The Flood We Should Not Forget
In the early hours of August 8, 1975, the Banqiao Dam in Henan Province failed after days of extraordinary rainfall brought by Typhoon Nina.
It was not the only dam to fail.
Research published by the American Society of Civil Engineers says two large dams, two medium-sized dams and 58 small dams failed during the storm. Five-day rainfall reached 1,631 millimeters in the hardest-hit area. The resulting floods inundated about 12,000 square kilometers.
The number of people who died remains disputed.
Chinese water-resources figures cited in later reporting put direct deaths from the 1975 reservoir failures at more than 26,000. A later meteorological study likewise estimated about 26,000 direct deaths, followed by roughly 100,000 additional deaths from famine and disease. Other historical estimates have been substantially higher, reaching as high as 230,000.
There is no reason to pretend that one of these numbers is universally accepted. What is certain is that this was one of the deadliest dam disasters in history.
The rain itself was extraordinary. But rain alone does not explain everything that followed.
Banqiao had been designed for an extreme flood, yet the storm exceeded its design assumptions. Its discharge capacity was insufficient for what arrived. Communication lines failed as the disaster developed, leaving dam operators increasingly isolated. Downstream warnings did not reach enough people in time.
That distinction matters.
Human beings cannot stop a typhoon.
They can decide how much uncertainty an engineering design allows for.
They can maintain spillways.
They can build redundant communication systems.
They can prepare inundation maps and evacuation plans.
They can make sure that people downstream know what to do before a dam begins to fail.
The lesson of Banqiao is therefore not that engineering is useless.
It is almost the opposite.
Engineering works only when people remember that every design has limits.
After the 1975 disaster, China carried out major reservoir safety inspections and reinforcement programs. Banqiao itself was eventually rebuilt and passed national acceptance in 1993. Chinese engineering research continued to study the disaster and revise flood-control and reservoir-management practices.
That history should be included because it would be misleading to say nothing changed after 1975.
But remembering the improvements does not require forgetting what failed.
More than fifty years later, extreme rainfall remains a serious problem. Forecasting is better, communications are faster and reservoir monitoring is far more advanced. In 2026, Chinese authorities continue to operate formal dam-safety registration and inspection systems.
None of that makes risk disappear.
The most important lesson from Banqiao is simpler than a political slogan:
When the possible consequence is mass death, safety cannot depend on the belief that the worst probably will not happen.
The public should know the risks.
Warnings should have more than one way to reach people.
Evacuation plans should work before the emergency.
Critical infrastructure should be designed not only for what is expected, but with respect for what may exceed expectations.
And when disaster happens, the record should remain public.
The death toll should be studied honestly, including disagreements over the numbers.
The engineering failures should be studied.
The decisions should be studied.
The successful reforms afterward should also be studied.
That is how a disaster becomes a lesson instead of a legend.
The Banqiao flood happened in 1975.
The people who died cannot be protected now.
The only thing the living can still do for them is make sure the next warning is heard before the water arrives.