Matti Kummu and Juha Sarkkula

Impact of the Mekong River Flow Alteration onthe Tonle Sap Flood Pulse

Rapid development in the upper reaches of the MekongRiver, in the form of construction of large hydropowerdams and reservoirs, large irrigation schemes, and rapidurban development, is putting water resources understress. Recent studies have concluded that these devel-opments will lead to flow alterations in the Mekong River.These flow alterations would threaten the sensitiveecosystems downstream, particularly Tonle Sap River,Tonle Sap Lake, its floodplain, and its gallery forest andprotected areas, by changing the flood-pulse system ofthe lake. This article estimates the changes in parametersof the Tonle Sap flood pulse due to the aforementionedflow alterations. The impacts on the flooded area and lossof gallery forest and protected areas were analyzed usinggeographic information system–based methods. Rela-tively small rises in the dry-season lake water level wouldpermanently inundate disproportionately large areas offloodplain, rendering it inaccessible to floodplain vegeta-tion and eroding the productivity basis of the ecosystem.It is highly important to maintain the natural hydrologicalpattern of the Mekong River, particularly the dry-seasonwater levels, to preserve Tonle Sap Lake’s ecosystemproductivity.

INTRODUCTION

During the last decades, the human impact on water resourceshas increased remarkably, and it will continue to increase,although probably not as rapidly as it has in the twentiethcentury (1). The standing stock of natural river water hasincreased sevenfold due to 8400 km3 of water stored inreservoirs behind dams (1). The water renewal time (3) of theworld’s rivers has increased dramatically, from 20 to 100 d, as aconsequence of dam and reservoir consumption (4). Accordingto Skole et al. (5), land-cover change may be the mostsignificant agent of change in hydrology, climate, and globalbiogeochemical cycles. Furthermore, climate change may play avery important role in those changes in the future (e.g., 6, 7).

Wetlands, including floodplains, are important parts ofrivers and other freshwater ecosystems. On a global scale,wetlands cover between 2.0–5.3 million km2 (8) and 12.8 millionkm2 (9), corresponding to ;3%–9% of the land surface. Thedegradation and loss of wetlands occur more rapidly than thatof other ecosystems (9, 10). Maintenance of the hydrologicalregime of a wetland and its natural variability is necessary tosustain the ecological characteristics of the wetland, includingits biodiversity (9). According to the Millennium EcosystemAssessment (9), dams play a major role in fragmenting andmodifying aquatic habitats by altering the flow of matter andenergy.

Existing, large-scale utilization of the Mekong’s resources islimited compared to almost all large river basins in the world.The present flow regime in the main stem of the Mekong isessentially natural (11); although, in many parts of the basin,water resources are used intensively at a small-scale, local level.However, the Mekong Basin is under rapid development. Theconstruction of large-scale hydropower dams and reservoirs in

the basin began on the tributaries in the 1970s in Lao People’sDemocratic Republic (PDR) and Thailand, and in the main stema decade ago in China. Numerous additional dams are envisagedin China, Lao PDR, Cambodia, and Vietnam (12–16).Population growth and economic development have led to rapidland-cover changes, mainly deforestation, in many parts of thebasin (17). Development in the river and the basin will alter theflows and floods in the basin. This is also the conclusion of thecumulative impact assessments (CIAs) that have been made (17–19). The impact of climate change on the hydrology has not beenincluded in the CIAs, but it is considered to have an importantimpact on the Mekong and Tonle Sap hydrology, especiallyduring the latter part of the 21st century (20).

Each CIA report (17–19) made for the Mekong Basin,mainly for hydropower impact assessment, concludes that dry-season water levels would rise and wet-season water levelswould be lower than at present. The flow alterations would bemore significant close to the dam and gradually decrease withdistance in the lower Mekong Basin. The flow alterations in theMekong main stem would directly impact the flood pulse ofTonle Sap Lake. This is because around 60% of the Tonle Sapflood water originates from the Mekong, and the water level inthe lake is controlled by the water level in the Mekong mainstem (21).

Flood pulses in rivers and lakes can be described by anumber of characteristics that are important from an ecosystemproductivity point of view. The list of flood-pulse parametersused in this analysis is based on Welcomme and Halls (22) andwas adapted to Tonle Sap Lake by Lamberts (23). The list is asfollows: flood duration, amplitude, flood volume, timing,rapidity of the water-level change, continuity, and smoothness.

Lamberts (24) presented a detailed description of the TonleSap flood-pulse parameters susceptible to anthropogenic flowalterations in the Mekong River and discussed the possibleimpacts of these alterations on ecosystem productivity.

This paper aims to assess the impacts of the flow alterationon the flood characteristics, gallery forest, and protected areasin Tonle Sap Lake. The paper consists of three parts: i)statistical analysis of the present Tonle Sap flood characteris-tics, ii) assessment of the flow alteration impacts on floodcharacteristics, based on the recent CIA studies, and iii) spatialanalysis of the flow alteration impacts on the flooded area,gallery forest, and protective areas.

TONLE SAP LAKE

Tonle Sap Lake (Fig. 1) is an integral part of the Mekong River,and it is the largest freshwater lake in Southeast Asia. TheMekong River is among the largest rivers in the world and isranked as the 10th largest by volume, with an annual dischargeof 475 km3 (10). The importance of the lake is unquestioned forCambodia and the lower Mekong Basin (25–29). Over onemillion people depend on the natural resources of the lake. Thevalue of Tonle Sap Lake has also been recognized internation-ally, and the lake has three Biosphere Reserve core areas underthe United Nations Educational, Scientific and CulturalOrganization (UNESCO) Programme on Man and the Biosphere(30) and one Ramsar site under the Ramsar Convention (31).

Ambio Vol. 37, No. 3, May 2008 185� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

Water Balance

Water-balance calculations for Tonle Sap Lake have shown thatmost of the inflow into the lake (57%) originates from theMekong main stem, either by discharge through the Tonle SapRiver (52%) or by overland flooding (5%) (21). Tributaries toTonle Sap Lake contribute about 30%, and precipitationdirectly into the lake contributes some 13% of inflow. Theaverage annual inflow is 79.0 km3, ranging between 44.1 km3 in1998 and 106.5 km3 in 2000. Around 88% of the receding lakewater returns to the Mekong River through the Tonle Sap River(87%) and overland flooding (1%), while 12% evaporates fromthe lake. The average annual outflow is 78.6 km3, ranging from43.5 km3 in 1998 to 104.8 km3 in 2000. The flow in the MekongRiver is the principal factor determining the flood pulse ofTonle Sap Lake (21).

Flood Pulse and Floodplain Gallery Forest

According to Junk (32), ecosystems that experience fluctuationsbetween terrestrial and aquatic conditions are called pulsingecosystems and can be described in terms of the flood-pulseconcept. Junk’s flood pulse concept (32) has been widelyaccepted as describing the highly productive floodplainenvironments and the ecology of pulsing systems. Tonle SapLake and floodplain ecosystem is well described by the flood-pulse concept, where the annual monsoon floods of the MekongRiver are a key driver of its productivity (23).

The floodplain vegetation is one of the most importantelements of the Tonle Sap ecosystem (23, 28). Among it, the tallgallery forest stripe in the immediate vicinity of the permanentlake shore (illustrated in Fig. 1) covers only a small part of thefloodplain, but it constitutes an important physical barrierbetween the open lake and the floodplain, creating favorableconditions for effective sedimentation within the forested zone(33). In addition to the permanent lake edge and the river banks,isolated stands of trees are scattered throughout the floodplain.The total area of the gallery forest of Tonle Sap Lake’sfloodplain is around 198 km2, calculated from the land-usemap created by Japan International Cooperation Agency (JICA)in 1999 (34). McDonald et al. (35, 36) provide more informationabout the gallery forest and other floodplain vegetation.

DATA AND METHODS

Data for Analysis of the Present Flood-Pulse Characteristics

Reliable water-level data from Tonle Sap Lake are availableonly for the years 1997–2006 due to the unstable history ofCambodia. Water-level data exist also for the period 1923–1965(26), but these data were not used in this analysis because theaim was to analyze the current flood-pulse characteristics withcomparable data. The data were used for statistical analysis ofthe present flood-pulse parameters. Daily water-level data areavailable from the Kampong Luong station, located on the edgeof the permanent lake (station location in Fig. 1). The water-level data in this article are always expressed as elevation abovemean sea level (amsl) in Hatien, Vietnam.

Cumulative Impact Assessment Studies

Due to the considerable variety and ambiguity of differentdevelopment plans, the prediction of cumulative impacts ofongoing and planned development is extremely challenging. Forexample, existing cumulative impacts assessment (CIA) studiesfocusing on flow changes have applied different approaches,and used different values, and therefore provide differentestimates of the potential changes in flow. Three differentCIA studies were analyzed and used for flow-alterationpredictions in this article:

i) CIA 1: The Mekong River Commission (MRC) hascompiled a basinwide CIA under the Integrated Basin FlowManagement (IBFM) project by using Decision SupportFramework (DSF) modeling tools (19).

ii) CIA 2: The Asian Development Bank (ADB) conducted abasinwide CIA within the Nam Thuon 2 environmentalimpact assessment study (18).

iii) CIA 3: Adamson (17) compiled analyses of the downstreamhydrological impact of the Chinese cascade of dams.

The results of the aforementioned CIAs were used to analyzethe flow alteration due to upstream development and itsimpacts on the Tonle Sap flood pulse. A summary of the CIAstudies is presented in Table 1, and each CIA and itsbackground are briefly summarized here.

Figure 1. Map of Tonle Sap Lakeand its location, including thedistribution of gallery forest andthe protected areas.

186 Ambio Vol. 37, No. 3, May 2008� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

CIA 1: Mekong River Commission

The Mekong River Commission compiled the CIA as part of itsIntegrated Basin Flow Management project by using DecisionSupport Framework modeling tools (19). Altogether, sixscenarios, including the baseline, were developed. The scenariosrepresent the possible development alternatives occurring by2020. Here, the High Development Scenario has been selectedfor the analysis so as to be as comparable with the other CIAsas possible, especially with the ADB (18) CIA. In summary, thisscenario predicts that the overall character of the hydrograph ismaintained. However, the low flows are systematically higherthan the historically observed values, and the dry-season waterlevel will rise 0.15 m in Tonle Sap Lake. High flows are reducedin frequency, and the flood peak of the lake will be reduced by0.36 m (19).

CIA 2: Asian Development Bank

The ADB conducted a basinwide CIA within the Nam Thuon 2environmental impact assessment study (18). Two differentscenarios were constructed: 5 y and 20 y scenarios. Here, theresults from the 20 y scenario were used for further analysis.This scenario was selected so that the time span would beroughly equivalent to the MRC scenario. In the 20 y scenario,the ADB assessment (18) predicts that the average annualmaximum level of Tonle Sap Lake will be lowered in averageby 0.54 m and the dry season water levels increased by about0.60 m.

CIA 3: Adamson

Adamson (17) compiled an analysis of the downstreamhydrological impact of the planned Chinese cascade of dams.However, Adamson (17) did not considered other basindevelopment in the analysis. Two different scenarios wereanalyzed for selected sites on the main stem, one with 10% andthe other with 20% regulation of the dams. A 20% regulation inChina would result, according to Adamson (17), in a 50%increase in average discharge in Kratie (location in Fig. 1) inMarch. By this point, the potential impacts on the wet-seasonhydrology of the main stem are modest (17). The DanishHydraulic Institute (37) analyzed the impact of flow changes inKratie on the Tonle Sap water level by using the resultspresented by Adamson (17). The 20% regulation was predictedto increase the dry-season water level in Tonle Sap Lake by 0.3m (37).

Spatial Data

We used the available land-use, bathymetric, and elevationdata, as well as other related spatial data sets, for geographic

information systems (GIS) analysis to assess the impacts of flowalteration on the flooded area, including the gallery forest andprotected areas in the floodplain.

The following GIS data were used for the spatial analysis:

– The digital bathymetry model (DBM) for Tonle Sap Lake,Tonle Sap River, and their floodplains is based on thecombined data sets of the Certeza survey (38) of the TonleSap floodplain and the Mekong River Commission Hydro-graphic Atlas for the lake proper and Tonle Sap River (39).

– Land-use data around Tonle Sap, from which we use onlythe gallery forest layer, were compiled by JICA (34).

– Protected areas—the location of Tonle Sap BiosphereReserve and Ramsar site—are based on the Mekong RiverCommission spatial databases (40).

– Tonle Sap catchment boundaries, Tonle Sap tributaries,Mekong main stem, and the Lower Mekong Basin inundatedareas are based on the Mekong River Commission spatialdatabases (40).

– Country boundaries are based on the Mekong RiverCommission spatial databases (40).

Method Used for Analyzing the Present Flood-Pulse

Characteristics

Water-level data from the Kampong Luong station were used tocalculate the present flood-pulse characteristics. First, the lakearea (km2) and lake volume (km3) were expressed as functionsof water level (m) above the mean sea level (amsl), based on theDBM of the lake (21). Next, the maximum and minimum waterlevels and related flooded areas and volumes were calculatedfrom the available water-level data.

In this paper, flooding is considered to occur during theperiod when water level is higher than 1 m above the average 30d minimum water level of 1.44 m amsl, i.e., when the water levelis above 2.44 m amsl. The start, end, peak, and duration of theflood were analyzed for each water year in which data wereavailable. The water year starts on 1 May and ends on 30 Aprilto coincide with the flood timing.

The rate of water-level change was calculated based on dailywater-level changes. The average for each month was thencalculated.

Method Used for Assessment of Flow-Alteration Impacts on

Flood-Pulse Characteristics

The flow-alteration data from each CIA were first documentedand analyzed. The changes in the hydrograph due to flowalteration for the selected scenarios were then analyzed and usedto calculate the new water levels for the ‘‘average’’ water year inTonle Sap Lake. For the average year, the daily average water

Table 1. Summary of the flow alteration according to the CIA studies.

CIA 1 CIA 2 CIA 3

Basic assumptions and methods

Increased storage in the Mekong Basin 49.5 km3 54.9 km3 22.7 km3*Increased irrigation þ53% — —Other development activities taken into

accountIncreased domestic and industrial

use of water, and basin diversionsIncreased domestic and industrial

use of water—

Method used to estimate the developmentimpact on flow alterations

Hydrological and hydrodynamicmodel

Water balance and hydrodynamicmodel

Statistical analysis

Flow alteration impacts on Tonle Sap water levels

Wet-season water level �0.36 m �0.54 m N/ADry-season water level þ0.15 m þ0.60 m þ0.30 m

* Only reservoirs planned to be constructed in China are considered.

Ambio Vol. 37, No. 3, May 2008 187� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

levels were used. Finally, the flood-pulse parameters (methodspresented in previous section) were analyzed and compared tothe measured water levels. However, due to the lack of timevariability, it was not possible to analyze all of the parametersfor each CIA.

Method Used for Assessment of Flow-Alteration Impacts on

Flooded Area, Gallery Forest, and Protective Areas

Using the GIS data listed previously and ESRI’s ArcGIS 9.2software and its extensions, an assessment of the impacts causedby flow alteration on the flooded area—gallery forest andprotected areas—was achieved. The inundated areas were firstanalyzed for each water level by using the DBM constructed forTonle Sap Lake and its floodplain. This resulted in polygons ofinundated areas. The inundated areas were then mapped, andtheir extent was calculated using the polygons. Thereafter, theareas of gallery forest inundated due to the increased dry-seasonwater level were mapped. The same was done with the locationof the protected areas.

RESULTS

Current Flood-Pulse Characteristics

According to the statistical flood analysis, the timing of theflood peak is very regular in Tonle Sap Lake. However, the startand end dates of the flood (41) vary significantly, depending onthe timing of the flood on the main stem Mekong and local

rainfall in the Tonle Sap tributaries. The bottom of the lake is atan altitude of 0.6–0.7 m amsl. Thus, during low water, the lakewater depth is around 0.8 m.

On average, the flood starts on the 22nd of June, has its peakon the 7th of October, and ends on the 5th of March (Table 2).The average flood duration is 258 d, or 71% of the year. Duringthe analysis period, it varied from 205 d (1998) to 299 d (2000).The start of the flood varied from 27 May (2000) to 15 June(1998), and the end date varied from 04 February (1998) to 10April (2004), 49 d and 65 d, respectively. The peak of the floodoccurred more consistently around early October, varying from28 September to 17 October, i.e., within 20 d. Figure 2Aillustrates the timing of the flood, including start date, peakdate, and end date.

The water level varied from the lowest (daily minimum)value of 1.32 m amsl to the peak flood of 9.14 m amsl. At thesame time, the flooded area varied from 2215 km2 to 13 258km2, and the volume varied from 1.6 km3 to 59.7 km3. All thisinformation is summarized in Table 2.

Figure 2B illustrates the rate of water-level change, which isplotted along with the average water level over one flood cyclestarting in May. The rapidity of the water-level change wasgreatest during July, around 8 cm d�1. During the recedingflood, the most rapid water level change occurred duringDecember, around �6 cm d�1.

Flow-Alteration Impacts on Flood-Pulse Characteristics

The general results for each CIA, a comparison of the flood-pulse parameters calculated from the observed values (dailyaverage value for the entire study period), and parameters basedon the CIA scenario results are presented in Table 3. The floodduration would be reduced by 14 d, while the floodplain area,total flood volume, and amplitude would be reduced by 7%–16%.

The predicted average annual hydrograph, based on themonthly average water levels under present situation and CIA2, is presented in Figure 3. The figure shows a clear drop in theflood peak and increase of the low water levels in the mostextreme estimation of flow-alteration impacts on the waterlevels of Tonle Sap, i.e., the amplitude of the flooding woulddecrease. According to Lamberts (24), the water depth in thefloodplain determines the volume of the euphotic zone in whichaquatic photosynthesis can occur. Thus, smaller floodingamplitude would decrease the ecosystem productivity.

Flow-Alteration Impacts on Flooded Area

The dry-season water-level rise in Tonle Sap Lake due toupstream development on the Mekong River has beensummarized in Table 3. The results of the spatial analysisillustrate how the increased dry-season water level would impact

Table 2. Flood characteristics of Tonle Sap Lake. Flood is considered the time when water level (WL) is higher than 2.44 m above mean sealevel (amsl; 1 m above the average 30 d minimum water level of 1.44 m amsl). Years shown are flood years (01 May of year indicated to 30 Aprilof next year).

YearMin

WL (m)Area(km2)

Volume(km3)

MaxWL (m)

Area(km2)

Volume(km3)

Amplitude(m)

Startdate

Peakdate

Enddate

Duration(d)

1997 1.40 2307 1.7 9.33 13 542 61.5 7.9 07/07 08/10 13/02 2221998 1.24 2120 1.4 6.86 9637 33.0 5.6 15/07 12/10 04/02 2051999 1.38 2284 1.6 8.97 12 950 56.8 7.6 28/05 17/10 15/03 2932000 1.48 2402 1.8 10.36 15 278 76.1 8.9 27/05 28/09 21/03 2992001 1.42 2331 1.7 9.89 14 478 69.2 8.5 14/06 05/10 15/03 2752002 1.19 2061 1.3 10.10 14 834 72.2 8.9 20/06 06/10 16/03 2742003 1.27 2155 1.5 8.26 11 805 48.1 7.0 03/07 08/10 08/02 2212004 1.25 2131 1.4 9.20 13 327 59.8 8.0 21/06 04/10 10/04 2942005 1.26 2143 1.5 9.29 13 475 61.0 8.0 09/07 11/10 28/02 235Avg 1.32 2215 1.6 9.17 13 258 59.7 7.8 22/06 07/10 05/03 258

A woman selling noodles from her boat in the floating village of PrekToal, Tonle Sap Lake (Photo: M. Keskinen).

188 Ambio Vol. 37, No. 3, May 2008� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

the permanently inundated area (Fig. 4). The average 30 dminimum water level during the analysis period of 1997–2006was 1.44 m amsl, and this was used as a reference level for thedry-season water level. The lake area corresponding to a waterlevel of 1.44 m amsl is 2300 km2. Rises of 0.15 m, 0.3 m, and 0.6m, representing each analyzed CIA, would result in a permanentlake area of 2700 km2, 3000 km2, and 3200 km2, respectively.Thus, the permanent lake area would increase between 400 and1000 km2 (17%–40%).

The analysis based on CIA studies also predicts that the peakwater level would decrease and thus reduce the inundated area,i.e., size of the floodplain and of the lake. Thus, the area of thefloodplain would decrease, depending on the CIA data used,between 7% and 16% (Fig. 5 and Table 3). This would havedirect impact on the ecosystem productivity: the smaller thearea that becomes flooded, the smaller the area between aquaticand terrestrial phases (the ATTZ) and the smaller the potentialtransfer of floodplain terrestrial organic matter and energy intothe aquatic phase (24).

Flow-Alteration Impacts on Gallery Forest and Protected

Areas

The predicted dry-season water-level rise of 0.15–0.60 m wouldmean permanent inundation of large areas of gallery foreststripes located in the vicinity of the lakeshore (illustrated inFig. 6). This significant reduction, along with the 0.60 m dry-

season water-level rise total extinction of the gallery foreststripe, could have a significant impact on the entire Tonle Sapecosystem.

Two types of protected areas and in total four core sites arelocated within Tonle Sap Lake and its floodplains (locations inFig. 1):

i) Tonle Sap Biosphere Reserve (TSBR): 3 core areas(established in 1997), and

ii) Ramsar site (established in 1999).

The impact of the dry-season water-level rise on the TSBRand Ramsar core sites is illustrated in Figure 6. Of the total areaof 149 km2 of the Ramsar core site, 6%, 31%, or 83% (CIA 1 vs.CIA 3 vs. CIA 2) would be inundated by an increase in dry-season water level. Rise in water level would inundate 13%,23%, or 42% (CIA 1 vs. CIA 3 vs. CIA 2) of the TSBR core areas(total area of 423 km2). Thus, the impacts of the dry-seasonwater-level rise could be dramatic on these internationallyrecognized protected areas.

DISCUSSION

The flood-pulse parameter analysis is based on water-level datafrom Tonle Sap Lake. The quality of these data has beenchecked, but still there is the possibility of some inaccuracybecause only one measurement site has been used in a largelake. During periods of strong winds, the water level might

Table 3. Comparison of flood-pulse parameters calculated from the observed values and CIA scenario results (WL is water level).

MinWL (m)

Area(km2)

Volume(km3)

MaxWL (m)

Area(km2)

Volume(km3)

Amplitude(m)

Startdate

Peakdate

Enddate

Duration(d)

Observations 1.44 2386 1.8 9.09 13 138 58.3 7.6 15/06 08/10 06/03 265CIA 1 1.59 2532 2.0 8.73 12 559 53.8 7.1 N/A N/A N/A N/ADifference 0.15 147 0.2 �0.36 �579 �4.5 �0.5CIA 2 2.04 3107 3.2 8.49 12 168 50.8 6.4 25/06 08/10 01/03 251Difference 0.60 721 1.4 �0.60 �970 �7.5 �1.2 10 0 �5 �14CIA 3 1.74 2712 2.4 N/A N/A N/A N/A N/A N/A N/A N/ADifference 0.30 326 0.6

Figure 2. (A) Summary of flood timing, start, end, and flood peak date in the Tonle Sap Lake. (B) Rate of water-level change (green spheres,left vertical axis) and mean monthly water level (continuous line, right vertical axis). The observed minimum and maximum monthly averagewater levels are illustrated with dotted bars.

Ambio Vol. 37, No. 3, May 2008 189� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

differ significantly from one end of the lake to the other.However, this does not seem to be a major source of error sincethe wind velocities are normally relatively low, and strong windevents last only a few hours (42). The observed water levelsshow no evidence of any significant seiche oscillation, evidentlydue to the high friction caused by the floodplain vegetation. Theflooded area analysis is based on the bathymetry data collectedon 1960s by Certeza survey company (38). Even though the dataused are rather old, their accuracy has been tested andreaffirmed during the MRC/WUP-FIN project (42), and it isgenerally agreed that it is the best data set available.

The flow alterations predicted by the CIAs have similartrends in each of the studies, but the magnitude of the predictedchanges in the Tonle Sap hydrograph differs considerablybetween the scenarios. This is due to the different methods andassumptions used in each CIA. There are also many uncertain-

ties in the CIAs, e.g., in the scenario definition, and assumptionsinvolved in the modeling and analysis. Adamson (43) gives anoverview of the hydrological models applied in the Mekong andreports the limited ability of the MRC’s DSF model system,used in the CIA 1 (19), in some water-resource developmentsimulations, such as for hydropower. Also, due to the differentscenarios used in each CIA, the results cannot be directlycompared with each other. Thus, since part of the analysispresented in this paper is based on the CIA studies, the resultsare subject to the possible inaccuracies in the source data.

Various studies highlight the importance of future climatechange impacts on precipitation through intensification ofmonsoon rains, which would lead to increased flooding inSoutheast Asia (e.g., 44, 45), especially after 2030 (45), andparticularly in the Mekong (20, 46). Thus, incorporatingpredicted climate change impacts into the basin developmentscenarios would change the outcomes of spatial modelingpresented here (20). Cumulative assessment modeling isrequired where direct human development impact (reservoirs,irrigation, etc.) is modeled together with predicted climatechange impacts on precipitation and temperature. However, thetime spans of the events are different: the time span in climatechange studies is normally several decades, while in the foreseendirect basin development, it is only 10–20 y. This should be

Figure 4. Inundated areas due toincreased dry-season water level.

Figure 5. Schematic presentation of the possible impacts on thefloodplain extent due to changes in flow regime.

Figure 3. Change in yearly hydrograph due to the upstreamdevelopment, based on CIA 2 (18) results. Present monthly averagehydrograph is based on the years 1997–2005.

190 Ambio Vol. 37, No. 3, May 2008� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

taken into account when comparing, or combining, the differentimpacts. A global analysis of the potential effect of climatechange on river basins indicates that rivers impacted by dams orextensive development will require more management interven-tions to protect ecosystems and people than basins with free-flowing rivers (47).

The impacts of flow alterations on the analyzed flood-pulseparameters indicate a decline in every parameter in regard toecosystem productivity. In this paper, the results have beenpresented separately for each parameter. It is possible that thecombined impact of the parameters will result in severe impactson the lake’s ecosystem. Work on quantification of the impactsof the predicted flood-pulse changes on the ecosystemproductivity of Tonle Sap is ongoing. Lamberts (23, 24, 48) isfocusing on the terrestrial and aquatic primary production ofthe floodplain. This is necessary to improve our understandingof the impacts of development and to make reliable estimates ofthe impact of flow alteration on the lake’s productivity andconsequently fisheries, and further link those results with thesocioeconomic livelihoods analysis (e.g., 27).

In this paper, the hydrological impacts of the large-scalewater-resources development have been discussed and analyzed.However, the impact on water quality can be also significant,although not yet well studied. The impact of future hydropowerconstruction on the basin sediment balance has been estimatedto be considerable as well, impacting directly on the nutrientsand ecosystem productivity (13). To understand better thepossible impacts of development on the water quality andsediment, further basinwide research is urgently required toguide decisions for minimizing the impacts of flow alterations.

CONCLUSIONS

The recent cumulative impact assessment studies of the MekongBasin are consistent, and they state that increased developmentactivities, especially construction of hydropower dams andreservoirs, large irrigation schemes, and rapid urban develop-ment, will result in higher dry-season water levels and lowerflood peaks. This paper aimed to assess the impacts of thesepredicted flow alterations on the flood characteristics, galleryforest, and protected areas of Tonle Sap Lake, by using the

results of the CIA studies, hydrological data-series analysis, andspatial analysis.

The flow alterations in the Mekong River, due to develop-ment activities in the upstream countries, will cause negativeeffects for ecosystem productivity, and thus also for livelihoodsof the inhabitants of Tonle Sap floodplain, who directly dependon the lake’s natural resources. Thus, it is extremely importantfor Tonle Sap, and other floodplains in the Lower MekongBasin, to maintain the natural hydrological regime of theMekong main stem. Changes in the dry-season water levels,estimated to increase the water level in Tonle Sap Lake by 0.15–0.60 m, would, in particular, be harmful to the presentecosystem of the lake.

Relatively small rises in the dry-season lake water levelwould permanently inundate disproportionately large areas offloodplain, rendering it inaccessible to floodplain vegetationand eroding the productivity basis of the ecosystem by reducingthe inundated area and duration and amplitude of flooding. Thelake extension would cause permanent submersion, in essencedestruction, of considerable areas of the gallery forest stripesurrounding the lake in the floodplain. The reduction of galleryforest area would mean loss of livelihood sources for asignificant number of people, due to both loss of gallery forestsper se and the consequent negative effects on aquaticproductivity (49). Large parts of internationally protected areasunder Ramsar Convention and the Tonle Sap BiosphereReserve Programme (under UNESCO) would also be lost bypermanent submersion in the lake.

Hydropower dam regulation will probably be the main causeof flow alterations in the near foreseen future. Other importantactors are land-cover changes and irrigation schemes. Climatechange may play an equally important role from the latter partof the century on. Integrated, cross-boundary planning,involving both downstream and upstream countries, andcumulative impact assessment are urgently required to minimizethe impacts of the flow alteration.

References and Notes

1. Vorosmarty, C.J. 2000. Anthropogenic disturbance of the terrestrial water cycle.Bioscience 50, 753–765.

2. Water renewal time can be defined as follows: ‘‘ratio between its [this case river] totalvolume and the daily volume flux . . . [ ] . . . leaving it.’’ Based on (3).

Figure 6. Impacted gallery forestareas (solid polygons) and protect-ed areas (hatched areas, TSBRareas in points and Ramsar site inlines).

Ambio Vol. 37, No. 3, May 2008 191� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se

3. Jouon, A., Douillet, P., Ouillon, S. and Fraunie, P. 2006. Calculations of hydrodynamictime parameters in a semi-opened coastal zone using a 3D hydrodynamic model. Cont.Shelf Res. 26, 1395–1415.

4. Golubev, G. 1993. Sustainable water development: Implications for the future. WaterResour. Dev. 9, 127–154.

5. Skole, D.L., Justice, C.O., Townsend, J.R.G. and Janetos, A.C. 1997. A land coverchange monitoring program: Strategy for an international effort. Mitig. Adap. Strat.Glob. Change 2, 157–175.

6. Nijssen, B., O’Donnell, G., Hamlet, A. and Lettenmaier, D. 2001. Hydrologic sensitivityof global rivers to climate change. Clim. Change 50, 143–175.

7. Kundzewicz, Z.W., Mata, L.J., Arnell, N.W., Doll, P., Kabat, P., Jimenez, B., Miller,K.A., Oki, T., et al. 2007. Freshwater resources and their management. In: ClimateChange 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II tothe Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Parry,M.L., Canziani, O.F., Palutikof, J.P., Linden, P.J.v.d. and Hanson, C.E. (eds).Cambridge University Press, Cambridge, UK, p. 173–210.

8. Kvist, L.P. and Nebel, G. 2001. A review of Peruvian flood plain forests: Ecosystems,inhabitants and resource use. For. Ecol. Manage. 150, 3–26.

9. Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-Being: Wetlandsand Water Synthesis. World Resources Institute, Washington DC, 74 pp.

10. Finlayson, C.M. and Spiers, A.G. 1999. Global review of wetland resources andpriorities for wetland inventory. Supervising Scientist Report 144/Wetlands Interna-tional Publication 53, Supervising Scientist, Canberra, Australia.

11. Mekong River Commission (MRC). 2005. Overview of the Hydrology of the MekongBasin. Mekong River Commission, Vientiane, 73 pp.

12. Dore, J. and Yu, X. 2004. Yunnan Hydropower Expansion: Update on China’s EnergyIndustry Reforms & the Nu, Lancang & Jinsha Hydropower Dams. Chiang MaiUniversity’s Unit for Social and Environmental Research & Green Watershed,Kunming, People’s Republic of China, 38 pp.

13. Kummu, M. and Varis, O. 2007. Sediment-related impacts due to upstream reservoirtrapping, the Lower Mekong River. Geomorphology 85, 275–293.

14. McCormack, G. 2001. Water margins: Competing paradigms in China. Crit. Asian Stud.33, 5–30.

15. Plinston, D. and He, D. 1999. Water resources and hydropower. In: ADB TA-3139Policies and Strategies for Sustainable Development of the Lancang River Basin. AsianDevelopment Bank, Manila, 35 pp.

16. Kummu, M., Varis, O. and Sarkkula, J. 2008. Impacts of land surface changes onregional hydrology—mainland Southeast Asia. In: A Common Need for Action: GlobalEnvironmental Change and Development in Monsoon Southeast Asia. Lebel, L.,Snidvongs, A., Chen, C-T.A. and Daniel, R. (eds). Gerakbudaya, Kuala Lampur,Malesia. (In press).

17. Adamson, P.T. 2001. Hydrological perspectives on the Lower Mekong Basin: Thepotential impacts of hydropower developments in Yunnan on the downstream flowregime. Int. Water Power Dam Constr. March 2001, 16–21.

18. Asian Development Bank. 2004. Cumulative Impact Analysis and Nam Theun 2Contributions. Final Report. Prepared by NORPLAN and EcoLao for AsianDevelopment Bank, Manila, 143 pp.

19. World Bank. 2004. Modelled observations on development scenarios in the LowerMekong Basin. In: Mekong Regional Water Resources Assistance Strategy. Prepared forthe World Bank with Mekong River Commission cooperation, Vientiane, 142 pp.

20. Penny, D. 2008. The Mekong at climatic crossroads: Lessons from the geological past.Ambio 37, 164–169.

21. MRCS/WUP-FIN. Technical Paper No. 5—Tonle Sap Lake Water BalanceCalculations, WUP-FIN Phase II—Hydrological, Environmental and Socio-EconomicModelling Tools for the Lower Mekong Basin Assessment. Mekong River Commissionand Finnish Environment Institute Consultancy Consortium, Vientiane, 32 pp.

22. Welcomme, R.L. and Halls, A.S. 2004. Dependence of tropical river fisheries on flow.In: Proceedings of the Second International Symposium on the Management of LargeRivers for Fisheries. Volume II. RAP Publication 2004/16. Food and AgricultureOrganization of the United Nations (FAO), Bangkok, pp. 267–283.

23. Lamberts, D. 2006. The Tonle Sap Lake as a productive ecosystem. Int. J. Water Resour.Dev. 22, 481–495.

24. Lamberts, D. 2008. Little impact, much damage: The consequences of Mekong Riverflow alterations for the Tonle Sap ecosystem. In: Modern Myths of the Mekong—ACritical Review of Water and Development Concepts, Principles and Policies. Kummu,M., Varis, O. and Keskinen, M. (eds). Water and Development Publications, HelsinkiUniversity of Technology, Finland, pp. 3–18.

25. Bonheur, N. 2001. Tonle Sap Ecosystem and Value. Technical Coordination Unit forTonle Sap, Ministry of Environment, Phnom Penh, Cambodia, 21 pp.

26. Campbell, I., Poole, C., Giesen, W. and Valbo-Jorgensen, J. 2006. Species diversity andecology of Tonle Sap Great Lake, Cambodia. Aquat. Sci. Res. Bound. 68, 355–373.

27. Keskinen, M. 2006. The lake with floating villages: A socio-economic analysis of theTonle Sap Lake. Int. J. Water Resour. Dev. 22, 463–480.

28. Kummu, M., Sarkkula, J., Koponen, J. and Nikula, J. 2006. Ecosystem management ofTonle Sap Lake: Integrated modelling approach. International Int. J. Water Resour. Dev.22, 497–519.

29. Sverdrup-Jensen, S. 2002. Fisheries in the Lower Mekong Basin: Status and perspectives.MRC Technical Paper No. 6. Mekong River Commission, Phnom Penh, 103 pp.

30. UNESCO. 2006. Biosphere Reserves–World Network. UNESCO-MAB (Man and theBiosphere) Secretariat, Paris, France. (http://www.unesco.org/mab/)

31. Ramsar Convention on Wetlands. 2007. The list of wetlands of internationalimportance. Ramsar Convention Secretariat, Gland, Switzerland. (http://www.ramsar.org/)

32. Junk, W.J. 1997. The Central Amazon Floodplain: Ecology of a Pulsing System.Ecological Studies Volume 126., Springer, Berlin, 525 pp.

33. Kummu, M., Penny, D., Sarkkula, J. and Koponen, J. 2008. Sediment: Curse or blessingfor Tonle Sap Lake? Ambio 37, 158–163.

34. JICA. 1999. Cambodia reconnaissance survey digital data. Ministry of Public Worksand Transportation (MPWT), Kingdom of Cambodia. Japan International CooperationAgency (JICA), Phnom Penh, Cambodia.

35. McDonald, A., Pech, B., Phauk, V. and Leev, B. 1997. Plant Communities of the TonleSap Floodplain. Contribution to the nomination of the Tonle Sap as Biosphere Reservefor UNESCO’s ‘‘Man in the Biosphere Program.’’ UNESCO, Phnom Penh, 30 pp.

36. McDonald, A. and Veasna, S. 1996. Floristic reconnaissance of Prek Sramoach lake andvicinity, Tonle Sap. Field document. Participatory natural resources management in theTonle Sap region project. FAO, Phnom Penh, 72 pp.

37. Danish Hydraulic Institute. 2004. Study on natural reverse flow in the Tonle Sap River.Prepared by DHI (Danish Hydraulic Institute) Water & Environment for the MekongRiver Commission Secretariat, Vientiane, Lao PDR, 26 pp.

38. Certeza Surveying. 1964. Final Report to the Committee for Coordination ofInvestigations of the Lower Mekong Basin. Certeza, Quezon City, The Philippines.

39. Mekong River Commission. 1999. Hydrographical Atlas for Tonle Sap. Mekong RiverCommission, Phnom Penh, Cambodia.

40. Mekong River Commission. 2006. Spatial databases of Mekong River Commission.Mekong River Commission (MRC), Vientiane, Lao PDR. (http://www.mrcmekong.org/spatial/data_catalog.htm)

41. The term flood used in this paper for the Tonle Sap Lake is defined in details in Sectiontitled ‘‘Method Used for Analyzing the Present Flood-Pulse Characteristics.’’

42. MRCS/WUP-FIN. 2003. WUP-FIN Phase I: Modelling Tonle Sap for EnvironmentalImpact Assessment and Management Support. Final Report. Mekong River Commissionand Finnish Environment Institute Consultancy Consortium, Phnom Penh, 107 pp.

43. Adamson, P. 2006. Hydrological and water resources modelling in the Mekong Region:A brief overview. In: Exploring Water Futures Together: Mekong Region WatersDialogue. Resource papers from regional dialogue. The World Conservation Union,Thailand Environment Institute, International Water Management Institute, MekongProgram on Water Environment and Resilience, Vientiane, pp. 69–74.

44. Bhaskaran, B. and Mitchell, J.F.B. 1998. Simulated changes in Southeast Asianmonsoon precipitation resulting from anthropogenic emissions. Int. J. Climatol. 18,1455–1462.

45. Hu, Z.-Z., Latif, M., Roeckner, E. and Bengtsson, L. 2000. Intensified Asian summermonsoon and its variability in a coupled model forced by increasing greenhouse gasconcentrations. Geophys. Res. Lett. 27, 2681–2684.

46. Falloon, P.D. and Betts, R.A. 2006. The impact of climate change on global river flow inHadGEM1 simulations. Atmos. Sci. Lett. 7, 62–68.

47. Palmer, M.A., Reidy Liermann, C.A., Nilsson, C., Florke, M., Alcamo, J., Lake, P.S.and Bond, N. 2007. Climate change and the world’s river basins: Anticipatingmanagement options. Front. Ecol. Environ. 6, 81–89.

48. Lamberts, D. and Koponen, J. 2008. Flood-pulse alterations and productivity of theTonle Sap ecosystem: A model for impact assessment. Ambio 37, 178–184.

49. Keskinen, M., Kakonen, M., Prom, T. and Varis, O. 2007. The Tonle Sap Lake,Cambodia: Water-related conflicts with abundance of water. Econ. Peace Sec. J. 2, 49–59.

50. The authors are grateful to the WUP-FIN team, especially Yin Savuth and JormaKoponen. The staff at the Water Resources Laboratory of Helsinki University ofTechnology, particularly Olli Varis, Marko Keskinen, Pertti Vakkilainen, and TuomoKarvonen are highly acknowledged. We thank also Dirk Lamberts for his continuoussupport in the ecosystem productivity part of the research, Michael Huggins forproofreading the manuscript, and Des Walling and Dan Penny for their insightfulcomments on the manuscript. This work has received funding from the Academy ofFinland project 211010. The first author was funded by the Academy of Finland project111672 and Maa-ja vesitekniikan tuki ry. The work was partly completed within theWUP-FIN project under the Mekong River Commission.

Matti Kummu is a researcher and PhD candidate at the WaterResources Laboratory of Helsinki University of Technology,Finland. His research concentrates on human impact onhydrology and sediment transport in the Mekong region. Hehas several years of work experience from the Mekong Basin,under the WUP-FIN Project. His address: Helsinki Universityof Technology, Water and Development Research Group, POBox 5200, FIN-02015 Espoo, Finland.E-mail: matti.kummu@iki.fi.

Juha Sarkkula is a senior research scientist at the FinnishEnvironment Institute. He has worked as a team leader for theWUP-FIN project and has been working with the Tonle SapLake for over six years. He has long-term experience inhydrological, hydrodynamic, sediment and water-quality mod-eling projects for water resources management and environ-mental impact assessment. His address: Finnish EnvironmentInstitute, Mechelininkatu 34a, 00260 Helsinki, Finland.E-mail: juha.sarkkula@environment.fi.

192 Ambio Vol. 37, No. 3, May 2008� Royal Swedish Academy of Sciences 2008http://www.ambio.kva.se