There is a specific wrongness the first time you feel it, and you never quite lose it. You climb a levee in flood season from the dry side — up out of a town, past the backs of houses, up an earthen slope — and you reach the crown and turn, and the river is there, wide and brown and moving, its surface somewhere around your knees. Then you look back the way you came and the rooftops of the town are below the water you are standing beside. The river is higher than the homes. The only thing disagreeing with gravity is the ridge of dirt under your boots, and you, a person who reviews the drawings for that dirt, are standing on top of it in the dark.
I have felt it on the Yangtze. I felt the mirror of it the first time I read John Barry's account of 1927 on the Mississippi and realised, with a small chill, that the American engineers had felt exactly the same thing, and had exactly the same nightmare.
The clear water and the muddy water 清水与浊水
On the dry side, at the toe of the levee, the flood does its most patient work. It pushes under the wall through the pervious ground and surfaces as a sand boil — 管涌 — a little spring welling up in a farmer's field a hundred metres behind a levee that is not, by any visible sign, failing. The old hands on both rivers read a boil the same way: by the water it throws. If it runs clear, the ground is only leaking, and you can hold it. If it runs muddy, the boil is carrying the levee's own foundation up through itself, hollowing a pipe under the wall, and you have very little time.
And what you do about it is the thing that has never left me, because it is the same act performed by two peoples who never met. You do not plug the boil — plug it and the pressure simply finds the next weak spot, unseen. You build a wall of sandbags in a ring around it and let the ring flood, until the column of water you have raised inside presses back down hard enough to calm the spring. You fight water with a small tower of the same water. In 1927 they ringed boils along the Mississippi by lantern. In 1998 I watched — and helped — hundreds of thousands ring them along the Yangtze by floodlight. The same ring. The same held breath. The same reading of clear against muddy.
Two rivers, two languages, one gesture at two in the morning: a ring of sandbags around a spring of rising water, and everyone watching to see if it runs clear. 两条河，两种语言，凌晨两点却是同一个动作——在一眼上涌的泉水周围垒起沙袋，所有人盯着看它，是清，还是浊。
The promise, renewed every July 每年七月的续约
What I have come to believe, standing on both crowns, is that a levee is not really a wall. It is a promise renewed annually — and the terrible arithmetic of it is that every year you keep the promise, you make the price of breaking it higher. The river drops its silt inside the walls and climbs; the town behind grows, sure of its safety; the day of failure, when it finally comes, is bigger than any flood the town would ever have seen if there had been no levee at all. This is not an argument against levees. My whole plain would drown without them. It is only the reason the people who understand them best sleep worst in July.
Which is why both rivers, in the end, made the same grim peace: the floodway, the diversion area — a place chosen in advance to be drowned so that the cities are not. It is the most honest thing in the whole trade, and the least comfortable: an engineer drawing a line on a map and saying, if it comes to it, the water goes here, and these people bear it for those. I have signed off on that kind of line. I think about it more than I think about any dam.
Read the structural companion to this note: The Wall That Must Not Fail.
The Far Bank · 对岸