One of the fears about the sustainability of agriculture is that erosion will remove so much topsoil that crops can't grow. Indeed land with good topsoil is better for agriculture than land with poor topsoil. This difference is reflected in the price per acre of land.

One of the main viewers with alarm about erosion is David Pimentel of Cornell University. He wrote that erosion costs U.S. agriculture $40 billion per year. See (D. Pimentel, et al., Science 267, 1117 (1995)), although I'm not sure this is the article that has the cost estimate.

In Science for 1999 August 20, Stanley W. Trimble of the University of California at Los Angeles wrote that erosion has greatly decreased since the 1930s in the area he studied in Wisconsin. Here is Trimble's abstract.

The total measured rate of alluvial sediment accretion in the agricultural Coon Creek Basin for the period 1975-93 was only about 6 percent of the rate that occurred in the 1930s, but the distributed changes within the basin were highly variable and systematic. Sediment yield (efflux), however, remained relatively constant despite large stream and valley changes within the basin. These observations demonstrate (i) that sediment sources, sinks, and fluxes vary widely over time and space and (ii) that, although improved soil conservation measures have decreased soil erosion, the downstream effects are complex.

According to an article by Richard Monastersky in Science News for 1999 August 21, Pimentel "responds that Trimble's study looked specifically at sediments in a watershed, not at what was happening on fields." Pimentel was quoted as saying, "It really is not a good, sound study on agricultural croplands or pasturelands. We really don't know how much has left the fields."

I'm sorry not to have more on how serious the erosion problem is.

Here's [2000 February] a lot more information. D. Gale Johnson in his article The growth of demand will limit output growth for food over the next quarter century in the colloqium Plants and population: Is there time?, sponsored by the National Academy of Sciences and held in December 1998 refers to the work of Peter Lindert of U.C. Davis as follows:

Has the world's cropland been seriously degraded?

It is alleged that the quality of the world's land resource has been degraded through erosion, loss of organic matter, and other forms of loss of potential productive power (Footnote 10). There are claims that enormous quantities of top soil are lost each year to water and wind erosion. Does the world enter the next millennium with soil that is badly degraded and of lower productive capacity than it was when we entered the 20th Century or in 1950 when the recent surge in agricultural productivity started? There are many allegations that this is the case.

Such allegations emanate from the Worldwatch Institute and the World Resources Institute, for example. Some aspects of these claims were endorsed by the presidents of the Chinese and the United States academies of sciences in a statement issued in Beijing on January 16, 1997. They agreed that "The need for improvement is urgent, since all resource indicators - changes in the atmosphere, loss of topsoil, loss of forests, the extinction of organisms - have moved sharply and continuously downward during the second half of the twentieth century, while both world population and levels of consumption continue to rise. Globally, these trends are not sustainable over the long run." Among the impending disasters, I shall consider only the loss of topsoil.

I agree with Peter H. Lindert of the University of California at Davis that the claims that there have been serious loss of topsoil from agricultural lands or other forms of widespread deterioration of the productive capacity of farm land are without foundation - are not based on evidence of change over time. Lindert states the following: "Lacking an actual quantitative history of soil conditions beyond experimental plots, scholars have fixed on the useful but risky task summarized by the title of this section (Footnote 11). The literature has three main shortcoming: (1) using crude indicators that prove little about human impacts on the soils, (2) using trends in cultivated land area as clues about land quality, and (3) using single-snapshot predictions as if they were time-series data." (Lindert 1998, p. 3) (Footnote 12).

Lindert utilized long run data from two large developing countries - China and Indonesia - in an effort to measure changes in the quality of the land resource. The data he used were soil surveys, dating from the 1930s to the relatively recent past, the 1980s in China and 1990 in Indonesia. This is not the place to provide a detailed summary of his results, but his following brief summary will provide a sharply different perspective than those that are commonly given (Lindert 1996, p. 37):

"The broadest outlines of the interaction of soil and agriculture are now somewhat clearer for two developing countries. We have some idea which dimensions of soil quality have improved and which have not. Soil organic matter and nitrogen appear to have declined on cultivated lands in both China and Indonesia. Total phosphorous and potassium have generally risen. Alkalinity and acidity have fluctuated, with no over overall worsening. The topsoil layer has not gotten thinner.

"Some of the mixed trends revealed here have more effects on yields than others. China's patterns show that the decline in soil organic matter and nitrogen makes little difference, presumably because fertilizers can substitute for the soil endowment. More relevant are the pH and total potassium, for which the trends are better."

Based on comparisons of soil surveys over a period of 50 years, Lindert reaches the remarkable conclusion for China and Indonesia: "The topsoil layer has not gotten thinner." This conclusion is wholly inconsistent with the statements from the presidents of the two academies of science and from most other commentators on the subject.

This is not to deny that erosion exists in China - after all, the Yellow River didn't get its name by accident. But to note that there are obvious signs of erosion does not tell us from whence it comes or why it occurred. The data assembled by Lindert indicates that it has not come to any considerable extent from farm land. I find it hard to believe that farmers are as careless with a resource that is very important to them as is implied by much of the opinion expressed concerning the alleged enormous loss of topsoil. I have long held that farmers are at least as smart as the rest of us and I believe that they know how to act in their own interest. It is not in the interest of farmers to prevent all erosion since there are likely to be costs involved. But in the range where the benefits and costs of preventing erosion are approximately equal or the benefits exceed the costs, it seems reasonable to assume that farmers are acting in their own interest, at least until there is solid evidence to the contrary. Lindert's analysis of historical data for two important countries indicate that there may not be evidence to the contrary. Soil surveys do exist in other countries. It is perhaps time that more use is made of this neglected source of information about the state of the world's land resource and less reliance is placed on information that lacks a time dimension.

Here are Johnson's references to Lindert.

Lindert, Peter H. (1996), "Soil Degradation and Agricultural Change in Two Developing Countries," Working Paper Series No. 82, Agricultural History Center, University of California, Davis.

"The Bad Earth: China's Agricultural Soils Since the 1930s," forthcoming in Economic Development and Cultural Change.

Soil Change in General

Soil changes in many ways over time, with and without human intervention. An important new book on the subject is Understanding Soil Change: soil sustainability over millenia, centuries and decades by Daniel Richter and Daniel Markewitz, published in 2001 by Cambridge University Press. The authors' detailed research concerns a farm in South Carolina, on which farming stopped in the early 1950s and which has grown up as an unfertilized and unharvested forest of loblolly pines since then. They carefully measured the changes in soil at various depths over more than 40 years.

Here are some conclusions I drew from their work.

  1. There are several levels of soil between the surface and bedrock, and each level changes over time in different ways.

  2. The book has information about changes occurring over different time scales - millenia, centuries, and decades.

  3. Farming was abandoned in the upland Piedmont areas of the Southeastern US, because it couldn't compete with farming the same crops elsewhere, even though farming was still feasible. Thus cotton farming in the Southeast can't compete with cotton farming in Texas and California. Farming could be resumed, but productivity elsewhere is growing too rapidly to make this likely in the near future.

  4. Repeated cropping of trees in the Southeast will require some nutrients, especially nitrogen, to be supplied by fertilizer.

  5. The study of soil change is in its infancy. Systematic studies over many decades are needed in many climates and basic soil conditions. The authors propose international co-ordination of such studies.

  6. Unlike Pimentel, the authors don't predict disasters.

I hope to include more information from this interesting book.

The long term

The studies mentioned above are relevant to at most the next few hundred years. Longer term agricultural progress may require more extensive measures. Here are some considerations.

  1. Topsoil can become a commodity on a large scale. Here are some preliminary calculations. Some years ago land with good topsoil in Idaho sold for $4,000 per acre and land with poor topsoil $1,000 per acre. The topsoil apparently has to be 6 inches deep. Taking into account the cost of trucking, it would be marginally profitable to truck topsoil short distances from places with deep topsoil (river bottoms) to places with poor topsoil. I concluded that it won't happen until the price differential is considerably larger.
  2. The Mississippi delta mainly consists of topsoil eroded from higher in the Mississippi Valley. At considerable cost, the delta could be partly dug up and barged back up the river.
  3. Volcanic ash grows good crops within 20 years of the eruption that created the ash. This phenomenon challenges the statement that good topsoil takes hundreds of years to develop.
  4. Erosion is one of a few factors affecting long term sustainability of progress about which not enough is known to make confident predictions. No-one seems to be motivated to carry out studies needed to answer these questions. It would be nice to know more, but there isn't any emergency. Our great great great grandchildren will do it if they need to by then.

Up to: Sustainability of Material Progress

Send comments to

The number of hits on this page since 1999 August 27.