In the early morning of January 23, 1556, an earthquake struck the Wei River Valley in Shaanxi, in Ming-dynasty China. No seismograph existed to measure it, but catalogues compiled by the U.S. Geological Survey list it at around magnitude 8 — and by the traditional count, it is the deadliest earthquake in recorded history, with an estimated toll of around 830,000 people.
It is also called the Jiajing earthquake, for the emperor on the throne at the time. Nearly five centuries later, its lesson is not really about magnitude. It is about what people were living inside when the ground moved.
The epicentre was near Huaxian — now Huazhou District, in Weinan — close to Weinan and Huayin, in the Wei River Valley at the foot of the mountains. Historical accounts place the maximum shaking at intensity XI, extreme damage, and describe destruction across dozens of counties in Shaanxi and neighbouring Shanxi. In the worst-hit places, contemporary records suggest around 60 percent of the population was lost.
The valley sits in the Loess Plateau, a vast blanket of fine, wind-blown silt laid down over thousands of years. Loess is soft enough to carve and, in dry weather, firm enough to stand in cliffs. For generations, millions of people in the region lived in yaodongs — homes dug directly into those loess cliffs. A yaodong is warm in winter and cool in summer. It is also, in strong shaking, a roof made of earth.
When the earthquake struck before dawn, people were asleep inside those cave homes. The shaking collapsed yaodongs, triggered landslides in the loess, and buried homes under the very cliffs they were carved from. Heavy stone buildings in towns failed too. Accounts describe deep ground fissures opening, ground rising in some places and sinking in others, and hills and valleys changing shape — a landscape rearranged in minutes.
That 830,000 figure needs honest handling. It comes from Ming-era records and population registers, not a modern count. Modern publications from the China Earthquake Administration put direct deaths from the shaking itself at roughly 100,000 or more, with the rest of the roughly 830,000 population loss coming from what followed — famine, disease and people fleeing and never returning in a bitter winter. Even on that narrower reading, it was catastrophic. On the traditional reading used by the USGS, it remains unmatched.
Much of what we know about living through that morning comes from Qin Keda, a scholar who survived the earthquake and left a detailed account. He described what the shaking felt like, what collapsed, and — remarkably modern — what to do in the moments after: stay alert for aftershocks, and do not rush blindly for a doorway in a failing building. It is one of the earliest surviving attempts to turn an earthquake from an omen into a set of practical lessons.
The faults matter too. The earthquake is associated with the fault system that bounds the Wei River basin, a block of crust that has been sinking between the Loess Plateau and the Qinling Mountains. The same kind of basin-edge faulting that built the valley's farmland also made it dangerous: soft sediment can amplify shaking, and loess that stands firm for centuries can flow and slide when it is shaken hard or wet.
Shaanxi 1556 is not the strongest earthquake ever recorded — the 1960 Valdivia earthquake in Chile, at magnitude 9.5, released far more energy and killed far fewer people. Deadliness is not magnitude alone. It is magnitude plus depth, plus soil, plus housing, plus the hour of the day.
Yaodongs are still lived in on the Loess Plateau today, though modern versions are often reinforced with brick or stone arches — a direct, quiet acknowledgement of 1556. The principle generalises far beyond Shaanxi: the safest earthquake technology is rarely exotic. It is a home that does not turn into its own burial ground when the ground it stands on moves.
Sources: U.S. Geological Survey / National Earthquake Information Center deadliest-earthquakes list (January 23, 1556, magnitude around 8, estimated 830,000 deaths); Encyclopaedia Britannica, Shaanxi province earthquake of 1556; China Earthquake Administration publications on direct versus indirect losses, as summarised in modern historical catalogues. Magnitude and toll are estimates from historical records — no instruments existed in 1556 — and are hedged accordingly.