Previous Page Table of Contents Next Page

Institutional conditions for effective water delivery and irrigation scheduling in large gravity systems: Evidence from Pakistan

D.J. Bandaragoda, International Irrigation Management Institute, Lahore, Pakistan


This paper highlights the need to consider the necessary institutional conditions for introducing effective water delivery and irrigation scheduling in developing countries such as Pakistan. In Pakistan's large canal irrigation systems, the scheduling of water delivery is normally limited to the main system. When this limited operation is not performed well, water flow fluctuations occur in the main system, which are readily transferred to the secondary and tertiary systems. In the older canal commands of the Indus Irrigation System, largely as a consequence of the increasing flow variability, the rigid warabandi water distribution method is no longer functional in apportioning the scarcity of water equally among all the users. The magnitude of this linkage, however, is not yet fully appreciated. The practice of warabandi is becoming increasingly flexible, with the time schedules frequently modified by farmers, apparently by their mutual agreement. This flexibility seems an outcome of complex social dynamics that represent 'survival of the influential', with some farmers satisfying their mil water requirements at the expense of others, disregarding the equity objective of warabandi. A low institutional accountability in water management often accentuates the inequity in distribution. In the recently established irrigation systems in northern Pakistan, the farmers respond differently to the lack of institutional accountability. They enjoy the advantage of new design features aimed at increased water allowances, by increasing cropping intensities and changing cropping patterns, or otherwise, overirrigating crops. This behaviour has undermined the design intentions for introducing water delivery scheduling in these systems. Thus, with both traditional as well as modem system design, there are institutional impediments to the realization of design objectives. Successfully introducing effective water delivery and irrigation scheduling in this context requires a strengthening of accountability, motivation and awareness regarding water use efficiency, and organized collective behaviour as a strategy to improve equity.

The need to use scarce water resources judiciously has been a major concern from the early days of irrigation development in Pakistan. Several characteristics of Pakistan's irrigation system, related to water availability, soil conditions and diverse cropping patterns, make it a potential beneficiary of effective irrigation scheduling. However, the situation is a little clouded by an important objective of the system design, which requires that the country's scarce water resources be equitably distributed (in proportion to the land to be irrigated) at all levels of the system: main canals, distributaries, outlets, and among the individual farms. The purpose of this paper is to discuss some institutional issues related to the present water allocation practices, and their possible implications for introducing improved water delivery and irrigation scheduling in Pakistan. (The statistics given in the following paragraphs have been extracted from the Ministry of Food and Agriculture (1988), a WSIP Study (1990), and Pakistan's National Conservation Strategy document).

Despite having an impressive natural resource base, Pakistan can be considered a water-short environment. Under arid to semi-arid conditions, a low average annual precipitation of around 200 mm makes Pakistan's agricultural effort heavily dependent on irrigation. The overall conveyance and field application losses reduce the total quantity of water from all sources by almost 65 %, leaving a mere 85 billion cubic metres for crop use over an irrigable area of about 12 million hectares. The design itself imposes a major constraint on water availability at the farm gate. Pakistan's canal network was designed to provide 'protective' irrigation extensively over a large command area, at an average 'water allowance'1 of 0.28 l/s/ha on a 100% cropping intensity. This works out to a meagre irrigation depth of 2.4 mm/day from canal water.

1 Water allowance is the design discharge assigned to the head of a distributary or a watercourse on the basis of the area to be irrigated and is given in litres per second per hectare in this presentation.

Another constraint is that the water resources of the country are unevenly distributed in time and space. About 84% of the total annual river flow mentioned above occurs during the full kharif season, whereas 36% of canal head withdrawal takes place during the rabi season. Thus, the mismatch between the streamflow in the major rivers and the pattern of water requirements of the main kharif and rabi cropping seasons causes seasonal shortages of canal irrigation supplies.

Declining annual growth rates in farm gate water availability, of 3.9% for 1960-67, 2.7% for 1968-78, and 1.6% for 1978-86, confirm the limitations of the overall water supply in Pakistan. The expansion in the irrigated area has increased at even lower rates of 2.7%, 1.3% and 1.5%, respectively, for the same periods. These trends have to be viewed in the light of Pakistan's rapidly increasing population, which is growing at an annual compound rate exceeding 3%, and is estimated to be about 148 million in 2000. The increased demand for water could exacerbate the water shortage in the near future.

This increasing shortage of irrigation water in terms of seasonal crop requirements clearly makes Pakistan's irrigated agriculture a potential beneficiary of irrigation scheduling. However, the institutional conditions that are required for a successful introduction of sophisticated scheduling methods become a critically important consideration.


In Pakistan, the warabandi2 water delivery schedules were designed to apportion the water shortage equitably, allowing the farmers to respond to the available water with appropriate cropping patterns and farming systems. It is essentially a supply-oriented delivery system, primarily designed to ensure equity in water distribution. However, a combination of institutional factors and their technical implications have contributed to make the flow conditions in the canal system increasingly variable. With the variable flow, warabandi schedules have ceased to become an equitable method of water distribution.

2 Warabandi is a rotational method for equitable distribution of the available water in an irrigation system by turns fixed according to a predetermined schedule specifying the day, time and duration of supply to each irrigator in proportion to the size of his landholding in the outlet command (Malhotra, 1982).

Variability caused by post-design structural changes

The water allowances for the older main canal systems in Pakistan are in the range of 0.20 to 0.30 l/s/ha, averaging at 0.23 l/s/ha. The secondary level distributaries also have a similar uniform pattern of water allowances. However, a contrasting situation exists in the tertiary level watercourses where the variation in water allowances appears to be very significant. In a study of a sample of six secondary canals, the water allowances assigned for individual watercourses along each of them showed an unexpectedly high variability (Table 1).

TABLE 1 - Variability of water allowances given to watercourses (n) in six selected distributary/minor canals


Range of water allowance (l/sec/ha)

Average water allowance (l/s/ha)

Coefficient of variation (%)



Mananwala (n=74)





Karkan (n=47)





Pir Mahal (n=47)





Junejwala (n= 19)





Azim (n=75)





Fordwah (n=87)





Source: Bandaragoda and Saeed ur Rehman (1995)

The above data seem to imply that a few outlets have been given exceptionally high water allowances. Considering the equity imperatives of warabandi theory, it is hard to believe that this much variation in water allowances for different watercourses within a given distributary was part of the original design. Such variation points towards the possibility of subsequent informal influence, what is commonly known is 'mogha tampering' (unauthorized modifications to the outlet or the mogha).

Table 2 gives the monthly average discharges into six selected distributaries/minors as measured once a day for the 1993 kharif season.

The warabandi theory assumes that each distributary canal, by and large, maintains a flow not less than 75% of the full supply level in order to feed all its outlets. Often, this technical requirement cannot be fulfilled due to the neglect of operational efficiency in the main system on the one hand, and many institutional aberrations downstream of the distributary head on the other. The six distributaries/minors remained fairly consistently below the design supply level. Monthly averages and their standard deviations show considerable variations in the actual discharge. The consequences of distributary water-flow variability for warabandi practice are two-fold. First, the flow variability during the season imposes severe inequity in water distribution within the watercourses, as the irrigation time allocation per hectare does not change according to this variable flow. Second, when the flow drops substantially, say below 70% of the design discharge, some watercourses receive very little water or no water all, causing inequitable water distribution among the watercourses. This is sometimes circumvented by rotations along the distributary, which again causes a disruption of the warabandi schedules. Also, variability at the distributary level can be a major reason for 'mogha tampering'. This was evident from the data on high variability of the discharges at the watercourse head.

Field studies also show that the officially sanctioned warabandis are not generally adhered to in practice, and are found to be superseded by 'agreed warabandi' schedules, which are substantial modifications of original schedules effected by farmers, supposedly by mutual agreement. The role of the influential farmers seem to be significant in effecting these modifications. Most of the official warabandi schedules have not been updated for a number of years despite the fact that the number of water users has increased substantially since the last official amendment. This delay itself could lead to unofficial modifications of the schedules by the water users themselves. As long as they are not disputed by an individual water user, or a group of water users, the procedure does not allow any official intervention. This situation explains the present high prevalence of 'agreed warabandi'.

TABLE 2 - Monthly averages of actual discharges during kharif 1993 for six sample secondary canals


Design discharge (m3/s)

Average actual discharge in m3/s and as % of design discharge



























Karkan Minor




















Pir Mahal




















Junejwala Minor




























































() gives the standard deviation for each average value.
Source: Bandaragoda and Saeed ur Rehman (1995)

TABLE 3 - Variability of water allocation in hours per hectare basis through warabandi turns (n) in Watercourse No. 89-L, Pir Mahal Distributary


Official Warabandi (n=36)

Agreed Warabandi (n=56)

Actual Warabandi (n=49)


0.42 - 0.93

0.47 - 3.29

0.65 - 6.30





Standard Deviation




Coefficient of Variation




(n = the number of water turns)

Field observations of the actual application of water turns by farmers show that even the 'agreed warabandi' is not strictly followed, and frequent changes take place on timing and duration of turns almost on a daily basis. While the reasons for introducing some flexibility in developing a more functional warabandi by mutual agreement can be easily understood, the divergence between the official warabandi schedules and what is actually practised in the field is unexpectedly large.

Data collected from official records, farmer interviews, and direct observations in the field showed that, for one selected watercourse (Watercourse No. 89-L, Pirmahal, at 89 250 feet on the left bank of the Pir Mahal Distributary), the official, agreed and actual warabandi schedules are significantly different from one another (Table 3). The changes made in the official warabandi have generally resulted in an increase in the average irrigation time per unit of land (which can be defined as 'water allocation'1), and an increase in inequity.

1 The "water allocation" is defined, for the purpose of this presentation, as the irrigation time per unit of land on the basis of a constant discharge to the watercourse. The water allocation in the warabandi system is usually understood in hours per acre. It varies from one watercourse to the other, depending on the command area to be irrigated. For equitable distribution, the water allocation measure should not vary too widely among the different farm plots in a given watercourse.

Attempts to increase the average time allocation means a major disruption in the warabandi rotation schedule and a situation in which some farm plots do not receive any water at all. There is no evidence to indicate a group decision by the farmers to deliberately irrigate a smaller area. The major institutional issue in this instance relates to the subversion of warabandi's primary objective of equity, which can be attributed to the combined effect of a value change and a general breakdown in the overall institutional accountability for the management of irrigation systems.

In the context of several such changes in the application of warabandi, it is the technical deficiency related to underirrigation through warabandi that attracted policy attention. In the North West Frontier Province of Pakistan, two canal irrigation projects incorporated higher water allowances in their design for modernized irrigation practices, with the expectation that some form of crop-based water delivery scheduling could be achieved.


The case of the Lower Swat Canal (LSC)

The LSC is a recently remodelled irrigation system with a much higher water allowance relative to its original design. Before rehabilitation, the average irrigation allowance of the LSC was about 0.43 l/s/ha, already in the higher range when compared with most of the canal systems in Pakistan. By the early 1970s, the LSC system, like many others elsewhere in Pakistan, was experiencing a shortage of water for its increased cropping intensity in the command area, and being independent of the Indus River system, was considered for the needed capacity enhancement. The remodelling of the system was completed in 1993, with an increased capacity to carry a water flow of 0.77 l/s/ha for peak requirements.

This is the first attempt in Pakistan's long history of irrigation to incorporate into a major system design the possibility of some form of demand-based irrigation schedule. However, the ambiguity at the planning stage in the timing of this shift from the traditional supply-oriented operations to a demand type greatly contributed to the uncertainty associated with the post-construction implementation problems. A rapid appraisal (Bandaragoda et al., 1993) conducted on two selected secondary canals in the LSC system in 1992/93 presented the following findings:

(1) The average ten-day flow data at the head of the Lower Swat Canal kept by the Irrigation Department for the period 1988-1992 showed that the supplies had gradually increased during the period. In 1992 the supplies at the head of LSC were about 60% more than those in 1988. This increase in supplies did not reflect the full supply level, as in 1992 the construction of the aqueduct in the main canal was still in progress and was completed only in late 1992.

(2) There was not only inequity, but also some inconsistent trend in the outlet discharge in relation to the inflows of the channels. For example, a tail-end outlet (No.15358-R) on Sheikh Yousaf Minor could draw up to three times of its design discharge, as against the performance of many upstream outlets, indicating the possibility of unauthorized intervention.

(3) There is strong evidence that the farmers are responding to increased supplies by trying to increase cropping intensity, as well as shifting to cash crops. The cropping pattern had changed substantially from 1980 to 1991.

(4) Farmer interviews and field observations indicated that the warabandi system had collapsed after the project as a result of increased water supplies.

(5) The new technology installed in the system for water delivery scheduling, including the gates at the watercourse head, has been rejected by both the farmers and the operating agency, as together, they could not cope with its operational responsibilities.

With the substantially increased water allowance (about 0.77 l/s/ha), the system is now being operated in the traditional supply-oriented management mode, causing concerns over problems of waterlogging despite the investment in subsurface drainage in the area, and the problems of water wastage.

The case of the Chashma Right Bank Canal (CRBC)

As a major deviation from the average type of canal in Pakistan, the CRBC was designed with a water allowance fixed at 0.60 l/s/ha to meet peak crop water requirements for a projected annual cropping intensity of 150% (90% in rabi and 60% in kharif). To increase the flexibility of water delivery, the CRBC was also designed to operate over a range of discharges at the Chashma Barrage, with the intent of using 'crop-based irrigation operations', in which water was to be supplied more in response to crop water requirements than was done in the traditional supply-oriented rotation system.

A study conducted in two distributaries of CRBC Stage I revealed that there was a considerable difference between planned and actual operations in the CRBC system (Garces et al., 1994). The project design was based on an assumed cropping pattern and cropping intensity that would just meet the maximum permitted diversion from the Chashma barrage. In periods when demand was lower than this maximum demand, there would be commensurate reductions in discharge into the main canal and its distributary channels. Yet, field observations during the study indicated significant deviations from these initial assumptions. The most important of these findings are:

(1) A much larger area is under rice, averaging around 20% of total area and ranging from 5-50% in sample watercourses, compared with an initial design estimate of 2% rice; the area under rice is steadily increasing.

(2) There is no evidence of systematic variations being made in discharges into the main canal at the Chashma barrage or at distributary head gates in response to the changing crop demands during the year; there is limited capacity in the operating agencies to undertake discharge measurement at key locations, or to obtain real-time information on crop patterns and changes in demand, that would permit a more systematic response to crop water demand. This operational lacuna results in extended periods of excess water deliveries along distributaries, as well as large volumes escaped into the drain at the end of Stage I.

(3) The observed farmers' behaviour in closing and opening the outlets depending on their requirements indicates that the system is delivering excess water; and that farmers are very adept at assessing crop water requirements.

The consequences of this situation are that the area under rice cultivation is far too high and yet increasing, and sugar cane extent is also likely to increase dramatically once sugar processing facilities are completed. Moreover, groundwater levels in Stage I and even in some portions of Stage II are rising rapidly and will soon pose a serious threat to the concept of diversified cropping. If these trends continue, there will not only be difficulties in sustaining the productive environment in the upper portions of Chashma Right Bank, but also tail-end portions of Stage III may not receive adequate water once the physical infrastructure is complete. This means that the attempt at changing the design to achieve water saving through irrigation scheduling has failed.


For over a century, the large canal irrigation systems in Pakistan functioned in a supply-driven mode using the warabandi allocation method. Considering the social tension that is normally associated with the shortages of irrigation water, the continued imposition of scarcity through warabandi over such large areas, and for such a long time, has been considered 'little short of a miracle' (Malhotra et al., 1984). The 'image' of equity in warabandi, partly the result of its transparent application within the watercourse, appears to have enabled warabandi's sustainability. Since a careful comparison of the discharges into different outlets along a given distributary is not an easy task for farmers, the inequity built into the water allowances normally escapes their attention. However, the water flow variability in the canal system is an issue they can easily observe, and therefore, increasing inequity in water distribution cannot pass unnoticed by the water users. The tolerance of this situation can be attributed to an equilibrium that may exist in the prevailing socio-economic forces, which can bear the present levels of inequity. It is unlikely that this socio-economic situation would allow the majority of water users to benefit from a more complex and flexible irrigation scheduling.

Field interviews indicate that during early days of irrigation development, the operating agencies, the water users and policy makers were collectively geared towards the objective of equitable and productive water delivery and use. A strong sense of accountability was the basis for such collective behaviour. Effective rule application and adjudication systems resulted in a good law and order situation. As long as these institutional conditions existed, the large irrigation systems were functional. With a gradual erosion of these conditions, the systems started to become dysfunctional, and a vicious circle set in with the physical sub-system and the human sub-system acting on each other towards a rapid performance decline.

To arrest this trend towards low performance, there have been attempts to inject the ailing systems with technological inputs, but the technical solutions became increasingly ineffective when confronted by low accountability and policy inertia regarding institutional reform. In this region, the irrigation institutions have a great deal of 'staying power' (Merrey, 1993). This has resulted from a complex process of formal and informal interactions among technology, foreign aid, social environment and governance mechanisms. In this complex process, a few resourceful individuals play a larger role, and gain a larger share of the benefits of any innovation.

Where the innovations favour a more broad-based gain, they usually fail, as in the cases of experimentation in the CRBC and the LSC systems.

The cardinal mistake of the planners of the CRBC and LSC irrigation systems was in their failure to anticipate the required institutional changes for coping with the new irrigation management technologies they wished to introduce. The LSC design in particular was based on some broad assumptions for a shift from the traditional irrigation operations.

One assumption related to the present institutional capacity, or even adaptability, to undertake an immediate shift to crop-based irrigation scheduling. The demands of the individual farmers were to be articulated by the 'Water Users Association' (which was non-existent), in the form of detailed 'demand water orders' specifying time and place for delivery. While there was no plan to promote social organization for this purpose, the intentions were not clear as to how the agency staff and farmers would collaborate to determine the periodic water requirements. A second assumption was that the users were flexible in their daily or weekly irrigation practices, and in their related habits and customs. It was assumed that the farmers would quickly adjust to a pattern of variable supply with irrigation scheduling and to a system under those conditions where the demands of all the irrigators on the watercourse would be met. The third assumption was that the social background in the LSC area was harmonious or at least self-sustaining in dispute resolution. There was no mention of how the adjustments were to be made in the individual demands to make them realistic or about the procedures to be followed to resolve competing demands. Thus, the increased intensity of physical infrastructure in the remodelled LSC system, and the operational intentions of the CRBC, were not matched with an enhancement of the required institutional framework for its operation and maintenance. The main reason for this unfortunate missed opportunity to try out something new and useful can be attributed to the uncertainty during planning stages in making the institutional change a specified project objective.

Given the constraints of deep-rooted social preferences, a substantial change from present patterns of behaviour occurring so quickly can be a grossly overestimated expectation. A truly dedicated effort in social organization, persistent over an adequate period of time, seems essential to overcome this constraint. While there is an urgent need to capitalize on the investment already made, a move forward under the new approaches would not be possible without some attempts at institutional change as well.


Lessons from Pakistan's experience suggest that the introduction of irrigation scheduling in this region will have to be preceded by a change in the attitudes, policies, awareness and the institutional framework for irrigation management. This change will not occur easily or quickly. Above all, a strong vested interest among influential individuals to keep the existing institutional arrangements intact seems to be the greatest obstacle to change.

Nonetheless, alternative strategies need to be developed for introducing new technologies to replace or modify the traditional water delivery systems which are crumbling under the stress of changing socio-economic conditions. However, these strategies should ensure that the benefits of new technologies of water delivery and irrigation scheduling reach the more disadvantaged groups among the water users. Encouragingly, there are some efforts in both policy and research currently under way in Pakistan to develop viable institutional strategies for more productive and sustainable irrigated agriculture.


The author wishes to thank Professor Gaylord V. Skogerboe, Director IIMI Pakistan, for the valuable suggestions given in the preparation of this paper, and Dr. Chris Perry and Douglas J. Merrey, Program Leaders of IIMI, for their comments on the earlier drafts of the paper.


Bandaragoda, D.J., Badruddin, M., Afaq, R.M. and Rehman, S. 1993. Institutional and management implications of system modernization: The rehabilitation of the lower swat canal irrigation system. Paper presented at the Internal Program Review. International Irrigation Management Institute. Colombo. Sri Lanka.

Bandaragoda, D.J. and Rehman, S. 1995. Warabandi in Pakistan's canal irrigation systems: widening gap between theory and practice. Forthcoming IIMI Country Paper - Pakistan No. 7. Colombo, Sri Lanka: International Irrigation Management Institute.

Garces, C., Bandaragoda, D.J. and Strosser, P. 1994. Final Report on Crop-based Irrigation Operations Study in the North West Frontier Province of Pakistan. Volume I: Synthesis of Findings and Recommendations. International Irrigation Management Institute. Lahore. Pakistan.

Kijne, J.W and Kuper, M. 1995. Salinity and sodicity: a threat to sustainability of irrigated agriculture. J. Water Resources Development March 1995.

Malhotra, S.P., Raheja, S.K. and Seckler, D. 1984. A performance monitoring study of the warabandi system of irrigation management. In: Productivity and Equity in Irrigation Systems. Niranjan Pant (ed.). Delhi: Ashish Publishing House.

Merrey, D.J. 1993. Institutional context for managing irrigated agriculture. Paper presented at the DSE/IIMI Strategy Workshop on Institutional Framework Irrigation, 1-5 November 1993, Chiang Mai, Thailand.

Ministry of Food and Agriculture, Government of Pakistan. 1988. Report of the National Commission on Agriculture. Pakistan.

WSIP Study. 1990. Water Sector Investment Plan (1990-2000). Lahore: Federal Planning Cell.

Previous Page Top of Page Next Page