Who gains from saving nature
Ledger 1 : 33 — For each dollar financing nature-based solutions in 2023, about $33 financed activities harmful to nature. (UNEP)
Cases / Wind, hydro and other renewables · Cambodia · 1962–2026

Mekong upstream dams and the decline of Tonle Sap flow reversal

Nature's voice · report
My Tonle Sap reverse flow fell by 56.5% between 1962–72 and 2010–18.
Did nature suffer? Yes. Harm is on record. Disputed
Recorded
What the record lacks: 3 none on record · 3 insufficient · 3 disputed · 1 unclear See the gaps
Accounts of the harm differ
Did upstream Chinese dams cause the low Mekong flows of 2019?
  • Researchers Stimson Center (US-funded study): Concluded that Chinese dams held back water in 2019. Stimson Center (Eyes on Earth)
  • Government and regulators China: Rejected the conclusion.
  • International bodies Mekong River Commission: Said the study used calculated flow and the record low came from rainfall; accepted that the dams raised dry-season flows and lowered wet-season flows.
What explains the decline in Tonle Sap reverse flow?
  • Researchers Chua et al. (2022): Dams explain about 21% of the 56.5% decline; irrigation and riverbed incision are the other important drivers; no significant rainfall trend over 1960–2019.
  • Researchers Earth Observatory of Singapore and NTU: In the Vietnamese delta, sand mining accounts for 25–30% of riverbed erosion and 16–30% of the rise in salinity. Earth Observatory of Singapore
Is sand extraction behind the erosion of Mekong riverbanks in Cambodia?
  • Government and regulators Ung Dipola, Cambodia's Director General of Mines: Said the erosion does not come from sand and that dredging has taken about 4% of the river. Mongabay
  • Press and investigations Mongabay: Reported 49 licences on the Mekong and Bassac rivers and extraction in 2020 at 16 times natural replenishment. Mongabay
  • Researchers Singapore Earth Observatory and NTU compilation: Reported about 20 million tonnes a year permitted against more than 50 million tonnes a year extracted on the main stem of the Mekong, about 60% of it illegal. Earth Observatory of Singapore

Each monsoon the Mekong pushes the Tonle Sap River backwards into Cambodia's Tonle Sap lake. A 2022 peer-reviewed study found this reverse flow fell by 56.5% between 1962–72 and 2010–18, and attributed about 21% of the fall to dams.

56.5%Decline in Tonle Sap reverse flow, 1962–72 to 2010–18
Chua et al. 2022 (Hydrology and Earth System Sciences). The study attributed about 21% of the decline to dams.

Status As of July 2026: studies attribute the decline to dams, irrigation abstraction and sand-driven riverbed incision; the attribution to Chinese dams was disputed by China in 2019–2020 and no ruling or settlement was found.

What?
The reverse flow from the Mekong into Cambodia's Tonle Sap fell by 56.5% between 1962–72 and 2010–18, and the flood season shortened by 26–40 days.
Where?
Tonle Sap lake and river, Cambodia; upper Mekong (Lancang), Yunnan, China; Mekong Delta, Vietnam.
When?
Measured between 1962–72 and 2010–18; dam dispute from the 2019 low flows.
Who?
Gained
  • Operators of upper-basin (Lancang) hydropower in Yunnan, China — Operate 11 dams with more than 47 billion m³ of storage.
Lost
  • Tonle Sap fishers and lower-basin communities — Fish provide more than 60–70% of animal protein in Cambodia; annual catch has been put at 150,000–550,000 tonnes.
Harmed
  • Residents of the Vietnamese Mekong Delta — 1,808 houses lost to the river or sea in 2018–2020; more than 600 hectares of coastal land lost each year. Eco-Business (Mekong Eye)
Why?
Dams, irrigation abstraction and riverbed incision from sand mining are the drivers named in the sources; rainfall showed no significant trend over 1960–2019.
How?
Dams store and release water; irrigation takes water from the river; sand extraction lowers the riverbed.
6 of 6 questions answered from the record

What happened

The record, in order.

The Tonle Sap in Cambodia is a lake whose flow reverses each year. During the monsoon the Mekong pushes back the Tonle Sap River and the lake grows to more than three times its dry-season size, above 15,000 km². The Mekong River Commission has put the lake's annual fish catch at 150,000–550,000 tonnes, and fish provide more than 60–70% of animal protein in Cambodia. For the lower Mekong countries, the Commission's 2007 paper put the share of animal protein from fish at 47–80%.

A study by Chua and colleagues (2022, Hydrology and Earth System Sciences) found that reverse flow into the Tonle Sap fell by 56.5% between 1962–72 and 2010–18, and that the flood season shortened by 26–40 days. It attributed about 21% of the decline to dams. It named irrigation abstraction and riverbed incision as other important drivers, and found no significant trend in rainfall over 1960–2019.

In the upper basin, the Lancang (the upper Mekong in China) carries 11 dams with a combined storage of more than 47 billion m³. After the low flows of 2019, a US-funded study published through the Stimson Center in April 2020 said Chinese dams had held back water. China rejected this. The Mekong River Commission said the study used calculated rather than measured flow and that the record low came from low rainfall, but it accepted that the dams raised dry-season flows and lowered wet-season flows.

Sand extraction is the other driver the sources describe. In the Vietnamese Mekong Delta, a study published in Science Advances in July 2026 estimated average extraction of 48 million m³ a year in 2017–2021, 6–15 times natural replenishment, and attributed 25–30% of riverbed erosion and 16–30% of the rise in salinity to sand mining. In Cambodia, satellite-based work reported extraction of the Mekong rising from 24 to 59 million tonnes a year between 2016 and 2020.

A Reuters and EOMAP analysis reported in 2022 that four dams retain 81% of the Mekong's sediment. A Mekong River Commission scenario for 2040 projects a 97% fall in sediment reaching the delta and a loss of between 40% and 80% of fish biomass.

  1. 1962Start of the 1962–72 baseline period used to measure Tonle Sap reverse flow
  2. 2007Mekong River Commission paper: fish provide 47–80% of animal protein in the lower Mekong countries
  3. 2016Cambodian Mekong sand extraction measured at 24 million tonnes a year (later satellite work)
  4. 2017Cambodia bans sand exports; the ban is lifted in 2020
  5. 2019Low flows in the Mekong; dispute over the role of upstream dams begins
  6. 2020Stimson Center publishes the study saying Chinese dams held back water (April); Cambodian extraction measured at 59 million tonnes a year
  7. 2022Chua et al. report a 56.5% fall in Tonle Sap reverse flow; Reuters and EOMAP report that four dams retain 81% of sediment
  8. 2026Science Advances study quantifies sand mining in the Vietnamese delta (July)

What was said

Claims, in the words of those who made them.

Where accounts differ

Each party's account, side by side, with its source. The record does not settle these points; it shows them.

Did upstream Chinese dams cause the low Mekong flows of 2019?
Researchers
Stimson Center (US-funded study)

Concluded that Chinese dams held back water in 2019.

Stimson Center (Eyes on Earth)
Government and regulators
China

Rejected the conclusion.

International bodies
Mekong River Commission

Said the study used calculated flow and the record low came from rainfall; accepted that the dams raised dry-season flows and lowered wet-season flows.

What explains the decline in Tonle Sap reverse flow?
Researchers
Chua et al. (2022)

Dams explain about 21% of the 56.5% decline; irrigation and riverbed incision are the other important drivers; no significant rainfall trend over 1960–2019.

Researchers
Earth Observatory of Singapore and NTU

In the Vietnamese delta, sand mining accounts for 25–30% of riverbed erosion and 16–30% of the rise in salinity.

Earth Observatory of Singapore
Is sand extraction behind the erosion of Mekong riverbanks in Cambodia?
Government and regulators
Ung Dipola, Cambodia's Director General of Mines

Said the erosion does not come from sand and that dredging has taken about 4% of the river.

Mongabay
Press and investigations
Mongabay

Reported 49 licences on the Mekong and Bassac rivers and extraction in 2020 at 16 times natural replenishment.

Mongabay
Researchers
Singapore Earth Observatory and NTU compilation

Reported about 20 million tonnes a year permitted against more than 50 million tonnes a year extracted on the main stem of the Mekong, about 60% of it illegal.

Earth Observatory of Singapore

How it looked

What was shown, beside what was recorded.

Shown

    Recorded

    • Tonle Sap reverse flow fell by 56.5% between 1962–72 and 2010–18, and the flood season shortened by 26–40 days.
    • The Chua et al. study attributed about 21% of the decline to dams, named irrigation and riverbed incision as other drivers, and found no significant rainfall trend over 1960–2019.
    • Four dams retain 81% of the Mekong's sediment (Reuters and EOMAP, 2022).
    • Sand extraction in the Vietnamese delta averaged 48 million m³ a year in 2017–2021, 6–15 times natural replenishment.
    • In the Vietnamese delta, 1,808 houses were lost to the river or sea in 2018–2020, and more than 600 hectares of coastal land are lost each year.
    • Cambodia's Mekong and Bassac rivers carried 49 sand licences; extraction in 2020 reached 16 times natural replenishment.

    Who gained

    Money that reached nature, beside money that stayed elsewhere.

    Reached nature or the ground

    Not published
    No figure published.

    Gained elsewhere

    Not totalled
    Currency: USD · Window: . All figures cover the same period. * Estimate built from cited figures.

    Responses on record

    • China (government) (2020): Rejected the 2020 study's conclusion that its upstream dams held back water in the 2019 low-flow period.
    • Mekong River Commission (2020): Said the 2020 study used calculated flow and that the record low came from low rainfall, while accepting that the dams raised dry-season flows and lowered wet-season flows.
    • Cambodia, Director General of Mines (Ung Dipola) (2022): Said riverbank erosion does not come from sand and that sand dredging has taken only about 4% of the river. · Mongabay

    In court and with insurers

    Every proceeding and every insurer decision on record.

    Lawsuits and rulings

    No court case on record. No court has ruled on this case.

    Insurers

    No insurer decision on record.

    What nature saw

    Outcomes measured on the ground, and nature's note.

    Nature's report opens this page. Below: further outcomes measured on the ground.

    Tonle Sap reverse flow
    -56.5%
    1962–72 to 2010–18 (Chua et al. 2022); dams about 21% of the decline
    Flood season
    26–40 days shorter
    Same study
    Sediment retained at four dams
    81%
    Reuters and EOMAP, 2022
    Sand extraction vs replenishment, Vietnamese delta
    6–15 times
    2017–2021 average of 48 million m³ a year (Science Advances, July 2026)

    What we don't know

    Undisclosed, unproven, untried: recorded as data.

    Nature

    • Disputed Did upstream Chinese dams cause the low Mekong flows of 2019?
    • Disputed What explains the decline in Tonle Sap reverse flow?
    • Disputed Is sand extraction behind the erosion of Mekong riverbanks in Cambodia?

    Court

    Figures

    • Insufficient Reverse flow decline bed lowering pct: Riverbed lowering 1998–2018, reported from a paper that could not be checked.

    Record

    • None on record Share of the Tonle Sap decline caused by each driver: dams account for about 21% in the Chua et al. study; the remainder is attributed to irrigation and riverbed incision without a published breakdown in the sources reviewed.
    • None on record Who holds the sand licences and extracts sand on the Tonle Sap and Cambodian Mekong stretches: the owners are not named in the record for the Tonle Sap decline.
    • Insufficient A 2025 Nature Sustainability paper (Southampton) is reported to say that riverbed lowering in 1998–2018 cut reverse flow into the Tonle Sap by 40–50% and that this would fall further by 2038. The paper could not be checked and the figures are reported but not independently verified.
    • Insufficient A figure of about 6.18 million tonnes of sand entering the delta each year against about 50 million tonnes extracted is reported from a summary of a 2020 Nature Sustainability paper; it is reported but not independently verified.
    • Unclear Whether the Mekong River Commission's 2040 sediment and fish-biomass scenario figures (97% and 40–80%) have been independently checked is not stated in the sources reviewed.

    Did the harm stop?

    The sources reviewed record no measure that has stopped the decline; the cited studies describe the driver trends as continuing.

    Who paid?

    The sources reviewed point to lower-basin fishing communities and delta residents, including 1,808 houses lost in the Vietnamese delta in 2018–2020.

    Sources

    Voices on record: — Company or project— Government and regulators— Courts✓ Press and investigations✓ Communities and civil society✓ Researchers✓ International bodies

    Communities and civil society

    1. New evidence: how China turned off the Mekong tap
      Stimson Center (Eyes on Earth) · 2020-04

    International bodies

    Connected cases

    Related matters, and cases linked to this one through the same organisations, with the role each played in both.

    Science and environment

    Biomass, root-to-shoot ratio, carbon fraction
    Biomass is the dry weight of living plants per hectare. The root-to-shoot ratio adds the weight of roots. The carbon fraction is the share of that weight that is carbon, about 47% in tropical forests.
    Hectare (ha)
    A unit of area equal to 10,000 square metres, about 2.5 acres. 100 hectares make one square kilometre.

    Carbon and conservation finance

    Baseline
    The projected scenario of what would have happened without a project, such as an expected rate of deforestation. Credits are issued for the difference between the baseline and what actually happened, so an inflated baseline produces more credits.