Sunday, December 27, 2009

Dirt: The Erosion of Civilizations

I'm still on Christmas vacation, but have been reading dirt: The Erosion of Civilizations by David Montgomery as my idea of cheery relaxation reading. I wanted to offer a few comments as a mini-review (meaning I want to give my impressions without being held to any particular standard - I am still on vacation after all!)

Firstly, to state my prejudices on this issue - soil erosion is an issue that I've found difficult to form a clear opinion on (you can read as far as I'd gotten in the "Soil Loss" section of this Oil Drum piece. It's clearly important in the long term, it doesn't seem to be an acute crisis globally (though it's certainly serious in some places), and it's not easy to sort out the science and decide exactly how acute a problem it is. Conventional agriculture clearly discounts the problem and is focussed on short term profit over long-term health of the soil. On the other hand, there is clearly an over-alarmist wing of the environmental movement (epitomized by people like Lester Brown) that is prone to overstating the severity and immediacy of the issue. In between these extremes, I haven't really been able to figure out the situation to my satisfaction.

I'd like to understand the issue better, particularly since in one likely future of continued high oil prices, it seems probable that global civilization will try to greatly increase its take from the biosphere by making a lot of biofuels, which will presumably require using more fragile lands than we have used hitherto. In that scenario, soil erosion will become more important as a potential limitation on the growth and sustainability of civilization. Hence my picking up and reading dirt.

In terms of coverage, the book starts with a very short introduction to the natural processes of soil formation and erosion, and then almost immediately turns to an account of the role of soil problem in past civilizations, beginning with the Sumerians and the salinization of what is now Southern Iraq, then moves onto the role of the loss of soil in the demise of Greek and Roman civilization, the Mayans, neolithic Europe and various Pacific islands. There is also an extensive discussion of the history of soil problems in the early United States, and the modern era, beginning with the dust bowl, as well as problems in the Soviet Union and the Sahel of Africa.

I have pretty mixed feelings about the book. On the plus side, it's accessible - it's a clear and entertaining read that I had no difficulty finishing. It covers a lot of ground, and the author obviously knows a great deal about his field.

On the downside, if you've read books like Jared Diamond's Collapse, and Clive Ponting's New Green History of the World, a fair amount of this stuff is familiar ground. That's fair enough I guess - not all readers of dirt will have read those books and there's something to be said for being comprehensive. Still, it did mean that for me, the most interesting sections were the ones covering things I didn't know much or anything about. Like the evidence for boom-and-bust cycles associated with soil erosion in prehistoric Europe. For example, from pp85-86
German soils record periods of agriculturally induced soil erosion from hillsides followed by periods of soil formation lasting roughly five hundred to a thousand years. Soil profiles and alluvial sediments in southern Germany's Black Forest record several periods of rapid erosion associated with increased population. Neolithic artifacts in truncated soil profiles show that initial erosion after the arrival of agriculture about 4000 BC culminate in extensive soil loss by 2000 BC. Declining cereal pollen and a period of soil formation characterized a thousand years of lower population density until renewed erosion in Roman times peaked in the first centuries AD. A second cycle of agricultural decline, soil formation, and forest expansion followed until renewed population growth in the Middle Ages initiated a third, ongoing cycle.

I could have used a far longer exposition of the evidence for the collapse of neolithic agricultural societies in Europe.

A longer and fascinating-to-me account was the discussion of the neglect of soil conservation measures in the early post-revolutionary United States. Apparently, techniques of crop-rotation, fallowing, and manuring that were commonplace in Europe were all neglected in the US because it was cheaper to exhaust the soil and move west than it was to look after the soil properly. His hypothesis of the role of soil exhaustion in the start of the Civil War is particularly fascinating. For example, from p137:

But this still doesn't explain the vitriolic southern opposition to Lincoln's proposed territorial limitation on the spread of slavery. After all, slavery in the South was not itself directly at issue in the the election of 1860. Consider that slaves moved west along with their owners. At the time of the first national census in 1790, Maryland, Virginia, and the Carolinas held 92% of all the slaves in the South. Two decades later, after a ban on importing more slaves, the coastal states still held 75 percent of the South's slaves. By the 1830s and 1840s many of the slaveholders in the Atlantic states were breeding slaves for western markets. For plantation owners who stayed behind to work exhausted fields, the slave trade became an economic lifeboat. In 1836, more than one hundred thousand slaves were shipped out of Virginia. One contemporary source estimated that in the late 1850s slave breeding was the largest source of prosperity in Georgia. Census data for 1860 suggest that the value of slaves directly accounted for almost half the value of all personal property in the South, including land. By the start of the Civil War, almost 70 percent of the slaves in the South toiled west of Georgia.

Whether Missouri, Texas, and California would become slave states was a make-or-break issue for plantation owners moving west. The labor-intensive plantation economy of the South required conscripted labor. And for all practical purposes, the rapid soil erosion and soil exhaustion produced by slave-based agriculture condemned the institution of slavery to continuous expansion or collapse. So if slavery was banned in the West, slaves would lose their value - wiping out half of the South's wealth. Lincoln's election threatened slave owners with financial ruin.
It's certainly a compelling and plausible narrative!

But this brings me to my greatest problem with the book - I can't figure out how much of it to believe. Most of the book is structured as narrative exposition, with relatively light attention to detailed evidentiary discussion, and minimal guidance to the reader as to where the author is advancing ideas that are far from a scholarly consensus (a critique I would strongly make of Clive Ponting's books also).

For example, Montgomery discusses the role of soil exhaustion in weakening Rome prior to it's final collapse. He doesn't claim it's the only factor, but he does claim it's an important one. But what he doesn't give his reader much clue about is the fact that mainstream historians don't by and large believe this. For example, if you read something like Peter Heather's The Fall of the Roman Empire : A New History of Rome and the Barbarians you'll find internal Roman environmental problems largely dismissed as irrelevant, and instead a focus on the expansion of the Persian and Hunnic empires placing excessive stress on a Roman empire which had long-standing vulnerabilities due to the need to have more than a single emperor to cover an empire that was large compared to contemporary travel distances.

Where does the truth lie? I don't know, but Montgomery didn't give me much compelling reason to accept his account over that of others. Similarly, in his discussion of more current issues, I found more than a hint of an author who had accepted relocalist ideology without enough critical thinking (p243)
Similarly, as oil becomes more expensive it will make less sense to ship food halfway around the world: the unglobalization of agriculture will become increasingly attractive and cost effective. The average piece of organic produce sold in American supermarkets travels some 1500 miles between where it is grown and where it is consumed. Over the long run, when we consider the effect on the soil and on a post-oil world, markets for food may work better (although not necessarily more cheaply) if they are smaller and less integrated into a global economy, with local markets selling local food. As it becomes increasingly expensive to get food produced elsewhere to the people, it will become increasingly attractive to take food production to the people -- into the cities.
While this sounds superficially plausible, the whole idea that "food miles" is a useful measure of the environmental impact of producing and distributing food, or even the amount of oil required, doesn't really stand up to much quantitative analysis. And the idea that producing food in the city is ever going to happen at much scale sounds implausible to me: the whole point of cities is to concentrate all the non-food-producing trade-and-manufacture functions of society in a small area where they can happen more easily and efficiently. This has been true since Uruk was the largest city in the world, was true all through the era when society had no petroleum at all to power transport, and I'm sure will be true in the future: land in the city is too expensive to use very much of it for food production.  We're a long way from soil science here, and I think the fact that the author hasn't actually thought this through very deeply is showing.

So at the end of the day, although I'm glad I read it, I found myself unsatisfied by the book. It's enjoyable and interesting, but I'm looking forward to the beefier more thorough book that I hope Montgomery will write in ten years after he's thought harder about all the issues, engaged more thoroughly with opposing points of view and gained a deeper understanding of disciplines outside of his own that bear strongly on the points he's trying to make.


Mike Aucott said...

Dirt looks like an interesting book, albeit not a particularly festive choice for the Holidays. I too have found apparent hyperbole mixed with probable truth in discussions of soil, soil quality, and soil's role in civilization. This mix of likely fact and exaggeration goes back a long time, to Sir Albert Howard and J.I. Rodale (e.g., the latter's 1946 book, Pay Dirt).

There are some other good books on food and soil that you and your readers might appreciate. Especially good are Michael Pollan's Omnivore's Dilemma and Vaclav Smil's Enriching the Earth. And the famous soil science textbook, The Nature and Properties of Soils, by Nyle C. Brady, is a wonderful read.

It's clear to me that nitrogen is generally the limiting nutrient in soils, and that, without fossil fuel inputs (in the form of nitrogen fertilizer made mostly from natural gas), maintaining organic matter in the soil is essential to maintaining the ability of soils to supply nitrogen.

My understanding is that soils in temperate regions lose about 2% of their organic matter per year when cultivated (exposure of soils to air and sun accounts for most of the loss). If not replaced (in the form of manures or crop residues) the organic matter content of cultivated soils steadily declines. Virgin soils have maybe 8 or 10% organic matter. If this percentage declines below about 2%, soils become hard to work, infertile, and erosive.

I've been a gardener for many many years and operated a small farm for 7years. It's clear to me that maintaining and building organic matter content in a soil leads to healthier plants and better crops. (It's also clear to me that judicious use of chemical fertilizers and pesticides is not a problem - but this is another subject.) In my view, as fossil fuels become increasingly expensive, maintaining soil organic matter content will become a growing and probably critical challenge.

(By the way, I took your example and started a blog of my own - it's at


Mike "Pops" Black said...

Hi Stuart, just a comment about relocalization of food production and why I think it is important.

Food is transported long distances because plants grow best in climates/soils they prefer and as their cultivation becomes more concentrated so does the infrastructure needed for production giving a better profit.

For example, according to the FAO just a few counties in Central California (where I grew up btw) now account for 68% of the worlds almond production and just one variety accounts for almost half of that production.

Concentrated geographically and genetically, irrigated and highly mechanized, with a concentrated, specialized production infrastructure, it's a perfect example of the increasing vulnerability of our monocultural food supply.

Obviously almonds aren't a staple, basic food but they are a great example of how geographic and genetic monoculture is the real problem behind long distance food transport.


Datamunger said...

Thanks for sharing your thoughts, Stuart. Here's hoping for more mini-reviews. :-)

Unlike all those of the past, ours is not primarily an agricultural civilization. (Though of course we depend on it) As much as mistreating the land disturbs me, I'm grateful for attempts to get perspective on it.

I'd guess though that as fossil fuels deplete, the ag portion of GDP will grow (not just on account of biofuels) and that soil and water will emerge much more as big issues.

kjmclark said...

That's funny. I thought Heather was pretty much dismissive of just about all the other theories of the fall of Rome, and didn't give a lot of evidence debunking any of them. I'll have to go back and re-read it. I was only interested in the third-century crisis at the time, and this was a couple of years ago, so maybe I missed something.

I've never been particularly interested in the idea of the fall of Rome due to agricultural problems, as opposed to a problem of decline of wood resource/quality leading to an economic crisis. Two other books on similar topics you may want to check are "A Forest Journey" by John Perlin (you'll want to skip the foreword!) and "Man and the Mediterranean Forest" by JV Thirgood.

One of these books on soil erosion had a really tremendous picture of an erosion gully somewhere in the Carolinas. Another had really interesting maps of erosion siltation of ancient harbors. If those weren't Montgomery's, then there are other good books on the topics out there. I'll have to go back and see if I can figure out which was which.

Stuart Staniford said...

Kjm - my point was not so much to claim the Heather book was perfect, or even that his theory is better, but just to illustrate that there is controversy about the role of environmental factors. Peter Heather is certainly a high status expert on that period of european history, and so the fact that there are such people who do not at all accept the role of soil exhaustion in the decline of Rome is significant. So to be persuasive, I think someone like Montgomery needs to acknowledge that he is making a controversial claim, outline what the opposing case is, and then make his arguments for why we should accept his account rather than opposing ones. I didn't think he really did very much of that. So then in other areas he covers where I know less, I don't know how much credibility to give his account.

Stuart Staniford said...


I think the link to your blog should be And welcome to the blogosphere!

Edward Brown said...

I just thought I would say "Thank you" during this holiday season. I have been reading and enjoying your blog for about a month now, and I check it each day.

I just want you to know that your work is appreciated by folks like me who probably don't comment much, but lurk and read daily.

Happy New Year.


kjmclark said...

I did a lookup on google books for "deforestation ancient rome", and found a really good summary with some references in "A Companion to Ancient History", 2009, by Andrew Erskine, a professor of Hellenistic and Roman History at the University of Edinburgh. He covers the erosion/deforestation discussion well. Here's a bit:

"The question of the causes of soil erosion leads us to the problem of deforestation, the most controversial issue in the Mediterranean environmental history. Many historians have believed that extensive deforestation occurred in the Mediterranean in antiquity, leaving a denuded landscape. Undoubtedly by the late republic the city of Rome, for example, had huge requirements for building work, heating houses and baths, industrial activities and many other purposes, which could not be met locally (Meiggs 1982; 218-59). These requirements were mainly met by floating timber down the Tiber to Rome. Wood was also important for metal smelting, for example silver ores in Attica and Spain, copper in Cyprus, iron from Elba near Populonia in Tuscany, and for shipbuilding.

Despite the importance of the timber industry in antiquity, the "ruined landscape theory" of Mediterranean deforestation has attracted criticism (Grove and Rackham 2001 contra Hughes 1994 and 2005). There is a great diversity of opinion about this controversial issue among scientists who are specialists in Mediterranean ecology (e.g. Blondel and Aronson 1999; 2001-6 contra Grove and Rackham 2001). Similarly, those historians who do believe in large-scale deforestation have differences of opinion regarding its chronology. One study concluded that there has indeed been extensive deforestation in five mountain zones of the Mediterranean world (McNeill 1992). However the conclusion was reached that it occurred principally in the early modern period, not in classical antiquity."

There's a lot more. I've put a request in for some of the works cited.

woodsy_gardener said...

Fascinating post, thanks.

Couple of points:

Soil is full of life; from bacteria and fungi to my beloved earthworms. Plants grow in soil. Too often soil is treated like dirt--something to hold the plant upright. To grow a plant in dirt you need fertilizer, pesticides, herbicides, miticides, fungicides. These pollute aquifers, rivers, and ultimately create dead zones in the ocean.

NPK N is not the limiting factor there are plants that fix nitrogen into the soil. Thunderstorms create a fertile rain of N. The limiting factor is P. We've flushed so much P rich detergent as well as fertilizer down the river that there will be a shortage. In the 1800s England maintained their P levels by stealing bones from cemeteries around the world.

Stuart Staniford said...

Kjm - thanks, that does look interesting. Keep us posted if you find any other particularly interesting resources.

Also, I did a Google Scholar search for David Montgomery and went through the first ten pages of results. He has a large number of widely cited publications, but to judge by the titles, the bulk of his influential work is on technical aspects of soil formation and erosion, and none of it is on ancient history.

I also went over the publication list to see if there was significant more recent research on some of the book's more controversial claims, that hadn't had a chance to become widely cited yet. There's a fairly short survey paper here, but I see no sign of publications in ancient history.

So my feeling here is that we have a very good soil scientist who hasn't absorbed or engaged with much of the perspective of historians on these issues. So his conclusions on the historical significance of particular episodes of soil erosion may be intriguing but not necessarily very reliable yet.

Stuart Staniford said...

Ed - thanks for saying so!

Stuart Staniford said...

woodsy_gardener - I was going to say that the peak phosphorus theory doesn't pass the smell test to me and refer you to the comments I wrote briefly about it in the "Fertilizer" section of Food to 2050: roughly, the USGS reserve base numbers for phosphate are enormous, and there's been no sign of a spike in phosphate prices, so I think a serious problem any time soon is unlikely. The I double checked phosphate prices to see if more recent data changed the picture and found this extraordinary graph with a huge price spike. So I guess I should shut up and do some more research.