China - Second Tarim Basin Project : environmental assessment

E2112 INTEGRATED ENVIRONMENTAL MANAGEMENT IN THE TARIM BASIN Prepared by OPCV for the World Bank A 'l'echl~icalAssistance Project Funded by AUSAID JULY 1999  INTEGRATED ENVIRONMENTAL MANAGEMENT IN THE TARIM BASIN Prepared by OPCV for the World Bank OPCV: Hug11Cross (report leader) Jcff Ball Dugald Black Lee Bowliilu, Jean Hung Ary van der Lely Chinese counterparts are acknowledged in each Working Paper Overseas Projects Corporation of Victoria Limited - Australia Level 1, 590 Orrong Road, Armadale, Melbourne, Victoria 3143 Australia A 'Technical Assistance Prqjcct Fundcd by AUSAID Australian Age~~cyfor International Development Australian Embassy, 21 Dong Zhi Men Wai Dajie, BEIJING, P.R. China 100600  PREFACE This Technical Assistance (TA) is entitled, "Integrated Environmental Management in the Tarim Basin (IEMTB), Xinjiang Province (XP or Xinjiang Uygur Autonomous Region), China". Funding was provided by AusAid in response to a request by the World Bank (WB) to facilitate its support of the Chinese authorities' wholistic planning of water resource rehabilitation and use in the Tarim River Basin (TRB) in relation to the WB hnded "Tarim Basin I1 Project" (TB 11). The TA is to ensure as far as possible that the sustainable environmental management goals of the Governments of the Peoples' Republic of China (PRC), XP and the relevant Prefectures, can rest on the best available information and an objective assessment of the value of the Lower Green Corridor (LGC) of the Tarim River (TR) in ecological and socio-economic terms. ABBREVIATIONS BEPB Baying01 Environmental Protection Bureau GC Green Corridor from Aler to Taitema ha hectares (10,000 m2 = 15 mu) XIDRIWRHP Xinjiang Investigation Design Research Institute for Water Resources and Hydro-Power IEMTB Integrated Environmental Management in the Tarim Basin LGC Lower Green Corridor or GC from Qiala to Taitema. mu mu, the Chinese unit of area equal to 1/15 of a hectare m3 cubic metre (1000 litres) OPCV Overseas Projects Corporation of Victoria Pty Ltd PD Populus diversijolia (Schrenk), commonly known as diversiform-leaved poplar, or poplar diversifolia PRC People's Republic of China TA Technical Assistance TB I1 Tarim River I1 Project TR Tarim River TRB Tarim River Basin .TMB Tarim Management Bureau TBWRC Tarim Basin Water Resources Commission WB World Bank XP Xinjian Province (or Xinjian Uygur Autonomous Region) XPCC Xinjiang Production and Construction Corps (generally refers to Division II)  TABLE OF CONTENTS EXECUTIVE SUMMARY ................................................................................................................ 1 1. INTRODUCTION ........................................................................................................................ 1 2. WHAT IS THE C0NL)ITION OF THE LOWER GREEN CORRIDOR? ............................2 2.1 .................................................................................................................... 2 '['LIE "l~?\l<lhlI<l\!t:I< IIO'I'I'O I"' "' 2.2 \!11~'1' IS 1~141:C'III<I<I;N'I(.ON111 I ION 0 1. '1 111' 1 .OhI:I< Cil<l..1.N C.OKKII)OR'.'.....................................................2 2.3 WI~~I.I'.AKI. HIF;(.AI!S,II . ~ I I ; ~ ~ I I I I N I......................................................................................................4 ~ ! V I ~ ? 2.4 Li"l1.4,l'IS 'rll1; M/\KA(il:LIEN~I SI'I'II,\'I I( )N'! ...................................................................................................7 3. WHAT METHOD WAS USED TO VALUE THE LOWER GREEN CORRIDOR? ...........8 4. WHAT IS THE VALUE OF THE LOWER GREEN CORRIDOR? .................................. 10 4.1 S ~ C I O - ~ : C O N\'A1~ ~ I C....................................................................................................................... 10 O LIES .. 1.1.I We/it.ccn-I1~~~chIi~l/rle-~!fls ..........................................................................................................10 4.1.2 (.'ui.i~eilfSocio-l..'conornic I'nllie .................................................................................................. I 0 4.1.3 I'ofcnfiul .\`o~~io-l~.'c~onon~i~~krllics........................................................................................ I 2 4.2 I.cr)l.ocilc.il VALIILS ...............................................................................................................................I h 4 2.1 (;cn~rtrll'01lic.v. ............................................................................................................. 16 4 2.2 I.1.'/111/(JIIL'I ~ C.S/?C(. I"~~nc/ion.~h I.(.(.? .................................................................17 i/ic (/ r ~ 5. WHAT CAN BE DONE TO SAVE THE LOWER GREEN CORRIDOR? .........................19 5.1 M'.\.I'f.f< 1111 OCi\'l`lotis.............................................................................................................................19 5.2 LY'l1.2'1-r \ ~ < . ) ~ . ~ . l1-111:~) ~. 0 l ~ l~~; lI < E ~ C:01<1<11)01<C,l\ii 13t SAVI:~)'! ' ~ . : N ........................................................20 . . 5 2 1 /'i.ri~iurrsI:.sf!i~lotc.r....................................................... .......... .......................................20 5.2.2 Ke~ieu:of I:\isling I<.rliiilafe.s.............................................................................................. . 21 5.3 L\l 11:rui C"ZK'1'1 I E WAI'EICt 3 ~SI.I'I'L.IEL) : I.I<oM'.' ........................................................................................ 23 5.3.1 Ne~:ie,t!c?f'l'~.iori/ie.r.(111d lichabilitn~io~ikaiciol~.ir.s....................... ......................................... 33 5.3.2 I'otenliul ,\'u~r~~ce.s ................................................................................................................ 24 6 . MONITORING FRAMEWORK ..............................................................................................25 6.1 KI:Y 1 ~ ~ 1 .Al .NsI ) C:~NCI:I~TI~!\~ . MOI)EII)I~VI:I.OI~MEN~` . ...........................................................................2-5 0.2 Cjl:OMoI<l'l ILL' I'.4I<AMtS'I'l;I<S ........................................ .............................................................................28 6.2.1 KL'\; I.SSIICS ..................................................................................................................... 28 6.2.2 .\./oiritoi.iirl: I..izr17ls1r.c~1.k. ............................................................................................... 28 6.3 1 l ~ n n o 1 ocirc: I'AI~!\MI<I'I.I<S .....................................................................................................................29 6.3.1 .J'ii~.jbcc ~?.'UI(?I ......................................................................................................... 21) 6.3.2 .Sl1rfi1cr I.lilf<t.Ql~(llity ......................................................................................................... 20 6.3.3 Soil rznd (;l.ounrh~.u/c>l-Purutn~~rei:\ .......................................... .. ............................................32 6.4 I i c o l . o c ; l c ~ l .~ ' , ~ I L \ ~ ~ I I - " I ~ ~ ' K s ...................................................................................................................... 33 6.4.1 Key 1.s.tUC'S............................................................................................................................. 33 6 4.2 .l.loni/oring I'i.otn~.i.l.ork...... .......................................................................................................33 6.5 S o c ~ o - l ; c ' o ~ oI'AJIAM~I'EIIS ~ ~ c ..............................................................................................................36 7 . CONCLUSlONS ......................................................................................................................... 37 8 . RECOMMENDATIONS ...........................................................................................................39 9 . REFERENCES ........................................................................................................................... 41 10 . ANNEXES 1 .3 ........................................................................................................................... 44 .. 11  TABLES FIGURES ~I(iOl<l.:1 : -rill! l.OC~/\'l'll~)N 'l'lll: [dO\irl'!.l<<il<l:l!~c O R ~ l l ~ i );\Nl> Kl:Y l ~ l ~ ~ \ " l ' l ~ l < l ~ S . 01: l < 3 r:lc;l:rilr1:I'i\RIM RIVEIIhllN1:l<:ZI.lSr~~fION:2f AI;E:.R. XINQIMAN AND C)IAl.i\ 5 Flc;~;i<t!3: 'I.!\I<IM I'<lvt.~<!I\/II<!I(~I I:I.(:)U' vol .L;\:I~.:.s 1 3 I)I;CAI~I:, ~ 6 FI(;L:RE4: I.YI.>I:sV,II.LJI!Ss[!I>I)I)I< '1'111':l2o~j1:l<GRI:I:SCOI<I<II)C)I< 01: 1.1.1)IIY 9 F1c;r IRE 5: CC:)LIPONLNT SYSTLM P)IA(ilZAblS 1:OR 'I'IIIi "NATI::RALPRC)CPSSCS" C(.)NC:~':I"~I.~~L.27 Mon1:r. WORKING PAPERS I. Project TOR 2. Ecological Iafornlittion and Natural Features 3. Surface Hydrology 4. S~rfitceWatcr Qnality 5. Groundnitter Resources 6. River Modelling Requirements 7. Socio-cconomic Valuation and Kcstoration Evaluatioli 8. Worltshop Findings iii  EXECUTIVE SUMMARY The Lower Green Corridor (LGC) is that area of the Tarim River and its floodplain, downstream of Qiala. It is approximately 428 km long and generally 3 to 10 km in width. By comparison, the entire Tarim River is 1,320 km long and has a floodplain up to 30 km wide. The LGC used to support about YI of the Tarim River's total area of poplar diversifolia (Populus diversifolia, Schenk), in addition to Tamarix woodlands and grass and shrub lands. These woodlands and grasslands were important resources for the Uygur herders (about 2000 in the 1950s) who depended on the natural cycle of flooding to maintain these ecosystems in a healthy condition for grazing, fishing, wood, herbs and other needs. There are now more than 40,000 people on State Farms in the LGC, but very few herders because of the almost total loss of grazing land over the past 20 years. Reductions in the average annual volumes of water supplied to the LGC are the main cause of the decline in its ecological health and grazing land condition. Although this has been occurring along the entire GC the impact is most noticeable downstream of Qiala and particularly downstream of Daxihai. Within the LGC almost all trees between Alegan and Taitema Lake, the lower third of the LGC, are dead and those between Yinsu and Alegan are severely stressed or dying. Taitema Lake is dry and there is no pasture. In the middle third, between Daxihaizi Reservoir and Alegan, only the trees and pasture near the reservoir are in moderate condition, due to groundwater fed by seepage from the reservoir and irrigation. The upper third, between Qiala and Daxihaizi Reservoir, is also in moderate condition, again due to seepage to groundwater from the Qiala Reservoir and irrigation. Secondary impacts of the flow reductions are over- grazing of the degraded pastures and secondary salinisation due to lack of soil flushing and increased river salinities. It is not possible to save the entire LGC without reinstating annual flow volumes that approach the magnitude of those that occurred in and before the 1950s. Upward trending river salinity levels also need to be stabilised and reduced. Grazing pressures need to be matched to the reduced carrying capacities. Significant rehabilitation, however, can be achieved if average annual flows of 300 x 106m3 can be reinstated as provided for in the recent Water Allocation Agreement, agreed to by the Prefectures at the "Second Standing Committee for the Tarim River Prefecture Water Use" held in January, 1999 - a committee of the Tarim Basin Water Resources Commission. Although an annual average flow of 300 x lo6m3 is about 'A of the average flow in the 1950s and perhaps 118 of the flow prior to any irrigation use in the Tarim Basin, it is estimated that 300 x lo6 m3 would largely restore the condition of the trees and pasture from Qiala to Yinsu, with partial recovery between there and Alegan. There would be little benefit downstream of Alegan and effectively no ecologically useful change at Taitema Lake. Initially such flows would need to be augmented to raise groundwater levels that have been reduced over the past 20 years since flows to the LGC effectively ceased. If these initial flows cannot be augmented, then it will take longer for the restoration to be achieved. In addition to restoring flows, there is a need to spread and hold water on the floodplain for sufficient time for the natural flora to re-establish. Natural floods had the high discharges needed to do this. The restoration flows envisaged do not, and artificial structures in the Tarim River coupled with flood runners will be required if the height of the water in the river is to be raised and the water is to be spread over the flood plain. Executive Summary 1  Quantifying the socio-economic value of the LGC, however was restricted due to limited and incomplete data. In socio-economic terms sufficient evidence was gathered to suggest that the case to restore, save or even sustain the LGC downstream of Daxihaizi is very weak. None of the options investigated for delivering water to the LGC had a positive NPV at the discount rates considered. In effect all options will impose a cost on the Chinese economy if implemented. That cost is significant, in two ways. First, there will be lost socio-economic opportunities. It will cost the Chinese economy to save or partially restore the LGC and these finds could be better used to benefit the Chinese economy if they were used in other productive ways. Second, and due to transmission inefficiencies, unless the water required is sourced from the Kongque River, attempting to save or partially restore even some of the LGC, will cause more GC to be lost in higher reaches than can be saved or restored in the LGC. Sustaining the GC along the higher reaches of the TR, however, has a relatively strong socio- economicjustification Analysis of the current condition of the LGC, its value in ecological and socio-economic terms, and the underlying causal factors is severely constrained by the available data. No quantified flow records exist prior to 1957 anywhere on the Tarim River but large scale irrigation water use existed prior to the 1950s. Hydrologic modelling is therefore needed to estimate the scale and nature of pre-1950 flows and link these with post 1950 data. There is only one flow recording station in the LGC and this is not a National Level station. 'The quality of data from this station might not match that of the National Level river monitoring stations at Aler and Xinqiman. There is even less information on water quality and biological parameters. Data constraints are not unusual in river basin studies but in this case finding arrangements have caused data to become a "commodity" to be purchased to augment agency funds. This has given rise to a lack of interagency data exchange and a degree of adversarial competition between agencies to protect their perceived areas of responsibilities and expertise. This is not an efficient use of data or resources and results in a large, long-term economic cost to the Tarim Basin and the XUAR through inefficient policies and poor use of available natural resources. It seriously constrained this analysis and prevented the development of a better understanding of how the system works and how the prevailing condition of the GC in general and the LGC in particular came to be. It is an important issue because river basin management, by nature, is multi-disciplinary and requires a great deal of cooperation and interaction that at present simply does not exist. The greatest need for monitoring in the future is to first correct data access and sharing arrangements so that a multi-disciplinary, inter-agency approach can make f i l l use of past and hture data. Without such changes there will be little value in collecting more data. There is aneed for much stronger coordination and oversight of water quality monitoring in the Tarim Basin. The TBWRC should have a much stronger role (similar to the Murray- Darling Basin Commission in Australia) in this regard and be responsible for coordinating Basin-wide data collection and management. The actual monitoring should be conducted by existing agencies within this coordinated holistic framework provided by the TBWRC. Second, new data need to be collected with specific issues and hypotheses of the biophysical, socio-economic and management /administrative systems in mind. All three of these Executive Summary 2  hypotheses need to be developed in an interactive manner involving all relevant stakeholders and managers. The TMB will require additional support for it to manage such processes, coordinate data gathering and archiving, and use the analytical modules and systems needed to support TBWRC operations. Executive Summary  1. INTRODUCTION 1.1 PROJECT CONTEXT The "Lower Green Corridor" (LGC) is the area naturally and artificially irrigated by the Tarim River, downstream of Qiala Reservoir to Taitema Lake, i.e. it is the Tarim River and its floodplain, including flood channels and "dry rivers". This area is surrounded by the Kuluk Desert to the east and the Taklimakan Desert to the west (Figure 1). A full description is provided in Section 2. The World Bank funded Tarim Basin I Project (completed in 1997) and Tarim Basin I1Project (in progress) are intended to achieve socio-economic benefits for poor farmers through sustainable rehabilitation and development of irrigated agriculture. The projects also aim to improve the greatly deteriorated environment of the "Green Corridor", which, at least at one time, had considerable socio-economic and ecological values. This Technical Assistance (TA), "Integrated Environmental Management in the Tarim Basin, Xinjiang Province, China", relates to whether and 'how water saved through the structural and management measures of the Tarim I1 Project could be used to improve the environmental condition of the LGC and thereby socio-economic conditions. The TA was also to quantify the ecological and socio-economic values, so that rational and objective decisions can be made about the reallocation of the scarce water resource saved through component projects of the Tarim Basin I1 Project. The five objectives of the TA are listed in Section 1.2. Package 1 of this TA is to provide a preliminary assessment of these values and criteria, and establish a framework for an environmental monitoring program that can be conducted through the Tarim I1 Project. It also recommends specific environmental research studies to be conducted as part of Package I1 of this Technical Assistance. Package I1 is tentatively scheduled to begin in late 1999. This report is part of Package 1. 1.2 OBJECTIVES The Tarim Basin Green Corridor Environmental Management component of Package I has the following objective: To identifi the broad constraints within which current andfuture development can occur without further destroying the Tarim River's environmental assets and to identlfi opportunitiesfor restoration or enhancement. Through the preparation of an environmental baseline study, five specific objectives will be addressed: 1. To determine the major changes in the extent and character of the Tarim River and its "green corridor" since the inception of large scale water use and river regulation. 2. To make a preliminary assessment of the value of the river and "green corridor" in ecological and socio-economic terms. 3. To determine the broad water quality and quantity conditions that will maintain or restore the ecosystems of the Tarim River and its "green corridor" to levels that match agreed local and regional expectations. 4. To input environmental values, constraints and management options to the development of the Tarim Basin Master Plan primarily at the Insight Workshop.  5. To identify a framework for the development of monitoring needs and make recommendations for the environmental studies to be conducted in Package 2. 2. WHAT IS THE CONDITION OF THE LOWER GREEN CORRIDOR? 2.1 THE "TAIIIM RIVER HOTPOT" The Tarim River (TR) can be likened to a frog in a pot of cold water. The story goes that if you heat it slowly enough the frog will not notice that the water is getting warmer and will remain oblivious to the danger until it is cooked. Similarly, changes to the TR have been very gradual. The damage was not done "yesterday". It has developed progressively over much of this century and started even before river flow and other records began to be kept. And now the "Tarim River Hot Pot" is more than half cooked. It is almost forgotten that the TR used to flow, not just to Taitema Lake, but joined by the flows of the Qarqan River, flowed a considerable distance further, at least in wet years, to Lop Nor Lake (Figure 1). That lake was two to three times the size of Bosten Lake earlier this century, but by the 1970's was dry. The reduction in TR and Kongque River flows were both responsible'. The deterioration of the Lower Green Corridor (LGC), i.e. that section from Qiala to Taitema ~ a k eis~just the most visible sign of the problem. The remainder of the Tarim River's (TRYs) , Green Corridor (GC),from Aler downstream and those of its tributaries, are also significantly degraded, but because of the "s1owness of the cooking", it has largely gone unnoticed and consequently there has been little retrospective action. There is a danger in concentrating on the obvious deterioration in the LGC, the complete lack of water downstream of Daxihaizi Reservoir, the dead and dying trees, the encroaching desert, and ignoring the effects of increasing river salinities, land salinisation and deteriorating groundwater levels and quality. These are secondary causal factors in the deterioration of the whole GC, not just the LGC, due to the primary cause of reduced water availability. It is as important to understand the mechanisms of deterioration, as it is to know the extent and severity of the consequences. By understanding the mechanisms, predictions can be made and early warning signs recognised in areas less severely affected. Management can then be targeted to address these problems knowing the full scale of the issues, what integrated management is needed and what trade-offs are necessary. It also provides the advantage of proactive management in being able to prevent further deterioration, rather than attempt restoration after the damage has been done. 2.2 WHAT 1S THE CUlIRENT CONDlTlON OF THE LOWER GREEN CORRIDOR? The lower one third of the LGC, between Alegan and Taitema Lake is already largely dead. There is no pasture and most of the poplar diversifolia trees are dead or nearly so. Mobile ' The Kongque River flowed separately to Lop Nor Lake via a more northerly route, but also used to contribute flows to the lower TR in the vicinity of Qiala, via Aksupu Swamp and Tiganlike, via the Ailik River (now dry). The former extent to Lop Nor is being ignored in this study in terms of any attempts to recover it. The volumes of water required would approach those volumes that originally flowed. The re-allocation of water on that scale is outside the terms of reference of this study and is, to all intents and purposes, impractical.  dunes have formed over some of the 7,000 ha of abandoned irrigated farmland and in the bed of the TR at Alegan. BOSTENM E KONGQUE RIVER LOP NOR M E - m PREFECTURE FARQAN RIVER a Ruoqiang LEGEND: .... lake existing monitoringlocation '. ,. '. swamp / wetland - ,,. 0 dry lake 0 proposedmonitoringlocation -river . - - -dry river Figure 1: The localion of 11ieL.o\ver Cireen Corridor and key features. In the middle third of the LGC, between Daxihaizi Reservoir and Alegan, the pasture has deteriorated to the extent that only six of the original 28 grazed species remain and there is only temporary occupation of Yinsu each year by a few families. Grazing is now restricted to 11,000 ha of the 278,000 ha of grass and bush area. It is being maintained almost entirely by groundwater. However, the groundwater levels are progressively becoming deeper and most of this remaining area, including the tree cover, is likely to be lost to within a short distance of Daxihaizi Reservoir if regular flooding is not provided within 5 to 10 years4. Water levels in wells in the river bed at Yinsu are more than 6 m deep, or more than 10 m below the general floodplain level5. Between Yinsu and Alegan tree condition is poor and there is evidence of wind erosion of alluvial sediments, reflecting the lack of grass cover and absence of flooding. Currently the mature trees are still in moderate condition upstream of Yinsu, but there is essentially no establishment of new trees. ' This "original" area is calculated from the "Map of the Landscape of the Tarim River Drainage Area". A full reference is provided in the References section. The date of the photography or satellite imagery that this map is based on is not known, but is likely to be relatively recent. Therefore the original extent of grassland in particular (less so for forest), could have been greater, i.e. the areas quoted in this report based on this map are likely to be conservative. This is a project estimate based on the biological information provided and the time which it has taken for the currently observed deterioration in tree health to occur at Yinsu, i.e. within 10years of the last decent flood. Obsen~ationduring a TA field trip 7 April 1999.  The upper third of the LGC, between Qiala and Daxihaizi Reservoirs, is in moderately good condition due to both surface and groundwater flows being maintained by irrigation water, albeit with different time and spatial patterns relative to the natural flows and floods that once characterised the TR. The flow of surface water in the river and in the main and subsidiary irrigation canals, all assist in maintaining groundwater levels, as do the major area-based recharge sources of the Qiala and Daxihaizi Reservoirs and the irrigation areas of the five State Farms. River salinity levels range between 1 - 3 g/l (measured as "mineralisati~n"~).They appear to have increased since the 1960's. Land salinisation is a significant issue. It is perhaps less prevalent in the upper third of the LGC than in the middle GC although relatively high salinity levels in the upper third of the LGC have forced the use of deep groundwater for drinking water. 2.3 WHAT ARE THE CAUSAL RIECHANISMS? The entire TR has an extremely arid continental climate and therefore experiences high summer temperatures of 30 to 4 0 ' ~daily maximums and evaporation in excess of 2 myand winter temperatures down to -30'~. There is less than 50 mm of average annual precipitation. Therefore all water is derived from river flows and the observed environmental changes have been primarily driven by the reduction in these flows. Land management issues, such as tree cutting and grazing, have exacerbated the situation brought about by the water shortage as the area and lushness of the GC has reduced. River flow records since 1957 show that flows in the 1990's at Qiala are only one third of what they were in the 1950's and virtually all of this is used by the State Farms. Therefore there is no flow currently downstream of Daxihaizi Reservoir. In fact the reduction in flows is even greater, as significant areas of irrigation were already present in the tributaries prior to 1957. Cheng Qichou (1993) records in The Research on the Tarim River, that average annual flow of the Tarim River (at Aler), prior to any irrigation development, was likely to have been 10.0 - 12.0 x 10' m3, i.e. about double the average over the last 50 years (4.60 x lo9 m3) and more than double the average for the last decade (4.17 x 109 m3). River modelling, using historical daily data, is required to determine the true situation of natural flows and to determine changes induced by human water use (Working Paper 6). The situation at Qiala is much worse. Relative to the 195OYs,the current flows are only 116th of the original and relative to the pre-development flows are perhaps only 1/12. In fact of the 211 x lo6 m3 currently supplied to the LGC, about 150 x lo6 m3 is water purchased from the Kongque River by the State Farms and diverted via the Kuta Main Canal. Apart from two small releases in 1995, totalling 28 x lo6 m3, which were largely confined to the river and didn't reach Yinsu, there have been no floods downstream of Daxihaizi Reservoir since the mid 1980's. Hence the remaining ecosystems have been relying on progressively decreasing groundwater levels. The term "mineralisation" was not explained in the material provided to the TA. It has been assumed that it represents Total Dissolved Solids, TSD, or something close thereto, despite the following reference citing it as being the sum of the cations only: Xiangcan, Jin 1995 Lakes in China: Research oftheir Environment. Vol 1. China Ocean Press. pp 278-3 19). 4  The reduction in river flows has been accompanied by increasing river salinity, particularly in spring and early summer before the summer flood dilutes the saline irrigation drainage water and groundwater. Salinity levels peak at up to 6 g/l in spring at Aler due to return groundwater and surface irrigation drainage. However, spring levels peak at only 2.8 g/l at Qiala, apparently due to flow attenuation and dilution by Kongque River water. Mineralisationat sites on Tarim River I I I I I I I I jan feb mar apr may jun jul aug sep oct nov dec Month 1 Figurc 2: Tarim Rivcr ~nincralisatiorlilt Aler, Xinqiman and Qiala There are insufficient data to conclusively quantify the changes over time, particularly as the diversion of water from the Kongque River has included releases to the Tarim River at unknown times since the mid 1970's and particularly since 1990. However, the data do indicate that salinity levels appear to have increased since the 1960's. The supply of water to the LGC is contingent on the total volume supplied to Aler and also the transmission efficiency of the upper and middle Tarim River reaches. As previously stated, all flows in the Tarim River are supplied by its tributaries; the Hotan, YarkantKashgar Rivers and the Aksu River at its head, and the Kongque River at the upstream end of the LGC. The Weigan River once contributed relatively small volumes (200 x lo6 m31yr)but no longer does so. As discussed below, average annual flows to the Tarim River are now less than half those which are estimated to have prevailed prior to any irrigation in the TB. Much of this reduction took place prior to the commencement of flow records in 1957. The middle and upper reaches of the GC, ie. from Aler to Yinbazar and thence to Qiala, have shrunk laterally and become drier in most locations. The exceptions are relatively small areas now receiving more water than previously due to enlarged overflow locations made by excavations. These diversions have been dug by local people, including herders, in response to the reduced flooding frequency and resultant decrease in pasture extent and health. However, because these diversions have no regulators they result in much higher water losses at low and medium flows than previously. As discussed below, the increase in low flow losses is almost certainly more than equalled by the reduced losses at high flows, due to the decrease in flood size and frequency caused by irrigation use in the tributaries.  Figure 3 shows the reduction in annual flows by decade from the 1960's to the 1990's. Flows reaching Aler have decreased by about 1 x 10' m3 over the period. Taking into account the 150 x lo6 m3 transferred from the Kongque River (included in the volumes presented in Figure 3), there has been a similar annual reduction at Qiala. Therefore, although it appears as if losses have increased between the two stations, because the proportion of flows at Qiala is now less relative to Aler, this is only a proportional difference and is not a true change in the relationship in annual flows between the two. Hence the apparent increase in the volume lost in the upper and middle reaches, often cited by some stakeholders as being due to increased floodplain losses, is in fact largely due to irrigation use upstream of Aler. The upper and middle GCs are as much a victim of reduced tributary flows to the Tarim River, as the LGC. Working Paper 3 provides a more detailed temporal and spatial analysis of the flow relationships between the stations along the TR. The pattern of losses within years could not be analysed using the available monthly data. Daily data exists for Aler, Xinqiman and Qiala from 1957, but was not provided to the project team because of "confidentiality", "intellectual property" and the size of the requested payment of RMB 1.6 x lo6 yuan (for Aler and Xinqiman data only). Daily flow data is required to determine the changes in low, medium and high flow losses, as individual events would need to be "tracked" downstream to resolve the loss component. However, given that some excavated floodouts (distributary channels) have inverts lower than the river bed, it is reasonable to conclude that losses at low flows are now greater than previously. However, these are unlikely to equal the reduction in high flow losses due to reduced flood size caused by upstream water use. In fact it is this reduction in high flow losses that has in part hidden the total increase in consumptive water use of 2.2 x 10' m3 upstream of Qiala. In addition to the 1 x 10' m3 increase upstream of Aler, discussed above, most of the Aler to Qiala allocation of 1.25 x 10' m3 between has developed since the 1960's. If were not for a reduction in overall annual losses, annual flows at Qiala would be negative, even with the average annual transfer of 150 x 1 o6m3from the Kongque River to Qiala. TARIM RIVER AVERAGE nowVOLUMES BY n E C l A n E 60