China - Yellow river basin investment planning study (Vol. 1 of 2) : Main report

Report No. 11146.CHA China Yellow River Basin Investment Planning Study (In Two Volumes) Volume I-Summary June 30, 1993 MICROGRAPHICS Agriculture Operations Division China and Mongolia Department East Asia and Pacific Regional Office Report No: 11146 CHA FOR OFFICIAL USE ONLY Type: SEC Document of the World Bank- This document'has a irestricted distribution ard may be used by recipients oniy in1~ the performnance of their official dut'es. Its contents may not otherwise be d8iscoed' whutWrd Bantk authoization.  4 CURRENCY EQUIVALENTS Currency Unit: Yuan (Y) $1.00 Y S.4 Y 1.00 $0.1851 FISCAL YEAR January 1 - December 31 WEIGHS AND MEASURES Metric System Area 1 ha = 15 mu 1 mu = 0.06666 ha ACRONYMS AND ABBREVIATIONS BLM Basin-level model CCPN Central China Power Network GOC Government of China CYJV Canadian Yangtze Joint Venture GVAO Gross Value of Agricultural Output GVIO Gross Value of Industrial Output MoM Ministry of Mining and Underground Water M&I Municipal and Industrial MOA Ministry of Agriculture MOC Ministry of Communication MOE Ministry of Energy MOF Ministry of Finance MOT Ministry of Transport MURC Ministry of Urban and Rural Construction MWR Ministry of Water Resources NCGWR National Coordination Group on Water Resources NCPN North China Power Network NEPA National Environmental Protection Agency NWT'M Northwest China Power Network RBC River Basin Conservancy Commission RDPI Research, Design and Planning Institute SEPC State Environmental Protection Commission SPC State Planning Commission WRB Water Resources Bureau YRCC Yellow River Conservancy Commission  1OR OmCIAL USE ONLY CONTENTS Preface ........................................ iii ExecutiveSummary ......... ............................ v 1 Background . ................................. 1 2 Criticalsksues In 'Imudmg ........ ......................... . 4 3 Objectives and Methodology ................ ........ 7 4 Economic Uses of Yeow River Water ................. 9 5 The Current Situation . ...... ...................... . . . . .14 6 Emergig Shortages ............... .............. 16 7 Ihe Investment Progra ............................. 19 8 Projected lmpactso' Selected nvestmets .............. 21 9 l tionland Polky Alternatives ................... 24 10 Conduslons and Recommendations ................... 27 TABE IN TEa 1. Basin Cropped Area, Grain Production and Returns to Water Use . 12 2. Basin-WideWater Balances, 1990 ..... . . . . . . . 15 3. Basin-Wide Water Balances, 2000 .... .. . . . . ..... 17 4. Financial Costs of Popo os nvestments 20 5. Key Multipurpose Dam Projects . 20 FIGUREs IN EXT 1. SedimentConcentrationin Rivers.. . 3 2. Fatalities in Major Floods, 1920-90 ................... . 5 3. Runoff versus Dend for Water. . 10 4. Intrcountry Comparison of Prices and Opportunity Costs. . 14 5. Water Supply andDemand, 1987and2000 . . . 23 6. ValueAddedduetuo Wr... 25 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  - ii - MPS No. 24547 Water Resources Regions No. 24545 Exisdng and Proposed Dams/Hydropower Projects No. 6 Proposed Dams and Warping Sites No. 24549 Current and Planned Irrigation Areas  - iii - PREFACE One in nine Chinese live within the Yellow River Basin, and most of these people depend, directly or indirectly, on the river for their livelihood. The World Bank has helped finance eight irrigation development projects in the basin to date, and will scon appraise a major multipurpose dam, Xiaolangdi, on the Yellow River. Dozens more projects have been proposed, with a total cost exceeding $10.5 billion. This study was motivated by a desire to place the investment options in a consistent, basin-wide economic framework, and to begin to evaluate them with respect to the development goals and the physical and economic constraints under which thev will operate. Although local planning efforts have been intense, they have suffered trom inadequate and inconsistent agricultural and economic data, and from the absence of such an economic framework. As a result, the proposed program is not always consistent with the resource base, and developmental priorities sometimes get lost in the effort to harness the Yellow River. To avoid such problems, this study has developed a basin-level model to enforce consistency upon estimates of the available resources and the projects proposed to augment or utlize them, and to do so while maximizing economic benefits from the ultimate utlization of Yellow River water. Given the constrints on the resources available, the limitations of the data available, and the Yellow River Conservancy Commission's and Ministry of Water Resources' lack of experience at this point in performing basin-leve! economic analysis, the study could not hope to provide a definitive statement of the planning problems in the Yellow River Basin. Rather, the study is intended to initiate and promote a learning process. A number of critical issues remain that could be usefully explored in the proposed second-stage Yellow River Basin Water Resources Economic Modeling Study to be funded by the United Nations Development Program. Among these are national water allocation, water pricing policies, cost recovery of benefits, short-term system management, and institutional responsibilities. This report is based on a mission that visited Beijing and Zhengzhou during April 1992. Mission members were D.J. Gunaratn (EA2AG, mission leader), B. Trembath (ASTEG, power engineer), G. Kutcher (consultant, economic modeler), S. McGurk (consultant, agricultural economist/modeler), S. Sud (consultant, power planner), Jiang Ping (consultant, irrigation specialist), and Zhang Weizhen (consultant, groundwater specialist).1/ The mission worked closely with the planning departments 11 Mae awlen Shu and Li Qua provided vatuable assistac in prepara of tho mups tbrough digitizing and usi geographic informaon systems. Ms. R. HR=s lso provided valuable as in editing to aeport.  - iv - of the Yellow River Conservancy Commission and the Ministry of Water Resources. Excellent support was provided by both departments. The mission alsco benefited from interaction with the staff of Water Resource Management Inc., who are at present consultants developing the simulation studies of the Yellow River Basin.  -V. EXECUTIVE SUMMARY i. The average annual flow of the Yellow River is almost identical to Egypt's share of the Nile. Yet it supports a population twice that of Egypt, and irrigates an area twice that of the Nile. It does so in the face of the twin menaces which have charcterzed the river throughout history: sediment and floods. The Yellow River contains the highest sediment concentration of any river in the world-at times and in places it resembles a mud slide a mile or more in width. However, sediment is not the main concern of Chinese planners, although it is high on the list. Nor is it the river's ability to support an ever- growing number of people, both rural and urban, and industries, mines, oitfields, and fisheries. The primary concern is flood control. Floods along fte YeUlow River have killed millions, have disrupted the lives of tans of millions, and have caused incalculable poperty and economic loss since the earliest civilization appeared along its banks. Planning for economic sunvval along the Yellow River is a challenging task. Planning for economic dewlopment is another story altogether, of which this report may only constitute a preamble. U. Yellow River planners are undoubtedly correct in focusing thei highest concerns on flood control, followed closely by sediment control. Floods may strike virtually anywhere in the basin, both in late summer as the result of intense rainstorms or even in the dead of winter as ice formations dam the river. Flood and sediment control are closely linked and without sediment control, effective flood control may be impossible. The aggraded lower reach is already dangerously suspended, in places more an 10 meters above the surrounding plain. Without sediment control, the typical uses of rver water-irrigaton and water supply-are hampered by expensive and tme-consuming desiltation. iii. Planners and politcians look to the Yellow River for irrigation to feed the more than 120 million people residing in its environs. Without irrigation, farming in the basin is a marginal activity at best, with very low yields and incomes. Three large multipurpose reservoirs (Longyanxia, Liujiaxia, and Sanmenxia) and several hydroelectric power plants haness only a fraction of the river's potential energy. Each of te concerns-flood prevention, sediment contml, irrigation, power-is not necarily in harmony with the others, leading to severe managerial conflicts. For example, use of the limited reservoir storage for flood control typically reduces the ability of those acilities to supply irrigation water at the optimum time, and to maximize enery output. Multipurpose operation of the reservoirs for flood control and power may also exacerbate  - vi - the sediment problem, as evidenced by the disastrous experience with Sanmenxia reservoir.l/ On the other hand, operation of the system for maximum irrigation benefits (in the densely populated lower reaches)-the largest single source of economic benefft from the river-will not only reduce the ability to control floods and sediment, it will also induce income losses for millions of upper and middle reach farm families who are among the poorest groups in China. iv. To cope with t'hese problems, and plan for the future development and 'harnessing' of the river, a variety of institutions have grown up. The Ministry of Water Resources (MWR) is responsible for ovemall basin-level planning, but must coordinate all investments with the State Planning Commission. MWR's local arm, the Yellow River Conservancy Commission (YRCC), plays the leading role in carrying out regional studies, identifying and designing projects, and promoting sound water use policies. It also undertakes a coordinadng role among the nine provinces that the basin encompasses. These provinces also plan, invest, and malce policy, sometimes in unison, and sometimes at odds with one another and YRCC. Complicating matters still funher is the jurisdiction that the bMinistry of Energy has over all hydropower installations, including some key reservoirs. v. With the exception of water, the resources available for the promotion of basin-level economic development-including humans, land, and the impressive industil and mining base-are sufficient to accomplish virtually any dev'elopmental goals one might envisage. However, Yellow River water supplies tend to be not only sediment-laden, but ilso highly variable and unreliable in supply. Most of the runoff occurs in the late summer months, while the winter and spnng, when crops most need water, are typically dry. Year-to-year varation is also large, with five- or ten-year Oroughts not uncommon. Groundwater has been used to great advantage to supplement ri er supplies for inigation -and water supply, but there is evidence that the demands may be rapidly outstripping recharge rates in some areas. vi. Clearly, Yellow River water plays a dominant role in the regional economy. Unless water development is carefully planned and mar-aged, water shortages will inevitably restrain economic development. Our efforts to place economic values on the different uses of Yellow River water were hampered by the inadequate and inconsistent coverage of the basic data. Even irrigated crop yields are unknowns which must be estimated, even though irrigation accounts for 60-70 percent of all crop production. The average value of water used in irrigation is found to be about 24 fen/m3, while the marginal value, at critical times and in the most productive regions, is about 50 fenlm. However, farmers seldom pay more than 2 to 8 fen, and even municipal and industrial consumen only pay about 20 to 30 fen af treatment costs are taken into consideration. Clearly, a more rational pricing system reflectig the true opportunity cost of Yellow River water is called for. Among the more important opportunity costs seemingly ignored by 11 Wihn five yea of its completion in 1960, sme's stog aea of 6 bio wn was virtaily filod with adiet endins the proJet usdess for flood t orol and power geamad A costly sebuildg and flusig progSam bough the stonage capacity back to about 2 billiho m.  -vii- local plannes is the value of hydroelectric power that could be generated. In terms of forgone power, the opportunity costs of increasing diversions to irrigation or other uses is as high as 20 fen/m' in some ulpsteam regions. vii. This study only touches upon pricing and other policy issues facing the basin. The primary focus is on the projects-some costing sever billion US dollars-which have been proposed to harness the Yellow River and make the utilization of its waters moreefficient. However, the study of these investments must be prefaced first with all understanding of the physical and economic environment, in which the component projects will operate, and second with an understanding of the factors constraining development, which those projects are desiped to alleviate. To this latter end, a special Basin-Level Model (BLM) has been prepared. The BLM takes advantage of the latest computational technology to combine simulation and optimization techniques in a flexible yet powerful planning tool. viii. We have attempted to pull together the available data to construct a conidstent picture of the current situation. This effort too was hampered by incomplete data: the latest available water use data pertain to 1987. Earlier YRCC projections to 1990 are used to the extent possible, then the BLMW is used to fill in the gaps. The BLM demonstrates that, despite rapidly growing water demands in the decade of the 1980s, water availability is not yet an overtiding concern. Shortages are probably localized under most runoff scenarios. Only in a P90 (1-in-10) year are basin-wide shortages severe. Municipal and industrial (M&I) demands remain relativey small, and irrigated agriculture can provide an adequate diet, although not up to the level of China-wide standards. ix. However, little was accomplished during the 1980s toward alleviating the prime concerns of flood and sediment control. Efforts toward the former were mostly confined to river training works and dike raising and maintenance. There was no measurable reduction in the probability of a catastrophic flood. Sediment control efforts were also piecemeal and centered at the primary source of sediment-the Loess Plateau. Vigorous and expensive programs to control erosion through planting trees and constructing check dams are only now showing some measurable effect after a period of 30 years. x. All told, the investments proposed by YRCC, the provinces, and the central agencies have a value which probably exceeds $10.8 billion today. They include a variety of projects and programs, each designed to address one or more of the issues discussed above. Rivet harnessing and reguladon projects (8.4 percent of the total between now and 2000) will continue to raise and improve dikes, stabilize embankanents, construct weirs and spur dikes, and improve diversion works. Conservation efforts (6.8 percent) will continue to address erosion in the loess region, while at the sarne time seeking to improve the welfare of the most impoverished residents of the basin. xi. The largest component (36.8 percent) of the portfolio is targeted for rehabilitation and development of irrigation. About half of the existing irrigated area  - viii - requies ilnprovement, both in the land and in the faiiHties deliverilno water to it. This component also Includes massive expansion of t'he irrgated areas-by 40 percent or more. xi. The most sonspicuous (and ultimately, the most costly) components are the five multipurpose projects proposed for the next 20 years. First would be Xiaolangdi, the only one for which reliable cost data are available ($2.8 bitlion). It would provide the primary means of flood control for the most vulnerable lower reach, trap most of the coae sedinment entering the channel for a 22-year period, generate substantial power, ind vastly iprove the reliability of irrigation supplies to the highly productive lower reach areas. xiii. Next would be Wanjiazhai Dam in the nuddle reach (power and water supply) followed by Daliushu Dam in the upper reach. Comparable in size to Xiaolangdi, Daliushu would generate power and irrigation benefits, control flooding caused by ice jams and regulate flows for several existing as.. uture downstream run-of-river power plants. Qikou and Longmen are fute possibilities for the middle reach. Each would serve many functions similar to those of Xiaolangdi, in particular, trapping additional sediment, controlling flood surges, and providing irrigation to the chronically short middle reach regions. The remainder of the proposed projects, about which this study has little to say, deal with hydrodectric instaLlations, navigation works, and environmental protection. xiv. By the yea 2000 M&I demands will approximately double, and this, combined with the expansion of irrigated area, will lead to severe basin-wide water shortages in all but the wettest inflow scenarios. The BLM projects shortages between 5 and 12 billion m3 per year compared with total annual flows of 58 billion &n, and severe crop stress. YRCC also projected shortages for 2000, but less than 2 billion m3. Although the two sets of projections agree on the basic pameters of runoff and upper and middle reach consumption, YRCC's projections ignore about 80 percent of the lower reach demands, even though this region currently accounts for nearly a third of total water diversion and nearly half of irrigated production. xv. Altenative simulations for 2000 reveal that it is imperative that the projected im proiments in efficiency of water delivery, a component of the projects planned between 1990 and 2000, be achieved. Otherwise, much if not most of the proposed new irrigated area cannot be served with water. The analysis also suggests that about half of the proposed expansion probably cannot be supported by Yellow River supplies, and should be postponed until (and it) transfers from the Yangtze are realized. The resources could be better v1sd to improve delivery efficiency (in some regions it is as low as 25 percent), and to establish tighter controls. The BLM also demonstrates that there could be substantdal gains from moving toward an economically optimal diversion patter;!, although a purely economic allocation is not ecommended for two reasons. First, the physical controls to achieve such an optimum do not exist and their construction would be costly, and second, many of the poorst farm families would suffer. xvi. The single most important parameter affecting the basin's performance in 2000 is the volume of water required to flush sediment from the lower reach. YRCC  - ix - maintins that 22 billion mn (38 percent of total annual flow) are required under current conditions. Howe".er, it is estimated that this can drop by 25 percent in 2000 if soil conservation efforts are vigorously pursued on the Loess Plateau. If not, then the future for development of the basin looks grim in termn of flood control and water availability. On the other hand, substantial reduction of the sediment problem at the source could free up sufficient water to meet most projected demands. xvii. Between 2000 and 2010, the final year of analysis of this study, demands on the system will continue to grow, particularly those from the M&I sector. In addition, over 4 billion m3 of planned extrabassin transfers will take effect because the diversion structures will be in place. As expected, water shortages will intensify even more, and considerably restrain growth. This is due not so much to the absolute magnitudes of the shortages, but to the fact that water will not be available where and when It is needed most. About half of the irrigated area to be created between 1990 and 2010 cannot be fully used under most inflow scenarios. The single most effective program to reduce water snortages after allowing for sediment flushing is thdat concerned with improving distribution efficiencies. xviii. The program to constrtct several more large run-of-river power plants will male a substantial contribution to meeting energy demands, and appears warranted. Those plants situated above the proposed Daliushu reservoir will produce significantly more energy if Daliushu is available to reregulate flows. xix. Xiaolangdi Dam is the only viable option available to meet the primary concern about the river-flooding in the lower reach. Xiaolangdi is also the only realistic option for sediment control and regulation in the lower reach within any practical planning horizon. By trapping large amounts of sediment, it would fre.- "lushing" water for other productive uses throughout the basin. It is probably justified on these grounds alone. In addition to its hydropower, Xiaolangdi would alleviate some of the water shortages in critical months in Henan and Shandong, producing enough benefits to justify it on purely economic grounds. xx. Similar arguments pertain to the Daliushu Dam. It would produce significant power of its own, and go far in reducing seasonal water shortages in the poor middle reach regions. It would also reduce flooding caused by ice jams in the northern bend of the river and improve navigation. Insufficient analysis has been undertaken to fully justify Daliushu, but indications are that it should follow Xiaolangdi among the major reservoirs in priority. xxi. The main conclusions of this study are: (a) MW)'s 1988 Water Financial Directive, which would price water at marginal costs and ensure complett cost recovery by 1997, should be implemented as soon as possible. In vew of the severe water shortages expected In the years 2000 to 2010, MW) should seriously consider introducing, In addition to marginal cost pricing, a portion (30 to  SO percent) qo scarciy rt In water prices to rflect opportnity costs. hse prices would apply to new consumptve Investments In the water sector. (b) The plans for expading the Ifrrgated area are undoubtedly far too ambidous given water availability and rapidly growing M&I and extrabasin demands. Only about half of the planned addition of 26 million mu (1.7 million ha) I justiled. (c) Given the severity of ifure water shortages and the high marginal value of water In Irrigadon, efforts to Improve the efflciency of water delivery and use must be given the highest priority In some 43 million mu (2.9 million ha) of exdsng lrrgadon. Current loss rates are woafly high In the upper basin areas, andfainners are without doubt misusing water because its * st to ten Is a smallfraction of its economic value. () The fJuture water balance picture depends more on how much water Is neededforflushing sediment than on any other single factor. Conservaion efforts at the source must be promoted vigorously for about 60,000 to 80,000 Ame of the critical eroding areas In the middle reach of the basin. Xiaolangdi will be required to trap much of the coarse sediment and to regulate the flow to flush sediment to tk& ocean. WIthout such a dual program of sediment control, a change in river course will occur with catastrophic economic and social consequences. (e) XiaolangdI Is the most urgent, and probably the most economically viable multipurpose project proposed. Without It, the specter of a majorflood v11 continue to overhang the entire lower reach, including its most productive farm land, China's major oiftelds, millions of people living or farming inside the dikes with no flood protection whasoewr, and tens of million; more who are wdnerable to a dike breach. (f) Daliushu is probably the second most bportant multijurpose project, given the Impoverished regions It would serve, and Its impacts on energy production andflooding. (g) YRCC should dovelop a more detailed econondc model like the Basin-Level Model developed In this study and collect addidonal data to evaluate in detail the impacts of changes In basin operadon rules or new projects proposed for Implementation.  -1- SUMMARY REPORT 1. BACKGROUND 1. The Yellow River Basin. The Yellow River Basin is the cradle of Chinese civlization; irrigatioa has been piced in the basin for thousands of years, generating agricultural surpluses that permitted the development of ancient societies that were the envy of the world. The river is China's second longest, at 5,464 km. Its source is in the foothills of the Bayankala MAountains on the Qinghai-Tibet Plateau. It passes through nine autonomous regions and provinces (Qinghai, Sichuan, Gansu, Ningxia, Inner Mongolia, Shaanxi, Shanxi, Hen , and Shandong) before emptying Into the Bohai Sea, supporting on its way an agricultural population of over 100 million people (Map 24544). 2. The basin is divided into three distinct sections: the upper, middle and lower reaches (Map 24547). The upper reach, with a drinage area of 384,000 lin2, lies between the source and the gauging station Hekouzhen i Inner Mongolia (the start of the Yellow River's great turn to the south). It is characterized by mountain gorges and wide, lake-filled high valleys near the source, followed by high mountains, deep valleys and narrow gorges that empty into the great alluvial Ningxia and Inner Mongolia plains. The vertical drop in this region is 3,231 m, making the reach ideal for hydropower generation. On average, 56 percent of the basin's runoff is to the upper reach. The two largest reservoirs in the oasin, Longyanxia and Liujiaxia, are located in its upper environs, and two more large reservoirs are planned. Minimal precipitation, 300-600 mm per year, falls on most of this reach, which has limited water demand, while the lower stretch of the upper reach, from Lanzhou north to the Mongolian steppelands, has large and growing Irrigation demands. These demands are difficult and costly to meet: the distribution systems are poor and losses are high, and in places water must be pumped up several hundred meters. 3. The middle reach, with a drainage area of 344,000 km2 between Helouzhen and Huayuankou, lies between 1,000 and 400 m above sea level and encompasses the Loess Plateau, and two of the Yellow River's major tributary basins, the Fen River Basin in Shanxi and the Wei River Basin in Shaanxi, both of which rely on the Yellow for muc': of their irrigation demands. Water supplies to this reach are invariably short; groundwater extraction is often excessive, and water diverted from the Yellow must often be pumped tens of meters or more. The middle reh also contains large concentrations of population, industral complexes, and mining, all significant and growing water consumers.  -2- 4. The lowe reach begins near Huayuankou and stretches to the Bohai Sea. Most of the area is alluvial plain brought about by the meandering of the Yellow River. The lower reach contains most of the irrigated land in the basin, and on the whole is the most productive. Its temperate climate permits year-round cropping, and even paddy production is significant in its southernmost areas. Rainfall is sporadic, but often sufficient to grow most crops and obtain adequate yields without irrigation. However, irrigation is widespread, involving complex systems to make maximum use of Yellow River water when it is available (and at acceptable silt concentrations) as well as groundwater. Both are used conjunctively and are often stored in field ditches along with reusable drainage water. The lower reach is disadvantageously situated on three counts: (a) in periods of shortage, it is last in line to receive water; 0) it is most prone to suffer the consequences of flooding; and (c) it must deal with most of the silt deposited in the river from upstream regions. S. The unique feature of the Yellow River is its unparalleled sediment content. The Yellow River has the highest concentration of sediment in the world-nine times that of -its closest competitor (see Figure 1). Sudden brief but intense storms cause concentrated inflows into the Yellow River system, carrying with them the easily erodible soil. -Most of the sediment in the lower reach of the river originates from the world's largest loess plateau, located in the middle reach. From 1919 to 1960, the average annual sediment load passing Huayuankou was 1.60 billion tons, for an average annual flow of 47.5 billion .e. Twenty percent of thi3 sediment, primarily consisting of coarse sediments with a particle sze larger than 0.05 mm, was deposited on the river bed downstream of Ruayuankou. Most of the deposition of sediments occurs over the lower reach of the river, resulting in a naturally aggrading system. At places the river is suspended 10 meters above ground as a result of stabilization of the river course through centuries of dike construction; the main embankments stretch 798 km from Huayuankou to the Bohai Sea. 6. Inextricably linked to sediment are floods. Uncontrolled dike breaches have killed millions, and devastated large areas of productive land (Figure 2). Floods typically originate in the middle reach or upper lower reach during intense summer storns. Floods bring large amcl) its of sediment, the deposition of which makes flood control more difficult and more costly. Peak historical discharges have been estimated at over 33,000 m3/sec, while the existing structures can accommodate only a 22,000 m3/sec discharge. Flood control on the lower reaches can only be maintained at present relatively low protection standards (a recunrence of one flood in 60 years) through increasingly costly and massive work programs. In addition, there are at present about 1.4 million people living between the dikes in 2,000 villages. These people are subjected to frequent flooding. About 1.7 million people are living in the detention basins, and these basins have to be flooded in the event of floods greater than 1-in-60 years. 7. The greatest potential damage from a flood would occur if the main dikes were overtopped or breached. These dikes now protect about 120,000 kIm2, with a populadon greater than 100 million. Fatality rates could be as high as 1-3 percent with a piajor flood.  -3 - Fre 1: SEDmNT CONCENRAON IN RIVERS 40 25 20 25 j2~~~0 15 1 0 0 Auzkon Yerngtze Heithe 0mmwutr4 I ausu Ielo POrI Mimsleslopi Nile Koehi Gang"s 8. OrganizalUouaRepnsb 'noe h Minisfty of Wate Resurce OM%R) is responsble for overall, basin-level planning, but must coordinat all investments with the State Planning Commission (SPC). MWR's local arm, the Yellow River Conservancy Commission (YRCC), plays thie leading role in carrying out regional studies, identifying and designing projects, and promoting sounid water use policies. It also undertakes a coordinating role among the nine provinces which the basin encompasses. These provinces also plan, invest, and make policy, sometimes in unison, and sometimes at odds with one another and YRCC. The Ministry of Energy (MOE) has jurisdiction over all hydropower insalltios,including some key reservoirs. 9. YRCC"s last attempt at a comprehensive basin-wkide plan was carried out In the early 1980s and was based lagely on 1980 or earlie data.11 The study has been updated In piecemeal ahAion since then, in part using data for 1987. Although comprehensive and factual, the YRCC study suffers on sevemal counts: (a) the Methodology employed is now obsolete; (b) the data used are often incnsstent and j/ -no lwe of ewail vmansi is "Revised Sumayna of the Repoa on Yellow River Hammning and Develpmen Planing" (YRCC, 1988).  -4- incomplete; and (c) the results may be overly optimistic both in terms of future water supplies and in what can be accomplished hough structural changes (an early version called for no less than 29 major reservoirs). Furtmore, the YRCC study reflects the hydrologic and engineering bent of that institution, and is viuay devoid ot economic rationale. 10. World Bank Investments In the Basin. To date, the Bank Group has helped finance eight irrigation and rural development projects in the Yellow River Basin. A large multipurpose reservoir, Xiaolangdi, has been appraised. In addition, dozens of projects costing more than $10.8 billion (1987) have been proposed for the 8th (1991-95), 9th (1996-2000), and 10th (2001-05) five-year plans. Given this level of commitment to the basin's development, it is time to undertake an investment planning study. 2. CRMTICAL ISSUES IN PLANNNG 11. EmerIng Water Shortages. At present, aggregat water supply and demand in the basin are roughly in balance in all but the driest years. Shortages do occur in most years, but these are seasonal and local and due to either insufficient reservoir storage or insufficient diversion capacity from the main stem. i12. YRCC expects consumption of Yellow River water to grow at a 2.7 percent annual rate during the 1990s. By their calculations, shortages will be 1.65 billion m3 in a PSO (median)- runoff scenario in 2000. However, that analysis ignored much of the lower reach's demands, supplying only about 2 billion mn out of a projected demand of over 10 bilion mn. ,RCC may also have excluded some of the planned extrabain transfers, which could total 4 billion ml by 2000 or soon thereafter. When these assumptions were corrected, we found that the shortage will likely be 5.75 billion m' (12 percent) in a PSO scenario, and 8 billion ms (17 percent) in a P75 scenario. 13. Beyond 2000, the shortages are likely to intensify further. Municipal and Industral (M&) demands, including extrabasin transfers, are projeced to be more than 10 billion mn. The planned expansion of the irrigated area is targeted to reach 87 million mu, from about 59 million mu at present. Even under the most optimistic scenarios tsted, with significant improvements in water distribution efficiencies and up to 14 billion mn3 of additional storage facilities, serious shortages remain in all but the years of highest runoff. 14. Equity Cons deraUons. With a population far in excess of the carrying capacity of the land, the remote, heavily eroded Loess Plateau and the cold, high, and ard upper eaches of the Yellow River have long been among China's principal poverty- stricken regions. They are also the least efficient water users. Delivery and field losses are extremely.high, and many areas must be served by energy-intensive pumping schemes. With increasing frequency, Yellow River water will be in short supply, and the water used in dt regions will have an ever higher oppounity cost downstream. Sooner e than later, YeUow River planners wiU have to squarely face equity vs. efficiency issues.  -- Mm 2: FATALmES IN MAJOR FLOODS, 1920-90 I o,, 0.. 0.3 O Ja 0.4 0 00 ~flet~1 )aint1meU yet IeW S 15. Energy. Both nm-of-rver and storage reservoir hydropowa plants have a instaUed capacity of 3,660 MW and an average annual energy capability of 179750 GWh. However, this is only 7.6 percent of the demand from the three major grids supplied, a demand that is growing by 7 to 10 percent per year. Energy from hydro costs far less than the alternatives of coal and gas; thus, another 14,260 MW of capacity are scheduled to be installed in Yellow River facilities by 2010. 16. Although hydropower does not constitute a consumptive use of Yellow River water, it does generate severe conflicts in management of the systems. Run-of-river plants require a relatively uniform base flow, hence regulation through storage in larger upstream facilities Is required for their optimal use because the intrayear patterns of runoff and irrigation demands are far from uniform. Furthermore, maximum energy output from stor reservoirs requires iltl reservoirs and constant releases. This is in direct conflict with (a) irrigation releases, which require reservoirs to be emptied during the dry spring months; (b) flood control, which requires that specific amounts of storage be available during the flood season; (c) sediment control, which necessitates flushing the river with heavily silt-laden storm flows; (d) ize-run control, which requires the release of large flows prior to freezing to expand the cross-sectional area of the channel, and reduced flows after icing and prior to thawing to prevent ice jams from forming; and (e) navigation and water supply, which require relatively constant releases. 17. Filood Control Needs. Because of intense development in the main detention basins-Beijindi and Dongpinghu-it is now felt that use of these basins for flood diversion would result in unacceptabl: economic losses. There are at present no flood storage facilities in the middle reach, and the reconstructed Sanmenxia has very limited flood storage capacity (about 1.5 billion m3). Reasonable protection of the lower reach  - 6 - areas must await the constu of the Xlaolangdl Dam. Likewise, protecdon of the lower middle reach must await the construction of the Qlkou and/or Longmen dams. 18. The use of reservoirs for flood control engenders losses in irrigation supplies as well as hydropower output Water supply for irigation is most critical in the spring and the reservoir must be flled in summer months when water is abundant-precisely the time when reservoirs must be kept low in order to receive incoming floods. Floods of an entirely different origin than intense summer rains ae caused by ice jams both near the estuary (where the river tums north) and in the great northern river bend near Hekouzhen on the middle reaches. To minimize damage, flows must be carefully regulated from November to March. Too litde flow will induce a hard freeze while too much may result in ice jams. This factor reduces the water available for irrigation during early spring for all areas in the middle and lower reaches. A proposed reservoir at Daliushu, approximately midway between Uujaxia and Hekouzhen, will improve ice run control on this river section, permit full use of the upstream six-plant Longyangxia-Daxia hydropower system currently constrained by ice-run control, and reregulate upstream power system dischargs for downstream irrigation, run-of-rver and water supply needs. The proposed rervoirs at Longmen and Qikou, between Helouzhen and Sanmenxia, will alleviate middle reach irrigation and M&I shortages; they will also contain middle reach floods and control sediment. 19. I o nal Couflicts. lTe differences in the need for Yellow River water by season and by reach often lead to conflicts. Ice-run control on the Great Bend of the Yellow River in Inner Mongolia constrains not only upstream power generation in liujiaxia Reservoir during the thawing season (perhaps by as much as 200 MW of continucus power), but also irrigation and water supply in southern Gansu and Ningxia, and irriation and navigtion below Hekouzhen. Sediment control requires tradeoffs between middle reach sediment delivery and downstream power generation, reservoir regulation, ice-run and flood control, water supply, irrigation, navigation and estuary control. Within the middle reaches, there are significant problems in balancing power geneation, sediment regulation, wate supply, flood and Ice-run control, navigation and irrigation. On the lower reaches, flood and sediment control operations also create conflicts with other uses. Finally, regional priorities, for example, costly high-lift irrigation or coal washing on the upper and middle reaches may not agree with aggregate basin-wide water supply priorities. 20. In the shor term, contradictions between upstream power generation and irrigation supply, and downstream water supply and navigation and ice-run control on the Inner Mongolia Plain will continue. Responses to conflicts between sediment control and downstream needs include soil and water conservation programs in the Loess Plateau, improvements to lift irrigation schemes on both banks of the main stem between Yumenkou and Tongguan, improvements to river training worls on the Longmen to Tongguan river section, and investments in navigation works on the Fugu-Yumenkou river section. In the medium to long term, only with construction of regulafting fcilities at Daliushu can flood and ice-run control, drought management, and water supply be ensured in the middle reaches; only with the completion of the Qikou Reservoir can sediment delivery to the  -7- middle reach be reduced, allowing relatively sediment-free water to be supplied to M&I use In the Taiyuan region of Shanxi Province; and only the completion of the Xiaolangdi Dam will permit significant improvements in flood and sediment control in the lower reach without sacrificing other downstream demands. 21. The Planning Problem. Competing and often conflicting demands pose complex planning and management problems. Without massive additional investments in the Yellow River system, future consumptive demands from agriculture and M&I simply cannot be met with reasonable reliability. If sediment and flood control efforts are to predominate, as local planners believe they should, economic losses in agriculture, urban water supply, and industrial energy output will loom ever larger. It is extremely difficult to minimize these losses while managing the conflicts induced by competition for scarce Yellow River resources. 3. OBJECTIVES AND METODoLOGY Study Focus 22. The primary focus of this study is on the investments-some that will cost sever billion US dollars-which have been proposed by the various entities to harness the Yellow River and make the utilization of its waters more efficient. We attempt to ascertain which investments are economically viable, which of these should have priority given the urgent need for flood and sediment control as well as the demands of expanding production and incomes, and which types of investment are inconsistent with the future water balance andlor economic objectives. 23. A study of investments must be prefaced first with an understanding of the physical and economic environment in which the projects will operate, and second with an understanding of the factors constraining development, which those projects are designed to alleviate. To this latter end, a special Basin-Level Model (BLM has been constructed. The BLM aes advantage of the latest computational technology to combine simulation and optimization techniques in a flexible yet powerful planning tool. Using the BLM, this study attempts to pull together the available data to construct a consistent picture of the current situation, and then use that same tool to estimate the economic benefits from structural changes such as adding new reservoirs, expanding the irrigated area, and malkng water conveyance and use more efficient. 24. This study can only touch upon pricing and other nonstructural issues which might help achieve many of the goals for the basin's development. Such a study is warranted, but is beyond the scope and resources of this work. However, by bringing together the available data in a consistent framework, and placing economic viability in the forefront of the analysis, the current study provides the groundwork for such a policy study.  Basin-Level Model 25. Tnm Basin-Level Model (BLM) was designed to provide (a) a consistent framework to resolve missing and inconsistent data issues and force water supply-demand balances; and (b) a logical framework within which to test the implications of future development strategies and thereby estimate the impacts of certain types of projects. The available data do not permit derivation of a full water balance for any recent year, and some of the numbers YRCC used to project water balances to 1990 are inconsistent with data reported for 1987. This study has made it clear that the database for planning is a major concem and must be improved, both in terms of technical and hydrologic data, and in terms of related ag-icultural, economic and social data. Plans for water development and use must rest on accurate estimates of future water supplies. 26. The BLM is a nonlinear optimization model of approximately 1,000 equations. It maximizes the value added from Yellow River water subject to a variety of physical, hydrological, agronomic, and, if desired, policy constraints. The value added originates from irrigated agriculture in ten regions, hydropower output from six existing reservoirs and run-of-river plants, and from diversions to municipal and industrial users. Te BLM also provides for, but does not attempt to count, benefits from flood control, prevention of ice jams, and improved navigation through restrictions on reservoir operations and channel flows. The physical and hydrological constraints ensure that water demands do not exceed water supplies, and that the limitations of water storage and delivery capacities are respected. The agronomic relationships relate irrigated crop output to available water supplies, subject to the stock of irrigable land and other resources. 27. A constrained optimization approach is used because it is assumed that the relevant Chinese authorities wish to obtain maximum economic benefit from the operation of the system as well as investments in the system, subject to a variety of hydrologic, physical, and agronomic constraints. The core of the BLM is a network of nodes and connecting arcs common to simulation models. Water enters the system at selected nodes as exogenously given runoff, and is directed through the system according to predefined paths. Certain nodes represent offtake possibilities for irrigation and/or municipal and industrial demands. Some nodes represent run-of-river hydropower plants. Reservoirs are unique nodes in that inflows may be stored for later release; each reservoir also has an associated hydropower plant, and most have offtake possibilities. Other nodes ire included because reports on water balances are desired for comparison with similar data available elsewhere. Each node is constrained to be in balance on a monthly basis. That is, the sum of inflows from upstream nodes plus runoff plus return flows from previous diversions plus releases from storage (if a reservoir) must equal flows to the next downstream node plus diversions to agriculture and M&I plus retained storage (if a reservoir) plus losses. 28. The four major irrigated crops-wheat, corn, cotton, and paddy-are reprsented in each of the 10 regions. Based on available irrigated area, a fixed cropping pattern, and given monthly crop water requirements net of effective rainfall, a "desired" diversion pattern is determined. According to the availability of water in the system and its opportunity value elsewhere, three outcomes are possible: (a) full yield is obtained and  -9 - the entre irrigated area is cropped; (b) the entire area is cropped, but yields are reduced because of water shortages; and (c) yields are reduced and part of the irrigated area is operted as rainfed. Thus the model endogenously determines both the irrigated area of each crop and its resultant yield given the water supply. M&I uses are specified as demands and include certin extrabasin ansfers; they are given first priority among demands. Hydropower energy output is computed endogenously for each reservoir and major run-of-river plant, and exogenously for other run-of-river plants, in each month. For reservoirs, energy output is a function of the net head and the discharges through the powerhouses. Net head is endogenously determined by functions relating reservoir storage and elevation, and discharges and tailwater elevations. Discharge through the powerhouses is determined as releases from the reservoir up to the limit of powerhouse capacity. Net head is assumed to be given for run-of-river plants, thus energy output from them is computed directly, given the flows in the relevant river reaches. 29. Flood protection, sediment control, and environmental protection measures are included in the model as either restrictions on reservoir operating procedures, or as constaints on reach flows. For each reservoir, upper and lower operating limits constrain the storage permitted in each month. Flood protection measures imply that the downstream reservoirs must be near dead storage at the onset of the flood season, and must remain low until the flood season has passed. In the reaches prone to ice jams, river ftcws must be restricted (through reservoir control) in the coldest months to limit flooding. Minimum flows in each month are attempted in the last reach to minimize ecological damage to the estuary. 4. ECONOMiC USES OF YELWW RiVER WATE Irrigation 30. The most important economic use of Yellow River water, by far, is for irrigation. From a very low base in 1949, nearly 39 percent of the cultivated area was irigated by 1987. The expansion of irrigation was not without its problems. Many of the early programs were carefully conceived and well managed, and function effectively to this day. However, some of the poorly planned and overly ambitious programs implemented during the late 1950s had disastrous consequences. Without adequate drainage, waterlogging and salinity made large areas nonproductive. In other areas, groundwater mining led to rapidly falling water tables. During the early 1960s, a drought combined with these failures led to China's greatest agricultural catatrophe. Reportedly, millions died of starvation, particularly in the lower reach. Since then, the lessons have been taken to heart. Most projects are carefully managed from an ecological, as well as hydrologic, point of view. Not all, however, are managed to maximize economic output from limited water supplies. 31. Today, there are over 110 major irrigation projects in the basin, covering over 85 million mu (39 percent of the total cultivated area). The importance of irrigatioa to fte basin's economy cannot be questioned. The basin is roughly self-sufficient in grains  - 10- and many other commodities, although per capita grain production is stll less than the national average. Irrigated grain production accounts for over 68 percent of total grain production and most of this is from Yellow River sources (see Table 1). In most areas, irrigated grain yields are estimated to be nearly double those of rainfed. In the poorly endowed Loess Plateau, they are nearly triple. 32. The differences in productivity between irigated and rainfed areas are due not so much to the different average volumes of water available as to the timing of the water that is applied. Most of the rainfall occurs in the late summer months, while water requirements peak in the spring, early summer, and late fall (see Figure 3). These patterns lead to very high returns to water applied at critical times, and hence translate into high returns to irrigation. Flgure 3: RuNOFF VESUS DEmAND FOR WATR 10 j ~ ~~~~~ n gRunoft V D