1. What a riverbank stone wall is protecting against
When a flood makes a river run fast, the water scrapes at the soil of its banks. If nothing protects the bank, it can wear away bit by bit, and in the end the whole embankment can fail. A Kyoto Prefecture river glossary says that a revetment protects embankments and banks from being eroded by flowing water [2]. Stones are one common material. Concrete blocks are another.
The wall looks simple, but the way it is built varies a lot. The biggest difference is whether concrete holds the stones together.
2. Glued with concrete or stacked without it: the two main types of stone wall
The Kyoto glossary splits stone walls into two kinds. In a neri-ishizumi, which we can call a mortared stone wall, the spaces between the stones are filled with concrete. In a kara-ishizumi, a dry-stone wall, the stones are fitted together skillfully and stacked without concrete. The glossary adds that the stone walls of old Japanese castles are dry-stone walls too [2].
Dry-stone walls have been used for a long time as a traditional way to protect Japanese riverbanks. A 2018 paper by Kyushu University researchers says people still want them for their look and for the wildlife they can shelter. It also says that little is known about their strength and how to design them, so engineers do not often choose them [1].
That brings us back to the question in the title. If water can flow through the gaps, why doesn’t it carry the soil behind the wall out with it?
3. How water can pull sand out through the gaps of a dry-stone wall
The Kyushu University team studied one way a dry-stone wall can fail. Water moving through the wall can drag sand from behind it out through the gaps. The Japanese word is suidashi, and we can call it sand washout. The team recreated the situation in a water-flow experiment. Most of the sand washed out near the bottom of the wall, and the wall tipped over when it failed [1].
So the gaps themselves are not the whole problem. What matters is what sits behind them.
4. Packing small stones behind the wall stops the sand from escaping
The same experiments showed that filling the space behind the wall with a packing material (uragome-zai in Japanese) can prevent the washout. Smaller grains in the packing and a wider packing layer both made it work better [1]. The abstract does not give the exact sizes that were tested.
American erosion-control agencies use a similar idea for loose stone linings. The North Carolina state standard for riprap (a lining of large loose stones) says the stones need a filter blanket under them. This is a layer between the stones and the soil that keeps soil from moving into or through the stones. The filter can be graded gravel or sand, or a fabric made for the job, and a gravel filter may have several layers growing coarser step by step, from sand up to the large stone [4]. The two are not the same thing: one is a lab study of a Japanese wall and the other is a practice standard. But both rely on small grains that hold the soil while water passes.
A simple way to picture it is that the large stones are like armor and the small stones behind them act like a filter that holds the soil back. That picture is mine, not a claim from the sources.
One limit to keep in mind: the Japanese result comes from a lab experiment under set conditions. It shows a trend, but it does not tell an engineer what size and width to use on a real river, where the flow and the soil differ from place to place.
5. How a modern mortared stone wall is built to hold its stones in place
A construction page from the Jinzu River sabo (erosion-control) office of Japan’s Hokuriku Regional Development Bureau shows how a mortared stone wall is built today in the mountains of the Okuhida area. First, crews pour concrete into a form to make a foundation that receives the stones. They wash mud and ice off each large stone so it does not slip while a crane lifts it. They place the stones with the flat face up, which they say also helps the look of the landscape. Then they drill holes in the stones and fix them with steel reinforcing bars so they will not tip over. Finally they pour concrete between the stones as a kind of glue [3].
This is one project’s method, so other sites may differ. The pages I read give no numbers on which type lasts longer, so I cannot say whether the dry-stone wall or the mortared wall is better.
6. How the steepness of a stone-covered slope changes its name
The Kyoto glossary also names stone work by steepness. A slope steeper than a set cut-off is called ishizumi (stone stacking), and a gentler slope is called ishibari (stone facing, or stone paving) [2]. The glossary sets the cut-off at a gradient of “1 wari”, a Japanese slope unit that I have not converted here.
So there are two things you can often judge from a distance: whether the gaps are filled with concrete, and how steep the slope is.
7. How to look at a riverbank wall in your own town
If there is a stone wall near a river, look at it from a bridge or from the top of the embankment, with an adult, and never go down to the water. Stay away when it has been raining or the river is high. Ask yourself three questions. Are the gaps between the stones filled with concrete or open? Are the stones natural shapes or cut to match? Is the wall steep or gently sloped?
You cannot see what is hidden behind the stones, but now you know it is probably there. To go further, look up the pages on bank protection from your own river or highway agency, and see how they describe stone linings and filters.