This is an electronic reprint of the original article.This reprint may differ from the original in pagination and typographic detail.

Powered by TCPDF (www.tcpdf.org)

This material is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form. You must obtain permission for any other use. Electronic or print copies may not be offered, whether for sale or otherwise to anyone who is not an authorised user.

Varis, O.; Kummu, M.; Lehr, C.; Shen, DajunChina's stressed waters: Societal and environmental vulnerability in China's internal andtransboundary river systems

Published in:Applied Geography

DOI:10.1016/j.apgeog.2014.05.012

Published: 01/01/2014

Document VersionPublisher's PDF, also known as Version of record

Published under the following license:CC BY-NC-ND

Please cite the original version:Varis, O., Kummu, M., Lehr, C., & Shen, D. (2014). China's stressed waters: Societal and environmentalvulnerability in China's internal and transboundary river systems. Applied Geography, 53, 105-116.https://doi.org/10.1016/j.apgeog.2014.05.012

https://doi.org/10.1016/j.apgeog.2014.05.012https://doi.org/10.1016/j.apgeog.2014.05.012

lable at ScienceDirect

Applied Geography 53 (2014) 105e116

Contents lists avai

Applied Geography

journal homepage: www.elsevier .com/locate/apgeog

China's stressed waters: Societal and environmental vulnerability inChina's internal and transboundary river systems

Olli Varis a, *, Matti Kummu a, Christian Lehr a, c, Dajun Shen b

a Water & Development Research Group, Aalto University, PO Box 15200, 00076 Espoo, Finlandb School of Environmental and Natural Resources, Renmin University of China, Beijing 100872, Chinac Institute of Landscape Hydrology, Leibniz Centre for Agricultural Landscape Research, 15374 Müncheberg, Germany

Keywords:ChinaRiver basinSocio-economyEnvironmentSustainabilityVulnerability

* Corresponding author. Tel.: þ358 503516484.E-mail address: olli.varis@aalto.fi (O. Varis).

http://dx.doi.org/10.1016/j.apgeog.2014.05.0120143-6228/© 2014 The Authors. Published by Elsevier

a b s t r a c t

China is undergoing a rapid transition from rural to urban dominated economy. Economy is booming,social structures are changing, ecosystems are stressed, and sustainability is challenged. We analysed thesocioeconomic and environmental vulnerability of river systems that are entirely or partly located in thecontinental part of China. One-third of the mankind inhabit the area covered by this study. Six stressfactors (governance, economy, social issues, environment, hazards and water stress) were analysedseparately and in combination as an overall vulnerability. China's most vulnerable parts were found to besituated in the lower Hai and Yellow River basins, with their high population density, low water avail-ability and high human footprint. The other water-stressed areas in the northwest showed highvulnerability, too, and so did the water-rich coastal areas due to high population density, natural hazardsand high human footprint. We went beyond existing water stress and vulnerability studies in threedimensions. First, our perspective was highly multidimensional and thus very relevant in addressingChina's water challenges in a realistic and multifaceted way. Second, we combined administrative andriver basin scales and used an essentially higher spatial resolution than done so far. Third, we includedthe transboundary dimension, which is not customary. This is highly important since one billion peopleChina's neighbouring countries, in basins that are partly in China.© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Introduction

China has been undergoing stunning economic and socialdevelopment for several decades now, and the trend continues.Between 1980 and 2010, the country's Gross National Income (GNI)grew 15-fold, the poverty headcount ratio (1.25USD per dayadjusted with purchasing power parity) fell from 60% to 15% andilliteracy rate from 22% to 6% (World Bank, 2014). Meanwhile,China's urban population grew by 119%, CO2 emissions increased2.4-fold and industrial water withdrawals 3-fold (World Bank,2014). China being the world's most populated country withextreme population densities in large areas, this developmentcontinues to set the sustainability (in the sense of balancing envi-ronmental, social and economic development) e or harmony be-tween nature and man, as Chinese often say e in question (Cao,Chen, & Liu, 2007; Economy, 2004; Zheng & Dai, 2013).

Ltd. This is an open access article u

China has a long history of seeking harmony between humansand the nature (Cao et al., 2007, 2013; Zheng & Dai, 2013). Therespect to nature in China has ancient roots and dates back at leastto the Zhou Dynasty (1115e1079 BCE). At that time, the mostimportant leadership talent for an officer was to be able to skilfullymanage forests, rivers, mountains, birds, and other animals(Economy, 2004). The equally central role of water resourcesmanagement in present days, too, is clearly reflected in the fact thatvarious recent key political leaders have been water engineers. Thecontemporary political weight of water resources management inChina continues to be extremely high. This was demonstrated in2011 when China's most important annual policy document, theNumber 1 Document, was focused onwater (Gong, Yin, & Yu, 2011;Liu&Wang, 2012; Varis, 2011). Quadrupling the water conservancyinvestment from the past decade's level was proposed in that keypolicy document as the main handle to better water future (Liu &Wang, 2012).

Despite this long tradition in seeking harmony between manand nature, the country's water systems and aquatic environmentare highly stressed (Economy, 2004; Gleick, 2009; Liu & Wang,2012; Ran & Lu, 2012; Varis & Vakkilainen, 2001; Zhang, Chen,

nder the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Delta:1_given nameDelta:1_surnameDelta:1_given nameDelta:1_surnamehttp://creativecommons.org/licenses/by-nc-nd/3.0/mailto:olli.varis@aalto.fihttp://crossmark.crossref.org/dialog/?doi=10.1016/j.apgeog.2014.05.012&domain=pdfwww.sciencedirect.com/science/journal/01436228http://www.elsevier.com/locate/apgeoghttp://dx.doi.org/10.1016/j.apgeog.2014.05.012http://creativecommons.org/licenses/by-nc-nd/3.0/http://dx.doi.org/10.1016/j.apgeog.2014.05.012http://dx.doi.org/10.1016/j.apgeog.2014.05.012

O. Varis et al. / Applied Geography 53 (2014) 105e116106

Chen, & Xu, 1992). Today's gigantic challenges (Bawa et al., 2010;Jiang, 2009) are due to factors such as rapid urbanization, intensi-fication of agriculture, massive industrial development andbooming energy sector development. They all contribute togrowing pollution, watershed degradation and growing pronenessto natural hazards (Jiang, 2009; Ran & Lu, 2012; Varis &Vakkilainen, 2001). These challenges are boosted by climatechange: certain historical trends have contributed to the polariza-tion of China's water problems: arid and water-scarce areas,particularly the North China Plain, have become even drier thanbefore, and precipitation has increased in China's flood-pronesouthern part (Shen, 2010; Shen & Varis, 2001; Wang et al., 2012;Xu, Milliman, & Xu, 2010).

China is geographically a vast country with a high diversity inclimate, population density, economic prosperity, ecosystems andproneness to natural hazards. Moreover, the upstream parts ofseveral major Asian transboundary river basins (Red River, Mekong,Salween, Irrawaddy, GangeseBrahmaputraeMeghna (GBM), Indus,Ili, Ob-Irtysh and Amur) are in China's territory, making China'swater sector stresses and activities particularly relevant to itsneighbours.

As the analysed river systems portray a high diversity and areextremely intricate and multifaceted systems, the data to describetheir status, and the factors affecting that status, are not trivial is-sues. At the same time, there is an urgent need to produce infor-mation that is easily accessible to a wide range of audiences,particularly at the policy level (Asia Society, 2009; AsianDevelopment Bank, 2007). Obtaining a systematic and analyticview on the importance of various sources of vulnerability, weargue, is of high importance.

We aim to analyse China's river basins in light of the above-outlined array of entangled change processes and to identify therelated major bottlenecks to sustainable development of thecountry's waters. Moreover, we attempt to bring the produced in-formation into a form which is maximally accessible and useful inpolicy-making. For that purpose, an approach is required whichallows the analysis of the triple-bottom-line of sustainable devel-opment (social, economic and environmental aspects) and relates itto the political and governance capacity.

We chose to approach the above-outlined complexity from thedirection of vulnerability assessment. This is because Chinesewaterpolicy discourse increasingly includes the logic of adaptive man-agement, according to which adaptation policies should focus onthe reduction of vulnerabilities. We used the river basin vulnera-bility approach developed by Varis, Kummu, and Salmivaara (2012),which allows a joint analysis of six stress factors (or sources ofvulnerability): social, economic, environmental, governance-related, natural hazards and water stress. We base our analysislargely on a river basin classification system that is used extensivelyby China's ministries, by other major policy actors and by manyChinese scholars working on China's river basins (see e.g. Jiang,2009; World Bank, 2006; Xie et al., 2009). On top of that, we donot only use the river basin classification but include also juris-dictional boundaries into our analysis, since most policy-makingoccurs in jurisdictions and only in some special cases in river ba-sins. These extensions, we hope, will facilitate the usefulness of ourfindings and our approach, besides in scholarly work, in policy-making.

Thus far, vulnerability of China's river basins has not beenaddressed in such a multifaceted manner although internationalpolicy agendas and recommendations call for looking at water re-sources challenges and policies in a comprehensive and integratedway (Biswas, 2005; Varis, 2005; WWAP, 2009). The existingvulnerability studies concentrate on specific subject areas such asriver discharge (Lu, 2003), water scarcity (Huang, Cai, Zhang, & Cai,

2008; Xia, Qiu, & Li, 2012), droughts (Wang, He, Fang, & Liao, 2013;Zhang et al., 2013), groundwater (Yin et al., 2013), water manage-ment scenarios (Wu, Li, Ahmad, Chen, & Pan, 2013), urban areas(Strohsch€on et al., 2013) and climate change impacts on ecosystems(Ni, 2012).

Given China's highly dynamic economic and social situation, aswell as massive challenges with the sustainability of water re-sources management, we aim at providing a comprehensive andcomparable view of vulnerability of the river basins that are locatedentirely or partly in China, including a broad array of aspects thatwould allow addressing the present (and historical) quest in Chinatowards harmonious relation between man and the nature.

Materials and methods

Delineation of China's river basins: the CARU system

We analysed China's continental territory as 21 river systems(Fig. 1 and Table 1; Fig. S1, Tables S1 and S2 in Online Supplement),of which 16 drain to oceans (we call these subsequently ‘open ba-sins’) while five are endorheic (interior/closed) basins. Their totalsurface areas (according to Water Resources eAtlas, 2003) are5,133,062 km2 and 738,208 km2, respectively.

The point of departure of the delineation of our river basin unitswas the conventional spatial grouping of China's river systems intoso-called ‘planning units’. Most planning of China's water resourcesoccurs presently in nine units (Songhua-Liao, Hai-Luan, Yellow,Huai, Yangtze, Pearl, SE Rivers, SW Rivers, Inland Rivers; see e.g.Jiang, 2009; World Bank, 2006; Xie et al., 2009). We call those unitssubsequently as CPUs (Chinese Planning Units). We enhanced theresolution of this system considerably by using altogether 21 riversystems, based on two datasets of river basin divisions (USGS,2001; Water Resources eAtlas, 2003). At the same time weattempted to maintain the compatibility with the CPUs as far aspossible, and, besides, included administrative borders as anadditional layer in our delineation. These were done to maximizethe applicability of our approach and results in Chinese policymaking. As we combine administrative areas with river basins wecall subsequently our novel delineation as CARU (Chinese Admin-istrative River Basin Units).

Some of the major open river systems of China include smallclosed basins (Wuyur and Baicheng in Amur basin, Upper Yangtzeclosed basins in Yangtze Basin, and South Tibet closed basins in theGBM basin). Those basins were considered so small that they wereincluded in the surrounding major river system. The only exceptionwas the Ordos basin e surrounded by the Yellow River basinand sometimes seen as a part of it e that was included in theGansueInner Mongolia closed basins. Both these definitions are inalignment with the CPUs.

Again, to concur with the CPUs, the following configurationswere made to our CARU delineation. First, even though some partsof our GansueInner Mongolia closed basins may also be seen asparts of the Yellow River basin, wemaintained the boundaries usedin the CPUs. Second, the Hexi corridor in Gansu is sometimes seenas a part of the Tarim basin but we maintained it in the Gansu-eInner Mongolia basins. Third, the Hai River basin's boundaries areextremely difficult to define precisely, and numerous definitionsexist. Wemaintained the one used in the CPUs, except in the case ofthe Liao basin, which we excluded from the Hai system andconsidered it together with Northeast Coastal Rivers. This was donebecause we want to produce globally comparable river basinvulnerability information for the major global rivers, and, for thatpurpose, we chose to separate the Hai basin from the Liao basin.Both basins are extremely populated and in many ways quitedifferent in character. Due to reasons similar to those above, we

Fig. 1. Map of the 21 river systems under study. Detailed information of the river basins is presented in Table 1 and Table S1. Note: only the Chinese part of the transboundary riversis included in this study; CB refers to a closed basin.

O. Varis et al. / Applied Geography 53 (2014) 105e116 107

excluded the coastal basins of the Leizhou Peninsula from the Pearlbasin and included them to the SE River Basins.

Transboundary river basins

Over half of the river systems (11 among the 21) are trans-boundary (Sharedwith one or more neighbouring countries), while10 are entirely in China. The total surface area and population sizeof the transboundary basins outside China's territory is8199 � 103 km2, and 1044 million, respectively (Table 1). The

Table 1The investigated river systems: their area and population. See the map of the basins in F

Drainage Code Basin

East China Sea, Okhotsk 1 Amur2 Liao and NE Coastal Ri3 Hai4 Yellow5 Shandong Coastal Rive6 Huai7 Yangtze

South China Sea 8 SE River Basins9 Pearl

10 Red River11 Mekong

Indian Ocean 12 Salween13 Irrawaddy14 GBM15 Indus

Arctic Ocean 16 Ob-Irtysh

Interior (endorheic basins; i.e. closed basins) 17 Gansu-Inner Mongolia18 Qinghai CB19 Tarim-Junggar CB20 Ili CB21 Tibetan CB

TOTAL

Note: SE River Basins stand for Southeast River Basins, GBM stands for GangeseBrahmaData sources: Water Resources eAtlas (2003), LandScanTM (2007).

corresponding figures for all the river systems within China are9356 � 103 km2 and 1313 million people (Table 1). Adding up thepopulation figures, our research area thus covers the homes of over2.35 billion people, being roughly one-third of the mankind.

In the transboundary analysis part of our study, we included allof these basins except Ob-Irtysh, Amur, and other basins that aresharedwith Russia or People's Republic of Korea. This exclusionwasdone for two reasons. First, we have performed vulnerability ana-lyses of the other transboundary basins in previous studies (Varis&Kummu, 2012; Varis et al., 2012), but results are not available for

ig. 1 and Fig. S1A.

Area (103 km2) Population (millions)

In China Other countries In China Other countries

895.2 1201.9 64.4 4.1vers 418.9 e 73.4 e

243.3 e 105.2 e789.8 e 146.7 e

rs 52.1 e 29.9 e263.3 e 164.4 e

1770.5 e 439.1 e316.3 e 123.1 e447.9 10.3 113.6 1.484 73.5 11.7 18.1

168.3 647.4 6.7 64.1139.8 124.1 3.6 4.421.5 392.9 1.9 34.3

316.9 1320.3 1.7 645.686.1 1058.9 0.04 240.750.4 2960.6 0.4 28.2

CB 807.1 e 7.7 e319.3 e 0.5 e

1411.3 45.1 17.0 0.0757.7 364.3 2.2 2.9

696.5 e 0.2 e

9356.2 8199.3 1313.4 1043.9

putraeMeghna basin, and CB for closed basins.

O. Varis et al. / Applied Geography 53 (2014) 105e116108

Amur, Ob-Irtysh and other basins shared with Russia or People'sRepublic of Korea. Second, the latter-mentioned basins will be atopic of a separate analysis, in which they will be set within thecontext of other major North Eurasian rivers.

Data

As we combine river basins with administrational units in ourCARU delineation, we face a situation, in which a simultaneousoperation at two geographical divisions is essential, namely, naturalunits of water management (i.e. river basins), and administrativeunits (i.e. jurisdictions). This arrangement is important in practice,since the river basin context is highly relevant in the managementof water resources, whereas policy making and administrationoccurs typically at national, provincial, or some other jurisdictionallevel, which only in rare instances follow boundaries of river basins(Shen & Varis, 2000; Varis et al., 2012).

Consequently, two types of data were used: spatially griddeddata were available for environment, water resources and some ofthe social data while for policy, governance issues, and macroeco-nomic aspects data were available at, or scaled to, the administra-tive scale of provinces (see Table 2).

We refer to the article by Varis et al. (2012) for detailed docu-mentation of data and its selection rationale. Four modificationswere made to the approach by Varis et al. (2012), regarding to the

Table 2Data used for six vulnerability indices including the information of data description, resoluindex (see Table 3).

Source of vulnerability (indicator) Data description Res

Data by administrative regionsGovernance (Political instability) Political instability index consists of 12

indicators for Underlying vulnerabilityand three for Economic distress (EIU,2011).The data were not directly available forprovinces and thus, for obtainingprovincial estimates, we used nationalaverages for all other indicators exceptInequality, Ethnic fragmentation andLevel of income per head. Those wereobtained from China StatisticalYearbook (2011).

Cou

Economy (GNI (PPP) per capita) Gross National Income per capita(adjusted with Purchasing Power Parityto International Dollar)

Prov

Social (Multidimensional Poverty Index) The index (originally developed byAlkire & Santos, 2010) represents thenature and intensity of poverty at theindividual level in education, healthoutcomes, and standard of living.The MPI data were not available forprovinces. Provincial values wereobtained by scaling China's nationalvalue by province-specific humandevelopment Index components ofeducation and health (UNDP, 2010) aswell as level of water services (urbanand rural separately; China StatisticalYearbook, 2011).

Cou

Spatially gridded dataEnvironment (Human footprint) Global Human Footprint Dataset from

the Last of Wild v 2 project coveringhuman population pressure, humanland use, and human access

Grid

Hazards (Multihazards) Natural disaster hotspots: GlobalMultihazard Frequency andDistribution, classified

Grid

Water stress (Water stress) Net water demand divided by amountof renewable blue water resources

Grid

data issues, due to the limitations of sub-national data availabilityfrom China, and due to availability of more recent data.

First, provincial data was used to replace national averages forseveral of the components that are used to calculate the PoliticalInstability Index (EIU, 2011). For details, see Table 2.

Second, the Multifaceted Poverty Index (MPI; Alkire & Santos,2010) was calculated by province from the national average byscaling the latter with the province-specific values of the threecomponents of the MPI, namely education, health outcomes, andstandard of living. The two former ones were obtained from theeducation and health indices of the Human Development Index(HDI) which were available for Chinese provinces by UNDP (2010).The MPI indicator used for standard of living was the access toimproved water source (separately by rural and urban areas) fromChina Statistical Yearbook (2011).

Third, for human footprint, we used the most recent, version 2.0data of WCS/CIESIN (2005) instead of the previous version sincethis was now made available.

Fourth, for water stress we used the data provided byWada, vanBeek, and Bierkens (2011, 2013) instead of the less recent GWSPDigital Water Atlas (2008) data.

These changes have a relatively small impact on the overallvulnerability maps due to fairly limited year-to-year variability inthe indicator values. Anyhow, theymust be taken into considerationwhen comparing our results to those by Varis et al. (2012).

tion, sources, acronym used in the text and equations of calculating the vulnerability

olution Source Acronym

ntry/Province EIU (2011); China StatisticalYearbook (2011)

PSI

ince China Statistical Yearbook(2011)

GNIpop

ntry/Province Alkire and Santos (2010); UNDP(2010); China StatisticalYearbook (2011)

MPI

: 30 arc seconds (z1 km � 1 km) WCS/CIESIN (2005) HF

: 2.5 arc min (z5 km � 5 km) Dilley et al. (2005) MH

: 0.5 arc degree (z50 km � 50 km) Wada et al. (2011, 2013) WS

O. Varis et al. / Applied Geography 53 (2014) 105e116 109

River basin vulnerability index approach

We used here the river basin vulnerability (RBV) analysisapproach that was developed by Varis et al. (2012) within thecontext of analysing the vulnerability of ten major river basins inAsia-Pacific. This was because the approach has been tailoredparticularly topurposes such as this one, and comparable river basinvulnerability results were already available from 16 major riverbasins of Asia-Pacific (Varis et al., 2012) with a population thatcovers well over one-quarter of the mankind. The approach, as wellas the concept of vulnerability, is only briefly introduced belowwhile details can be found from Varis et al. (2012).

The RBV method is designed to be applicable for addressingriver systems when analysed as highly complex soci-etaleenvironmental systems. The approach includes six sources ofvulnerability which are governance, economy, social issues, envi-ronment, hazards, and water stress. They have been selected so,that they cover maximally the common key dimensions of theentity where water resources management and adaptation policiesare being performed. They are also aligned with the philosophy ofcommon policy frameworks in the water sector, particularly Inte-grated Water Resources Management (IWRM; Rahaman & Varis,2005).

The RBV approach is able to produce results which are compa-rable with any other geographical area in the world. The approachthus relies maximally on already published composite indicators.This is because those indices have been thoroughly investigated inthe global context, are easily available for most locations across theplanet and allow the analysis of temporal evolution of the vulner-ability, since they are constructed from indicators that are chieflypublished on an annual basis by various organizations.

In the approach, a geospatial mesh of administrative regions andbasins was first created using the ArcGIS software (see the map inFig. S1C). Then, using the gridded population data fromLandScanTM (2007) dataset, the total population and its proportionof the total basin population were calculated for each mesh unit(see Table 2; Fig. S1D). Each indicator valuewas thenweightedwithpopulation within each mesh unit and finally the river basin valuewas calculated by dividing the sum of weighted indicator values bytotal basin population. The spatially gridded data were aggregateddirectly to the basin areas (Fig. 1).

Then the data were scaled between 0 and 1 based on globaldistribution of each analysed vulnerability components (Table 3).We used the 5-percentile threshold of the global distribution torepresent the 0 value and 95-percentile to represent the 1, in orderto avoid one or few outliers to have a dramatic impact on thescaling of all other data points. The exceptionwasWater Stress. It iscalculated as the ratio of net water demand and the availablerenewable water resources. If this ratio exceeds 1, then the

Table 3Equations for calculating the indicators for River Basin Vulnerability Index (RBVI).The final RBVI is calculated as the average of the six components shown in the table.See Table 2 for the abbreviations and sources of the components.

Component Integrated index

Governance PSI0popEconomy (1 � (log(GNIpop) � log(GNIpop max))/(log(GNIpop min)

� log (GNIpop max)))0Social MPI0popEnvironment HF0

Hazards MH0

Water scarcity min(WS,1)

0 ¼ scaled value x0 ¼ (x � xmin)/(xmax � xmin) where xmax is the maximum 95% fractalof the global dataset in question, and xmin is the minimum 5% fractal of the same set.pop ¼ population weighted basin average value.

indicator receives the value 1. Finally the six components werecombined to the River Basin Vulnerability Index (RBVI) as theaverage value over them (for more details see Varis et al., 2012).

Cluster analysis

In order to identify the areas with a similar RBV profile, we per-formed a k-means cluster analysis with SPSS v20. In the analysis, wegrouped China's territory into four characteristic groups in terms ofvulnerability profiles. We centralized (zero mean) the data beforeapplying the k-means clustering. Datawere already standardized inthe scaling (see above and Table 3) and, thus, normalization (otherpart of the normal standardization procedure with centralizationneeded before clustering) of the data were not needed.

Results

The results on the six dimensions of vulnerability included inour analysis are documented below for river systems that areentirely or partly in the continental part of China. For those notentirely within China, only the part of the system that is located inChina was included in the analysis. A vulnerability profile for eachbasin is presented thereafter. At the end, the results of the clusteranalysis are presented.

Indicator values

GovernanceThe western and southern parts of China as well as the coast up

to Jiangsu have higher governance-related vulnerability thanChina's other parts (Fig. 2A). The most vulnerable part of thecountry is Guizhou. The municipal cities under the central gov-ernment, including Beijing, Shanghai and Tianjin, have the lowestgovernance-related vulnerability. The range in relation to the globalvariation is relatively large: from 0.15 to 0.53.

EconomyMost of the coastal China, in particular Beijing, Shanghai and

Tianjin, as well as Inner Mongolia represent the lowest economicvulnerability (Fig. 2B). The economically most vulnerable provincesinclude Anhui in the east, Gansu, and the south-western stripe fromGuizhou toTibet Autonomous Region. The range is relatively similarto that of the Governance index.

Social issuesThe range of internal variation of China in this regard is fairly

small (Fig. 2C). By and large, the coastal provinces appear moreadvanced in this regard compared to the rest of the country,whereas Tibet Autonomous Region appears the less advantageousin this regard.

EnvironmentThe human footprint (indicating environmental vulnerability) is

lower in thewest than in the east, with the exception of the Ili basinwith a relatively high value (Fig. 2D). In most parts of Tibet, thefootprint is at the lowest, whereas the highest indicator values arefor the Hai, Huai and lower Yellow River systems, as well as theShandong Peninsula. High values exist also in certain areas alongthe southern coast. The internal range of China's various parts isparticularly high in this respect, as large as that of the entire world(ranging thus from 0 to 1).

HazardsThe multihazard indicator shows particularly high vulnerability

for south-eastern coastal areas between the Yangtze River Delta

Fig. 2. The scaled indicator values mapped for China's river basin systems. A: Governance (Political Instability Index); B: Economy (Gross National Income per capita adjusted withPurchasing Power Parity); C: Social issues (Multidimensional Poverty Index); D: Environment (Human Footprint); E: Hazards (Multihazard Index); and F: Water stress. Note: thesame hue for different colours means similar index value. There are 10 steps for each tone of a colour, i.e. 0e0.1; 0.1e0.2; etc., as indicated in the legends. The larger the value, thelarger is the vulnerability. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

O. Varis et al. / Applied Geography 53 (2014) 105e116110

and the China-Vietnam border (Fig. 2E). The lowest values arerecorded on the large inland stripe from Tibet (excluding the GBMsystem) towards InnerMongolia. As in the case with environmentalvulnerability, the range is as large as that of the entire world.

Water stressA stripe of highwater stress crosses China in the northwesteeast

direction (Fig. 2F). It is most pronounced in large parts of North-Central China, and in. Besides, the Xinjinag province together withlowest parts of the Hai, Yellow, and Huai systems large areas inCentral China, are highly water-stressed. In the case of water stress,China's internal range is as big as the global range (0e1).

The overall vulnerability

The overall vulnerability, i.e. RBVI, is highest in central and lowparts of the Yellow river system and in southeastern part of the Hairiver system (Fig. 3). The very high population density combinedwith low water availability and high human footprint yield aparticularly challenging combination. This is despite lower eco-nomic vulnerability than in China's most other parts. Quite inter-estingly, large parts of the rest of China show relatively evenvulnerability level, although the combination of vulnerabilitysources differs quite remarkably (compare to Fig. 2 for the sixsources).

Fig. 3. The overall vulnerability (RBVI) of China.

O. Varis et al. / Applied Geography 53 (2014) 105e116 111

As an exception, most parts of Tibet and Qinghai (particularlyTibetan closed basins and the upper parts of the Yangtze basin)show the lowest vulnerability. Here a combination of low popula-tion density, low human footprint and low hazard level are themost important factors. The economic affluence level and socialdevelopment that are lower than in most other parts of the countrydoes not change the situation.

Vulnerability profiles

We grouped the river systems into open and closed ones. Wefurther grouped the open systems with regard to the sea they drainto. The river systems were thus classified into four groups (Fig. 4A):those draining (1) to the East China Sea, (2) to the South China Sea,(3) to the Indian and Arctic Oceans, and (4) to closed basins.

The systems draining to the East China Sea have all a fairly highoverall vulnerability: Yellow has the highest and Amur the lowest.The systems draining to the South China Sea are also highlyvulnerable. The systems draining to the Indian and Arctic Oceansrepresent low vulnerability levels, with an exception of Irrawaddy.The closed basins show a very heterogeneous group of vulnerabilitywith Tibetan CBs scoring reasonably low.

The open river systems show an interesting geographicvulnerability pattern (Fig. 4B). Starting from the north, from theAmur, there is a growing vulnerability tendency towards the Huai,and then gradually decreasing tendency all the way to Indus. Boththe environment- and hazards-related vulnerabilities grow to-wards the Huai, and thereafter, the direction is opposite. There arefew exceptional river systems that scoremarkedly higher than theirsurrounding systems. Those are the Hai, Yellow, Red and Irrawaddysystems. This anomaly is explained by water scarcity in all othercases besides the Irrawaddy, in which environmental vulnerabilityis exceptionally high.

Clustering the administrative-basin mesh

The total number of investigated river systems being fairly high(21), we clustered them into five groups in order to obtain acondensed view of the areas with similar vulnerability character-istics (Fig. 5).

Group a consists of the west, excluding the Ili system (seeFig. 2F). Here, the overall vulnerability is relatively low. Comparingto the population density map (Fig. S1D) is revealing; the

population density within this cluster is quite low in comparison tothe area to the east and south-east of this cluster. We name thisgroup as the Western Cluster.

Group b includes the water-stressed areas in north-centralChina including Inner Mongolia. This group covers quite well thenorthern part of the arid area of China, which is outside themonsoon influence. Their overall vulnerability is highest in waterstress and environment. We name this as the Dry Cluster.

Group c includes most of the central and southern parts of China(excluding the coastal stripe, the northeast, and the Ili basin). Here,environment and economy are important sources of vulnerability.We call this as the Humid Inland Cluster.

Group d consists of most of the Yellow and Huai river systems.For those, water stress and environment appear particularlyimportant vulnerability sources. The area is highly crowded andshort of water. We name this as the North China Plain Cluster.

Group e covers most of the coastal area of China. Its challenge isvery high vulnerability due to environment and hazards. We callthis as the Coastal Cluster.

Discussion

River system vulnerability: China's geographic features

China is one of the planet's largest countries in land area, and thelargest one in population. Therefore, it is not surprising that Chinarepresents a high level of spatial heterogeneity with regard to allthe aspects that were included in our vulnerability analysis. Thisheterogeneity is important, we argue, to understand and to identifyriver system vulnerabilities in a systematic and pragmatic way.

Closed river basins cover one-third of China's surface area(Table 1). In contrast, large areas in the eastern and southern partsof the country produce excessive discharge to the oceans. Economicincome level, population density, proneness to hazards and envi-ronmental situation all have large spatial differences inside China.

One could expect that closed basins and arid areaswould bemostvulnerable in terms of water resources. In China's case this appearsnot be true; our results indicate that the areas with highest riversystem vulnerability are not in dry but humid areas. The reason ispartly in the spatial distribution of population;whereas around one-third of China's territory is arid, only 2.1% of the population live inclosed basins (Table 1). In addition, closed basins are less prone tohazards and they have a lower human footprint than most of thehumid areas. Therefore, large dry, hydrologically closed areas such asXinjiang, Inner Mongolia and Qinghai's closed basins have not beenincluded in most vulnerable areas in our results, despite of highwater stress.

In many studies (Varis & Vakkilainen, 2001; Xia et al., 2012)China's most serious water related challenges have been addressedto the North China Plain, which has low precipitation but is stillmostly humid, but has a high population density. Our study is inaccordance to this. Yet, often the so-called 3H basins (Hai, Huang(Yellow) and Huai), have been used somewhat synonymously toNorth China Plain and attributed to be most challenged areas interms of water resources (Berkoff, 2003; Jiang, 2009; Xia et al.,2012). Our study does not fully accord with this view. The mostimportant reason to this disagreement is that we calculated thewater stress on the basis of water use in relation to its availability,and not on basis of water withdrawals or water availability percapita as is often done (see e.g. Jiang, 2009). The two latter ap-proaches show the 3H basins as more critical in terms of waterscarcity than the former (cf. the data provided by Wada et al., 2011,2013). Consequently, it is crucial to distinguishwater use fromwaterwithdrawals in this context, as well as in more general terms whenaddressing water related challenges. Our approach relates water

Fig. 4. Vulnerability profiles of China's river systems.

O. Varis et al. / Applied Geography 53 (2014) 105e116112

stress towater use andnot towaterwithdrawals since per definitionpart of thewithdrawnwater returns to the basin, whereas all water-use is away from the basin. We propose concernwhen selecting themetric as an indicator of water scarcity since for instance thewater-availability-per-capita metric ignores the demand of water e if anarea does not have much water demanding industry or agriculture,it can survive quite well with water resources that would be quiteshort for an area with intensive, irrigated agriculture. Instead, thewater scarcity should in our view bemeasured through the relationof supply and demand of water as done here (for more discussion,see Falkenmark, 2013; V€or€osmarty, Green, Salisbury, & Lammers,2000; Wada et al., 2011, 2013; Wu et al., 2013).

It is interesting to note, that large areas with ample water re-sources, particularly in the upper-middle Yangtze basin, such asSichuan, Chongqing and Hubei, are modestly water-stressed. This isbecause of their highpopulation density. In addition, they are hazard-prone and have a high human footprint. Therefore, they classifyequally or evenmore vulnerable than for instance Xinjiang and InnerMongolia in our study. There is plenty of evidence in history on their

vulnerability as for instance several serious floods and earthquakeshave caused massive societal and property damage in these areaseven inpast several years. This fairly high level of vulnerability of thatpart of the Yangtze basin is important to note when discussing theSoutheNorthwater transfer schemewhich drawswatermainly fromthe Yangtze basin towards the north, particularly to the 3H basins(Berkoff, 2003; Jiang, 2009; Xu et al. 2010).

After all, our vulnerability map shows less spatial differences indifferent parts of China than the studies that we refer to in theintroduction would suggest, and accordingly China's vulnerabilitylevel is relatively even in most parts of the country. This is becausein many areas the six components of vulnerability balance out oneanother; for instance in very hazard-prone coastal areas in south-east water stress is low and so is the social vulnerability.

Major transboundary basins

China shares several major continental-scale river basins withits neighbours (He, Tang, & He, 2000). In this study, we analysed

Fig. 5. Cluster results: A. Map of the five clusters of China's river basins with regard totheir vulnerability profile; B. Profiles of each cluster.

Fig. 6. Comparison of the results between this study (dark tone; name without asterisk) andof the whole basin is calculated (lighter tone; names with one asterisk refer to basins shar

O. Varis et al. / Applied Geography 53 (2014) 105e116 113

their vulnerabilities for the part that is located inside China's ter-ritory. In previous studies, nine of the total 11 transboundary basinshave been analysed with the RBVI method as a geographic entity(Varis & Kummu, 2012; Varis et al., 2012). The comparison of riversystems inside China's boundaries and the 9 transboundary basins(Fig. 6) shows that, whereas the level and range of vulnerability ofthose basins that are in China is remarkable, there are several largebasins which appear to be essentially more vulnerable than anybasin in China. Those include, above all, Indus and GBM. These twobasins have a total population of around 900 million people(Table 1; see also Varis et al., 2012). Therefore, when expressing thevery justified concern of the status and vulnerability of China'swater systems, one should indeed relate those challenges to theones existing in some of the surrounding river basins e some ofwhich China also shares as an upstream country.

In more general terms, transboundary basins are crucial whentalking about China's waters, albeit they are too often ignored inanalyses such as those mentioned in the above sections. Notableinstitutional and legislative challenges exist in transboundary co-operation on shared river basins of China and its neighbours, yetthe situation is highly dynamic for the time being and potentiallydeveloping towards the wider acceptance of international princi-ples and law (Wouters & Chen, 2013). Although the size of thepopulation and vulnerability level of the Chinese sections of thosetransboundary basins tend not to be quite high, China influencesand modifies profoundly many of these basins, causing impacts tothe downstream countries (Pearce, 2012; Ran & Lu, 2012). Massivehydropower construction in LancangeMekong (Keskinen, Kummu,K€ak€onen, & Varis, 2012; Molle, Foran, & K€ak€onen, 2009) can beused as one example and the emerging pressures of various eco-nomic activities upstream the Brahmaputra (Grumbine & Pandit,2013; Rahaman & Varis, 2009) and Salween (Magee, 2011) riverbasins as other examples. Despite of these major impacts on China'sneighbours, and disregarding of perennial recommendations byvarious policy documents (e.g., Jiang et al., 2009;World Bank, 2006;

the studies of Varis et al. (2012) and Varis and Kummu (2012), where the vulnerabilityed with China while names with two asterisks refer to basins outside China).

O. Varis et al. / Applied Geography 53 (2014) 105e116114

Xie et al., 2009), China's water policy remains thin in transboundarywaters (Liu &Wang, 2012). Consequently, when addressing China'swater related challenges, we argue that the inclusion of trans-boundary waters should be done far more rigorously than done inmost of the contemporary cases.

Vulnerability analysis approach

We produced vulnerability maps of the six aspects of vulnera-bility and their combined overall vulnerability. No such mappingand analysis have been available for China before. Besides having afiner resolution for river systems, we overlaid the basins withadministrative borders in our CARU river system delineation. Thisaspect is typically missing from water resources studies of China,although it is highly relevant to policy making and policy analysissince most policies are implemented through jurisdictions. Equallynovel in the context of China's water resources studies is ourcombination of social, economic, governance, hazards and envi-ronmental indicator data with water availability data.

Therefore, the current analysis provides a new level of spatialresolution andsystematizationof challenges andpressures to China'sriver systemswhen compared to the existing studies (e.g. Bawa et al.,2010; Economy, 2004; Gleick, 2009; Huang et al., 2008; Jiang, 2009;Lu, 2003; Ni, 2012; Varis& Vakkilainen, 2001;Wang et al., 2012; Xiaet al., 2012; Zhang et al., 1992). This is the case with basins that areentirely in China as well as with those shared with China's neigh-bours. In our division, the basin borders are defined precisely asspatial data, and this obviously helps in performing comparablestudies on China's river basins in the future. Besides, our analysisreveals quite clearly that water related vulnerability is a far morecomplicated issue than themerewater stress. Inmost of the analysesthat we have referred in this paper, water stress has not been lookedtogether with coping capacity of the society to tackle with waterchallenges, nor with other stress factors to the environment. Wemaintain that this should be done more often than done today.

The vulnerability analysis approach by Varis et al. (2012) that weused here has been developed for the analysis of large-scale riverbasins in which much of the administrative-unit based data is at afairly coarse scale e often only at national level. When applyingsuch an approach in a sub-national scale, as was done here, there isa certain challenge to find corresponding data for local jurisdictionssuch as China's provinces and autonomous regions. Therefore,complicated indices such as PSI and MPI are easy to use by country,but certain challenges emergewhen used in a sub-national context.It remains somewhat obscure whether the indicators used are ableto capture the internal variability in China.

Another source of inaccuracy to the analysis stems from the factthat the data stem from several different years. This type of data, as itcomes from various auxiliary datasets, is not available for each year.This is unfortunate, yet we do not believe, that year-to-year changesor fluctuations in these data are large enough to cause any notableinaccuracy to this analysis, since annual changes are incremental innature and remain relatively small with regard to all aspects studied.

Accordingly, certain reservations are justified, particularly whencomparing vulnerabilities related to governance and social issuewithin and outside China. However, we do not believe that this is amajor shortcoming of the approach.

Future research directions

We have excluded several important factors that could beaddressed in more refined studies. These include China's islands,water quality (Xie et al., 2009), groundwater (Jiang, 2009; Qiu,2011; Xie et al., 2009), impact of hydraulic constructions such asdams and water transfers (Ran & Lu, 2012; Yan et al., 2012), climate

change (Shen & Varis, 2001; Wang et al., 2012; Xu et al., 2010),climatic variations and variability (Xie et al., 2009; Xu et al., 2010)and particularities of urban areas (Bao& Fang, 2011; Finlayson et al.,2012; Shen, 2009; Wang, 2011). These aspects were excluded fromour study, due to the already large dimension of our analysis. Weencourage inclusion of those aspects in future analyses. Forinstance, in the case of the North China Plain e an area which wasclassified as the most vulnerable part of China according to ourmethod e inclusion of groundwater, water quality and other con-cernsmentioned abovewould be highly valuable (cf. Xia et al. 2012;Xie et al., 2009).

The methodology employed allows a straightforward extensionof the analysis to cover other geographic areas to obtain compa-rable results. An extension up to a global analysis is a feasibleoption.

Conclusions

In this article, we have documented an analysis of the socio-economic and environmental vulnerability of China's river systems.China's water resources challenges have frequently been empha-sized as one of the bottlenecks of the future development possi-bilities of the country (Bawa et al., 2010; Gleick, 2009; Gong et al.,2011; Jiang, 2009; Liu &Wang, 2012; Ran& Lu, 2012; Yu, 2011). Weshare this major concern. However, it is worth recognizing that thevulnerability level in many of the river systems in China's neigh-bouring areas (e.g., Indus, GBM, Hari Rud, Helmand and AmuDarya)is clearly higher than in any river system inside China.

It is also well-known that, given the vast geographical size,China represents notable spatial heterogeneity. The range of het-erogeneity with regard to environmental aspects (water stress,hazards, human footprint) equalled that of the entire globe, whilesocietal heterogeneity was substantially smaller in China comparedto the global context.

We found the highest overall vulnerability, calculated as thecombination of the six classes of vulnerability, in central and lowparts of the Yellow river system and in southeastern part of the Hairiver system. This area covers only roughly one-third of the mostwater-stressed parts of China; an areawherewater stressmeetswithhigh environmental stress and relatively high hazard level. Alsocoping capacity in terms of social and economic development level isfairly low in that area. Despite of goodwater availability, most coastalareas also appeared highly vulnerable due to high population densityand human footprint. Tibet andQinghai showed lowest vulnerability,largely due to low hazard level and low stress to water and envi-ronment. Quite interestingly, large parts of the rest of China showrelatively even vulnerability level, although the combination ofvulnerability sources is highly varied across the country.

Our study went beyond the existing water sector challengestudies of China in three dimensions:

1. Broader perspective. Broadening the perspective towards moremultidisciplinarity e particularly to societal direction as calledfor by Liu and Wang (2012) e is highly important whensearching for ways to develop river basin management into amore sustainable direction or towards a higher level of harmonyas often articulated by Chinese. These two articulations may notbe very far from one another, and the aspects included in ouranalysis aim at being relevant to both articulations.

2. Enhanced spatial resolution. A systematic, multidisciplinaryanalysis of China's water challenges combining jurisdictionaland river basin aspects has been missing thus far. Our studyprovided such an analysis and produced a novel delineation ofChina's river system units (the CARU system). The CARU hasbeen designed to be compatible with the existing Chinese river

O. Varis et al. / Applied Geography 53 (2014) 105e116 115

basin planning units (CPUs) while providing essentially higherspatial resolution together with inclusion of administrativeboundaries besides river basin boundaries.

3. Beyond China. Around one-third of humans live in the riverbasins that are entirely or partly in China, and almost half ofthem outside China's borders. Therefore, it is extremely impor-tant to better include the transboundary aspect to China's riverbasin policies, as emphasized by e.g. World Bank (2006), Xieet al. (2009) and Jiang et al. (2009). Our analysis covers thisentire geographic area, and enables the joint consideration ofChina's internal and transboundary river systems.

We argue that all these three aspects are important in under-standing and addressing China's water problems, particularlywithin the agenda of sustainable development (in the sense ofbalancing environmental, social and economic interests). Wesincerely hope that our analysis will be a useful step in advancingChina's e and more generally Asian e river basin development andmanagement towards higher level of balance and harmony be-tweenman and the nature. This would be very important due to thehighly demanding economic, environmental and social challengesthat this swiftly changing and developing part of the planet iscurrently facing.

Acknowledgments

The authors are grateful to the colleagues at Aalto University’sWater and Development Group for their thoughtful comments andinspiration. The thorough and insightful comments of Brian Fin-layson are greatly appreciated together with those of the editor andtwo reviewers. Olli Varis was funded by the Cultural Foundation ofFinland and Dajun Shen from base funds of Renmin University.Matti Kummu received funding from the postdoctoral funds ofAalto University and Academy of Finland project SCART (grant no.267463). This work also received funding from the Maa- ja ves-itekniikan tuki ry. and from Academy of Finland project 133748.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.apgeog.2014.05.012.

References

Alkire, S., & Santos, M. E. (2010). Multidimensional poverty index: 2010 data. Oxford:Oxford Poverty and Human Development Initiative. Available at www.ophi.org.uk/policy/multidimensional-poverty-index/.

Asia Society. (2009). Asia's next challenge: Securing the region's water future. HongKong: The Asia Society.

Asian Development Bank. (2007). Asian water development outlook. Manila: AsianDevelopment Bank.

Bao, C., & Fang, C. L. (2011). Water resources flows related to urbanization in China:challenges and perspectives for water management and urban development.Water Resources Management, 26, 531e552.

Bawa, K. S., Koh, L. P., Lee, T. M., Liu, J., Ramakrishnan, P. S., Yu, D. W., et al. (2010).China, India, and the environment. Science, 327, 1457e1459.

Berkoff, J. (2003). China: the SoutheNorth water transfer project e is it justified?Water Policy, 5, 1e28.

Biswas, A. K. (2005). Integrated water resources management: a reassessment. InA. K. Biswas, O. Varis, & C. Tortajada (Eds.), Integrated water resources man-agement in South and Southeast Asia (pp. 319e336). New Delhi: Oxford Uni-versity Press.

Cao, S., Chen, L., & Liu, Z. (2007). Disharmony between society and environmentalcarrying capacity: a historical review, with an emphasis on China. Ambio, 36,409e415. http://dx.doi.org/10.1579/0044-7447(2007)36[409:DBSAEC]2.0.CO;2.

Cao, Y., Wu, Y. H., Zhang, Y. J., & Tian, J. (2013). Landscape pattern and sustainabilityof a 1300-year-old agricultural landscape in subtropical mountain areas,Southwestern China. International Journal of Sustainable Development and WorldEcology, 20, 349e357. http://dx.doi.org/10.1080/13504509.2013.773266.

China Statistical Yearbook. (2011). China statistical yearbook 2011. Beijing: NationalBureau of Statistics of China.

Dilley, M., Chen, R. S., Deichmann, U., Lerner-Lam, A. L., Arnold, M., Agwe, J., et al.(2005). Natural disaster hotspots: a global risk analysis. In Disaster risk man-agement series. Washington D.C: The World Bank.

Economy, E. (2004). The river runs black e The environmental challenge to China'sfuture. Ithaca, New York: Cornell University Press.

EIU. (2011). ViewsWire: Business intelligence on 203 countries. New York: EconomistIntelligence Unit.

Falkenmark, M. (2013). Growing water scarcity in agriculture: future challenge toglobal water security. Philosophical Transactions of the Royal Society A,371(2002), 1471e2962. http://dx.doi.org/10.1098/rsta.2012.0410.

Finlayson, B. L., Barnett, B., Wei, T., Webber, M., Li, M., Wang, M. Y., et al. (2012). Thedrivers of risk to water security in Shanghai. Regional Environmental Change.http://dx.doi.org/10.1007/s10113-012-0334-1.

Gleick, P. (2009). China and water. In P. Gleick, H. Cooley, M. J. Cohen, M. Morikawa,J. Morrison, & M. Palaniappan (Eds.), The world's water 2008e2009 (pp.79e100). Washington: Island Press.

Gong, P., Yin, Y., & Yu, C. (2011). China: invest wisely in sustainable water use.Science, 331, 1264e1265.

Grumbine, R. E., & Pandit, M. K. (2013). Threats from India's Himalaya Dams. Science,339, 36e37.

He, D., Tang, D., & He, Q. (2000). International rivers in China. Beijing: Science Press(in Chinese).

Huang, Y., Cai, M., Zhang, D., & Cai, J. (2008). Freshwater under threat: NortheastAsia. In Vulnerability assessment of freshwater resources to environmental change.Nairobi: United Nations Environmental Programme.

Jiang, Y. (2009). China's water scarcity. Journal of Environmental Management, 90,3185e3196. http://dx.doi.org/10.1016/j.jenvman.2009.04.016.

Keskinen, M., Kummu, M., K€ak€onen, M., & Varis, O. (2012). Mekong at the cross-roads: next steps for impact assessment of large dams. Ambio, 41, 319e324.http://dx.doi.org/10.1007/s13280-012-0261-x.

LandScanTM. (2007). Global population database for 2007. Oak Ridge, TN: Oak RidgeNational Laboratory. Available at http://www.ornl.gov/landscan/.

Liu, J., & Wang, W. (2012). Water sustainability for China and beyond. Science, 337,649e650.

Lu, X. X. (2003). Vulnerability of water discharge of large Chinese rivers to envi-ronmental changes: an overview. Regional Environmental Change, 4, 182e191.

Magee, D. (2011). The dragon upstream: China's role in LancangeMekong develop-ment. In J. €Ojendal, S. Hansson, & S. Hellberg (Eds.), Politics and development in atransboundarywatershed: The case of the LowerMekongBasin. NewYork: Springer.

Molle, F., Foran, T., & K€ak€onen, M. (2009). Contested waterscapes in the Mekong re-gion. London: Earthscan.

Ni, J. (2012). Impacts of climate change on Chinese ecosystems: key vulnerableregions and potential thresholds. Regional Environmental Change, 11(Suppl. 1),S49eS64.

Pearce, F. (2012). Whose water is it? New Scientist, 8e9, 28 April.Qiu, J. (2011). China to spend billions cleaning up groundwater. Science, 334(6057),

745. http://dx.doi.org/10.1126/science.334.6057.745.Rahaman, M. M., & Varis, O. (2005). Integrated water resources management:

evolution, prospects and future challenges. Sustainability: Science, Practice &Policy, 1(1), 1e9.

Rahaman, M. M., & Varis, O. (2009). Integrated water management of the Brah-maputra Basin: perspectives and hope for regional development. Natural Re-sources Forum, 33(1), 60e75.

Ran, L., & Lu, X. X. (2012). Redressing China's strategy of water resource exploitation.Environmental Management. http://dx.doi.org/10.1007/s00267-012-0002-8.

Shen, D. (2009). Urban water management in China. International Journal of WaterResources Development, 25, 249e268.

Shen, D. (2010). Climate change and water resources: evidence and estimate inChina. Current Science, 98(8), 1063e1068.

Shen, D., & Varis, O. (2000). World water vision: balancing thoughts after TheHague. Ambio, 29, 523e525.

Shen, D., & Varis, O. (2001). Climate change in China. Ambio, 30(6), 381e383.Strohsch€on, R., Wiethoff, K., Baier, K., Lu, L., Bercht, A. L., Wehrhahn, R., et al. (2013).

Land use and water quality in Guangzhou, China: a survey of ecological andsocial vulnerability in four urban units of the rapidly developing megacity. In-ternational Journal of Environmental Research, 7, 343e358.

UNDP. (2010). China and a sustainable future: Towards a low carbon economy andsociety. China Human Development Report 2009/2010. Beijing: United NationsDevelopment Program, Beijing, Renmin University, Beijing and China Trans-lation and Publishing Corporation.

USGS. (2001). Drainage basins of the world, part of the global GIS database by U. S.Geological Survey (USGS). Reston, VA: United States Geological Survey.

Varis, O. (2005). Externalities of integrated water resources management in Southand South-east Asia. In A. K. Biswas, O. Varis, & C. Tortajada (Eds.), Integratedwater resources management in South and South-east Asia (pp. 1e38). New Delhi:Oxford University Press.

Varis, O. (2011). The Yangtze river and regional development e 4th Yangtze Forum,Nanjing, China, 18e20 April 2011. Conference Report. International Journal ofWater Resources Development, 27, 621e624.

Varis, O., & Kummu, M. (2012). The major Central Asian river basins: an assessmentof vulnerability. Water Resources Development, 28(3), 433e452.

Varis, O., Kummu, M., & Salmivaara, A. (2012). Ten major river basins in monsoonAsia-Pacific: an assessment of vulnerability. Applied Geography, 32, 441e454.

Varis, O., & Vakkilainen, P. (2001). China's 8 challenges to water resources manage-ment in the first quarter of the 21st Century. Geomorphology, 41(2e3), 93e104.

http://dx.doi.org/10.1016/j.apgeog.2014.05.012http://dx.doi.org/10.1016/j.apgeog.2014.05.012http://www.ophi.org.uk/policy/multidimensional-poverty-index/http://www.ophi.org.uk/policy/multidimensional-poverty-index/http://refhub.elsevier.com/S0143-6228(14)00107-6/sref2http://refhub.elsevier.com/S0143-6228(14)00107-6/sref2http://refhub.elsevier.com/S0143-6228(14)00107-6/sref3http://refhub.elsevier.com/S0143-6228(14)00107-6/sref3http://refhub.elsevier.com/S0143-6228(14)00107-6/sref4http://refhub.elsevier.com/S0143-6228(14)00107-6/sref4http://refhub.elsevier.com/S0143-6228(14)00107-6/sref4http://refhub.elsevier.com/S0143-6228(14)00107-6/sref4http://refhub.elsevier.com/S0143-6228(14)00107-6/sref5http://refhub.elsevier.com/S0143-6228(14)00107-6/sref5http://refhub.elsevier.com/S0143-6228(14)00107-6/sref5http://refhub.elsevier.com/S0143-6228(14)00107-6/sref6http://refhub.elsevier.com/S0143-6228(14)00107-6/sref6http://refhub.elsevier.com/S0143-6228(14)00107-6/sref6http://refhub.elsevier.com/S0143-6228(14)00107-6/sref6http://refhub.elsevier.com/S0143-6228(14)00107-6/sref6http://refhub.elsevier.com/S0143-6228(14)00107-6/sref7http://refhub.elsevier.com/S0143-6228(14)00107-6/sref7http://refhub.elsevier.com/S0143-6228(14)00107-6/sref7http://refhub.elsevier.com/S0143-6228(14)00107-6/sref7http://refhub.elsevier.com/S0143-6228(14)00107-6/sref7http://dx.doi.org/10.1579/0044-7447(2007)36[409:DBSAEC]2.0.CO;2http://dx.doi.org/10.1080/13504509.2013.773266http://refhub.elsevier.com/S0143-6228(14)00107-6/sref10http://refhub.elsevier.com/S0143-6228(14)00107-6/sref10http://refhub.elsevier.com/S0143-6228(14)00107-6/sref11http://refhub.elsevier.com/S0143-6228(14)00107-6/sref11http://refhub.elsevier.com/S0143-6228(14)00107-6/sref11http://refhub.elsevier.com/S0143-6228(14)00107-6/sref12http://refhub.elsevier.com/S0143-6228(14)00107-6/sref12http://refhub.elsevier.com/S0143-6228(14)00107-6/sref12http://refhub.elsevier.com/S0143-6228(14)00107-6/sref13http://refhub.elsevier.com/S0143-6228(14)00107-6/sref13http://dx.doi.org/10.1098/rsta.2012.0410http://dx.doi.org/10.1007/s10113-012-0334-1http://refhub.elsevier.com/S0143-6228(14)00107-6/sref16http://refhub.elsevier.com/S0143-6228(14)00107-6/sref16http://refhub.elsevier.com/S0143-6228(14)00107-6/sref16http://refhub.elsevier.com/S0143-6228(14)00107-6/sref16http://refhub.elsevier.com/S0143-6228(14)00107-6/sref16http://refhub.elsevier.com/S0143-6228(14)00107-6/sref17http://refhub.elsevier.com/S0143-6228(14)00107-6/sref17http://refhub.elsevier.com/S0143-6228(14)00107-6/sref17http://refhub.elsevier.com/S0143-6228(14)00107-6/sref18http://refhub.elsevier.com/S0143-6228(14)00107-6/sref18http://refhub.elsevier.com/S0143-6228(14)00107-6/sref18http://refhub.elsevier.com/S0143-6228(14)00107-6/sref19http://refhub.elsevier.com/S0143-6228(14)00107-6/sref19http://refhub.elsevier.com/S0143-6228(14)00107-6/sref20http://refhub.elsevier.com/S0143-6228(14)00107-6/sref20http://refhub.elsevier.com/S0143-6228(14)00107-6/sref20http://dx.doi.org/10.1016/j.jenvman.2009.04.016http://dx.doi.org/10.1007/s13280-012-0261-xhttp://www.ornl.gov/landscan/http://refhub.elsevier.com/S0143-6228(14)00107-6/sref24http://refhub.elsevier.com/S0143-6228(14)00107-6/sref24http://refhub.elsevier.com/S0143-6228(14)00107-6/sref24http://refhub.elsevier.com/S0143-6228(14)00107-6/sref25http://refhub.elsevier.com/S0143-6228(14)00107-6/sref25http://refhub.elsevier.com/S0143-6228(14)00107-6/sref25http://refhub.elsevier.com/S0143-6228(14)00107-6/sref26http://refhub.elsevier.com/S0143-6228(14)00107-6/sref26http://refhub.elsevier.com/S0143-6228(14)00107-6/sref26http://refhub.elsevier.com/S0143-6228(14)00107-6/sref26http://refhub.elsevier.com/S0143-6228(14)00107-6/sref26http://refhub.elsevier.com/S0143-6228(14)00107-6/sref27http://refhub.elsevier.com/S0143-6228(14)00107-6/sref27http://refhub.elsevier.com/S0143-6228(14)00107-6/sref27http://refhub.elsevier.com/S0143-6228(14)00107-6/sref27http://refhub.elsevier.com/S0143-6228(14)00107-6/sref28http://refhub.elsevier.com/S0143-6228(14)00107-6/sref28http://refhub.elsevier.com/S0143-6228(14)00107-6/sref28http://refhub.elsevier.com/S0143-6228(14)00107-6/sref28http://refhub.elsevier.com/S0143-6228(14)00107-6/sref29http://refhub.elsevier.com/S0143-6228(14)00107-6/sref29http://dx.doi.org/10.1126/science.334.6057.745http://refhub.elsevier.com/S0143-6228(14)00107-6/sref31http://refhub.elsevier.com/S0143-6228(14)00107-6/sref31http://refhub.elsevier.com/S0143-6228(14)00107-6/sref31http://refhub.elsevier.com/S0143-6228(14)00107-6/sref31http://refhub.elsevier.com/S0143-6228(14)00107-6/sref32http://refhub.elsevier.com/S0143-6228(14)00107-6/sref32http://refhub.elsevier.com/S0143-6228(14)00107-6/sref32http://refhub.elsevier.com/S0143-6228(14)00107-6/sref32http://dx.doi.org/10.1007/s00267-012-0002-8http://refhub.elsevier.com/S0143-6228(14)00107-6/sref34http://refhub.elsevier.com/S0143-6228(14)00107-6/sref34http://refhub.elsevier.com/S0143-6228(14)00107-6/sref34http://refhub.elsevier.com/S0143-6228(14)00107-6/sref35http://refhub.elsevier.com/S0143-6228(14)00107-6/sref35http://refhub.elsevier.com/S0143-6228(14)00107-6/sref35http://refhub.elsevier.com/S0143-6228(14)00107-6/sref36http://refhub.elsevier.com/S0143-6228(14)00107-6/sref36http://refhub.elsevier.com/S0143-6228(14)00107-6/sref36http://refhub.elsevier.com/S0143-6228(14)00107-6/sref37http://refhub.elsevier.com/S0143-6228(14)00107-6/sref37http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref38http://refhub.elsevier.com/S0143-6228(14)00107-6/sref39http://refhub.elsevier.com/S0143-6228(14)00107-6/sref39http://refhub.elsevier.com/S0143-6228(14)00107-6/sref39http://refhub.elsevier.com/S0143-6228(14)00107-6/sref39http://refhub.elsevier.com/S0143-6228(14)00107-6/sref40http://refhub.elsevier.com/S0143-6228(14)00107-6/sref40http://refhub.elsevier.com/S0143-6228(14)00107-6/sref41http://refhub.elsevier.com/S0143-6228(14)00107-6/sref41http://refhub.elsevier.com/S0143-6228(14)00107-6/sref41http://refhub.elsevier.com/S0143-6228(14)00107-6/sref41http://refhub.elsevier.com/S0143-6228(14)00107-6/sref41http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref42http://refhub.elsevier.com/S0143-6228(14)00107-6/sref43http://refhub.elsevier.com/S0143-6228(14)00107-6/sref43http://refhub.elsevier.com/S0143-6228(14)00107-6/sref43http://refhub.elsevier.com/S0143-6228(14)00107-6/sref44http://refhub.elsevier.com/S0143-6228(14)00107-6/sref44http://refhub.elsevier.com/S0143-6228(14)00107-6/sref44http://refhub.elsevier.com/S0143-6228(14)00107-6/sref45http://refhub.elsevier.com/S0143-6228(14)00107-6/sref45http://refhub.elsevier.com/S0143-6228(14)00107-6/sref45http://refhub.elsevier.com/S0143-6228(14)00107-6/sref45

O. Varis et al. / Applied Geography 53 (2014) 105e116116

V€or€osmarty, C. J., Green, P., Salisbury, J., & Lammers, R. B. (2000). Global water re-sources: vulnerability from climate change and population growth. Science, 289,284e288.

Wada, Y., van Beek, L. P. H., & Bierkens, M. F. P. (2011). Modelling global water stressof the recent past: on the relative importance of trends in water demand andclimate variability. Hydrology and Earth System Sciences, 15, 3785e3808. http://dx.doi.org/10.5194/hess-15-3785-2011.

Wada, Y., van Beek, L. P. H., Wanders, N., & Bierkens, M. F. P. (2013). Human waterconsumption intensifies hydrological drought worldwide. EnvironmentalResearch Letters, 8, 034036. http://dx.doi.org/10.1088/1748-9326/8/3/034036.

Wang, R. (2011). Environmental and resource sustainability of Chinese cities: a reviewof issues, policies, practices and effects. Natural Resources Forum, 35, 112e121.

Wang, Z. Q., He, F., Fang, W. H., & Liao, Y. F. (2013). Assessment of physicalvulnerability to agricultural drought in China. Natural Hazards, 67, 645e657.http://dx.doi.org/10.1007/s11069-013-0594-1.

Wang, G. Q., Zhang, J. Y., Jin, J. L., Pagano, T. C., Calow, R., Bao, Z. X., et al. (2012).Assessing water resources in China using PRECIS projections and a VIC model.Hydrology and Earth System Sciences, 16, 231e240. http://dx.doi.org/10.5194/hess-16-231-2012.

Water Resources eAtlas. (2003). Watersheds of the World_CD (Vol. 2005). IUCN-TheWorld Conservation Union, the International Water Management Institute(IWMI), the Ramsar Convention Bureau, and the World Resources Institute(WRI) Accessed through internet on November 2012 http://multimedia.wri.org/watersheds_2003/index.html.

WCS/CIESIN. (2005). Last of the Wild Project, v2.0 (LWP-2): Global Human FootprintDataset (Geographic). Palisades, NY: Wildlife Conservation Society (WCS), Centerfor International Earth Science Information Network (CIESIN), SocioeconomicData and Application Center (SEDAC) and National Aeronautics and SpaceAdministration (NASA) Accessed through Internet in March 2014 http://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-footprint-geographic.

World Bank. (2006). China water quality management e Policy and institutionalconsiderations. Washington D.C: The World Bank.

World Bank. (2014).World development indicators. Washington D.C: TheWorld Bank.Wouters, P., & Chen, H. (2013). China's ‘soft-path’ to transboundary water cooper-

ation examined in the light of two UN Global Water Conventions e exploringthe ‘Chinese way’. Journal of Water Law, 22, 229e247.

Wu, G. Y., Li, L. H., Ahmad, S., Chen, X., & Pan, X. L. (2013). A dynamic model forvulnerability assessment of regional water resources in arid areas: a case studyof Bayingolin, China. Water Resources Management, 27, 3085e3101. http://dx.doi.org/10.1007/s11269-013-0334-z.

WWAP (World Water Assessment Programme). (2009). The United Nations WorldWater Development Report 3: Water in a changing world. Paris and Earthscan,London: UNESCO.

Xia, J., Qiu, B., & Li, Y. (2012). Water resources vulnerability and adaptive manage-ment in the Huang, Huai and Hai river basins of China. Water International,37(5), 523e536.

Xie, J., Liebenthal, A., Warford, J. J., Dixon, J. A., Wang, M., Gao, S., et al. (2009).Addressing China's water scarcity: Recommendations for selected water resourcemanagement issues. Washington D.C: The World Bank.

Xu, K., Milliman, J. D., & Xu, H. (2010). Temporal trend of precipitation and runoff inmajor Chinese rivers since 1951. Global and Planetary Change, 73, 219e232.

Yan, D. H., Wang, H., Li, H. H., Wang, G., Qin, T. L., Wang, D. Y., et al. (2012).Quantitative analysis on the environmental impact of large-scale water transferproject on water resource area in a changing environment. Hydrology and EarthSystem Sciences, 16, 2685e2702.

Yin, L. H., Zhang, E. Y., Wang, X. Y., Wenninger, J., Dong, J. Q., Guo, L., et al. (2013).A GIS-based DRASTIC model for assessing groundwater vulnerability in theOrdos Plateau, China. Environmental Earth Sciences, 69, 171e185. http://dx.doi.org/10.1007/s12665-012-1945-z.

Yu, C. (2011). China's water crisis needs more than words. Nature, 470, 307.Zhang, Z., Chen, B., Chen, Z., & Xu, X. (1992). Challenges and opportunities for

development of China's water resources in the 21st Century. Water Interna-tional, 17, 21e27.

Zhang, B. Q., Wu, P. T., Zhao, X. N., Wang, Y. B., Gao, X. D., & Cao, X. C. (2013).A drought hazard assessment index based on the VIC-PDSI model and itsapplication on the Loess Plateau, China. Theoretical and Applied Climatology, 114,125e138. http://dx.doi.org/10.1007/s00704-012-0826-4.

Zheng, D., & Dai, E. F. (2013). Environmental ethics and regional sustainabledevelopment. Journal of Geographical Sciences, 22, 86e92. http://dx.doi.org/10.1007/s11442-012-0913-6.

http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://refhub.elsevier.com/S0143-6228(14)00107-6/sref46http://dx.doi.org/10.5194/hess-15-3785-2011http://dx.doi.org/10.5194/hess-15-3785-2011http://dx.doi.org/10.1088/1748-9326/8/3/034036http://refhub.elsevier.com/S0143-6228(14)00107-6/sref49http://refhub.elsevier.com/S0143-6228(14)00107-6/sref49http://refhub.elsevier.com/S0143-6228(14)00107-6/sref49http://dx.doi.org/10.1007/s11069-013-0594-1http://dx.doi.org/10.5194/hess-16-231-2012http://dx.doi.org/10.5194/hess-16-231-2012http://multimedia.wri.org/watersheds_2003/index.htmlhttp://multimedia.wri.org/watersheds_2003/index.htmlhttp://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-footprint-geographichttp://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-footprint-geographichttp://refhub.elsevier.com/S0143-6228(14)00107-6/sref54http://refhub.elsevier.com/S0143-6228(14)00107-6/sref54http://refhub.elsevier.com/S0143-6228(14)00107-6/sref54http://refhub.elsevier.com/S0143-6228(14)00107-6/sref55http://refhub.elsevier.com/S0143-6228(14)00107-6/sref56http://refhub.elsevier.com/S0143-6228(14)00107-6/sref56http://refhub.elsevier.com/S0143-6228(14)00107-6/sref56http://refhub.elsevier.com/S0143-6228(14)00107-6/sref56http://refhub.elsevier.com/S0143-6228(14)00107-6/sref56http://dx.doi.org/10.1007/s11269-013-0334-zhttp://dx.doi.org/10.1007/s11269-013-0334-zhttp://refhub.elsevier.com/S0143-6228(14)00107-6/sref58http://refhub.elsevier.com/S0143-6228(14)00107-6/sref58http://refhub.elsevier.com/S0143-6228(14)00107-6/sref58http://refhub.elsevier.com/S0143-6228(14)00107-6/sref59http://refhub.elsevier.com/S0143-6228(14)00107-6/sref59http://refhub.elsevier.com/S0143-6228(14)00107-6/sref59http://refhub.elsevier.com/S0143-6228(14)00107-6/sref59http://refhub.elsevier.com/S0143-6228(14)00107-6/sref60http://refhub.elsevier.com/S0143-6228(14)00107-6/sref60http://refhub.elsevier.com/S0143-6228(14)00107-6/sref60http://refhub.elsevier.com/S0143-6228(14)00107-6/sref61http://refhub.elsevier.com/S0143-6228(14)00107-6/sref61http://refhub.elsevier.com/S0143-6228(14)00107-6/sref61http://refhub.elsevier.com/S0143-6228(14)00107-6/sref62http://refhub.elsevier.com/S0143-6228(14)00107-6/sref62http://refhub.elsevier.com/S0143-6228(14)00107-6/sref62http://refhub.elsevier.com/S0143-6228(14)00107-6/sref62http://refhub.elsevier.com/S0143-6228(14)00107-6/sref62http://dx.doi.org/10.1007/s12665-012-1945-zhttp://dx.doi.org/10.1007/s12665-012-1945-zhttp://refhub.elsevier.com/S0143-6228(14)00107-6/sref64http://refhub.elsevier.com/S0143-6228(14)00107-6/sref65http://refhub.elsevier.com/S0143-6228(14)00107-6/sref65http://refhub.elsevier.com/S0143-6228(14)00107-6/sref65http://refhub.elsevier.com/S0143-6228(14)00107-6/sref65http://dx.doi.org/10.1007/s00704-012-0826-4http://dx.doi.org/10.1007/s11442-012-0913-6http://dx.doi.org/10.1007/s11442-012-0913-6

China's stressed waters: Societal and environmental vulnerability in China's internal and transboundary river systemsIntroductionMaterials and methodsDelineation of China's river basins: the CARU systemTransboundary river basinsDataRiver basin vulnerability index approachCluster analysis

ResultsIndicator valuesGovernanceEconomySocial issuesEnvironmentHazardsWater stress

The overall vulnerabilityVulnerability profilesClustering the administrative-basin mesh

DiscussionRiver system vulnerability: China's geographic featuresMajor transboundary basinsVulnerability analysis approachFuture research directions

ConclusionsAcknowledgmentsAppendix A Supplementary dataReferences