- Email: [email protected]
Management components in irrigation system design and operation
MANAGEMENT
COMPONENTS IN IRRIGATION DESIGN AND OPERATION
GILBERT
SYSTEM
LEVINE
Water Resources and Marine Sciences Center, Cornell University, Ithaca, New York, USA (Received: 1 October, 1976)
SUMMARY
Irrigation systems in developing countries are often ineficient, both in water use and in cropping effectiveness. Studies in the Philippines, Taiwan and other parts of Asia suggest that in design and operation management constraints are considered inadequately. Designs based on preconceived norms of eficiency often fail to recognise the role of water as afactor substitutefor such inputs as labour, capital andmanagerial skill. Similarly, theyfail to recognise thatpublic objectivesfor system performance are usually not congruent with farmer objectives or even with those of the irrigation bureaucracy. The studies and experience in Asia suggest that signtj?cant improvements can result when these diflerent objectives, with their management implications, are considered in design and operation, and when more effective feedback and response mechanisms are included. The lack of understanding of peasant farmer objectives and of appropriate linkages between such farmers and irrigation bureaucracies also suggests that the relative advantages of small irrigation systems should be considered carefully when decisions about new systems are made.
INTRODUCTION
The design and operation of many irrigation systems in the tropics, especially in developing countries, are often inefficient because the importance of the management component and of social constraints has been, or is, under-estimated. This paper considers these inefficiencies primarily in terms of the seasonal (wet-dry) tropics of Monsoon Asia. The argument of this paper is that, although every irrigation scheme is location specific, there are some important management constraints on irrigation efficiency which are of general application and which are not under the control of the 37 Agricultural Administration Printed in Great Britain
(4) (1977)-a
Applied Science Publishers Ltd, England, 1977
38
GILBERT
LEVINE
irrigation engineer, either in the design or operational stages, but which greatly influence the efficiency of water use in practice. Thus: (1) Our knowledge of the interrelationships between water and plant growth far exceeds our knowledge of the inter-relations between water and the human element in delivery and utilisation: in other words, irrigation engineers face the same social problems, as say, veterinary surgeons. (2) The efficiency concepts used in irrigation system design tend to understress the human component as a factor in water use crop production. (3) Irrigation systems, on the one hand, and the farmers they serve, on the other, have criteria of optimal efficiencies of water use which may not coincide. When they are far apart there is friction between the system and the farmers and/or between the farmers. (4) Within the resources available to the farmers and to the system, the operational optima for both parties can be brought closer together by effective liaison, e.g. feedback and response mechanisms. (5) As a result of (1) to (4) above, it is usually better for the irrigation engineer to ‘recognise’ probabilities initially and strive, through reasonably acceptable change, towards possibilities.
WATER
USE,
CROP
GROWTH
AND
THE
HUMAN
ELEMENT
These five points can be illustrated by considering water delivery, water use and crop growth. A key decision in designing an irrigation system is to establish its water requirement. The design water requirement determines either the maximum area that can be served or, if the area is specified, the amount of the required water supply. The water requirement is the basic parameter dictating channel water-carrying capacity and has implications for the physical, biological, economic and social environment locally. The estimated water requirement frequently is based on Potential EvapoTranspiration (PET) (sometimes termed T). Potential Evapo-Transpiration is the water potentially evaporated from the leaves of a crop and from the land or water it is growing in. When PET is satisfied, plant growth is near or at its maximum. This is well shown in Figs. 1 and 2. Potential Evapo-Transpiration has been, or can be, calculated for most parts of the world from published meteorological data. The water requirement for irrigation is expressed as acre-inches (hectare-centimetres) or,sometimes cubic metres (m3) of PET. The irrigation or net design water requirement is the additional water needed to supplement probable local rainfall or soil water reserves up to the level of PET. (Note that moisture exceeding PET can, with many crops, lower yields.) In practice, however, the amount of water delivered by an irrigation system to a ‘turn out unit’ (e.g. water gate) on a farm or village has to exceed the net design
IRRIGATION
SYSTEM
Water
DESIGN
reaching
AND
crop
OPERATION
39
(mm)
Fig. 1. The yield of wheat as a function of applied water. Tunisia. (Centro International Majoramiento De Maiz Y Trigo, CIMMYT Report, 1968-69.)
I /
IO-
De
SOII
1 I 0
continuously flooded ’ a-**
a-
t E : E
0
6-
0
/
/
a,
jr
42004’
, 500
-/ .
, 600
Water
I 700 reaching
I 900
I 800 crop
( mm
I 1000
)
Fig. 2. The yield of rice (IR 8) as a function of applied water. Dry season, 1969, Los Banos, Philippines. (R. Reyes, unpublished data, lnternational Rice Research Institute.)
40
GILBERT
LEVINE
water requirement because of the losses in water distribution after the water has left the system, i.e. been delivered or sold to the farmer or the local farmers’ organisation. (The gross design water requirement is the net requirement plus allowances for loss by seepage from, and evaporation during, storage or transit along the channels of the systems.) Our knowledge of the crop water requirements for maximum yield is good. But our knowledge at field rather than crop level is less satisfactory because of the variability of local physical conditions, particularly of soil and water table conditions. However, from appropriate on-site measurements, it is possible to make reasonably precise estimates for assumed operating conditions. Adding together the losses before and after ‘turn out’ (i.e. delivery to farm), the amount of water to be diverted is somewhat less than twice the amount needed by the crop in field (i.e. the overall efficiency is 55 x-60 %) in some recent systems. But in some arid areas the efficiency is nearer 30 % than 60 %. Examples from South-East Asia range from 2500 mm
Fig. 3. Typical irrigation
system water requirements (mm/season).
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
41
25 % to over 90 % (Levine’). (See Fig. 3.) With a basic water requirement of 650 mm per crop the Tou Liu system in Taiwan is over 90 % efficient and efficiency is less than 25 % in the Philippines. Why does the Tou Liu system in Taiwan have an efficiency over 90% (with an estimated basic water requirement of approximately 650 mm), and systems in the Philippines less than 25x? Why are the majority of the systems in Taiwan in excess of 60 %, while those in Malaysia are closer to 40 % ? Superficially, an explanation of the differences can be provided. The systems in Taiwan practice rotational irrigation’ within 50 ha units in accordance with a specified plan; the lateral distribution channels frequently are lined with concrete; there are control gates and Parshall flumes at each 50 ha turn out; there is an extensive system of farm ditches; on-farm water distribution is handled on a 24 hours basis. The Malaysian systems practice continuous irrigation; control within the primary and secondary conveyance channels is centralised and effective; water policy is specified by ordinance each year; turn outs serve relatively large farming areas; distribution beyond the canal systems is in the hands of the farmers and few farm ditches are used. The Philippine systems are based upon continuous irrigation;‘ there are few effective controls in the conveyance channels and turn outs, channel maintenance is at a low level; there are essentially no measuring devices; system control is exercised eight hours a day, five days a week; farmer co-operation in water distribution is variable and frequently of a low order. Thus, very substantial differences exist among the systems described. These can be visualised in terms of
2500 k
lncreasmg
Fig. 4.
The irrigation
control
water requirement for lowland rice as affected by the level of control inputs
42
GILBERT
LEVINE
the paradigm shown in Fig. 4 with the operative differences combined in terms of their impact on the ability to control the water. Thus, we can relate in very broad terms a combination of physical facilities and set of actions to water use efficiency, within the South-East Asian context, but we are not in a position to separate them nor to establish the reasons for their effective combination, although attempts are being made to do so.* The fact that facilities and actions are not combined with equal effectiveness is easily illustrated in many modern systems. An outstanding example of low effectiveness is the Dez Pilot Irrigation Project in Iran. Here the traditional systems, with minimal facilities, had water use efficiencies approximating 25 %. The Dez Pilot Irrigation Project is a comprehensive system, with a full range of controls, measuring structures, organisational structure, and all the other accoutrements of a large modern system. But the average water use efficiency in the pilot programme area, after six years of operation, was between 11% and 15 %. By contrast, the Tou Liu system in Taiwan, using the same basic facilities and policies as other systems in Taiwan, achieves substantially higher water use efficiencies than the typical system. A very high degree of farmer co-operation, reflected in the joint hiring of common irrigators to whom complete responsibility for water management is delegated, plays an important role in this difference. I think it is valid to say that we have used the concept of physical or engineering irrigation system efficiency as a substitute for knowledge of the effect of the human element in water management. That some mechanism was-and, at this stage in our understanding, still is-necessary to fill this gap can be accepted. Unfortunately, reliance on the efficiency concept may have seriously inhibited study of the human components of water management.
WATER
AS A FACTOR
SUBSTITUTE
In addition to the potentially inhibiting effect on the identification of relevant components in water management, reliance on the water use efficiency concept tends to mask the value of ‘excess’ water as a substitute for other resources in more limited supply. The South-East Asian examples illustrate this substitution clearly. The Philippines, up to recent years, has had land available for agricultural expansion, rice varieties with relatively low yield potential, easily developed water supplies and relatively low farm and governmental income. The investment necessary to use the water more efficiently (in terms of net water requirement/amount diverted) has not been considered justified. Within the past few years the availability of new land has decreased,4 rice varieties with greatly increased yield potential5 have been developed and new water sources have increased costs associated with them. There is now evidence that both governmental and farmer attitudes towards the more efficient use of water are changing.
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
43
In Taiwan, conditions only recently appearing in the Philippines have existed for almost 50 years. Limited agricultural land, relatively high yielding varieties and a scarcity of easily developed water sources have put a value on efficient water use that is high in relation to the costs of providing this efficiency. Even with these conditions, it was not until the 1954-55 islandwide drought that widespread measures were taken to increase water use efficiency to the current high levels.6 To accomplish the necessary changes in system and farmer action, subsidies and technical assistance were provided, but in some cases force was required. Thus, political commitment, as well as financial resources, was necessary in order to achieve the efficiency objective. It took very special conditions, however, to raise the value of the water in the Taiwan system to the point where efficiency is essentially 100%. Obvious among these conditions is a clearly limited water supply; less obvious are the personal relationships among the farmers and between the farmers and the system. Of significance, however, is the fact that system design, and the associated operational plan, are not based upon an estimated water use efficiency, but upon specific identification of the water used for the various stages of crop culture. This information has been gained within the context of the local situation (much in the way the early water use studies were conducted in the United States) and adapted with improved water use in mind.
OPERATIONAL
EQUILIBRIA
Implicit in the foregoing discussion is the idea that not only is water used as a substitute for other resour.ces in more limited supply, but that the evolved use is a reasonably optimum use of the combination of available resources. If this argument is tenable, then it must be recognised that the supply of these resources is usually not available to the farmer and to the irrigation systems in equal amounts, at least when the systems are not private. It also must be recognised that the evolved system may not (and probably does not) represent an optimum use of the resources available to each, when considered independently. In addition to this difference in available resources, there are usually differences in objectives. While the public systems are assumed to have goals of serving the farmers in the most effective way, other considerations relating to system performance (apparent command area, cost per unit water delivered, etc.) may actually be the objectives towards which the system strives. Superimposed upon both the farmer and the system may be the government, with a different set of available resources and objectives. The net result of the interaction among these three sets of resomces and objectives is the pattern of practices and results that can be seen to be existing. But with the advent of new high yielding cereal genotypes, fertilisers, pesticides, etc., the situation for those farmers in a position to use, or willing to use, the new technology has dramatically changed their attitude to water and irrigation.
44
GILBERT
LEVINE
When the range of appropriate crops and the productive potentials of these crops were limited, the farmers were operating with a production function that was relatively flat, and the utility of their resources changed little over a wide range of water service. Thus, system policies that resulted in a relatively low level of service to the individual farmer had a relatively small impact on relations between the farmer and the system. In addition, since the deliveries were, in the case of wheat, usually in the linear range of a single production function (Fig. 1) there was relatively little adverse effect on aggregate production from the available supply of water. However, with the development of these crops with high productive potential, an individual farmer now has the opportunity to shift to entirely new-and steeper -production functions. His investment of labour and other resources can now be of much greater utility with higher levels of irrigation service; thus, his optimal level may be substantially different from that of the system on which he depends for water, if this system has equity or command area objectives. If the farmer’s optimum is not recognised (and included in the design and operation of the system) a combination of events can be predicted : aggregate production from the area served will be less than the potential; those farmers with sufficient resources will develop private systems (frequently tube wells where physically possible); those farmers without these resources will put increased pressure on the public system. This discussion could be extended to consider, for instance, the lowland rice case, which has a different set of physical factors operative (e.g. the rice production function has much more of a ‘threshold’ than a ‘linear’ relationship), or the effect of different land tenure patterns on the optimum water service for the farm operator. Without doing this, however, I hope the main points have been made: what is optimum from an irrigation system point of view is not necessarily optimum from the farmer’s point of view; these optima must be reasonably close for efficient use of the resources available to each; a change in the resources available to either results in a change in the water use level. Change in the water use level may in its turn require changes in the organisation and administration, i.e. in the control capability of the irrigation system. Efforts to increase this capability usually require significant investm.ent of financial and other resources. In some Taiwan situations, for example, it was necessary to jail farmers for opposing diversion of water serving their areas, even though the amounts remaining were substantially in excess of the physical environment requirements. SYSTEM-FARMER
INTERACTION
In most systems the actual amount of water available for delivery is substantially in excess of the basic crop requirement, though there may be considerable variability in the short time supply of many diversion, or run-of-the-river, systems. In many, if not most, of these systems, even the potentially adequate supply does not serve
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
45
to encourage farmers to invest in new technology because, as has been suggested before, the ‘total’ system (water source to crop continuum) does not operate with an overall optimum use of resources in mind. In Uttar Pradesh, India,’ i by comparison to the private tube wells, the state tube wells were less effective in meeting the needs of the farmers, and the farmers reacted by neither adopting new practices nor investing financial resources. But the picture is not entirely clear. The state tube wells provided a basic water supply of approximately 420 mm/ha/y, with the farmers served from these wells using about 365 mm/ha for the &si wheat crop. The private tube wells provided a water source of approximately 500 mm/ha/y, with the farmers using 485 mm/ha for the same crop. These ratios of use to supply represent very high levels of water use efficiency-about 85 % and 95 % for the state and private wells, respectively. However, from a net income point of view, the utility of the different water sources was very different. The question arises, ‘to what extent was the difference in net income (approximately 790 rupees per hectare per year) due to the difference in quantity of water available to the individual farmer, and to what extent was it due to differences in other elements of irrigation service?’ The data show major differences in the use of new dwarf wheat by the two groups of farmers, as well wasdifferences in other practices reflecting more intensive cropping, with private well farmers being more intensive in their operations. If we confine our considerations to wheat alone, and assume the production function shown in Fig. 1 to be reasonably applicable to the Indian environment, we can see that the delivered water supply to priiate well farmers is close to the peak of the production function, while state well farmers would be operating approximately 25 % lower on the yield scale, if they grew dwarf wheat. This, however, would still be substantially higher than the yield of traditional varieties, and more than sufficient to pay for the additional costs associated with dwarf wheat production. Some of the state well farmers did grow the new varieties, but ‘Farmers using state tube wells and growing dwarf wheat were the more influential farmers with larger farms located near the wells; they were, therefore, able to provide an extra irrigation for their dwarf wheat’. 7 It would seem, therefore, that there were problems of equity of distribution in addition to level of supply. In addition, the stress on efficiency of equipment utilisation (evidenced by a 4000 hours per year running time, in contrast to 1500 hours per year for the private wells) resulted in a number of equipment breakdowns and unpredictable delays in repair. Added to the problems of delivery equity and reliability of service was the complete lack of mechanism to adapt water deliveries from state wells to meet farmer needs. Water deliveries were on a rigid timetable with no provision for co-ordination with the farmers’ abilities to utilise water. This combination of elements made it unattractive for most of the state well farmers to take the risks accompanying the higher investment for dwarf wheat and for other forms of crop intensification.
46
GILBERT
LEVINE
By contrast, in Taiwan, when government policy emphasised water use efficiency, but also stressed productivity, equity and payment of irrigation costs by the farmer beneficiaries, a series of interacting efforts was conducted to implement the policies6 and interaction was built into irrigation system operating mechanisms. These latter included regular meetings between system personnel and farmers to explain broad policies, specific policies relative to water delivery and operational procedures, election by the farmers of an honorary group leader to act as the intermediary between small groups of farmers and the irrigation system (on a continuing and operational basis); performance rating of all system personnel; a programme of field performance data collection and a number of others. The impact of these types of interaction can be illustrated with a few examples. When the engineering personnel of one system planned a channel lining programme they anticipated major reductions in the losses of water in transit and recommended a 40 y0 reduction in the amount of water diverted into the channels. Management personnel of the system objected to the reduction pending actual measurement of anticipated savings. The field data collection programme revealed much smaller water savings, and the diversion reduction actually implemented reflected field information, rather than the projections. There was essentially no adverse effect on the farmers. The lining programme, however, did result in improved timing of deliveries and in reduced maintenance costs. The improved maintenance situation was noted by the farmers, who have responsibilities for maintenance of many of the smaller channels, and they embarked on lining programmes for the smaller channels, with the assistance of regular system personnel. In the Tou Liu system, previously mentioned, the normal pattern of farmersystem interaction was insufficient to provide effective service due to a very limited water resource. While the system could provide the limited water equitably to the 10 hectare turnout units, water distribution within the units (farmed by 8-10 farmers) was done with the individual farmer objectives uppermost. Discussions among the farmers, with the honorary group leader and system technical personnel participating, indicated possibilities for improved service provided within-unit water distribution was delegated to common irrigators. These individuals, responsible to and paid by the farmers, were trained by system personnel. The result has been a very effective water utilisation programme, with equity of water distribution being recognised as differing from equality of water distribution. For example, in the gently sloping areas of the system, upper paddies are irrigated first, with specified quantities. As water is ‘rotated’ to lower lying paddies in subsequent days of the irrigation cycle, the amount of water turned into these paddies is less, in recognition of seepage from higher paddies. Thus, the overall system operation provides a water environment that permits equity of crop growth, rather than providing equal amounts of water. This represents a sophisticated management practice which is not found in many systems in developed countries. All of this is
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
47
done within the crop constraint of relatively high sensitivity to water stress (by comparison to most upland crops), and evidences both dependable ,system performance and farmer acknowledgement of that performance. The interaction and response illustrated are not only to be found in water delivery aspects of system operation. Within a given system the special fee charges are based upon capital costs associated with individual channels to reach the areas of the system. Thus, farmers within one system will have different irrigation fees, depending upon their location. In some cases these charges may be as high as 5000 $NT ($125) per hectare/annum.’ Notwithstanding these rates, the percent of farmer repayment has been very high-frequently above 95 %. However, when it has not been possible to adequately serve an area, and a drop in production results, adjustments are made in the fee schedule. In much modern. irrigation system design the importance of feedback-response interaction is recognised, in principle, and a delivery on ‘demand’ programme is assumed to be the way to implement this interaction. Within a variety of constraints (water rights, requirements for multiple requests on the same channel, etc.) water is delivered upon request of the farmers. That this will not always lead to anticipated results is amply illustrated in the Dez Pilot Project. That it is not the only appropriate mechanism is clearly evident from the Taiwan experience. PROBABILITIES
AND
POSSIBILITIES
The problems experienced by farmers served by existing irrigation systems in tropical low income countries, and the problems experienced by the systems themselves are manifold and manifest. As a result, but not solely due to these problems, a philosophy has evolved to the effect that irrigation modernisation can take place only with radical departures from traditional practice. These changes usually are viewed from the perspective of engineering efficiency: efficiency in use of water, efficiency in mechanics of operation and maintenance; efficiency in irrigation system costs. Where causes of problems are identified as being in the realm of the physical environment, design usually includes specifics for operation within that environment. Attempts are made to tailor a physical solution to a specific problem. In doing so, social constraints are either ignored or treated in a general, non-specific manner. The pattern of looking for general rather than specific solutions to social constraints, and for emphasising change rather than adaptation, is evident even when problems are identified as being primarily in the social area. Almost all new irrigation projects recommend some form of public institution similar to that found in developed countries. There is very little attempt to work within existing institutions or to develop adaptations based upon indigenous social relationships. The argument frequently is made that adapting to existing social constraints reinforces the existing social structure. From a development point of view, or from
48
GILBERT
LEVINE
an equity point of view, this may be undesirable. In principle, there is no disagreement. However, I would argue that moderate adaptation with specific equity or development goals in mind is a more effective and efficient mechanism for achieving these goals than is the introduction of foreign institutions with a high probability of failure. Compounding the problem is the emphasis on large projects. While there frequently are major economies of scale associated with irrigation projects, there are attendant diseconomies associated with the added and more stringent requirements for data, for accurate projections and for specialised skills, as well as with the more complex nature of relationships within large projects.
CONCLUSIONS
(1) There are serious gaps in our understanding of irrigation in the low-income tropics. (2) These gaps exist not only in our knowledge of the physical environment, but even more broadly in the social and economic environment. (3) Given these gaps, research to bridge them should be given high priority. (4) Recognising that irrigation investment will continue to be made before substantive answers are obtained, the role of smaller systems vis-ir-vis the larger should be evaluated with the tools at hand. (5) Existing irrigation systems represent research infrastructure. Comparative research using this infrastructure may be the most efficient route to identification of pertinent problems, and towards their solution, in the technical engineering and biological fields as well as in the social field.
REFERENCES
1. LEVINE, G., Paper presented to a seminar sponsored by the United States Agency for International Development (USAID), 1971. 2. TSAI, TSUI-YUAN, Rotational irrigation practice, Chianan Irrigation Association, 1964. 3. MIRANDA, S. M., The effects of physical water control parameters on Philippine lowland rice irrigation performance. Unpublished PhD. Thesis,, Cornell University, USA. 4. GOLAY, F. II. and GOODSTEIN, M. E., Philippine rice needs to 1990. Output and input requirements, USAID Mission, Manila, Philippines, 1967. 5. CHANDLER, R. L., The basis for the increased yield capacity of rice and wheat and present and potential impact on food production. In: Poleman, T. T. and Freebairn D. K. (Eds) Food, The impact of the Green Revolution. Praeger Publishers, New York, population and employment. 1973. 6. LEE, T. S. and CHIN, L. T., Development of rotational irrigation in Taiwan. Far East Regional Irrigation Seminar, Taipei, Taiwan, 1961. 7. MELLOR, J. W. and MOORTI, T. V., Dilemma of state tube wells, Economic andPoIitica1 Weekly, VI (13 March 1971) p. 27. 8. CHIN, L. T., Water Management Specialist, FAO, Rome. (Pers. Comm.)
COMPONENTS IN IRRIGATION DESIGN AND OPERATION
GILBERT
SYSTEM
LEVINE
Water Resources and Marine Sciences Center, Cornell University, Ithaca, New York, USA (Received: 1 October, 1976)
SUMMARY
Irrigation systems in developing countries are often ineficient, both in water use and in cropping effectiveness. Studies in the Philippines, Taiwan and other parts of Asia suggest that in design and operation management constraints are considered inadequately. Designs based on preconceived norms of eficiency often fail to recognise the role of water as afactor substitutefor such inputs as labour, capital andmanagerial skill. Similarly, theyfail to recognise thatpublic objectivesfor system performance are usually not congruent with farmer objectives or even with those of the irrigation bureaucracy. The studies and experience in Asia suggest that signtj?cant improvements can result when these diflerent objectives, with their management implications, are considered in design and operation, and when more effective feedback and response mechanisms are included. The lack of understanding of peasant farmer objectives and of appropriate linkages between such farmers and irrigation bureaucracies also suggests that the relative advantages of small irrigation systems should be considered carefully when decisions about new systems are made.
INTRODUCTION
The design and operation of many irrigation systems in the tropics, especially in developing countries, are often inefficient because the importance of the management component and of social constraints has been, or is, under-estimated. This paper considers these inefficiencies primarily in terms of the seasonal (wet-dry) tropics of Monsoon Asia. The argument of this paper is that, although every irrigation scheme is location specific, there are some important management constraints on irrigation efficiency which are of general application and which are not under the control of the 37 Agricultural Administration Printed in Great Britain
(4) (1977)-a
Applied Science Publishers Ltd, England, 1977
38
GILBERT
LEVINE
irrigation engineer, either in the design or operational stages, but which greatly influence the efficiency of water use in practice. Thus: (1) Our knowledge of the interrelationships between water and plant growth far exceeds our knowledge of the inter-relations between water and the human element in delivery and utilisation: in other words, irrigation engineers face the same social problems, as say, veterinary surgeons. (2) The efficiency concepts used in irrigation system design tend to understress the human component as a factor in water use crop production. (3) Irrigation systems, on the one hand, and the farmers they serve, on the other, have criteria of optimal efficiencies of water use which may not coincide. When they are far apart there is friction between the system and the farmers and/or between the farmers. (4) Within the resources available to the farmers and to the system, the operational optima for both parties can be brought closer together by effective liaison, e.g. feedback and response mechanisms. (5) As a result of (1) to (4) above, it is usually better for the irrigation engineer to ‘recognise’ probabilities initially and strive, through reasonably acceptable change, towards possibilities.
WATER
USE,
CROP
GROWTH
AND
THE
HUMAN
ELEMENT
These five points can be illustrated by considering water delivery, water use and crop growth. A key decision in designing an irrigation system is to establish its water requirement. The design water requirement determines either the maximum area that can be served or, if the area is specified, the amount of the required water supply. The water requirement is the basic parameter dictating channel water-carrying capacity and has implications for the physical, biological, economic and social environment locally. The estimated water requirement frequently is based on Potential EvapoTranspiration (PET) (sometimes termed T). Potential Evapo-Transpiration is the water potentially evaporated from the leaves of a crop and from the land or water it is growing in. When PET is satisfied, plant growth is near or at its maximum. This is well shown in Figs. 1 and 2. Potential Evapo-Transpiration has been, or can be, calculated for most parts of the world from published meteorological data. The water requirement for irrigation is expressed as acre-inches (hectare-centimetres) or,sometimes cubic metres (m3) of PET. The irrigation or net design water requirement is the additional water needed to supplement probable local rainfall or soil water reserves up to the level of PET. (Note that moisture exceeding PET can, with many crops, lower yields.) In practice, however, the amount of water delivered by an irrigation system to a ‘turn out unit’ (e.g. water gate) on a farm or village has to exceed the net design
IRRIGATION
SYSTEM
Water
DESIGN
reaching
AND
crop
OPERATION
39
(mm)
Fig. 1. The yield of wheat as a function of applied water. Tunisia. (Centro International Majoramiento De Maiz Y Trigo, CIMMYT Report, 1968-69.)
I /
IO-
De
SOII
1 I 0
continuously flooded ’ a-**
a-
t E : E
0
6-
0
/
/
a,
jr
42004’
, 500
-/ .
, 600
Water
I 700 reaching
I 900
I 800 crop
( mm
I 1000
)
Fig. 2. The yield of rice (IR 8) as a function of applied water. Dry season, 1969, Los Banos, Philippines. (R. Reyes, unpublished data, lnternational Rice Research Institute.)
40
GILBERT
LEVINE
water requirement because of the losses in water distribution after the water has left the system, i.e. been delivered or sold to the farmer or the local farmers’ organisation. (The gross design water requirement is the net requirement plus allowances for loss by seepage from, and evaporation during, storage or transit along the channels of the systems.) Our knowledge of the crop water requirements for maximum yield is good. But our knowledge at field rather than crop level is less satisfactory because of the variability of local physical conditions, particularly of soil and water table conditions. However, from appropriate on-site measurements, it is possible to make reasonably precise estimates for assumed operating conditions. Adding together the losses before and after ‘turn out’ (i.e. delivery to farm), the amount of water to be diverted is somewhat less than twice the amount needed by the crop in field (i.e. the overall efficiency is 55 x-60 %) in some recent systems. But in some arid areas the efficiency is nearer 30 % than 60 %. Examples from South-East Asia range from 2500 mm
Fig. 3. Typical irrigation
system water requirements (mm/season).
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
41
25 % to over 90 % (Levine’). (See Fig. 3.) With a basic water requirement of 650 mm per crop the Tou Liu system in Taiwan is over 90 % efficient and efficiency is less than 25 % in the Philippines. Why does the Tou Liu system in Taiwan have an efficiency over 90% (with an estimated basic water requirement of approximately 650 mm), and systems in the Philippines less than 25x? Why are the majority of the systems in Taiwan in excess of 60 %, while those in Malaysia are closer to 40 % ? Superficially, an explanation of the differences can be provided. The systems in Taiwan practice rotational irrigation’ within 50 ha units in accordance with a specified plan; the lateral distribution channels frequently are lined with concrete; there are control gates and Parshall flumes at each 50 ha turn out; there is an extensive system of farm ditches; on-farm water distribution is handled on a 24 hours basis. The Malaysian systems practice continuous irrigation; control within the primary and secondary conveyance channels is centralised and effective; water policy is specified by ordinance each year; turn outs serve relatively large farming areas; distribution beyond the canal systems is in the hands of the farmers and few farm ditches are used. The Philippine systems are based upon continuous irrigation;‘ there are few effective controls in the conveyance channels and turn outs, channel maintenance is at a low level; there are essentially no measuring devices; system control is exercised eight hours a day, five days a week; farmer co-operation in water distribution is variable and frequently of a low order. Thus, very substantial differences exist among the systems described. These can be visualised in terms of
2500 k
lncreasmg
Fig. 4.
The irrigation
control
water requirement for lowland rice as affected by the level of control inputs
42
GILBERT
LEVINE
the paradigm shown in Fig. 4 with the operative differences combined in terms of their impact on the ability to control the water. Thus, we can relate in very broad terms a combination of physical facilities and set of actions to water use efficiency, within the South-East Asian context, but we are not in a position to separate them nor to establish the reasons for their effective combination, although attempts are being made to do so.* The fact that facilities and actions are not combined with equal effectiveness is easily illustrated in many modern systems. An outstanding example of low effectiveness is the Dez Pilot Irrigation Project in Iran. Here the traditional systems, with minimal facilities, had water use efficiencies approximating 25 %. The Dez Pilot Irrigation Project is a comprehensive system, with a full range of controls, measuring structures, organisational structure, and all the other accoutrements of a large modern system. But the average water use efficiency in the pilot programme area, after six years of operation, was between 11% and 15 %. By contrast, the Tou Liu system in Taiwan, using the same basic facilities and policies as other systems in Taiwan, achieves substantially higher water use efficiencies than the typical system. A very high degree of farmer co-operation, reflected in the joint hiring of common irrigators to whom complete responsibility for water management is delegated, plays an important role in this difference. I think it is valid to say that we have used the concept of physical or engineering irrigation system efficiency as a substitute for knowledge of the effect of the human element in water management. That some mechanism was-and, at this stage in our understanding, still is-necessary to fill this gap can be accepted. Unfortunately, reliance on the efficiency concept may have seriously inhibited study of the human components of water management.
WATER
AS A FACTOR
SUBSTITUTE
In addition to the potentially inhibiting effect on the identification of relevant components in water management, reliance on the water use efficiency concept tends to mask the value of ‘excess’ water as a substitute for other resources in more limited supply. The South-East Asian examples illustrate this substitution clearly. The Philippines, up to recent years, has had land available for agricultural expansion, rice varieties with relatively low yield potential, easily developed water supplies and relatively low farm and governmental income. The investment necessary to use the water more efficiently (in terms of net water requirement/amount diverted) has not been considered justified. Within the past few years the availability of new land has decreased,4 rice varieties with greatly increased yield potential5 have been developed and new water sources have increased costs associated with them. There is now evidence that both governmental and farmer attitudes towards the more efficient use of water are changing.
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
43
In Taiwan, conditions only recently appearing in the Philippines have existed for almost 50 years. Limited agricultural land, relatively high yielding varieties and a scarcity of easily developed water sources have put a value on efficient water use that is high in relation to the costs of providing this efficiency. Even with these conditions, it was not until the 1954-55 islandwide drought that widespread measures were taken to increase water use efficiency to the current high levels.6 To accomplish the necessary changes in system and farmer action, subsidies and technical assistance were provided, but in some cases force was required. Thus, political commitment, as well as financial resources, was necessary in order to achieve the efficiency objective. It took very special conditions, however, to raise the value of the water in the Taiwan system to the point where efficiency is essentially 100%. Obvious among these conditions is a clearly limited water supply; less obvious are the personal relationships among the farmers and between the farmers and the system. Of significance, however, is the fact that system design, and the associated operational plan, are not based upon an estimated water use efficiency, but upon specific identification of the water used for the various stages of crop culture. This information has been gained within the context of the local situation (much in the way the early water use studies were conducted in the United States) and adapted with improved water use in mind.
OPERATIONAL
EQUILIBRIA
Implicit in the foregoing discussion is the idea that not only is water used as a substitute for other resour.ces in more limited supply, but that the evolved use is a reasonably optimum use of the combination of available resources. If this argument is tenable, then it must be recognised that the supply of these resources is usually not available to the farmer and to the irrigation systems in equal amounts, at least when the systems are not private. It also must be recognised that the evolved system may not (and probably does not) represent an optimum use of the resources available to each, when considered independently. In addition to this difference in available resources, there are usually differences in objectives. While the public systems are assumed to have goals of serving the farmers in the most effective way, other considerations relating to system performance (apparent command area, cost per unit water delivered, etc.) may actually be the objectives towards which the system strives. Superimposed upon both the farmer and the system may be the government, with a different set of available resources and objectives. The net result of the interaction among these three sets of resomces and objectives is the pattern of practices and results that can be seen to be existing. But with the advent of new high yielding cereal genotypes, fertilisers, pesticides, etc., the situation for those farmers in a position to use, or willing to use, the new technology has dramatically changed their attitude to water and irrigation.
44
GILBERT
LEVINE
When the range of appropriate crops and the productive potentials of these crops were limited, the farmers were operating with a production function that was relatively flat, and the utility of their resources changed little over a wide range of water service. Thus, system policies that resulted in a relatively low level of service to the individual farmer had a relatively small impact on relations between the farmer and the system. In addition, since the deliveries were, in the case of wheat, usually in the linear range of a single production function (Fig. 1) there was relatively little adverse effect on aggregate production from the available supply of water. However, with the development of these crops with high productive potential, an individual farmer now has the opportunity to shift to entirely new-and steeper -production functions. His investment of labour and other resources can now be of much greater utility with higher levels of irrigation service; thus, his optimal level may be substantially different from that of the system on which he depends for water, if this system has equity or command area objectives. If the farmer’s optimum is not recognised (and included in the design and operation of the system) a combination of events can be predicted : aggregate production from the area served will be less than the potential; those farmers with sufficient resources will develop private systems (frequently tube wells where physically possible); those farmers without these resources will put increased pressure on the public system. This discussion could be extended to consider, for instance, the lowland rice case, which has a different set of physical factors operative (e.g. the rice production function has much more of a ‘threshold’ than a ‘linear’ relationship), or the effect of different land tenure patterns on the optimum water service for the farm operator. Without doing this, however, I hope the main points have been made: what is optimum from an irrigation system point of view is not necessarily optimum from the farmer’s point of view; these optima must be reasonably close for efficient use of the resources available to each; a change in the resources available to either results in a change in the water use level. Change in the water use level may in its turn require changes in the organisation and administration, i.e. in the control capability of the irrigation system. Efforts to increase this capability usually require significant investm.ent of financial and other resources. In some Taiwan situations, for example, it was necessary to jail farmers for opposing diversion of water serving their areas, even though the amounts remaining were substantially in excess of the physical environment requirements. SYSTEM-FARMER
INTERACTION
In most systems the actual amount of water available for delivery is substantially in excess of the basic crop requirement, though there may be considerable variability in the short time supply of many diversion, or run-of-the-river, systems. In many, if not most, of these systems, even the potentially adequate supply does not serve
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
45
to encourage farmers to invest in new technology because, as has been suggested before, the ‘total’ system (water source to crop continuum) does not operate with an overall optimum use of resources in mind. In Uttar Pradesh, India,’ i by comparison to the private tube wells, the state tube wells were less effective in meeting the needs of the farmers, and the farmers reacted by neither adopting new practices nor investing financial resources. But the picture is not entirely clear. The state tube wells provided a basic water supply of approximately 420 mm/ha/y, with the farmers served from these wells using about 365 mm/ha for the &si wheat crop. The private tube wells provided a water source of approximately 500 mm/ha/y, with the farmers using 485 mm/ha for the same crop. These ratios of use to supply represent very high levels of water use efficiency-about 85 % and 95 % for the state and private wells, respectively. However, from a net income point of view, the utility of the different water sources was very different. The question arises, ‘to what extent was the difference in net income (approximately 790 rupees per hectare per year) due to the difference in quantity of water available to the individual farmer, and to what extent was it due to differences in other elements of irrigation service?’ The data show major differences in the use of new dwarf wheat by the two groups of farmers, as well wasdifferences in other practices reflecting more intensive cropping, with private well farmers being more intensive in their operations. If we confine our considerations to wheat alone, and assume the production function shown in Fig. 1 to be reasonably applicable to the Indian environment, we can see that the delivered water supply to priiate well farmers is close to the peak of the production function, while state well farmers would be operating approximately 25 % lower on the yield scale, if they grew dwarf wheat. This, however, would still be substantially higher than the yield of traditional varieties, and more than sufficient to pay for the additional costs associated with dwarf wheat production. Some of the state well farmers did grow the new varieties, but ‘Farmers using state tube wells and growing dwarf wheat were the more influential farmers with larger farms located near the wells; they were, therefore, able to provide an extra irrigation for their dwarf wheat’. 7 It would seem, therefore, that there were problems of equity of distribution in addition to level of supply. In addition, the stress on efficiency of equipment utilisation (evidenced by a 4000 hours per year running time, in contrast to 1500 hours per year for the private wells) resulted in a number of equipment breakdowns and unpredictable delays in repair. Added to the problems of delivery equity and reliability of service was the complete lack of mechanism to adapt water deliveries from state wells to meet farmer needs. Water deliveries were on a rigid timetable with no provision for co-ordination with the farmers’ abilities to utilise water. This combination of elements made it unattractive for most of the state well farmers to take the risks accompanying the higher investment for dwarf wheat and for other forms of crop intensification.
46
GILBERT
LEVINE
By contrast, in Taiwan, when government policy emphasised water use efficiency, but also stressed productivity, equity and payment of irrigation costs by the farmer beneficiaries, a series of interacting efforts was conducted to implement the policies6 and interaction was built into irrigation system operating mechanisms. These latter included regular meetings between system personnel and farmers to explain broad policies, specific policies relative to water delivery and operational procedures, election by the farmers of an honorary group leader to act as the intermediary between small groups of farmers and the irrigation system (on a continuing and operational basis); performance rating of all system personnel; a programme of field performance data collection and a number of others. The impact of these types of interaction can be illustrated with a few examples. When the engineering personnel of one system planned a channel lining programme they anticipated major reductions in the losses of water in transit and recommended a 40 y0 reduction in the amount of water diverted into the channels. Management personnel of the system objected to the reduction pending actual measurement of anticipated savings. The field data collection programme revealed much smaller water savings, and the diversion reduction actually implemented reflected field information, rather than the projections. There was essentially no adverse effect on the farmers. The lining programme, however, did result in improved timing of deliveries and in reduced maintenance costs. The improved maintenance situation was noted by the farmers, who have responsibilities for maintenance of many of the smaller channels, and they embarked on lining programmes for the smaller channels, with the assistance of regular system personnel. In the Tou Liu system, previously mentioned, the normal pattern of farmersystem interaction was insufficient to provide effective service due to a very limited water resource. While the system could provide the limited water equitably to the 10 hectare turnout units, water distribution within the units (farmed by 8-10 farmers) was done with the individual farmer objectives uppermost. Discussions among the farmers, with the honorary group leader and system technical personnel participating, indicated possibilities for improved service provided within-unit water distribution was delegated to common irrigators. These individuals, responsible to and paid by the farmers, were trained by system personnel. The result has been a very effective water utilisation programme, with equity of water distribution being recognised as differing from equality of water distribution. For example, in the gently sloping areas of the system, upper paddies are irrigated first, with specified quantities. As water is ‘rotated’ to lower lying paddies in subsequent days of the irrigation cycle, the amount of water turned into these paddies is less, in recognition of seepage from higher paddies. Thus, the overall system operation provides a water environment that permits equity of crop growth, rather than providing equal amounts of water. This represents a sophisticated management practice which is not found in many systems in developed countries. All of this is
IRRIGATION
SYSTEM
DESIGN
AND
OPERATION
47
done within the crop constraint of relatively high sensitivity to water stress (by comparison to most upland crops), and evidences both dependable ,system performance and farmer acknowledgement of that performance. The interaction and response illustrated are not only to be found in water delivery aspects of system operation. Within a given system the special fee charges are based upon capital costs associated with individual channels to reach the areas of the system. Thus, farmers within one system will have different irrigation fees, depending upon their location. In some cases these charges may be as high as 5000 $NT ($125) per hectare/annum.’ Notwithstanding these rates, the percent of farmer repayment has been very high-frequently above 95 %. However, when it has not been possible to adequately serve an area, and a drop in production results, adjustments are made in the fee schedule. In much modern. irrigation system design the importance of feedback-response interaction is recognised, in principle, and a delivery on ‘demand’ programme is assumed to be the way to implement this interaction. Within a variety of constraints (water rights, requirements for multiple requests on the same channel, etc.) water is delivered upon request of the farmers. That this will not always lead to anticipated results is amply illustrated in the Dez Pilot Project. That it is not the only appropriate mechanism is clearly evident from the Taiwan experience. PROBABILITIES
AND
POSSIBILITIES
The problems experienced by farmers served by existing irrigation systems in tropical low income countries, and the problems experienced by the systems themselves are manifold and manifest. As a result, but not solely due to these problems, a philosophy has evolved to the effect that irrigation modernisation can take place only with radical departures from traditional practice. These changes usually are viewed from the perspective of engineering efficiency: efficiency in use of water, efficiency in mechanics of operation and maintenance; efficiency in irrigation system costs. Where causes of problems are identified as being in the realm of the physical environment, design usually includes specifics for operation within that environment. Attempts are made to tailor a physical solution to a specific problem. In doing so, social constraints are either ignored or treated in a general, non-specific manner. The pattern of looking for general rather than specific solutions to social constraints, and for emphasising change rather than adaptation, is evident even when problems are identified as being primarily in the social area. Almost all new irrigation projects recommend some form of public institution similar to that found in developed countries. There is very little attempt to work within existing institutions or to develop adaptations based upon indigenous social relationships. The argument frequently is made that adapting to existing social constraints reinforces the existing social structure. From a development point of view, or from
48
GILBERT
LEVINE
an equity point of view, this may be undesirable. In principle, there is no disagreement. However, I would argue that moderate adaptation with specific equity or development goals in mind is a more effective and efficient mechanism for achieving these goals than is the introduction of foreign institutions with a high probability of failure. Compounding the problem is the emphasis on large projects. While there frequently are major economies of scale associated with irrigation projects, there are attendant diseconomies associated with the added and more stringent requirements for data, for accurate projections and for specialised skills, as well as with the more complex nature of relationships within large projects.
CONCLUSIONS
(1) There are serious gaps in our understanding of irrigation in the low-income tropics. (2) These gaps exist not only in our knowledge of the physical environment, but even more broadly in the social and economic environment. (3) Given these gaps, research to bridge them should be given high priority. (4) Recognising that irrigation investment will continue to be made before substantive answers are obtained, the role of smaller systems vis-ir-vis the larger should be evaluated with the tools at hand. (5) Existing irrigation systems represent research infrastructure. Comparative research using this infrastructure may be the most efficient route to identification of pertinent problems, and towards their solution, in the technical engineering and biological fields as well as in the social field.
REFERENCES
1. LEVINE, G., Paper presented to a seminar sponsored by the United States Agency for International Development (USAID), 1971. 2. TSAI, TSUI-YUAN, Rotational irrigation practice, Chianan Irrigation Association, 1964. 3. MIRANDA, S. M., The effects of physical water control parameters on Philippine lowland rice irrigation performance. Unpublished PhD. Thesis,, Cornell University, USA. 4. GOLAY, F. II. and GOODSTEIN, M. E., Philippine rice needs to 1990. Output and input requirements, USAID Mission, Manila, Philippines, 1967. 5. CHANDLER, R. L., The basis for the increased yield capacity of rice and wheat and present and potential impact on food production. In: Poleman, T. T. and Freebairn D. K. (Eds) Food, The impact of the Green Revolution. Praeger Publishers, New York, population and employment. 1973. 6. LEE, T. S. and CHIN, L. T., Development of rotational irrigation in Taiwan. Far East Regional Irrigation Seminar, Taipei, Taiwan, 1961. 7. MELLOR, J. W. and MOORTI, T. V., Dilemma of state tube wells, Economic andPoIitica1 Weekly, VI (13 March 1971) p. 27. 8. CHIN, L. T., Water Management Specialist, FAO, Rome. (Pers. Comm.)