Mekong Dam Monitor: Providing a Picture of How Dams Impact the Mekong River

13 July 2022

Mekong Dam Monitor data reveal impacts of upstream dams and point to ways transboundary cooperation can improve downstream outcomes during times of drought and crisis. Brian Eyler, Alan Basist, Regan Kwan, Courtney Weatherby and Claude Williams give more details.

The Mekong Dam Monitor’s (MDM) near-real time tracking of dam operations and river flow provides unique insight into the changes that dams have caused to the Mekong River’s natural flow. This article summarises three new findings from the first year of MDM activity using data produced by the MDM as well as the Mekong River Commission’s (MRC) historical database. Data shows that dams have exacerbated wet season droughts and in some places have inexorably altered the natural flow patterns of Southeast Asia’s most productive river. The MDM has also proven a useful early warning tool for communities to mitigate the immediate impacts of day-to-day upstream dam operations.

Key Findings

  1. Dam operations exacerbate wet season droughts: A severe lack of rainfall between 2019-2021 was the core driver of drought and low river flows, but dam restrictions during this period significantly impacted wet season flow and exacerbated drought conditions throughout the basin. More studies and additional years of data will help determine whether low flow from 2019 to 2021 represent a “new normal” or natural variability in weather patterns. In any event, an international water sharing agreement, which guarantees a minimum level of flow from upstream dams during period of drought would mitigate future crises.
  2. The impact is profound, especially farther upstream: At some parts of the Mekong Basin the impoundment of water and unnatural releases from dams have entirely altered the natural flow of the river. In other localities, the situation is more complicated. This matters because ecological processes above Vientiane, Laos, and the social practices that rely on these processes, have likely been inexorably altered. Further study is needed to determine the specific effects of dams on the Tonle Sap reversal and flooding process. Funds to support ecosystem integrity and other mitigation methods could offer opportunities to communities impacted by the loss of the Mekong’s natural flood pulse.
  3. The Mekong Dam Monitor is a useful early warning tool: MDM researchers are able to notify vulnerable communities living along the river of sudden changes in river level and provide them an early warning to mitigate the impact from these fluctuations. In 2021, we tracked and issued 22 notifications for instances when upstream dams in China suddenly altered the river level along the Thai-Lao border by 0.5 meters or more. These sudden shocks damage local communities, wetlands, and habitats of endangered species. These shocks can be largely mitigated by altering the operation patterns of Jinghong dam.

A note on data: China has yet to share sufficient data on the operation of its dam cascade. Together, Mekong Dam Monitor observations and Mekong River Commission data provide a fuller picture of what is happening in the Mekong. This picture could be even clearer if real-time data on sediment flow, fish migration, and other key indicators of impact were available. Nonetheless, historical data from the MRC and near-real time dam operational data from the MDM are sufficient to spearhead a push for more productive transboundary cooperation, promote transparency, and narrow accountability gaps.

1) Dams operations exacerbate wet season droughts

A severe lack of rainfall between 2019-2021 was the core driver of drought and low river flows, but dam restrictions during this period significantly impacted wet season flow and exacerbated drought conditions throughout the lower basin.

A recent Mekong River Commission report provides compelling evidence that a severe lack of rainfall between 2019 and 2021 was the core driver of low annual flows and drought in the Mekong Basin. MDM data can be used to support this finding. For example, the MDM’s natural flow models show that wet season peaks were considerably lower from 2019 to 2021 than most previous years. This indicates that the river would have flowed naturally at much lower levels during the wet season months of June to November given the actual amount of rainfall or “wetness” in the basin in these years.

However, when actual gauge data for Chiang Saen and Vientiane are compared to natural flow levels, observed river levels during the wet season across this period are still much lower than what natural conditions would provide. That difference is the result of upstream dams which reduce the Mekong’s wet season pulse. Even though there was a drought, the dam operators also severely restricted the already-limited water flow to fill their reservoirs. This greatly compounded the impact of the drought downstream.

Flow data

The difference between actual river gauge measurements and what the river flow would have been based on precipitation and wetness shows the impact that upstream dams have on river flow.

At Chiang Saen, Thailand

During the 2019-2021 drought years, actual water flow was lower than natural (undammed) river flow would have been. This shows the impact of dams.

As the chart shows, water restrictions behind upstream dams significantly reduced wet season flow peaks from 2019-2021.

Water levels at Vientiane, Laos

The pattern continues downstream at Vientiane. The drought would have brought lower than average flows, but upstream dam impacts reduce wet season peaks to much lower levels.

At Stung Treng, Cambodia

Lower flows are also observed far downstream at Stung Treng, Cambodia during the drought years.

During the wet season, floods drive important processes in the Mekong such as the Tonle Sap reversal, an event which makes the Tonle Sap Lake the world’s largest and most productive inland fishery. Moreover, natural wet season floods drive agricultural production in the Mekong Delta, which provides Vietnam with 20% of its annual GDP, half of its rice production, and most of its fruits and aquaculture. The Mekong River Commission continues to promote an evidence-based position that wet-season floods provide a value of US$8-10 billion annually for fisheries, agriculture, and freshwater availability that far outweighs the US$60-70 million in annual damages caused by extreme flood events.

The MDM’s virtual gauges calculate that the 45 largest dams in the Mekong Basin can hold about 47Bm3 of active storage, while the remaining 100+ smaller dams in the Mekong likely hold a fraction of this. Active storage is the variable amount of water that a dam’s reservoir can hold to produce hydropower or use for other purposes. The data shows that most large dams release some or all of their active storage during the dry season and recharge their reservoirs during the wet season by restricting the natural flow of the river. The chart below compares total active storage—-or the maximum amount of water that can be restricted by dams during the wet season—to wet season flow at the Stung Treng gauge in Cambodia. The 1910-2007 average wet season flow bar represents 97 years of natural wet season flow for the basin. China’s Xiaowan Dam (active storage 11.17Bm3) went into operation in 2008 and the larger Nuozhadu Dam (active storage: 11.35Bm3) became operational in 2012. Post-2008, flow during the wet seasons is impacted by large restrictions in upstream dams. In addition to wet season restrictions, some of this loss of flow since 2008 is related to dead storage in the reservoirs and recent drought conditions.

How much wet season flow can major dams hold?

The MDM’s virtual gauge data (see the Xiaowan and Nuozhadu charts below) show that large storage dams generally operate the same way year after year regardless of low flow and drought conditions downstream. In other words, they restrict just as much water in dry years as they do in normal years.

Water storage at the Xiaowan Dam over time

Water storage at the Nuozhadu Dam over time

To illustrate, Xiaowan and Nuozhadu restricted similar amounts of water in 2018, a year with above average rainfall, as they did in 2020, which was the driest year on record. In 2018, the two dams withheld 20.4Bm3 of wet season flow, and in 2020 the two dams withheld 20.1Bm3 of wet season flow. While the overall amount of water restricted during the wet season may not change much from year to year, during periods of drought in the lower basin, these upper basin restrictions make up a larger part of the overall flow. For example: In 2018, total wet season flow to Stung Treng would have been 5% higher without the Xiaowan and Nuozhadu dams. But water flow would have been 9.3% higher in 2020 without them because the year was so unusually dry. 

[Total wet season flow is calculated similarly to annual flow calculations, but only days from June to November are summed. We assume wet season water restricted by upstream dams would have passed by the Stung Treng gauge sometime during the wet season in question if it were part of natural flow.]

This impact is magnified the further upstream we go: The closer a location is to large dams, the more profound the impact, particularly during low flow years. China’s dam impacts are measurably larger in portions of the river, such as around Chiang Saen, Thailand, which are relatively close to the Xiaowan and Nuozhadu Dams. Without China’s upstream dam restrictions, Chiang Sean’s total wet season flow in 2020 would have been two-thirds greater than it was! Even in 2018, a year with greater than usual rainfall, total wet season flow to Chiang Saen would have been a third higher.

Another illustration of this effect can be calculated by combining MRC wet season flow data for 2021 with the wet season flow restricted in the 45 largest dams.  Without upstream dam restrictions, the flow at Stung Treng would have been 294 cubic kilometers, 11% higher than what was recorded at the MRC gauge. Xiaowan and Nuozhadu account for almost twice as much of the missing water compared to the 43 other dams tracked in the Mekong Dam Monitor.


Be cautious about declaring a "new normal."

Further studies are needed to determine how climate change is affecting weather patterns in the Mekong. Planners should be cautious in declaring a “new normal” of low wet season flow patterns. Historical records for the Mekong show periods of long drought followed by a return to normal flow patterns. Science predicts that as the climate changes, there will be more extreme flood and drought events, but there is currently no evidence that low wet season flows will become a permanent feature of the Mekong.

A negotiated crisis response mechanism is necessary to relieve future wet season droughts.

It is imperative to establish a constructive feedback mechanism so that during wet-season droughts, dam operators can alter operations to allow more water to pass through reservoirs instead of storing it. Allowing that water to pass through the dam instead of storing it in reservoirs would feed the river downstream and assist in maintaining the river’s integrity during a drought year. This can only be achieved through an international agreement on water sharing and by establishing basin-wide forecasting and early warning mechanisms. An agreement which provides incentives to alter the operation of the largest storage dams on the mainstream in China during times of need is the simplest physical solution.

The table below shows that a majority of the basin’s active storage is in China’s dams. In fact, the Xiaowan and Nuozhadu dams alone hold more than 50% of the total storage capacity in the Mekong basin and are both owned by the same parent company, Huaneng Hydrolancang. A financial incentive mechanism in which Huaneng Hydrolancang guarantees a minimum (and maximum) level of flow would be extremely beneficial in both the wet and dry season. An insurance policy is one method to offset the resulting hydropower revenue loss, though the details of underwriting such an agreement would need to be negotiated among all the impacted parties. Such an approach would cost money, but a negotiated solution would likely save the Mekong countries more than it would cost, given the billions of dollars of economic benefits that natural flow provides to the Lower Mekong.

A more difficult but more equitable option would be to establish a minimum flow requirement throughout the year for every dam in the Mekong, including those in tributaries, guaranteeing a minimum flow that supports the natural rhythm of the river. This would require a much higher degree of coordination than the previous option which involves only two dams. A third method to restore wet season flow during times of crisis would be to build large storage dams in the downstream for this specific purpose; however, this option comes with a very high price tag and introduces additional risk to the natural resources of the river. A cost/benefit analysis between building new reservoirs, compared to insurance underwriting, could be part of a feasibility study of each approach. 

The impact is profound, especially farther upstream

At some parts of the Mekong, upstream dams have entirely altered the natural flow of the river. In other localities, the situation is more complicated.

Over the last six years at Chiang Saen and Pak Chom, Thailand, the natural rise and fall of the river that can be clearly seen in the historical baseline is no longer apparent in the actual mainstream flow. The historic baseline chart shows low river levels during the dry season months that rise to a peak during the wet season months before falling again to next dry season. In contrast, during most years since 2017 the Mekong reaches some of its highest levels during the dry season and lowest levels during the wet season. This is caused by dams upstream that are releasing nearly all their active storage during the dry season and then filling their reservoirs during the wet season, artificially raising dry season river levels and decreasing wet season river levels.

Of the Mekong Dam Monitor’s six impact areas, Chiang Saen and Pak Chom are the closest to the largest upstream dams and most clearly reveal their impact. An exception to natural flow alteration can be seen in 2020, when dry season river levels are consistently low. MDM virtual gauges show the Nuozhadu dam did not release much water during the 2020 dry season, likely due to low rainfall in 2019 and low demand for hydroelectricity in China due to the coronavirus pandemic. Even though 2020 river levels at Pak Chom were lower than other post-2007 years, flows were still consistently higher in the dry season and lower in the wet season than the pre-2007 average. This shows that the operation of upstream dams was likely still the main cause of changes to seasonal flow patterns at Pak Chom, and those seasonal flow patterns have been entirely altered.

Historical vs recent flow: upstream

The blue historical baseline on each of the charts below show the reliable annual rise and fall of the Mekong before the impact of seasonal restriction and releases of large upstream reservoirs. 

The red recent average shows the flow over the last 5 years.

Chiang Saen river level

Pak Chom river level

These changes to natural flow have a pronounced impact on communities that rely on fishing and other resources for livelihoods and food security. The ecological processes in this part of the river, and the social practices that rely on them, have likely been inexorably altered by these changes. These processes range from fish migration to the flooding of wetlands to nesting patterns of birds that rely on exposed riverbanks in the dry season, among other yet-to-be documented ecological changes. All these changes to the Mekong’s ecology threaten the habitats and viability of numerous species. Higher dry season levels also wash out riverside gardens, which are an important source of food and marketable crops for locals.

Farther downstream, it is more difficult to conclude that upstream dams inexorably alter wet and dry season flow patterns.  MDM virtual gauges at Stung Treng and the Tonle Sap still show that a version of the historical of the wet season pulse is happening. However, most dry season river levels since 2017 are above the historic baseline and most wet season levels are below it, suggesting that dams still have an impact far downstream.

Historical vs recent flow: downstream

The blue historical baseline on each of the charts below show the reliable annual rise and fall of the Mekong before the impact of seasonal restriction and releases of large upstream reservoirs. 

The red recent average shows the flow over the last 5 years.

Stung Treng river level

Tonle Sap bottleneck river level

While impacts at the impact areas in Cambodia might not appear as profound as those along the Thai-Lao border, higher dry season levels are causing the rapid die-off of trees on islands in the middle of the Mekong between the Lao border and Stung Treng, which are designated as RAMSAR protected wetlands. These trees form important habitats for birds such as the critically endangered white shouldered ibis, and the trees’ root systems also provide habitat for fish species. This stretch of the river is also home to the Mekong’s dwindling and endangered dolphin population, and changes to natural flow likely impact dolphin habitat and food sources. Ultimately more study is needed to determine specifically how upstream dam operations are impacting river ecology along the course of the Mekong mainstream in Cambodia.

In addition, more study is needed to understand the impact of upstream dam operations on the Tonle Sap reversal. Analysis from the MDM above suggests that wet season flows at Stung Treng would have been 11% higher in 2021 without flow restrictions caused by the 45 large dams upstream. However, does that mean the Tonle Sap would have received 11% more water from the mainstream during the reversal period in 2021? This is not known. How would the reversal have happened differently if there were more and longer-lasting wet season flow? This is also not currently known. Getting these answers is important. We do know the longer the reversal period lasts, the more water goes into the Tonle Sap from the Mekong mainstream, and this corresponds with higher annual fish catch from the Tonle Sap. MDM data together with MRC data brings the science steps closer to understanding these impacts.


Further study is needed to determine the specific effects of higher dry season levels and lower wet season levels on the Tonle Sap

Keskinen et al used predictive models to highlight potential effects in a 2013 study, but now that impacts to the Tonle Sap reversal and flooding process are being realised, studies can point to specific impacts and provide recommendations for mitigation and adaptation. It is unlikely that upstream dams will change their operational behavior under normal weather conditions, and some changes to the Mekong’s hydrological cycle are likely permanent. Funds to support ecosystem integrity and other mitigation methods could bring opportunities to communities most impacted by the loss of the Mekong’s natural flood pulse.

The Mekong Dam Monitor is a useful early warning tool

The MDM is a useful early warning tool to notify vulnerable communities living along the river of sudden changes in river level and provide communities with time to adapt.

60-80 years of historical data

To demonstrate how the river would flow under normal conditions, the MDM includes historical baseline data for river levels at six impact areas along the course of the mainstream and Tonle Sap. To determine historical baseline river levels, we calibrated MRC data to these locations and chose a period ranging between sixty to eighty years, depending on data availability, and ending in 2007, the year before large storage dams began to impact the Mekong.

In 2021, the MDM project team issued 22 alerts to local communities, government agencies, and media outlets, providing at least 48 hours of advanced notice when sudden releases or restrictions from upstream dams would cause a change in river level of more than 50 centimetres. Two of those alerts were for sudden increases of a meter or more, enough to sweep away unmoored boats and farming and fishing assets lying along the riverbank. A one metre increase in water level is also enough to flood some riverside gardens commonly planted in the dry season when the river level is low. Additionally, four of those instances created a sudden decrease in river level of one metre or more.

While all 22 of these sudden alterations in river level created a shock to the river’s ecological processes, none of them were accompanied by a timely warning from China. Despite promises of early warning and notification of abnormal changes to flow to downstream countries, upstream dam operators have yet to fulfill their promises.

These sudden changes in river level mostly come from hydropeaking operations at the Jinghong Dam, which also passes water from the cascade upstream. Hydropeaking is a method of dam operation which maximizes energy production over short periods of high electricity demand by using sudden releases of water to generate electricity and then abruptly restricting water in order to quickly refill the reservoir so the pattern can be repeated. This practice delivers severe ecological shocks to fish populations, forests and wetlands along a river’s course and creates risks for communities downstream [1]. The combined effects of hydropeaking on the Jinghong Dam have been impacting lives and biodiversity along the Thai/Lao border since it went into operation in 2009. Sudden rises in river levels also come when upstream dams conduct extreme reservoir level drawdowns. This can happen if a reservoir is close to or above a normal maximum level and a major weather event threatens to inundate the reservoir and possibly create a situation where the dam wall fails.

In 2021, there were also scores of instances when China’s upstream dams delivered sudden changes in river level less than 50 centimetres for which the MDM did not issue an alert. The MDM team will continue to issue timely alerts to communities along the Thai/Lao border affected by upstream hydropeaking in addition to government agencies, media, and other stakeholders.

Recommendations include that further study can pinpoint exact impacts, but as a minimal best practice, upstream dam operators should provide timely and accurate notification of sudden releases to downstream communities. As a best practice, hydropeaking from the last dams of the cascade should be minimised.

The Jinghong Dam’s reservoir could be used as a buffer between the movement of water from upper basin dams into the lower basin. Or the Jinghong Dam could be operated in a way to re-introduce natural flow into the downstream. Or the planned Ganlanba Dam, to be located 30 kilometers downstream of Jinghong, could be operated specifically for this purpose. Concerns from Chinese NGOs and local economic interests over that dam’s local impacts continue to put the construction of the Ganlanba dam on hold.


[1] Atle Harby et al, ‘Ecological Impacts of Hydro Peaking in Rivers’ in Bjørn Honningsvåg, Grethe Holm Midttømme, Kjell Repp, Kjetil Arne Vaskinn and Trond Westeren (eds), Hydropower in the New Millenium (London: CRC Press, 2001), pp. 249–56.

[2] Marko Keskinen, Matti Kummu, Aura Salmivaara, Someth Paradis, Hannu Lauri, Hans de Moel, Philip Ward and Pech Sokhem (2011): Exploring Tonle Sap Futures: Baseline results from hydrological and livelihood analyses.



The authors are Brian Eyler, Senior Fellow and Director of the Southeast Asia Programme and Energy, Water and Sustainability at the Stimson Centre; Alan Basist, President, Eyes on Earth; Regan Kwan, Research Associate with the Southeast Asia Programme and the Energy, Water and Sustainability Programme; Courtney Weatherby, Deputy Director of the Stimson Center’s Southeast Asia Programme and a Research Analyst with the Energy, Water, & Sustainability Programme; and Claude Williams, a retired climate scientist.

This report was originally published on 3 March  2022 under the title "Mekong Dam Monitor at One Year: What Have We Learned" on the Stimson Center website:





Xiaowan dam China’s Xiaowan Dam and the Nuozhadu Dam hold more than 50% of the Mekong’s active storage. In 2020, the driest year on record, these dams could have been used to relieve drought during the wet season by increasing downstream flow by 63% to 9% at various parts of the river. Instead, their wet season restrictions were used to produce hydropower for markets in China during 2021’s dry season. Image: Brian Eyler
Mekong Dam Monitor sites Mekong Dam Monitor sites
Tonle Sap fish catch What is the Tonle Sap reversal? Every year, the wet season flood pulse is so significant that for five months, the flow direction of the Tonle Sap River reverses, filling the Tonle Sap Lake with 60 times its dry season water. This annual expansion makes the Tonle Sap the world’s largest inland fishery with a fish catch that feeds Cambodia’s population with up to 70% of its protein intake. Image: Brian Eyler
Stung Treng, 2018 to 2020 wet season Stung Treng, 2018 to 2020 wet season: This should raise concerns for policymakers because during an unusually dry wet season, the Mekong flood pulse needs all the water it can accumulate to promote the Tonle Sap reversal and drive productive flood processes in the Mekong Delta.
Chiang Saen, 2018 to 2020 wet season Chiang Saen, 2018 to 2020 wet season: Parts of the Mekong closer to China experience larger impacts
Wet season flow pie graph Wet season flow pie graph: The pattern continues despite the 2021 drought
Tonle Sap imagery High resolution satellite imagery from Planet Labs compares lake and river levels at the Tonle Sap Bottleneck in Cambodia in mid-August for 2018 (left) and 2021 (right). 2018 was a normal year for wet season flow while 2021 was the 9th lowest flow year on record.
Tonie Sap High resolution satellite imagery from Planet Labs compares lake and river levels at the Tonle Sap Bottleneck in Cambodia in mid-August for 2018 (left) and 2021 (right). 2018 was a normal year for wet season flow while 2021 was the 9th lowest flow year on record.
Mekong dolphin The Mekong’s Irrawaddy Dolphin population food source and habitat is threatened by upstream dam impacts and by other human activity along the river’s course. Currently about 90 individuals remain in Cambodia. Image: Tom Marshall]
Mekong habitats Giant trees which serve as the habitat for bird and fish species in northern Cambodia are dying off because of higher dry season river levels Image: Brian Eyler.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.