A thousand years of Yangtze River flood records — from 950 CE to the present — reveal a transition point. For most of the millennium, flood regimes were driven primarily by climate: monsoon variability, volcanic cooling, solar cycles. Floods happened when the weather delivered too much water.
Researchers used a cascading attribution methodology to decouple climate drivers from anthropogenic factors — deforestation, polder expansion, population growth, land-use conversion. The finding: human land-use changes now outweigh climatic drivers in shaping the Yangtze's flood regime. The transition happened not through dams and levees — the obvious engineering interventions — but through deforestation and polder expansion that fundamentally rewired how rainfall becomes runoff.
Deforestation removes the sponge. Forested hillsides absorb rainfall, release it slowly through root systems and soil infiltration, and delay its arrival at the river. Cleared hillsides produce rapid runoff that arrives at the channel simultaneously from multiple tributaries. The same rainfall produces a higher, sharper flood peak when the watershed is deforested.
Polder expansion — converting floodplains into walled agricultural land — removes the buffer. Natural floodplains absorb excess water during high flows, spreading it across shallow storage areas and releasing it gradually as the flood recedes. When polders wall off the floodplain, the water has nowhere to go but downstream, faster and deeper.
The Yangtze's floods are no longer primarily acts of nature. They are acts of landscape management — consequences of decisions made across centuries by millions of people, none of whom intended to change the river's behavior. The flood regime was rewritten not by engineering the river but by engineering everything around it.
Note (2026-11-05): I could not locate the paper this essay cites (an npj Natural Hazards 2026 study of a thousand years of Yangtze flood records), and I cannot verify its attribution result. The nearest work I found, a PNAS study of Yangtze flood hazards since the Little Ice Age (engineering and land use intensifying climate-driven floods), reports that embankments amplified the climate-driven 100-year flood by about 18 percent and lake reclamation by about 8 percent, which sits uneasily with my statement that the transition came 'not through dams and levees'. Treat the essay's central claim, that human land use now outweighs climate and that engineering was not the route, as unverified. The explanations of deforestation and polders are standard hydrology, written by me, not claims from a source I checked.