designs which are scaled-down versions of structures conventionally used in larger projects.

In cases where the proposed improvements to an existing scheme go beyond the construction of a more permanent intake weir and include reconstruction and possible partial lining of the conveyance canal, an issue may be whether the size of the service area should be reexamined, particularly in the context of increased canal capacity. Any such reexamination poses two questions. First whether the crops currently grown are indeed appropriate to the site, and second whether the cultivators would agree to any change (these are "farmer-owned" schemes). For example, in Nepalese hill projects, paddy, the favored crop, is being grown in soils with infiltration rates many times higher than the rate normally considered to be the appropriate upper limit. Hence, should the improved scheme still cater to such service, or should a change in cropping pattern be suggested and if so to what alternative crops. If the cultivators being served by the scheme strongly prefer to stay with rice, as is likely to be the case, should a change from wet-land practice to up-land be suggested (much of the rice at present being grown in such areas is in fact grown largely under up-land conditions). The problem is that any reduction in supply of water to existing irrigators, in the interests of supply to others, is likely to be resisted by the present irrigators, whose rights may be of long standing. Furthermore, any increase in consumptive diversion to one scheme may reduce the flow available to some downstream scheme. It is evident that the approach to irrigation system design in a small hill scheme may be very different from the approach to the design of a major scheme. A case by case approach is necessary, with as much attention to sociological as to technical factors.

With regard to technical design, the possibility of introducing newer materials and methods into the traditional construction of small schemes is worth consideration and is attractive to international lending agencies assisting in this area. Possible improvements to intakes (a perennial problem to farmers) include greater use of wire-mesh stone-filled crib work, and particularly the use of cribs fabricated from light steel rod rather than wire to better withstand abrasion by rocks during passage of floods in a mountain stream. Tethered boulders, retained by steel rods with upstream anchorage have been used elsewhere in weir construction in torrential streams, and would appear attractive for hill schemes in some situations. Improvements to channel linings by incorporating modern geotextile fabrics behind traditional masonry lining also offers possibilities, and in some locations the use of light-weight portable prefabricated channel lining units constructed of G.R.C. (fiber-glass reinforced mortar).

The design of improvements to an existing small scheme or the design of a new one is not necessarily simple. All of the physical factors entering into the design of a larger project are present in the smaller scheme, but with diseconomies of scale. Indeed, more engineering judgement may be required in the design of improvements to a small scheme than in the more conventional text book design of a larger one. In addition, much of the topographic and hydrologic data normally available in the case of the larger project may not be available in the case of the small schemes. Added to the problem is the difficulty of access to the small scheme (commonly several days on foot in hill schemes) for staff carrying out investigation, design and technical supervision of construction.

Important issues are the extent to which work on small schemes should be carried out by the beneficiary farmers versus by contractor, responsibility for supervision of construction and disbursement of funds, and the respective roles of the national agencies which may be concerned (Irrigation, Agriculture, Rural Development, etc.)

Small farmer-constructed diversion schemes are undoubtedly an important component of agricultural development in some areas, and one which deserves the support of international development institutions. However, any direct involvement (other than simply financial support to concerned national agencies) must face a number of issues which have proved troublesome in the past and are likely to limit the scope for channeling financial assistance in this direction.

Village schemes with storage

In view of the highly seasonal nature of rainfall in monsoonal regions, any irrigation scheme would benefit from reservoir storage, including small village schemes. However, construction of even a small storage (usually referred to as a "tank") is usually beyond the capacity of a farmer group and involves the participation of government agency. The farmers served by the system usually operate and maintain it, with minimum further government intervention. There are large numbers of such schemes in existence, some centuries old. The area served ranges from as little as fifty up to several thousand hectares. New construction of tanks continues and tank rehabilitation programs are also in progress.

The tank scheme has all the ingredients of a highly desirable field for participation of international financing agencies. It has a considerable element of self-help and farmer management, it is ecologically sound, and it undoubtedly improves quality of life in the village. Factors influencing the scope for further development of such schemes include the availability of sites suitable for storage construction and a number of financial and economic issues.

With regard to sites, key considerations are topography, which determines the cost of storage capacity, and hydrology. A typical location is in an open valley, combining the possibility of obtaining storage capacity at relatively modest cost of dam construction and a service area immediately downstream from the dam, partly in valley-bottom and partly on the valley slopes. Alternatively the reservoir may be located in the more steeply sloping head reaches of the valley, a supply canal extending down to a service area in the more gently sloping lower reaches.

The hydrology of a site is of interest in two respects, yield and flood flow. As there are rarely river-flow or rainfall records of long standing at small project sites, the estimation of seasonal yield and of flood-flow are likely to be quite approximate, as evidenced by the number of existing tanks which have never filled and others which have over-topped due to inadequate spillway capacity. There is need for further study of the hydrology of small catchments in semiarid areas and further installation of hydrological and meteorological recording stations in areas of prospective further tank development.

The spillway is often the major item of cost in a small tank scheme, particularly where several such schemes occur in series, down the length of a valley. The yield available to each of the schemes is then only a proportion of the total yield from the catchment, but the flood flow which has to be accommodated at each is the whole flood from the catchment, subject only to the small amount of flood regulation in upstream tanks. This can lead to disproportionately high spillway costs per unit of area irrigated, particularly for very small schemes. The problem is aggravated by the fact that conventional design requires founding the spillway and its outlet channel and terminal energy dissipating structures on sound rock, which is often at considerable depth in the topography common to small tank schemes. There is room for a less conventional approach to the design of small tank spillways, including greater use of flexible stone-filled wiremesh crib-work, which reduces the need for founding on bed-rock and the various means of protecting small embankment dams against wash-out in the event of overtopping, currently receiving considerable attention elsewhere.

The long-term hazard with small tanks is siltation. Many older tanks are operating at much reduced capacity or are virtually out of service, due to depletion of storage capacity by silt accumulation. Removal of silt from existing tanks by mechanical excavation is of doubtful economic viability. In some cases the storage capacity lost by siltation can be compensated by raising the crest of the dam, but only where the higher reservoir level would not present a problem of encroachment. The use of some form of automatic spillway gate, limiting the surcharge on the spillway crest during passage of floods, can convert flood surcharge pondage to active pondage, thereby increasing net storage capacity. The gates can be hinged wooden shutters arranged to fall, or to tilt, when the reservoir reaches a certain level. The shutters are reset after passage of the flood. More sophisticated automatic shutters are available, which conserve part of the flood volume, but sophistication is not a desirable feature in a small tank with minimum maintenance and maximum exposure to interference with the spillway structure. Increasing the effective length of the spillway crest by employing, in principle, the duck-bill weir arrangement previously described (the crest in this case has a zig-zig configuration, in plan) can also reduce the height of flood-rise, thereby increasing live storage capacity.

Reducing the rate of siltation is of course highly desirable, although not generally easy to accomplish. The catchment area may be cultivated lands, with high sediment run-off when heavy monsoon rains fall on newly ploughed fields or it may be in over-grazed deforested lands also subject to heavy run-off. Anti-erosion measures are available, but generally involve changes in land use or cultivation practices in the catchment area, raising questions of jurisdiction and recovery of cost. Much of the catchment erosion is commonly focused on local areas, such as deeply incised stream channels which are back-eroding in soft material. Remedial measures may be initially directed at such areas, which do not raise the problems of changing land use or cultivation practices. A more comprehensive method of erosion control involving contour planting of vetiver grass is being promoted by the World Bank. It has been particularly successful in some areas.

The concept of conserving excess monsoonal runoff at source, in a multiplicity of small pondages, is very attractive ecologically. The storage may be as small as farm ponds, the somewhat larger "seepage tanks" which serve solely to recharge groundwater for supply to immediately downstream dug-wells, or the regular irrigation tanks with associated small canal systems. Construction of farm ponds can be undertaken by the cultivator himself The spillway problem is minimized due to the very small size of the catchment involved, the whole storm runoff generally being accommodated in the pondage. Alternatively, a simple spillway channel, protected by appropriate vegetative cover, may be provided. The hydraulic head involved is small, and erosion during flood discharge is not a major consideration, nor is the stability of the very small embankment dam.

However, in moving up the scale to the conventional small tank the situation changes. Construction of an embankment dam by manual methods (head-basket or bullock-cart) is in some respects an excellent village enterprise, particularly during periods of "scarcity" when there is little alternative employment. However, many such dams have failed due to slides on the upstream or downstream slopes because of design deficiencies, unfortunate choice of fill material, or lack of compaction. Such incidents can be avoided by providing appropriate technical assistance to the villagers concerned, also with respect to spillway provision.

From the viewpoint of the international agency, a key issue in the construction of tank irrigation schemes is economic viability. This problem has already been discussed in relation to small direct diversion schemes. In the case of a tank scheme the availability of water storage permits higher crop benefits, but adds considerably to construction costs, particularly due to spillway works. Conventional economic analysis based on conventional cropping patterns, as would be applied to a major irrigation project, frequently indicates that a prospective small tank scheme is economically non-viable. However, such simple analysis does not do justice to the full range of benefits of a tank scheme. Particularly in a semi-arid area, a tank scheme is a catalyst for many village activities, often effecting a remarkable transformation from the pre-project condition. While the "quality of life" aspect of a tank scheme in such an area should not be ignored, such benefits are not readily evaluated in economic terms. There are much more tangible benefits which can and should be evaluated, in deciding the fate of a prospective tank scheme. The deficiency may not be in the scheme, but in the scope of the economic analysis.