Book Review: The Yellow River: A Natural and Unnatural History

Mostern, Ruth. The Yellow River: A Natural and Unnatural History (Yale University Press, 2021) hard cover.

It was once called the Great River, and flowed clear. Over time, human use, climate shifts, and political responses to floods and droughts led to the river becoming the Yellow River, sometimes called China’s Sorrow. How this happened is a story as convoluted as the river’s floodplains, and a fascinating lesson in parts vs the whole, and the limits of human power. People and water, and silt and sand, worked together to destabilize the great river over the course of a thousand years.

Asian environmental history has been relatively under-studied, in part because of problems with language, in part because of the enormous spans of time involved. European environmental history is easier to divide, and the archaeological pieces are gathered into tidier “heaps” of sources, so to speak. Only within the past 20 years or so have many works about the environmental history of China been published. This book builds on several classic works of that history, and expands the time-span of the history of the Yellow River.

Mostern argues that while climate shifts and weather pattern changes played a role in the changes observed in the Yellow River watershed, human activity played a far greater role, especially after roughly the year AD 600 CE. Differences in priorities between imperial governments and local officials, plus the focus on relatively free-market development and agriculture, led to Han Chinese culture expanding into regions not suited for intensive farming. By 1855, the Yellow River had become unusuable and impossible to manage (given the finances and technologies of the time), and what had once been a fertile and prosperous region turned into a salty, gravel and sand-choked series of barrens and wetlands. The Loess Plateau in the bend of the Yellow River transformed with ever-increasing speed from grasslands and mixed forests to a rugged, eroded near-desert that sent millions of tons of sediment to cover the floodplain downstream.

Warfare caused much of the damage to the ecosystem of the upper Yellow River, but stable imperial regimes could be just as bad for the environment. The region is one of conflicts – hot and cold air masses, desert winds and tropical moisture, herders and farmers, imperial centralization and tribal societies. Competing armies stripped the land of forests and grass, and the soldier-farmers of Imperial China denuded the land to build walls and grow food for their own survival. When the nomads chased the Han back to the river and farther south, they too removed forest cover, although long stable periods did allow for regrowth of grass and trees. Sometimes. The development of iron-bladed plows and intensive farming technologies caused further, faster, erosion. Demand for fuel and building wood in peace time as well as war devoured more and more forests, causing more erosion and more flooding downstream.

Some observers saw what was happening and argued that the erosion and loss of ground cover needed to stop at the source. When the capital city remained in the upper Yellow River, the government seemed—sometimes—more interested in considering those ideas. But once the government moved downstream, the focus shifted to coping with the results of the problem, not the sources. Huge floods in 1048 and other years devoured tens of thousands of farmland, displaced millions, and drained the imperial treasury. Only the Grand Canal made it possible to feed and supply Peking/Beijing as the land around it turned sandy and salty from inundation and sediment dumping. In 1885, efforts to keep the Grand Canal open failed, and sea transport became the only to move food to the city. Southern China refused to pay for the problems of the Yellow River.

The book is very well written with excellent illustrations, tables, and a long appendix of methodologies. It helps to have a background in overall Chinese history, but that is not needed. A bit of hydrology helps even more, otherwise the learning curve might be a touch steep in the introduction and first chapter. I found the book an easy read, but one with lots and lots to ponder and mull over. The author is even handed in her approach – people can’t know what they can’t know, and the imperial hydrocrats’ priorities made sense to them. They lacked the tools and the resources to see the entire watershed as a whole. Those who did pull back to see the larger picture lacked the will to sacrifice the imperial capital to floods in order to pour resources into the upstream lands.

The author’s use of some terms struck me as odd, enough so that it pushed me out of the story a few times. I disagree with using the term “Anthropocene,” although in this case there is some logic to it, given the importance of human influences on the life of the river. Other usages were literally correct, but jarring, almost as if the author were not a native speaker. I do not know, and it does not affect the overall readability and quality of the book.

I recommend the book to historians of water, historians of China, people interested in the interactions of government and the physical environment, and conservationists. The idea that “the problems caused by central control can be fixed by central control” rings all too true in the West today. I am reminded of an interview I did with a farmer about flooding on a small river. He shrugged and said, “Rivers flood. That’s what rivers do.” People can try to work around, with, or against floods and droughts, but only by looking at the watershed as a while, rather than reach by reach. This is an excellent addition to the literature in several disciplines.

FTC Disclaimer: I purchased this book for my own use and received no compensation or remuneration for this review.

Removing Dams: The Good, the Bad, and the Messy

“Set the river free” is a cry heard more and more often. It started in New England, when old, unused water diversions and containment structures were removed, often for good reasons, from smaller streams. The goal was to let the stream or river return to as close to original as possible, or to prevent a dam failure and the subsequent floods and damage. If the thing doesn’t serve a purpose, and might even have become a hazard, why not take it out? And again, in some cases, it made good sense. In other cases people took sediment samples from behind the dam and said, “Hold up a second! We don’t want this in the river.” And so the dam stays for now.

However, there’s a difference between a little mill-dam or a small diversion on a stream in what has reverted to forest, and a good-sized power dam, or flood control structure. And you can’t just “rip it out and let the river heal.” Depending on the dam, the river, and what’s down stream, that can lead to disaster, at least for the very aquatic life that people are trying to save/reintroduce.

So, a quick primer on what happens behind the dam when the water flow is greatly slowed or stopped. A large puddle forms, growing into a pond or lake. If the river carries lots of sediment (silt, sand), then the heaviest of that begins to settle out into the bottom of the lake. As the lake fills, the weight on the ground increases, as does pressure on the dam and the sides of the lake. If excitement is going to happen, this is when the first signs appear (Teton Dam is the horrible warning). A little seepage is pretty common, and should be clear. If muddy water starts appearing on the wrong side of the dam, then Things Need to Be Done. If all goes as planned, a lake fills in behind the dam. Hydropower dams use the pressure of that water to turn turbines that generate electricity, so they are always releasing water, or do so intermittently to meet peak demand. Irrigation dams draw-down water levels in the dry season for farmers to use, then allow levels to rise again in the wet. Recreational dams hold water at all times, unless there is a compelling reason to release water (controlled flooding as opposed to uncontrolled “it ate the dam!!!” flooding. See Johnstown for the horrible warning.)

Over time, if the stream carries enough silt, the lake might just turn into a mudflat with a waterfall on one end. That is very, very rare, at least in the modern era. You can see it with ancient structures in Arabia and a few other places. Eventually the dam wears away and you have to look very carefully to see that a water control structure ever existed. Most modern US dams have a life-time of about a century or so before something needs to be done about the silt, it silt deposition is a problem. Something like Hoover Dam or Glen Canyon? Not such a problem.

So, let’s say you want to tear out a mill-dam on a rocky stream that never had much sediment. You do a core check of what’s behind the dam, find nothing toxic and no huge slugs of sediment, and decide to get rid of the structure. Some sludge and gunk will flow downstream, so ideally you allow a slow release and everything keeps going until the sediment reaches the sea (or the next lake downstream) and you have a nicely restored river. You are happy, the land owners are happy, and the fish are happy. Good deal.

However, let’s say you look at a dam downstream of an industrial area. It needs to go. So you take a core of the yards’-deep sediment behind the dam and Whhoooooooah Nellie! Arsenic, dyes, lead enough to rearm the Ottoman Empire, and a few other things have accumulated. Do you really want that going downstream and settling all over the stream? Probably not. The better thing to do is to lower the dam, allowing increased flow-through but still keeping that “stuff” penned up and out of circulation until someone decides they want to pay for the removal and remediation.

So, what if there’s been no, zip, nada industrial development in the area, ever. Your core comes out clean, but full of sticky, silty, sludge. The flow in the stream is not what it used to be, for a number of reasons. If you rip out the dam, it will take a very, very long time for that material to get through the stream. Most likely, the stream will aggrade, getting shallower and wider, leaving the sediment over the current bed. It can also cause the shallower water to be warmer for a while, until the balance starts to return after more sediment is carried downstream. If you have really good trout fishing downstream, this is probably not the scenario you want. Now you start talking about dredging out the stuff, taking it somewhere else, and then opening the dam after the water settles again. $$$,$$$,$$$.00 can be involved.

So, OK, what if we just rip out Hoover, or Grand Coulee? Well, first, where is that electricity going to come from? Nuclear is the only close replacement, because wind and solar are not going to work. Second, that will be an enormous slug of very, very cold water racing down the Colorado River valley and taking a lot of stuff with it. Like downstream dams and diversions. Like any water-supply intakes. The fish will be in for a surprise with that temp, although the modern Colorado is far colder than it once was. And eventually the sediment will start to move, slowly, and you will get back to closer to the old Colorado, brown, wild or sluggish, and meandering. If you get/got drinking water from downstream of Hoover, well, buy a lot of filters and plan on changing them regularly. And get ready for floods, as the bed, once scoured, starts to rebuild with that sediment.

Did I mention floods? Annual or semi-annual floods will return. Any valuable infrastructure will have to be relocated, or turned sacrificial. You’re going to lose habitat for some creatures and gain it for others. Given the sediment that’s built up behind the dams of the 1910s-1950s, anything that needs a rocky bed with well-oxygenated water might have difficulty for a while, depending on how quickly the sediment is redistributed and filtered out as the balance in the stream resumes.

Note: All these are controlled removals. Nightmare fuel is if one of the big dams on the Colorado goes all at once, because then the downstream dams will likely go as well. The Colorado will reach the sea again, perhaps by a different route. Or it might refill the Salton Sea and a few other areas, then head out via LA or even double back and then go out to Baja. The loss of life would be tremendous, the loss of infrastructure eve more so.

So dam removal can be done, and done right. However, “rewilding the Columbia and the Colorado!” is probably an undertaking best left to fiction writers for now. Until nuclear reactors become far more common, and we know a lot more about what’s behind those structures and how to release the contents slowly, we could do far more harm to the environment than we ever did by building the dam.

Oxbows, Cut-offs, Meanders, and Snags

So, how do you lose an entire large-ish boat, that sinks intact, for a hundred years or so? Well, if the water is muddy enough, and the place is remote enough, you just don’t see it. Especially if that river likes to wander all over the place at whim, or at least at what to most people looks like whim. Welcome to the Mississippi and lower Missouri Rivers, two streams that snake more than the reptile house at the National Zoo.

What makes a river take a certain shape depends on a handful of elements. First is the geology under the river – is there room to move? Yes, over time the stream will cut a channel, but if the slope of the land is low, and the surface is flat, and the surface is also loose (lots of dirt, or sand, or small gravel), then the stream will tend to take on a series of S-curves, especially if the stream gets a decent amount of in-flow. Something like the Republican River in western and central Nebraska doesn’t get much inflow, so it tends to be straighter than similar streams east of the 40″ rainfall line. Glacial outflow streams carry lots and lots of gravel and smaller sediments, so they wander as well, even though they may be over bedrock (if you dig far enough). Even the Colorado River (western, not Texas) cut bends through very hard material, because it carried a lot of sediment that acted as an abrasive. It had room to move, until it cut that rather deep canyon.

Because of all the sediment, when the flow of water slows down, the sediment eventually settles out of the water and onto the bed. This builds up the bed, and if it happens long enough, the river moves to the path of least resistance. If the flow of the river rises and falls, as used to happen (still happens to an extent) with the Mississippi and Missouri, then the river also dumps “excess” energy by moving into an S-pattern. The bends shift location over time. This is why, after big floods, the riverboats on the Mississippi and Missouri had to be very wary of snags and changes in the river’s location. The flow of water that had carried small islands, entire trees, huge glops of dirt and sand, wasn’t there any more, so the river had less energy, and dumped the load. Now the boats got to find those lumps, and trees, and avoid them. You didn’t want to be the first boat up the stream after a big flood. Sort of like being the first rafting group down the Grand Canyon after a big storm or high water. Surprise!

DeSoto Bend in Nebraska is no longer a bend. It is fields around an ancient island. The river relocated. Depending on physics, weather, and inner perversity, a river may build up natural levees and terraces, like the levee that keeps the Garden District in New Orleans above many floods. That neighborhood was built on a natural rise, for good reason.

If the river moves far enough, the bend might get pinched off and an oxbow lake forms. Source of image used under Creative Commons Fair Use.

Over time, the oxbow may dry up. Unless you look at the soil from above, or study soil-type maps, you’d never know that there was once a river there. Fly over the Mississippi on a clear day in fall or winter, when the crops have been harvested, and you can see the ghosts in the fields, the long-gone bends and oxbows.

Fair use from:
Fair Use from:

Rivers wander. People try to keep rivers from wandering. This means dredging a channel, or building levees to constrain the wandering. Dredging can speed up the river, meaning it erodes more, or starts snaking more downstream of the deeper channel, or eroding upstream, or “Yes.” Levees can work, until you end the levee and the river starts dumping sediment and snaking more. All these mean that floods will spread farther below the dredged area or levee, unless the river decides to dump lots of sediment on the dredged area. This is called agrading, “building up the bed” as opposed to degrading “down-cutting the bed.” If a river has to be channeled, ideally, there will be areas nearby that are left as flood-zones, so that the river can release the energy and sediment that it needs to lose. And so that floods will happen in anticipated areas, and downstream floods won’t be as bad.

The problem is, when the Mississippi and Missouri and other streams were channelized, we didn’t know what we didn’t know. As much as I’d love to see those rivers revert to their old ways, I don’t think people are willing to sacrifice the infrastructure and good farm ground that has been developed along the streams. It’s like, oh, the plan to reintroduce certain fish to the Rhine and revert it to what it was before 1800s. You’d have to dismantle so many buildings, factories, power-plants and other things, and then tolerate floods and erosion and deposition and the mouth of the Rhine moving around and no more (or far less) shipping, and . . .

Eventually, the river is going to win. Rivers do that. The Old River Control Structure on the Mississippi is going to “fail,” meaning the Mississippi will route around it and shift back to an earlier channel and outlet to the Gulf of Mexico. Eventually, levees on other rivers are going to be overtopped so badly that they can’t be patched for months, or more likely years. Physics is on the side of the river.

As one older Midwestern farmer put it, “The river always flooded. That’s what rivers do.”

Why’s the Stream Doing That?

Part of my graduate work included learning to think like a hydroengineer. Which also means learning how streams work and don’t work, and how to “unfix” (as one prof put it) well-intended potential disasters.

To do this, first you have to understand why streams do what they do, or what they don’t do. The Platte River in Nebraska, or the lower Elbe in Germany, will never be trout streams. No amount of human effort will turn them into cold, clear, briskly flowing streams with stony beds. Likewise, the upper Cache le Poudre or the Rhone headwaters will not become slow moving, sandy, meandering lazy rivers. Continue reading