World Development Vol. 29, No. 12, pp. 2025±2041, 2001 Ó 2001 Elsevier Science Ltd. All rights reserved Printed in Great Britain www.elsevier.com/locate/worlddev 0305-750X/01/$ - see front matter PII: S0305-750X01)00087-0 The Manufacture of Popular Perceptions of Scarcity: Dams and Water-Related Narratives in ,

LYLA MEHTA * Institute of Development Studies, University of Sussex, UK Summary. Ð This paper critically examines some narratives of water scarcity in Kutch, western India. It argues that images of dwindling rainfall and increasing drought largely serve to legitimize the controversial Sardar Sarovar dam and manufacture dominant perceptions concerning scarcity. This manufacture has naturalized scarcity in the region and largely bene®ts powerful actors such as politicians, industrialists and large farmers. But the needs of the poor in water-limited areas are neglected. By exploring the various connotations of scarcity, the paper argues that scarcity is both a biophysical phenomenon as well as a powerful discursive construct. By distinguishing between the ``real'' and ``manufactured'' aspects of water scarcity, the paper attempts to enhance understand- ings of environmental change at the local level. Ó 2001 Elsevier Science Ltd. All rights reserved.

Key words Ð Asia, India, water scarcity, dams, narratives of scarcity

1. INTRODUCTION the panacea for water scarcity. The proponents of large dams tend to downplay the social and Water scarcity is one of the most pressing environmental costs of large dams against the problems confronting the well-being of bene®ts of hydropower and irrigation 9Biswas humankind in the 21st century 9Ohlsson, 1995; & El-Habr, 1993; British Dam Society, 1999). Postel, 1994; The World Bank, 1993). Scarcity These views are increasingly contested by a of water and drought are complex phenomena worldwide constituency comprising academics, that can be analyzed di€erently from social, scientists and members of voluntary agencies political, meteorological, hydrological and who have highlighted the problems of invol- agricultural perspectives. But technical and popular understandings of water scarcity have tended to be simplistic 9Falkenmark & Chap- * Earlier versions of this paper were presented at the man, 1989). There has been the tendency to annual conference of the Development Studies Associ- direct attention to the lack of supply of water ation, 1999 and at a seminar at the Department of Water due to natural forces rather than look at and Environmental Studies, University of Linkoping. I human-induced land and water use practices am grateful to the participants for their comments. Peter and at socio-political considerations. Real Davis provided encouragement and useful comments. causes of scarcity can be obscured leading to This paper draws from my doctoral thesis 9Mehta, inappropriate solutions. This paper explicitly 1998). Research was conducted in Kutch in 1995±96 for addresses the complexities surrounding water which funding was provided by the Overseas Develop- scarcity by taking the case of Kutch, a semi- ment Institute, London. I am grateful to Annette arid to arid region in western India. According Sinclair for helping in the data analysis and presenta- to Government of Gujarat documents, Kutch is tion, Dr Singh of the Institute of Desert Ecology for designated to bene®t from the controversial providing rainfall data in Kutch, P. Kaul for suggesting Sardar Sarovar Project 9SSP), a controversial Iread Herman and Chomsky 91994) and Rohan dam under construction in western India D'Souza and Sinnet Weber for their suggestions. Most 9Government of Gujarat, 1991; Raj, 1991). of all, Ithank my interview partners and the people of There are several polarized views regarding Kutch, in particular, those of the village, Merka, where I how the water needs for present and future conducted my detailed ®eldwork. Iwill always be generations of a country and region can be met grateful to them for their co-operation, understanding, 9cf. Mehta, 1997). Until a few decades ago, the warmth and inspiration. Final revision accepted: 23 July 1 large dam was universally considered to be 2001. 2025 2026 WORLD DEVELOPMENT untary resettlement and environmental damage social organization and questions concerning due to large dams 9e.g., Goldsmith & Hildyard, power 9cf. Bruns & Meizen-Dick, 2000; Mosse, 1992; McCully, 1996). The World Commission 1997; Potanski & Adams, 1998). By building on on Dams has recently completed its mandate to precedents set in work on environmental investigate the myriad aspects of dams narratives by authors such as Leach and concerning economic growth, equity, environ- Mearns 91996) and Roe 91991), the paper mental conservation and participation. It focuses on narratives connected with dwindling concluded that while dams have made a rainfall and increasing droughts which are considerable contribution to human develop- widespread in the district and the state. The ment, in too many cases unacceptable costs critical examination of these narratives leads to have been borne in social and environmental the conclusion that while water scarcity is a terms. The Commission also argues that often ``real'' enough problem with biophysical water and energy needs can be met through manifestations, it can also be ``manufactured'' alternative solutions that would fare better than in such a way to serve the interests of powerful dams on equity grounds 9WCD, 2000). actors such as politicians, bureaucrats and These views are not shared, however, by a irrigation farmers. Popular perceptions of large number of agencies and governments, scarcity, as represented in the mass media and especially in the South. Judging from the severe by politicians and advocates of the water reservations expressed by the Indian govern- question, have naturalized scarcity in Kutch. ment, the International Commission on Large They have also succeeded in manufacturing the Dams and the International Commission on dominant perception of water in the region, Irrigation and Drainage to the Commission's namely that there is no alternative to the Sardar Report 9Bhogale, 2001; , Sarovar Project 9SSP). Due to this ``manufac- 2001), it is clear that there are still very polarized ture'' which largely bene®ts powerful actors views on dams and the role that they play in such as politicians and large irrigators, mitigating water scarcity. Is this because large controversial schemes such as the SSP are dams are urgently required to solve the prob- legitimized. The paper also distinguishes lems in water-needy areas or is it because of between the physical phenomena of scarcity questions concerning a wider political econ- and their social explanations and manifesta- omy? Is there a need to investigate the rela- tions. The distinction between the ``real'' and tionship between discourses of water scarcity ``manufactured'' aspects of water scarcity, I and the vested interests in large-scale develop- argue, is a useful way to understand environ- ment projects? The paper demonstrates that mental change both at the level of discourse as that the pro and anti-dam dichotomy is largely well as at the biophysical realm. counterproductive for regions such as Kutch, Sections 2 and 3 present an overview of which are often a€ected by severe conditions of Kutch and its relationship with the SSP. aridity. Instead the paper suggests that concer- Section 4 examines the water-related narratives ted e€orts toward creating sustainable land and concerning dwindling rainfall and climatic water use practices might be the more desirable change. The paper then goes on to argue that alternative to waiting for a little water from a these popular perceptions feed into the controversial and distant dam. The paper draws perception of the SSP as a panacea and the only on a year's ®eldwork in Kutch and Gujarat in solution for Kutch. In the conclusion, the paper 1995±96. The research employed a combination argues it is not merely enough to discredit of methods such as participant observation at popular narratives of environmental change by the village level, semi-structured interviews with creating new counternarratives. Instead, the a host of rural and urban actors, the analysis of analysis of environmental change must capture rainfall data, discourse analysis and historical both discursive as well as material realities. The and archival research. Several months were paper attempts to achieve this by examining spent in a village Icall Merka, in eastern Kutch both the ``manufactured'' and ``real'' aspects of where diverse water and land-use practices were scarcity and the nature of their relationship. studied 9see Mehta, 1997). The paper follows a growing number of authors who argue for the need to see water 2. KUTCH AND ITS WATER RESOURCES resources as having symbolic as well as material characteristics. Such an approach calls for the is located in a crescent-shaped need to link water resources management with peninsula in the Gujarat State of Western MANUFACTURE OF POPULAR PERCEPTIONS OF SCARCITY 2027

India. It is the largest district in Gujarat and by the district level government. Many existing has an area of 45,612 sq. km constituting 23% schemes work under their potential capacity of the state. Kutch is like an island bound by due to a high rate of siltation 9between 15% and the sea in the South and West and by the Ranns 70%). 5 As the de-silting process is very costly, 9salt marshlands) in the East and North. Kutch it is cheaper, but less ecient to construct new was a ruled by the Maharao of dams downstream. The reservoirs of Kutch Kutch and was integrated into the state of also have a lower over¯ow rate than in other Gujarat only in 1960. parts of Gujarat 9Gujarat Institute of Area Kutch has a population of 1.2 million. It is Planning, 1989, p. 15). Due to the variable far more sparsely populated than the rest of rainfall, there is often not enough water for India and has witnessed a very marginal release in the canal networks, and therefore, the increase in the population in the ®rst few irrigation schemes of Kutch utilize only 28% of decades of the last century. Rapid population their entire potential. increase has, however, occurred in the decades The intensity of well irrigation is higher than since its integration into the Indian union that of canal irrigation because it is less depen- 9about 122%) 9see Mehta, 1998 for details on dent on the vagaries of the rains. But, the population dynamics and environmental recharge rate of the total precipitation is only 8% change). It has nine talukas or administrative in Kutch as opposed to the usual 20% 9Gujarat subdistricts: and Nakhatrana in the Ecology Commission, 1994, p. 34). These natu- North; Lakhpat and Abrasa in the West; ral constraints are intensi®ed by the uncon- , and Anjar in the South and trolled extraction and overexploitation of Bachau and Rapar in the East. Apart from its aquifers, not least due to the skewed pricing of very heterogeneous social and ethnic composi- electricity that uses a ¯at charge rate irrespective tion, the region has nine ecological zones of the water extraction. Instead of pricing for the 9Gujarat Ecology Commission, 1994). All of actual electricity consumed, the cost is estimated Kutch's 97 rivers are nonperennial and have a on the capacity of the motor. There is therefore high run-o€ rate. Rainfall is erratic and vari- no incentive for farmers to reduce water able. The average rainfall is about 380 mm consumption. These trends have led to the 9mm) ranging from 440 in southern Kutch to overexploitation of aquifers. This, combined 338 mm in western Kutch 9Raju, 1995, p. 10). It with seawater ingression, has led to salinity in the only rains a few days per year, 915 days on water and soils and to a falling ground-water average). Kutch has a semi-arid type of climate table 9Gujarat Ecology Commission, 1994). and accounts for 60% of the semi-arid tract in Groundwater resources are also tapped by Gujarat. Temperature ranges from 45 degrees the district's drinking water and industrial centigrade in the summer to two degrees in schemes. Government tubewells providing winter. Kutch is considered to be a drought- water to the regional water supply scheme run prone district as droughts take place every two dry regularly due to a falling water table in the to three years. 2 Fifteen percent of the area of surrounding area. The average life span of a Kutch is cultivable. Rainfed agriculture and tubewell in Kutch is about 50% less than in animal husbandry are the chief occupations in other parts of India. 6 Thus the drinking water Kutch. schemes have a limited life span and water is Water management began during the prin- prone to much wastage. Drinking water cely era that saw the creation of several medi- schemes also compete with the needs of indus- um-sized earthen dams, irrigation and drinking try. For example, within the Gandhidham± water tanks and wells. Seventy percent of the Kandla complex in southern Kutch there are present large schemes were built by the Kutch several industries such as the Kandla port, State. 3 Eleven percent of the net cultivated fertilizer plants and the Free Trade Zone. The area is irrigated. Privately owned shallow and rapid development of this area has led to tubewells account for just over 80% of all irri- tremendous pressure on the water resources in gation 9Mehta, 1998). No major dam is possible the villages that supply water to it. For example, in Kutch due to its topography. In 1996, the the water levels in the 42 tubewells supplying surface water schemes comprised 20 medium water to this area have declined from 12 to 60 m and 162 minor irrigation schemes. 4 Medium- in the recent past, and their salinity content is sized schemes fall under the jurisdiction of the increasing 9Mehta, 1998). The drinking water state of Gujarat and minor schemes are wells of local farmers in the area have also been bureaucratically and institutionally managed negatively a€ected. 2028 WORLD DEVELOPMENT

The implementation of government-spon- crop a year. As a long term prospect, we can envisage sored watershed projects and micro-level water a bene®cial change in the ecology of this area 9cited in schemes has only recently become a priority. Kutch Development Forum, 1993, p. B2). Watershed development falls under the juris- diction of the Gujarat State Land Development As will be outlined later, this did not happen. Corporation 9GSLDC) and state agencies The command area of the SSP is largely located working on rural development. They have in the central corridor of Gujarat, which not identi®ed 177 potential watershed sites in only has better water endowments than Kutch, Kutch, but due to limited resources only 40 but has the state's strongest industrial lobby. have been developed. Largely, micro-level The following account highlights how the needs solutions in Kutch had not been prioritized of Kutch and other drought-prone areas were adequately due to the focus on large-scale irri- by-passed in the planning processes of the SSP. gation and the extra-basin transfer of water. In 1965, the Government of India appointed the Khosla Committee to draw up a master plan for the allocation of water. The report of 3. KUTCH AND THE SARDAR SAROVAR the Khosla Committee awarded a high share of PROJECT 9SSP) water to Kutch. Out of the 44.56 million acres of irrigated land designated for Gujarat, 0.75 There is a widespread belief in Kutch that million acres were allotted to Kutch. This due to the harsh climate, erratic water supply, worked out to 16.45% of the total land to be declining groundwater sources and frequent irrigated 9Kutch Development Forum, 1993). droughts, the only solution is to get water from The scheme also included a plan for three the rivers of mainland India 9Kutch Develop- subcanals for Kutch: the ®rst in the Ranns and ment Forum, 1993). All hopes are being pinned Banni; the second through central Kutch and on the Sardar Sarovar Project 9SSP) a contro- the third along the coast. The Khosla versial multipurpose irrigation and hydroelec- Committee proposed a 500-feet dam to ful®ll its tric project under construction on the Narmada plans. This height was unacceptable to the River in Gujarat. If completed, the planned states of Madhya Pradesh and , 163-m dam is intended to bring drinking water which refused to bear the brunt of such massive to 30 million people and irrigate 1.8 million submergence for the bene®t of Gujarat. Hence hectares of land 9Raj, 1991, p. 11). It will also the Narmada Water Disputes Tribunal submerge 37,000 ha of forest and agricultural 9NWDT) was appointed by the Central land as well as the homes of at least 250,000 government in 1969. Deliberations went on for people. 7 about a decade. The Tribunal declared its Even though the project was conceived verdict on the project in 1978 and the SSP, as it almost a century ago, actual work has been is known today, was conceived. The dam's stalled due to interstate con¯icts regarding the height was ®xed at 460 feet 9163 m) against the height of the dam, the extent of submergence demands of Gujarat, which pressed for 540 feet and the sharing of bene®ts. Di€erent commit- and Madhya Pradesh, and Maharastra which tees and commissions were set up to resolve all were content with only 210 feet. Kutchis these interstate con¯icts. Some of the early maintain that the state of Gujarat did not commissions, such as the 1972 Irrigation represent their interests adequately before the Commission Narmada Waters, stressed the Tribunal. As a result, during the course of importance of providing Narmada water to the several rounds of negotiations, Kutch succes- scarcity areas of Gujarat: sively lost out. As opposed to the three canals that would have brought Narmada water to Irrespective of what share of the Narmada waters Kutch, only half of one of the canals along the might come to Gujarat, the ®rst priority in the use coast was sanctioned. Instead of allowing for of this water must be given to these areas where the the irrigation of 945,000 acres of land in Kutch, rainfall is scanty and irregular, rather than to Broach only 95,000 acres of land were to get irrigation and Baroda districts which have 762 mm of assured 9Kutch Development Forum, 1993). This rainfall. The bene®ts from the use of this water in makes up a only 2% of the total area of Kutch. the areas of North Gujarat, and Kutch, will be far greater than what would accrue in the SSP advocates in Kutch maintain that the two districts mentioned above. There will also be indi- Government of Gujarat did not contest the rect bene®ts from the replenishment of sub-soil water, Tribunal's allotment to the state and consider which will help the farmers to raise more than one its implications for Kutch 9Kutch Development MANUFACTURE OF POPULAR PERCEPTIONS OF SCARCITY 2029

Forum, 1993). The intrastate allocation plan increasing salinity of the soils and sea water for irrigation did not emerge through a ingression 9especially in southern Kutch and consultative process. According to the Kutch coastal areas); dwindling water aquifers 9all Development Forum 91993) no bureaucrats or around the groundwater belt in Kutch, engineers from the state irrigation departments including Merka); and species loss due to the or water resources ministry were consulted. planting of the exotic salt-resistant species Several Kutchis also believe that the decision Prosopis juli¯ora 9popularly known as ganda was based on something as trivial as petty bawal, literally wild or mad Acacia). 9 In real political rivalries between several decision- terms, cultivators experience diculties in makers. 8 According to a review committee set getting safe and plentiful yields and pastoralists up by the Indian government to investigate the face a growing paucity of fodder and forage. project and its controversies, the allocation of The causes of these phenomena are manifold water to Kutch was very meager and there has and Ishall turn to them shortly. Nonetheless, been a marked mismatch in ocial rhetoric the scapegoat is made out to be the lack of concerning the claims by project authorities rainfall and increasing droughts. Scapegoating about the bene®ts of SSP to Kutch and the climate change for increasing water scarcity is actual irrigation bene®ts going to the region not restricted just to Kutch. A study by Dahl- accorded to the region. berg and Blaikie 91999) also notes that dwin- Kutch has had a very marginal presence in dling rainfall is a common focus around which mainstream Gujarati politics, and thus this low explanations of environmental change turn. allocation of SSP is not very surprising. Usually Falkenmark, Lundqvist, and Widstrand 91990, important decisions concerning its future are p. 30) argue that the climate change often made in distant Gandhinagar where the referred to when explaining water shortages bureaucrats concerned are likely to be out of globally may be a myth. Indeed, a scrutiny of touch with location-speci®c realities. Kutch, the rainfall data of the past 60 years in Kutch however, continues to be used to legitimize the indicates that while there have been erratic scheme. The Government of Gujarat is known variations in the quantity of rainfall, there is no to have taken World Bank missions and other evidence to suggest that precipitation rates have international dignitaries to remote areas in changed. 10 A t-test, comparing the rainfall in Kutch to highlight the urgency of the need for Kutch over the last 30 years 91968±97) with the the project. Work on the dam has recently been previous 30-year period 91938±67), revealed resumed after a break of several years following no signi®cant di€erence 9tobt: ˆÀ0:28; p > a judgement by the Supreme Court of India. 0:052-tail, see Sinclair, 1998). Rainfall data were Still, it is unlikely that the water of the available for the talukas of Abrasa, Bhuj and Narmada will ever reach Kutch, and certainly Rapar for a longer period 9120 years). Hence, it not in the near future as Idemonstrate in was possible to conduct inference tests in order Section 5. Nonetheless, drought-prone Kutch is to examine whether di€erences existed between largely used to legitimize the dam's construc- four 30-year periods 91878±1907, 1908±37, tion and the popular narratives to which Inow 1938±67 and 1968±97). A repeated measures turn support this view. analysis of variance revealed no signi®cant di€erences over these periods 9Sinclair, 1998). Table 1 gives the mean rainfall over these 4. POPULAR NARRATIVES OF periods with their standard deviations. SCARCITY Why then does the perception of dwindling rainfall persist and what does it obscure? Roe There is unambiguous consensus in Kutch 91991) argues that the ``development narrative'' and elsewhere that climatic change, indepen- simpli®es ambiguity. Uncertainty at the micro dent of human intervention, exacerbates the and macro levels is overcome by being problems of water scarcity. This is exempli®ed subsumed into the broad explanatory narrative by the widespread perception at the district and 9Roe, 1991, p. 288). In the same way, the village level that, ``It is raining less and perception of dwindling rainfall and increasing droughts are increasing ...''. droughts obscures uncertainty and regional This is not just restricted to the perceptual variations. The study by Dahlberg and Blaikie level. Clearly, there are several symptoms and 91999) comes to similar conclusions. In their much empirical evidence of environmental study in northeast Botswana, interviewees also change in Kutch. Some of these symptoms are: perceived a dramatic drop in the rainfall, even 2030 WORLD DEVELOPMENT

Table 1. Mean rainfall in Kutch 'mm) Mean rainfall in mms 9S.D.) 1878±1907 1908±37 1938±67 1968±97 1878±97 Abrasa 329 9240) 314 9267) 384 9330) 354 9294) 345 9282) Bhuj 354 9171) 363 9239) 364 9256) 340 9246) 355 9228) Rapar 396 9212) 370 9177) 371 9203) 385 9294) 380 9223) Kutch N/a N/a 358 9204) 375 9259) 367 9221)a Source: Institute of Desert Ecology, Bhuj. a Based on data available for all talukas from 1932±97. though scientists discredited these views. With rainfall); moderate or kurwara in which one fall respect to Kutch, this narrative obfuscates of rain failed; and lean or dookal 9drought) several features intrinsic to rainfall and drought 9Burnes, 1831). This classi®cation, though o- patterns in Kutch and shies away from cially no longer in use, is still prevalent at the acknowledging human-induced forms of scar- practical level. Life in Kutch comprises both city. Let us explore these issues in greater depth. sookal and dookal. All years are never uniformly bad or dookal. It is common to use several years 9at least three, ®ve or even a 9a) High variability decade) as a unit of analysis. For example: ``On an average out of 10 years, approximately 3±4 While the results of Table 1 indicate that years are famine years, 4±5 years are lean years rainfall in Kutch 9and the talukas of Abrasa, and 2±3 years are better years...'' 9Government Bhuj and Rapar) is neither increasing nor of Kutch, n.d., p. 14). decreasing, rainfall is characterised by high This statement from an undated document of annual variability 9see Figure 1 for Kutch and the Kutch State in the 1950s is not very di€er- Figure 2 for Abrasa taluka). ent from today's village-level sentiment that a Local categories deal with this variation in decade comprises three good years, three bad di€erent ways. The traditional classi®cation of and four years, which are moderate. Hence, years was as follows: good or sookal 9good there is a tendency to take several years as the

Figure 1. Average annual rainfall in Kutch.