Editor's note

Agriculture is fundamental to human survival. As Wohlmeyer says, it feeds us. Organized agriculture (including managed forestry) now accounts for in the order of half of the total biomass of the earth. By the middle of the twenty-first century, to accommodate a doubled human population at (it is hoped) a higher standard of living, agriculture will necessarily be more intensive and probably more "industrial" than it is now. At least that is the conventional view. Some background data may be useful to put this set of problems in perspective. Wohlmeyer's contribution, which follows, addresses the possibility of an alternative approach to agriculture.

Agriculture and food production

Food production per capita in the developing countries increased 2() per cent during the 1980s. Yet, 800 million people in the world are undernourished and 500 million are chronically malnourished. Grain harvested per capita for the world as a whole rose 40 per cent from 1950 to 1984 (the peak year). However since then it has been falling at 1 per cent per year, with the biggest decline in the poorest countries. World grain planted area has been dropping since the early 1980s. Grain area planted per capita fell from about 0.24 hectares (ha) in 1959 to barely half that (0.13 ha) in 1992. This reflects loss of arable land to erosion and desertification.

Meanwhile, grain fed to livestock has doubled since 1950, aggregate meat production has more than tripled, and meat production per capita has increased by over 50 per cent (though not since 1988). The rising production of meat can be regarded at first blush as good news. However, it is not. It means that more and more basic plant calories are being diverted away from feeding people to feeding animals. In effect, the demands of the more prosperous segments of the world's population are aggravating the situation faced by the poorest.

Evidently, global food production is still rising but more slowly than in the past. Productivity gains per unit of cultivated land are continuing but becoming more and more difficult to achieve. High rice productivity per hectare in Japan is not a viable model for the rest of Asia - it is not only labour intensive, but also very energy and chemical intensive.

China is the focus of major controversy about the future of agriculture. In 1979 the leadership of China was taken over by Deng Xiaoping, a reformer. Under Deng, collective farming was gradually abolished. Though the land itself remained in the hands of the state, privatization of agriculture began to gain momentum during the early 1980s. Peasants were freed from central planning to plant, harvest, and sell as they chose. Food production jumped after 1980. By 1984 China had a record harvest; per capita production of rice in China was up to 90 per cent of Japan's level and the country was self sufficient. In 1985 China exported grain. Diets continued to improve. In 1990 another record harvest was achieved, 10 per cent above the 1984 level. The year 1993 saw still another record harvest, despite continuing conversion of agricultural land to other uses. Production in 1995 was 465 million metric tons, according to Prime Minister Li Peng's latest report to the National People's Congress. The target for 2000 is 500 million metric tons. Simply extrapolating the recent trend, many economists have concluded that China will have no problem feeding itself for the foreseeable future.

In China, urban population in 1950 was 61 million; currently it is around 330 million; by 2050 it is expected to be around 1 billion, according to the United Nations' population projections. The situation in India is, of course, comparable if not worse. This urbanization will, among other things, remove very significant amounts of land from cultivation. In China, for instance, cities, towns, and infrastructure took 324,000 km² of land in 1994, as compared with 1,364,000 km² that were cultivated.¹ In 1988/89 alone, 174,000 ha (1,740 km²) were lost to cultivation and converted to urban or industrial use. Assuming this modest rate of conversion continues unabated, only about 100,000 km² more land would be lost to cultivation in China by 2050. But if population increases as projected and the density of cities does not increase, urban area would have to increase by a factor of 3, i.e. to about 1 million km², largely at the expense of nearby land, most of which is cultivated at present. So the potential loss of cultivated land could be as much as 650,000 km² or close to half of the current total.

A closely related local/regional problem associated with urbanization and economic development is groundwater shortages. The cause of the water shortages now plaguing much of northern China is debatable, but the most likely cause is too many wells and too-intensive pumping of groundwater for industrial and municipal purposes. Wet rice cultivation and fish farming have both been abandoned in this region of China. The Heaven River no longer exists, and canals that formerly brought water to Beijing are now dry. In late 1993, the water resources minister of China (Niu Maosheng) said that 82 million rural Chinese were suffering water shortages, along with 300 Chinese cities, in 100 of which the shortages were "extreme" (Tyler 1993).

Projecting current trends, China will increase its per capita consumption of food, especially meat, while losing half its arable land to urbanization and industrialization by 2030. Unless agricultural productivity increases faster than it has in the past, China will necessarily become a large importer of food. For the above reasons, among others, Lester Brown of the World watch Institute has again raised the alarm about China's future ability to feed itself.² This would force China into the world's food market as a major importer. This shift will quickly exhaust reserve capacity and create the conditions for shortages and rapid and sharp price increases. But, whereas the gasoline shortage merely caused lines at filling stations, food shortages can cause mass starvation.

Vaclav Smil takes issue, mainly, with one point in Brown's thesis, namely his assumption that China has little scope for further increases in agricultural productivity (Smil 1996). Smil argues (1) that China has at least 25 per cent and possibly 45 per cent more cultivated land than official statistics admit (owing to underestimation of cultivated area to avoid taxes), whence (2) yield is actually much lower than official statistics assume, leaving more room for future improvement than Brown allows for. Smil also argues that there is enormous waste in the Chinese irrigation and fertilizer production systems, and also in the food distribution system, all of which offer significant further room for increased food supply to consumers in the future. Nevertheless, Smil concedes, rather reluctantly:

It would be a mistake to dismiss Brown's predictions as just another scare. Concerns about China's long term food production capacity are valid and many knowledgeable people, both Chinese and non-Chinese, are far from optimistic in their long-term assessment of it. (Smil 1996, p. 33)

The most optimistic (and mainstream) view is presented by Dennis Avery of the Hudson Institute. Avery notes that increased food production per capita in the third world has been based on a virtually constant cultivated land area. It was mainly due to the so-called "green revolution," which introduced new strains of rice that responded very well to increased fertilizer use. Avery expects further technological progress to continue this trend. He states that the world can feed "another billion people, right now, without stressing any fragile acres or putting on heavy doses of farm chemicals" (Avery 1995, p.388). He does not, however, suggest that there is room for another 5 billion people, which is what the UN population projections have suggested that we must expect.

The environmental impacts of industrial agriculture

The major mass flows and emissions associated with agriculture are nitrogen losses (as ammonia and nitrous oxide) via various processes, methane emissions from livestock, and soil erosion. Nitrogen losses must be made up by organic recycling, nitrogen fixation by legumes, or the application of synthetic nitrogen fertilizers. Erosion is the major cause of phosphorus loss, which must also be made up by the application of phosphate fertilizers.

A detailed nitrogen and phosphorus balance for five regions, including world croplands, and all agro-systems for both the United States and China, has been prepared (Smil 1993). According to this study, biofixation accounts for only 17 per cent of global crop nitrogen inputs (25 per cent in the United States, 11 per cent in China), with synthetic fertilizer supplying 43 per cent of the nitrogen for world croplands (31 per cent in the United States, 52 per cent in China). Smil also estimates the contribution from atmospheric deposition: 8.5 per cent of global Ninputs, 6.2 per cent of US N-inputs, and 3.7 per cent of Chinese N-inputs).³ The remainder is from organic recycling.

Future growth in food production in Asia and Africa will evidently be more dependent on synthetic ammonia production than on either organic recycling or legumes. Worldwide, about 87 per cent of synthetic ammonia production is used for fertilizer [Info. Chimie 346, 1993]. The World Bank's World Nitrogen Survey projected a 37 per cent increase in consumption in Asia between 1992 and 2000 (Constant and Sheldrick 1992).

Long-term projections of population growth and food production in the developing world suggest huge increases in fertilizer use. To put the problem in perspective, Food and Agriculture Organization (FAO) data for 1989 suggest that, in the wealthiest countries (Western Europe and North America), grain yield is about 4.4 tonnes/ha with an input of 270 kg/ha of fertilizer (cited in Huq 1994, fig. 8). For Eastern Europe and the former USSR, yields averaged 3.5 tonnes/ha with fertilizer inputs of 145 kg/ha. East Asia had yields of 3.3 t/ha with inputs of 135 kg/ha. Latin America had yields of 2.3 t/ha for fertilizer inputs of 90 kg/ha. South Asia got 2.25 t/ha with inputs of 77 kg/ha. Finally, sub-Saharan Africa enjoyed yields of 1.3 t/ha with only 26 kg/ha of fertilizer inputs.⁴

When these data are extrapolated, it is clear that fertilizer use increases nonlinearly with yield. There are strong indications that nitrogen fertilizer use in America and Europe may already be at saturation level. More fertilizer would not increase yields significantly, if at all. But, to raise grain yields in East Asia by 25 per cent - to current West European or US levels - fertilizer use in that region would have to double. For South Asia to double its grain yields, nitrogen fertilizer use would have to increase by something like 350 per cent. In the case of sub-Saharan Africa, low grain yields might conceivably increase three-fold, but fertilizer consumption would probably have to rise 10fold. In summary, if the world population doubles by 2050, as UN projections imply, most of the increase will be in the developing countries. To allow for urbanization and for increased food consumption per capita, fertilizer inputs in the developing world will almost certainly have to quadruple or more. This implies very significant increases in global nitrogen mobilization.

Nitrogen losses for the world (c. 1990) were about 95 million metric tons (MMT). Losses exceeded synthetic nitrogen fertilizer inputs - around 75 MMT (N)-by 20 MMT. Smil assumes the balance is made up by "mineralization" of 30 MMT (from fossil biomass) and atmospheric deposition of 15 MMT (Smil 1993). This seems unlikely. For instance, it is hard to see why the world average from atmospheric deposition should be higher than the US average, considering that both ammonia emissions (from animals) and NOx emissions from combustion sources are much higher in the United States than in the rest of the world. It seems more likely that Smil overestimated this source of nitrogen as a global input. Many experts believe the global system is out of balance and that the soils of the developing world, especially Asia and Africa, are being "mined" of nitrogen content - which must ultimately be replaced. This implies a growing worldwide nitrogen deficit. It also implies an enormous future demand for fertilizer.

Of the 95 MMT worldwide N-loss estimated by Smil, the United States accounts for 14 MMT. Western Europe probably accounts for somewhat less, perhaps 10 MMT (in proportion to agricultural output). Evidently most of the global N-emissions (roughly 70 MMT) are occurring outside the OECD countries. Subtracting the probable OECD contributions, the rest-of-world (ROW) breakdown of losses seems to be roughly as follows:

The denitrification process (due to bacterial action) generates mainly dinitrogen gas (N₂), but a small proportion (5-7 per cent) of the nitrogen is released as nitrous oxide (N₂O), which is a greenhouse gas. The fraction released as N₂O is not accurately known, and is certainly a function of local soil conditions and humidity. However, the best available estimate of the ratio is 16:1, which implies a global N₂O production of about 1.25 MMT, of which about three-quarters now comes from the ROW countries.⁵

This pattern of N-losses can probably be expected to continue for the foreseeable future, except that the totals can be expected to increase roughly in parallel with consumption of synthetic nitrogen fertilizer. In other words, nitrogen mobilization and N2O emissions may be expected to quadruple by 2050. Nitrous oxide currently accounts for something like 7-8 per cent of greenhouse gas emissions. This fraction can be expected to increase.

The other major air pollutant emissions associated with agriculture are ammonia and methane. Ammonia is associated with livestock urine and manure, especially on large-scale feedlots. Most of the ammonia released is volatilized and eventually returned to the land as nitrates and sulphates, which are essentially fertilizers. Little if any harm is done, except that some of the nitrogen is lost to non-agricultural land or oceans. Contamination of groundwater is a more serious problem, however, because ammonia itself and many organic nitrogen compounds are toxic. Grazing animals, especially ungulates, also produce methane in their stomachs - the work of anaerobic bacteria that help with the digestion of grasses. Methane is also generated in large quantities from wet rice cultivation; it is caused by anaerobic decay bacteria in the mud. The OECD countries account for the great majority of cattle and sheep in the world, but wet rice cultivation is primarily an Asian phenomenon. Overall, methane from agriculture (wet rice cultivation) accounts for about 8 per cent of global "greenhouse potential" (WRI 1990, table 2.4).

Pesticide use in the developing world is also rising faster than agricultural production, as measured in monetary terms (WRI 1994, fig. 6.11). The reasons for this are similar to the reasons for expecting fertilizer use to increase faster than food production. Many pesticides banned or tightly restricted in the OECD countries are being used, and in some cases manufactured, in developing countries. Chlordane and DDT are two cases in point. Pesticide use in the OECD countries has levelled off or even declined in recent years. By contrast, pesticide use in many countries of the Asia-Pacific region has been growing at 10 per cent per annum or more, notwithstanding some progress toward integrated pest management (IPM). In India, treated crop land increased from 6 million hectares in 1960 to 80 million in 1985; pesticide production in that country increased 13-fold between 1970 and 1980 and now meets 90 per cent of domestic demand (not to mention exports). India, Indonesia, and Russia are major producers of DDT, for instance, though use of that chemical was supposedly ended in the early 1970s (Ayres et al. 1995).

Agriculture is a major consumer of carbon dioxide and producer of oxygen. For instance, annual above-ground production of harvested crops in the United States is between 500 and 600 MMT, plus a slightly smaller amount of above-ground residues, making a total of 1,100-1,400 MMT but not including grass consumed directly by animals (another 200 MMT). About half of this mass is water; the rest is a combination of carbohydrates, proteins, and fats, and a small percentage of mineral substance (ash). Year-to-year variation depends on rainfall and climate factors. Much of the unharvested crop residue in the United States is left on the land and recycled. In Asia and Africa as much as two-thirds of it is collected and burned - rather inefficiently - for cooking (Smil 1993).

For every 100 units of dry biomass - taken as sugar (ribose) - produced by photosynthesis, 146.7 units of CO₂ are extracted from the atmosphere and 106.7 units of oxygen are returned thence. However, it must be recognized that biomass fed to animals for metabolic purposes reverses the photosynthesis process, consuming oxygen and generating carbon dioxide again. In principle, agriculture and forestry need not affect the global carbon cycle in equilibrium. However, in present practice, they do.

Despite increasing efforts to develop alternative modes of cultivation especially "no till" agriculture (now strongly promoted in the United States) world dry-land agriculture, as currently practiced, involves extensive ploughing of the soil, mainly for weed control. Ploughing exposes humus (partially decayed organic material) to rapid oxidation. On a net basis, global agriculture is losing humus faster than it is being replaced. No-till methods have a considerable potential to reduce this loss, but they require extensive reeducation of farmers.

Erosion, flooding, and topsoil loss are major environmental problems in all countries. Erosion losses in the United States have been carefully monitored (and have been gradually decreasing in recent years). However, despite this, annual topsoil losses in the United States currently average 1.5 billion metric tons or a bit less. In Asia, Africa, and Latin America the erosion problem is not under control. Topsoil losses each year in the non-OECD (ROW) countries, taken as a group, are at least 20-25 times the US level. The fact that erosion accounts for a very large fraction of non-OECD losses of N is significant. It is also the only significant mechanism for loss of phosphorus. Evidently, erosion control would be a very potent tool for reducing demand for synthetic fertilizers. By extension, it would also reduce the very serious pollution problems that are caused by phosphate rock mining, beneficiation, and fertilizer production.

The other source of disequilibrium is deforestation. Deforestation (at present) occurs mainly in the tropics, either for logging or to open new land for cultivation or cattle-grazing. Forests are major storehouses of fixed carbon. Cutting (or burning) forests releases this carbon to the atmosphere as carbon dioxide.

The rate of deforestation is variously estimated, and the larger estimates are somewhat disputed. However, the World Resources Institute (WRI) estimated several years ago that deforestation alone contributed 14 per cent to global climate warming potential. Of this, 10 per cent was attributed to carbon dioxide release and 4 per cent to methane release (WRI 1990, table 2.4). (For purposes of comparison, this implies that global deforestation contributes as much CO₂ as all the fossil fuels burned by the United States and the former USSR, combined.)⁶ The problem of deforestation is essentially limited to the tropical developing (non-OECD) countries of South-East Asia, Africa, and Latin America. In fact, Europe and the United States exhibit net reforestation, which partially counteracts the loss of tropical forests in terms of the global carbon cycle.