On 19 April 2024, the Yellow River, fed by an unusually heavy snowmelt off the Helan Mountains, breached a sandy embankment north of the town of Wuyuan in Inner Mongolia.

Within eleven days, it had carved a new channel 2.4 kilometers east of its old one, through what had been a sugar beet field and a stretch of disused irrigation canal.

The new channel was sixty to ninety meters wide, two to four meters deep, and ran for about fourteen kilometers before rejoining the old course south of the Bayan Nur city limits.

It was not the largest avulsion in the river's recorded history. The Yellow River, which the Chinese call the Mother River and also the Sorrow of China, has changed its course at least twenty-six times in the last three thousand years.

But it was the first such shift to be measured in near-real time by a single surveyor on foot.

The surveyor was Bao Wenli, then fifty-two, a senior cartographer for the Bayan Nur Bureau of Land and Resources.

She had been in the area for an unrelated boundary dispute when the breach occurred. She heard about it from a farmer named Lao Zhang at a roadside noodle stand on the morning of 20 April.

She drove out to the embankment that afternoon. By the next morning she had ordered her assistant in the city to send her two additional GPS units and a drone.

Over the next two weeks, she mapped the new channel as it formed, walking the muddy banks with a handheld total station, sleeping in her car or at a guesthouse in a village called Xinhu.

“Over the next two weeks, she mapped the new channel as it formed, walking the muddy banks with a handheld total station, sleeping in her car or at a guesthouse in a village called Xinhu.”

She filed daily updates to the bureau. The bureau, initially uncertain how to classify her work, eventually authorized it as an emergency hydrographic survey.

Her field notes from the period have since been published, in part, by the Inner Mongolia Hydraulic Research Institute.

They are detailed. She records, for instance, that on 24 April the channel widened by an average of 2.7 meters between 6 a.m. and 6 p.m., and that the flow carried away two willow trees and a section of concrete irrigation pipe.

She records that on 26 April, a herd of about thirty sheep stood on the new west bank watching the water for nearly an hour before their herder, a man named Erdene, drove them away.

She records that on 28 April she ate cold mantou and an apple for breakfast and her boots leaked.

The notes are not poetic. They are the notes of a person doing her job carefully, in conditions that did not favor poetry.

The new channel stabilized by early May. By August 2024, the regional water authority had decided not to try to force the river back into its old bed. The new course was accepted as the river's course.

This decision displaced 1,140 people across six villages and required the rerouting of a 31-kilometer section of provincial highway G110.

It also left Lao Zhang's sugar beet field, all twelve mu of it, on the wrong side of the river.

Bao filed her final survey report on 17 June 2024. It is 312 pages long. It includes maps drawn at five-day intervals showing the channel's progression.

The maps, viewed in sequence, look very much like a slow exhalation.

Bao retired from the bureau in October 2025, on a normal pension. She continues to consult occasionally for the hydraulic institute.

She has been asked, in interviews, what it was like to watch a river move. She has tended to say only that it was loud, and that the mud was bad for the equipment.

Once, to a documentary crew from Hohhot, she said that she had not felt as though she were watching the river do something unusual. She said she had felt as though she were watching the river remember something.

She did not elaborate. The crew did not press her. The remark made it into the final cut and then, on review, was cut for length.