|Ecology in Development: A Rationale for Three-dimensional Policy (UNU, 1984, 59 pages)|
|I. Irrigation in South-West Asia|
The environmental change caused by perennial irrigation in dry lands is spectacular. Both the change and the enormous investment it requires is more than justified by the similarly spectacular increase in productivity. In Pakistan today irrigation provides the major component of agricultural production, which is over a third of the country's Gross National Product. But the tremendous potential of the (natural) resource and the technology (on the drawing board) is frustrated in the application by ecology.
The growth of the ecological problem makes better sense when subjected to historical review. But the development record does not appear to have benefitted from such a review. The lessons to be learned from such a review become plainer when seen from the point of view of the individual farmer. The implications for policy are clear, but may be difficult to follow because we are so unused to integrating socio-cultural with economic, technological and ecological information.
Let us begin then with a brief description of the resource and the technology. The facts and figures are taken from Michel (1967) and the Water Management Technical Reports on Pakistan produced by the Consortium for International Development, especially Corey and Clima (1975), Eckert et al. (1975), Mirza (1975), and Radosevich and Kirkwood (1975). The cultural material is taken mainly from Merrey (1982).
The resource and the technology
The average total annual flow of the Indus River system in India and Pakistan is twice that of the Nile and 10 times that of the Colorado. It exceeds 170 billion cubic metres but, as is typical of dry lands, there is great seasonal variation in flow. Between November and February, the flow averages only one tenth of what is normal for the summer monsoon months. More direct rainfall adds an average 7.5 billion cubic metres to this resource each year and it is estimated that some 54 billion cubic metres can be taken annually from groundwater. Twenty four and a half billion of these are presently accessible, but because of greater salinity they must be diluted by mixing with river water before use. Generally the quality of river water is good, with 1 to 300 parts per million of dissolved solids. The whole system commands 15.5 million hectares in which soil quality varies from moderately fine and deep alluvial deposits in the flood plains to coarser deposits on the higher ground (bar) between the rivers. As is to be expected in dry lands, there is a general lack of organic material.
In 1947 the Partition of India and Pakistan caused a de facto division of this resource, which was formalized in the Indus Waters Treaty (1960) by which Pakistan received exclusive use of the Indus itself, plus the Jhelum and Chenab tributaries, leaving the Beas, Ravi and Sutlej (amounting to a little over twenty percent of the total average annual flow) to India.
The irrigation system that has been developed over the' last century and a quarter now diverts within Pakistan approximately 123 billion cubic metres of annual river flow and spreads it over 13.5 million hectares of cultivable land, of which nearly 9 million hectares can be irrigated throughout the year.
This controlled distribution is accomplished by means of 17 barrages and canal diversion works, 42 major canals, 6,000 kilometres of minor canals, 600 kilometres of link canals, and 78,000 watercourses. The total capacity is nearly 7,000 cubic metres per second, or 250,000 cubic feet per second (cusecs) as it is commonly measured. This flow is supplemented from 156,000 tube wells which raise 24.5 billion cubic metres from the subsurface water table. The overall pattern of flow is from one of the major rivers to major and minor canals through outlets (moghas) to watercourses (khals) to farmers'fields. What is not consumed as it passes through the system is either returned to the rivers or disposed of in some other (often more costly) manner or accumulates, resulting in waterlogging and salinity. Since 1955 a large network of surface drains has been created as part of a programme for the solution of this problem.
The purpose of this technology is to control the spatial and temporal flow of all the available water over the greatest area of cultivable land in order to achieve maximum distribution and optimum quantity and speed of flow. The speed must be slow enough to minimize erosion of the bed and banks of the canals. While the horizontal movement of the water within the system is controlled relatively efficiently, vertical movement out of the system has proved more difficult to manage. The system is vulnerable to seepage and evaporation. By spreading surface water over a much larger area or "command" than it would naturally cover between two points in a stream channel, and by causing it to spend more time in the commanded area than it would spend in the channel, any irrigation tends to increase the amount of recharge to the water table. Once the water has passed both below the root zone of the crops and below the level (approximately 3 metres in the sandy loams that predominate in the Punjab) from which capillary action can raise it to the root zone, it becomes valueless, unless it can be pumped out again.
Excessive recharge causes the water table to rise. If it rises to the level where it interferes with plant growth by waterlogging, or if capillary action combined with evaporation increases salt accumulation in the upper soil horizons or on the surface, productivity is reduced.
The interrelation of efficient control of water with crop requirements demands not only complex engineering but sophisticated organization of labour. Altogether this form of irrigation is at once the most large-scale, most investment-intensive, and most economically significant technology of food production in human history. Both as an economic or social and as an ecological or natural system, it is qualitatively different from what preceded it. It might be expected, therefore, that development would require complete reorganization of the human population that works it. In fact, however, there appears from the beginning to have been a conscious policy on the part of the developers not to interfere with local practice. Formal irrigation administration as it has evolved from the beginning reaches down to the level of the canals and their outlets only. Lower level officials report water flow, regulate distribution gates, and organize maintenance work. There are now also tube-well operators. But from the canal outlets onwards the farmer has always been left to his own devices. He has had to align, dig and maintain his own watercourse and develop a rotating water delivery or distribution system in cooperation with his neighbors, without the benefit of any outside assistance or advice. There is often only one outlet per village and no congruence of watercourse-sharing communities with other social or spatial groupings of the population. The development of this resource and of the technology to exploit it has not been complemented by attention to the human resource without which it cannot be exploited, let alone by a comprehensive plan to manage both the physical and the human resources as a means to improving the wellbeing of the society.
The ecological problem
The greatly increased level of productivity per hectare is not sustained. As a result, increase in gross sown area cannot keep pace with population growth. In fact, even by the time of Partition the Punjab had ceased to produce any substantial grain exports. Even though Pakistan inherited almost all the surplus-producing irrigated areas, the combination of population growth and ecological damage quickly - by the mid-1950s - made her a net importer of her major crop and food staple, wheat.
The principal problem arises from the loss of cultivable land through waterlogging and salinity as a result of seepage, poor maintenance of watercourses, and inefficient application of water to crops. About half of the total irrigated land is estimated to be affected to varying degrees. Until recently the process was counteracted only by bringing more land under cultivation. Paradoxically, a subsidiary but related problem caused by water loss through these same processes and through evaporation, is lack of water.
Before the development of the system, water-table depths over most of the area now irrigated were about 24 to 28 metres. Historical data indicate that the water table has risen an average of 15 to 35 centimetres per year since modern irrigation was introduced. Of the 123 billion cubic metres diverted annually, only about 71.5 billion cubic metres reach the heads of watercourses. it has been estimated that from 5 per cent to as much as 65 per cent per mile is lost in the watercourses. Altogether, less than 30 per cent of the water diverted from the rivers gets to the root zones of crops and is consumed. Further, a salinity of 1,000 parts per million is acceptable for virtually all crops, but groundwater of that quality which evaporates at a rate of half a metre per year, a typical rate where the water table is less than a metre deep, will in 20 years raise the salt content of the top metre of soil to about 1 per cent, which is too high for even the hardiest crops. Not only, therefore, are environmental problems causing loss of cultivable land, but the irrigation system is working at only 30 per cent efficiency, and this inefficiency is responsible for the disastrous loss of both land and water. These processes are almost certainly exacerbated by inefficiencies in actual cultivation. But these inefficiencies are disguised, because of the administrative segregation of irrigation from agriculture, which is characteristic of the way bureaucratic systems evolve (Cf. Spooner 1982a & b). In the 1960s it was estimated that between 20,000 and 40,000 additional hectares were being affected each year and, in the worst districts, 40 to 50 per cent of the cultivated land was already severely damaged.
Unfortunately, there is no exact method of quantifying waterlogging and salinity damage. Actual conditions vary from season to season, and year to year, depending partly on the strength of the monsoon and partly on other factors such as spatial variation in rainfall, groundwater recharge, and evaporation. Actual crop damage varies according to the sensitivity of the particular plant. Surface salinity has been compared to skin rash appearing in blotches which vary continually in intensity and extent.
Finally, it is possible that the ecological problem is caused by inefficient practice of the technology. Perhaps highlytrained farmers would be able to apply it without allowing excessive recharge to the water table. In this case the fault lies entirely in the neglect of the human component in the planning process. On the other hand, any such possibility remains to be proven. Perhaps the technology is deficient, and the fault lies entirely with the engineers and investors! The likelihood is, of course, that there are inadequacies in the technology and inefficiencies in the application, as well as incongruencies between the requirements of the technology and the perceptions of the farmers. But we do not know. Our knowledge is still almost exclusively technological and ecological. And what we do know about the people who use the technology has not been systematically related to their use of the technology. We tend habitually to keep our knowledge of resources and technology categorically separate from our knowledge of the way people behave and think.
The history of the problem
Before the middle of the 19th century, irrigation was confined to parts of the flood plains and was mostly seasonal. Pastoralism was the major form of land use, but was supplemented here and there by dry farming. Water was drawn during the summer when the rivers rose above the levels of canal inlets, and was used to irrigate lands which would not have received water by natural flooding. Such canals were, however, uncontrolled and did not allow exploitation of low river flows. Only relatively narrow strips of land along the rivers could be irrigated. The supply channels were inefficient: they depended on uncertain river flows and tended to silt up. There were also dangerous breaches during the flood season. In spite of these shortcomings, inundation canals constituted an important advance in the technology of irrigation. The system was subsequently improved during the Mughal period, especially in the 17th century, to the extent that limited perennial irrigation was possible in parks and gardens.
The traditional systems were designed to spread the water over as large an area as possible during the period of maximum flow. Limited engineering works maintained a constant level of water suitable to the level of the land to be irrigated, "heading up" the flow of water and distributing it though a system of canals. The modern system, construction of which began in 1851 and has developed steadily ever since, is designed for continuous control.
The motivation of the colonial government in embarking on this vast and innovative engineering scheme is in itself instructive. Like more recent motivations for further development of it, and for the development of similar projects elsewhere (for example, in Iraq, Afghanistan, Egypt and Soviet Central Asia), it was at least as much political as economic. The desire to appear to have improved on the engineering of earlier regimes provided the general motivation, while the immediate need to ensure against the threat of famine and to settle the recently-disbanded Sikh levies, were the specific motives. The guiding principle was uncautious optimism rather than careful research and planning, despite the complete lack of relevant experience. In addition, the colonial administrators wanted to expand the area under irrigation and bring new lands into cultivation so that they could be settled and taxed. Ancillary motives included re-settlement and relief of crowded conditions elsewhere,. the creation of a granary which could supply the famine-prone areas of north central India and, later, especially in Sind, creation of new areas for cotton production. The optimism inherent in these motivations hindered clear perception of the environmental problems that soon developed.
It is important to note the role played by perception. For decades there seems to have been a general tendency to ignore or misinterpret what now (with the benefit of hindsight) appears to have been obvious contrary evidence about the success of the technology. For example, on the Western Jumna Canal of the Ganges Basin (where the development of irrigation had begun, in what is now India, with the provision of permanent headworks in 1836), waterlogging and salinity problems had already appeared by 1859. Between 1870 and 1880, the irrigation channels were re-aligned and natural drainages cleared, with results which were encouraging but did not lead to a general policy for dealing with what was already becoming a general problem. When the Lower Chenab Canal, which opened in 1892, had produced serious waterlogging by 1908, some maintained the cause lay not in irrigation but in the fact that the canal, road and railway embankment were interfering with surface run-off, or even that the Punjab was in a rainy cycle. Others maintained that a high water table was actually an advantage, because it facilitated the operation of hundreds of Persian wheels in shallow wells and produced some regeneration of water supplies by seepage during the dry season. A Waterlogging Enquiry Committee was finally established in 1925, but still there was more interest in extending the system onto new lands than in reclamation, despite the growing awareness that the cost of bringing water onto new lands was increasing, the new lands had much coarser soils and lower initial fertility, their seepage and evaporation rates were higher, and even the Indus Basin would eventually run out of new land to replace the old in any gravity-fed surface-water irrigation system.
Partly as a consequence of the non-ecological motives and the unrealistic perception of the situation that they engendered, the cropping pattern was dominated by wheat which was the staple grain, and cotton, the obvious cash crop, which both made it feasible to spread the water thinly. However, even these crops, which have low water requirements, received much less than the optimum, and although sugar cane, which requires more water, was allowed to a limited extent, rice cultivation which would tolerate higher accumulations of salt as well as using more water, was generally discouraged, at least until the water table had risen close to the surface.
It is clear that even apart from the neglect of the social dimension there was a significant degree of wrong-headedness in the history of irrigation planning in the Punjab, and it is not surprising that it should take considerable time and intellectual pain to rethink it.
The development record
Since the 1950s efforts have been made to reduce evaporation and seepage but only to the extent that the cost appeared economically justifiable in the context of the perception of the problem. Apart from the policy of spreading the water thin, which had always carried an economic rationale, canals were re-aligned in badly leaking places. Some canals were lined and some surface drains were reconstructed. These practices have been continued up to the present. They have included no social component, and have had little impact on the general problem.
The most promising technical attack on the problem was a type of comprehensive control strategy. However, although such a strategy was proposed as early as 1927, it was not approved until 1944 and not put into full operation until 1952. In an effort to both lower local water tables and provide additional supplies of irrigation water, 21,257 tubewells were sunk along badly seeping canals in two of the interfluves. Although this scheme (known as the Rasul Scheme) was not particularly successful, because most of the wells were too close to the canals and actually accelerated seepage, it did lead to a better understanding of the problem. It was followed by a second similar scheme in 1953 - 4, and a third in 1957- 8. Finally, when the Water and Power Development Authority (WAPDA) was established in 1958, it was specifically entrusted with "prevention of waterlogging and salinity and reclamation of waterlogged and saline lands." At last the problem had been officially recognized, but the diagnosis was exclusively technological.
WAPDA's Salinity Control and Reclamation Projects (SCARP) have steadily increased the number of tube wells ever since. The factor which differentiates these projects from their predecessors lies in the concentration of these tube wells in fields of from 1,500 to 3,000 units, each with three to four cusecs (85-110 litres per second) capacity, and each serving approximately 250 ha. The capacity and spacing of the wells is designed to allow full control of the drainage in each project area.
Combined with supplies from an even greater number of privately owned one-cusec wells, and the enhanced surface water supplies made possible by the newly constructed Mangla and Tarbela dams on the Jhelum and Indus Rivers, the amount of water available for irrigation in Pakistan is now estimated at over 100 million acre feet or 1,233,438 million cubic metres. Two-thirds of this supply is from the surface water storage and distribution system, and almost one fourth from the government-owned tube wells. The total supply represents a substantial improvement over the 68 million acre feet or 838,772 million cubic metres available in 1965 and thus enhances significantly the capability for efficient irrigation of crops and for leaching of salts from the top soil.
This increased amount of water spread on the surface, however, would serve only to increase the waterlogging and salinity damage to soils and crops - except that the massive concentration of high capacity tube wells offers the hope of controlling the level of the water table. But a further problem is the quality of the ground water. Wherever the ground water is of usable quality (up to roughly 2,000 parts per million of total dissolved solids, depending on the chemical composition of the salts), its use for crops should produce a net gain, and through consumption and evapo-transpiration result in a gradual lowering of the water table. In other areas, saline ground water must be mixed with surface water of good quality before being applied to crops. To accomplish this mixing, canal capacities in certain areas need to be enlarged. In some areas, the ground water has proved too saline even for blending and must, therefore, be exported, either by the rivers (which will cause problems downstream), or via new wasteways constructed for this purpose.
Although the technical problems of the SCARPs have been overcome, and the projects have caused a significant improvement in the situation, they are nevertheless still inadequate. For example, in SCARP no. 1, which began in 1962, the water table has declined to an average of two to three metres. below the surface, and about 45 per cent of the affected area was reclaimed in the first nine years. Subsequently, however, progress was rather slow - a development which has been attributed to the sodicity of the soils. Generally, yields have improved as a result of land drainage, reclamation of considerable areas, and increase of water supplies from tube wells, together with additional agricultural inputs, such as fertilizers. In one experimental project area the gross value of agricultural produce - both crops and livestock increased by a factor of 2.5, but deterioration in groundwater quality is causing adverse changes in chemical characteristics of the soils and decline in the yield of sensitive crops.
Pakistan's groundwater and reclamation programme represents an extremely complex and costly effort to offset the consequences of inefficient surface-water
irrigation. For the periods of Pakistan's third and fourth fiveyear plans (1965-1975), the total cost of government owned tube wells, canal remodelling and drainage works (not including surface water storage) was set at about US $1,100 million, or slightly more than the cost of the Tarbeia Dam which itself represents roughly half the total cost of the Indus Basin project. It was expected that the gains achieved in Pakistan's agricultural sector, which grew at a healthy rate of 3-4 per cent per annum between 1960 and the early 1970s (though most of the gain was due to nonfood crops), would eventually more than compensate for these investments. But it was understood that these gains would depend not only on increased surface water and ground water supplies, but on further input of fertilizers, improved seed varieties, insecticides, and pesticides, and improved techniques of irrigation and cultivation. Generally, it was recognized that although reclamation programmes must be continued, the best hope for future progress lay rather in prevention. It remained, however, to develop a clear strategy for prevention.