FOOD, LAND, POPULATION and the U.S. ECONOMY
by David Pimentel Cornell University and Mario Giampietro Isiituto Nazionale dell; Nutrizione, Rome

Released November 21, 1994
Carrying Capacity Network 2000 P Street, N.W., Suite 240 Washington, D.C. 20036 (202) 296-4548

I. Introduction

This report focuses on the interdependency of land, food, and population in the U.S. economy. The United States is in a privileged situation compared to other nations in the world: the per capita endowment of natural resources is relatively high because of the relatively low population density. At the same time, the United States is seriously risking loosing this privilege if more attention is not given to the control of population growth (including immigration), the sustainable management of natural resources, and the development of alternative energy sources. The aim of this report is to increase the awareness of policy makers and the public of the importance of the interaction between population growth, self-sufficiency in food production, standard of living and, ultimately, national security.

The need to reconsider development strategies

Development strategies based on the economic theory of continuous growth imply a heavy reliance on stock depletion (e.g. coal, oil, minerals) and the option of imports from the international market. These strategies have worked well for the western world until now, since developed countries effectively managed to guarantee their citizens a high standard of living despite local shortages or temporary fluctuations in the supply of natural resources.

However, in the last few decades several points have become clear and suggest the need to seriously reconsider these strategies:

* Non-substitutability of natural resources

Technology is of immense importance in the effective management and exploitation of various natural resources, but technology can not increase the flow of natural resources available for exploitation (the 'raw materials'). For instance, increasing the size and efficiency of fishing vessels has enabled us to overfish the oceans but not to increase the quantity of fish produced per capita that, on the contrary, is steadily declining (Ehrlich et al. 1993). The limiting factor determining the fish catch is the reproductive capacity of fish populations. Indeed, biophysical constraints govern the speed of the biogeochemical cycles that regulate the productivity of all natural ecosystems.

Serious shortages of other natural resources, such as land, water, soil, and biota, prevent their use as substitutes for technology. For instance, technology can not double the world arable land, double the flow of the Colorado river or replace bees for pollination.

Land. Approximately 99% of the world food supply is derived from terrestrial ecosystems with the percentage from aquatic systems shrinking (Kendall and Pimentel, 1994). The availability of arable land at world level is less than 0.27 ha per capita, lower than it has ever been in history, and much less than the average of 0.7 ha per capita in the United States (WRI, 1994). Note that 0.5 ha per capita has been suggested as the minimum requirement for a diverse diet of animal and plant food products (Lal, 1989).

Not only is the availability of cropland per capita decreasing as the world population grows, but arable land is being lost due to excessive pressure on the environment. For instance, during the past 40 years nearly one-third of the world's cropland (1.5 billion ha) has been abandoned because of soil erosion and degradation (WRI, 1992). Most of the replacement land has come from marginal land made available by removing forests. Agriculture accounts for 80% of the annual world deforestation (Pimentel et al., 1992).

Biodiversity. The replacement of natural ecosystems, especially tropical forests, for agricultural purposes results in a loss of biodiversity. Ninety percent of the world's food is derived from just 15 plant and 8 animal species, while estimates of the existing number of species on Earth are in the millions. A large diversity of species is vital to agriculture and forestry, and plays an essential role in recycling the vital elements for the living system, such as carbon, nitrogen, and phosphorus, as well as in maintaining a quality environment (Pimentel et al., 1992). In agriculture, for instance, it would be impossible to produce the about $30 billion of fruits and vegetables annually without the free services of honey bees and wild bees for pollination. Because we need to maintain biodiversity in order to stabilize the structure and functions of the biosphere we can not transform the entire land available into agricultural fields.

Water supply. 44,000 Terawatts (1 Terawatt =1012 Joules/sec) of solar energy are used in the ecosphere only to maintain the water cycle, whereas worldwide human activities use no more than 12 Terawatts. These figures suggest that humans are totally dependent on the functional activities of the biosphere for maintaining the productive conditions of ecosystems. Technology can do little to recharge underground reservoirs, whereas agricultural production "consumes" (and requires) more fresh water than any other human activity. For example, to produce 1 kg of corn grain under irrigation requires the availability of about 1,400 liters of fresh water. Worldwide about 69% of the fresh water withdrawn is for the agricultural sector (WRI, 1994). Presently, 40% of the world's people live in regions that compete for short water supplies (Postel, 1989).

* Dependence on stock depletion

Several basic energy resources, such as oil, whose stocks are being rapidly depleted, require a timely substitution if the current pattern of energy consumption is to be maintained in the future for a growing population. The feasibility of such a substitution is doubtful. In fact, remaining fossil energy stocks are depleted at an exponential rate whereas alternative energy sources are not yet significantly expanding their share in the total world energy supply. For example, in the United States fossil energy accounts for 92.3% of the total energy consumption, whereas solar energy sources contribute only about 7.7% (hydropower 3.5% and biomass 4.2%) (Pimentel et al., 1994a).

Projections of the availability of fossil energy resources are discouraging. A recent report published by the U.S. Department of Energy (DOE, 1991), based on current oildrilling data, indicates that the estimated amount of U.S. oil reserves has plummeted. Instead of the 35-year supply of U.S. oil resources, that was projected about ten years ago, the current known reserves and potential discoverable oil resources are now limited to a 15-20 year supply at present rates of pumping (DOE, 1990; Lawson, 1991). Since the United States is now importing more than half of its oil, a serious problem already exists (Gibbons and Blair, 1991). Despite the rapid decline in U.S. fossil fuel reserves, the rate of fossil fuel use is expected to increase by 27% by the turn of the century, to 107 quads per year (DOE, 1991). (See Table 1 for energy units and conversion factors). This increase is attributed both to the growing consumption per capita and the expanding U.S. population.

The food supply worldwide is increasingly dependent on stocks of fossil energy, in the form of fertilizers, pesticides, irrigation and machinery. An increased demand of the U.S. economy for oil on the international market could increase oil prices. This would dramatically affect the economics of U.S. agriculture as well as the agriculture of many other developed and developing countries, all heavily dependent on fossil energy based inputs (mainly fertilizers). Clearly, there is a room for substitutability among fossil energy sources, and natural gas and coal are expected to increase their share as oil supplies decrease. However, gas supplies are not at all that much better than oil supplies. Similarly, the coal supply is finite and its use exacts a high environmental cost or a high price for pollution clean up.

* Dependence on food import

Most of the 183 countries of the world are now to some degree dependent on food imports. These imports come from cereal surpluses produced in only a few countries that have a relatively low population density and intensive agriculture. For instance, in the period 1989-1991, the United States, Canada, Australia and Argentina provided about 81% of net cereal export on the world market (WRI, 1994). As population density increases in these countries, internal grain demand will increase and arable land available per capita will decrease. Under these conditions the cereal grain surplus now exported on the international market will seriously erode.

Many developing countries rely heavily on fossil energy imports, especially in form of fertilizers, to sustain their internal food supply (Giampietro and Pimentel, 1993). A future slow down of fossil energy consumption-because of either a decline of oil supplies, increase in oil prices, or growing restrictions on fossil fuel use to limit its environmental impacts-will generate a direct competition between energy use in developed countries, to sustain a high standard of living, and that in developing countries, to provide an adequate food supply for survival.

The ability to import food from the international market implies that there are countries in which food production exceeds the internal demand and that we deal with a free market. However, when global biophysical limits to food production are reached and local surpluses are absorbed by growing internal demand, import is no longer an option.

* Degradation of environmental life-support systems

Even as the ecological processes underlying our environmental life-support system are taken for granted by many economists, they are increasingly jeopardized by human activities. Deforestation, urbanization, industrial growth and pollution are rapidly spreading throughout the planet. Worldwide, 24% of the land area is subject to high human disturbance. Only less than half (48%) of the Earth's land surface, excluding Antarctica, is in areas currently subject to low human disturbance. This figure includes nonvegetated land such as deserts and rocky mountain tops (WRI, 1994).

Some issues such as the greenhouse effect, ozone holes, and deforestation have won the attention of the general public. Indeed, carbon dioxide emissions from industrial processes have reached troublesome levels of 4.21 metric tons per capita worldwide and 19.53 metric ton per capita in the United States. The atmospheric concentration of carbon dioxide has climbed from an estimated level of 280 ppm in preindustrial times to 320 ppm in 1965 and reached 356 ppm in 1992 (WRI, 1994).

Regarding the environmental impacts of an affluent standard of living, Hall et al. (1994) have estimated that a baby born today in the United States will generate during her/his lifetime 10,355 tons of waste water, 2.5 tons of waste oil and solvents, 13 tons of waste paper, 3 tons of waste metals, and 3 tons of waste glass, as well as indirect wastes from manufacturing, including 439 tons of waste from agriculture, 419 tons from mining (coal excluded), 197 tons from the industrial sector, 83 tons of hazardous waste, 31 tons of demolition, 1,418 tons of carbon dioxide, and 19 tons of carbon monoxide. The same baby will 'consume' during his/her lifetime, among other things, 1,870 barrels of petroleum and 119 kg of pesticides.

* International conflicts over natural resources and social desegregation

Kaplan (1994), in his article "The coming anarchy", argues: "It is time to understand 'the environment' for what it is: the national security issue of the early twenty-first century." The denser the human population becomes, the more countries are forced to closely interact and compete for the shrinking endowment of natural resources. Intensification of such interaction may result in an emphasis of differences in cultural, religious, and political identities, and standard of living, and may precipitate international conflicts.

Examples of rising tension both within and between countries are becoming more frequent. The New Vision, Uganda, Wednesday, May 18, 1994 reported that large crowds of Bangladeshis rampaged through the city Dhaka, on Tuesday, May 17, 1994, expressing their anger at a water shortage they claim India caused by stealing water from the river Ganges. The protest came a day after Bangladesh launched a diplomatic campaign against its giant neighbor, India, charging it with 'unilaterally' drawing from the Ganges river. Early, in October 1993, Bangladesh Prime Minister, Begum Khaleda Zia, told the UN General Assembly that unilateral withdrawal of the Ganges water by India had brought over 40 million people face to face with water shortage and disaster.

West Africa is becoming the symbol of worldwide demographic, environmental, and societal stress, in which criminal anarchy emerges as the real "strategic" danger. Scarcity of resources, overpopulation, malnutrition, and disease are being followed by refugee migrations, increasing erosion of nation-states and international borders, empowerment of private armies, and unprovoked crime. West Africa provides an appropriate introduction to the issue, often extremely unpleasant to discuss, that will soon confront our civilization (Kaplan, 1994).

Similar problems are occurring elsewhere. Today, in 17 out of 22 Arab states the gross national product is declining while in the next twenty years, at current growth rates, the population of many Arab countries will double. These countries, like most African ones, will be ungovernable through conventional secular ideologies. Whereas the distant future will probably see the emergence of a racially hybrid, the coming decades will have us more aware of our differences than of our similarities. For the average person, politics will mean less than personal security.

Shifting from an "empty-world" to a "full-world" development paradigm

The industrialization and development of the United States and West European countries in the 19th and early 20th century was based on: (i) the availability of a large amount of natural resources per capita within their borders, such as fertile land, and stocks of minerals and fossil energy; and/or (ii) the importation of large quantities of natural resources through colonial exploitation and/or international trade. In other words, the industrial revolution of the last century occurred in a world that was still "empty" (Daly, 1992), in the sense that, at the world level, there was an abundant endowment of natural resources per capita.

Thus, being able to rely on the depletion of supplies of local resources and on import, developed countries managed to systematically remove any local ecological and economic constraints to the expansion of their population and its per capita resource consumption. Such a strategy is the basis of neo-classical economic theory that assumes: (i) that supplies of natural capital, such as land, energy and water, impose no limit to economic development; and (ii) that production factors, that is technological capital, labor and natural capital (e.g. land) are substitutes for one another rather than complements. Latter assumption implies that technology can overcome shortages in both natural capital and labor. The "empty-world paradigm" underlies both of these assumptions. In fact, it is true that a shortage in one of the production factors can be compensated to a certain extent by substitution with one of the others; for example, a larger input of technology and/or labor frequently can make up for a shortage in natural resources. However, the feasibility of such a substitution drastically diminishes as natural resources become scarce. Moreover, the substitution of natural resources has a price in terms of a lower biophysical efficiency of the economic process. When the shortage of natural capital becomes severe, substitution becomes practically impossible and biophysical constraints then limit further economic development.

This brings us to the present situation in which the world is full. The exponential increase in the demand for natural resources, due to demographic and economic growth, is rapidly eroding resource stocks and national food surpluses all over the world. As a result, the assumptions typical of the "empty-world development paradigm" are no longer valid. These trends have two important consequences:

1). There is a need to carefully study the effects that a reduction in natural resources available per capita, due to population growth, has on the economy and ecological balance;

2). Self-sufficiency in food production and a reduced dependence on foreign countries for basic resources can become, in the next decades, a determining aspect of sustainability, especially for countries enjoying a high standard of living, such as the United States.

II. The effect of population density and standard of living on agricultural performance and land use: An international comparison

The combined effect of demographic and economic pressure on agricultural performance

Demographic pressure

An increase in population reduces the per capita availability of natural resources, such as land, for a given society. The availability of land for U.S. agricultural production is described in Figure 3. In the United States, about 3.5 ha of land is available per capita However, biophysical constraints, such as climate, slope and soil quality, limit the amount of available land that can be used for food production (arable land). This area is further reduced by alternative land uses, such as urban development and roads. Given that the arable land available for agriculture in the United States is about 188 million hectare (USDA, 1992) and assuming a population of 260 million at the end of 1994, 0.7 hectare of arable land will be available per capita.

Economic pressure (labor productivity)

The socio-economic structure of society also imposes constraints on the labor productivity in agriculture. In fact, a country not only has a limited endowment of land, water and other natural resources to produce food (due to demographic pressure), but also a limited amount of labor hours available to manage and exploit these resources. Latter constraint is described in Figure 4 for the United States. The higher the standard of living, the smaller the fraction of human time available, at the level of society, for labor in the agricultural sector. For instance, out of the 2,277 billion hours of human time available per year to the U.S. economy only 223 billion hours are actually allocated to paid work. Since only 2% of the economically active population in the United States is employed in agriculture, a mere 17 hours per year per capita are available for work in agriculture.

Moreover, in order for U.S. agriculture to be economically viable, U.S. farmers need to have an income level that is comparable to the average income level in the United States, enabling farmers to reach "the average standard of living".

Role of technology in reducing demographic and economic pressure in agriculture

Technology, and the related use of fossil energy, in agriculture has the double role of increasing: (i) the productivity of land by means of fertilization, irrigation and pest control; and (ii) the productivity of labor through mechanization to till and harvest more land in less time. Thus, human-made, technological capital is used to substitute for natural capital to augment the yield per hectare, as well as to substitute for human labor to increase the yield per hour of labor. The combined effect of demographic and economic pressure on agriculture is shown in Figure 5.

Comparing agricultural performance worldwide

The feasibility and implications of substituting technology and fossil energy for land and labor inputs are closely related to the socio-economic structure of society, the standard of living, and the endowment of natural capital. We compare here the agricultural performances of 20 countries that were selected to include different levels of internal and external pressure (Tables 2-5). [Details on calculations, definitions, sources, conversion factors, and assumptions are provided in the Appendix].

For instance, Japan and the Netherlands are examples of societies that face a high economic (high GNP) and demographic pressure (little arable land per capita). Consequently their agriculture is characterized by a high productivity of both labor and land. The United States, Canada and Australia also have a high economic pressure (high standard of living) but a relatively low demographic pressure, that is more land and natural resources per capita. Because of the high GNP per capita, farmers in these countries (as in the Netherlands and Japan) must have a high labor productivity to achieve an income and standard of living comparable to the average in society. Otherwise farming in these countries would not be economically viable. On the other hand, the lower demographic pressure in the United States results in the availability of more arable land per farmer than in the Netherlands or Japan.

Densely populated countries, such as China, Egypt or Bangladesh, with a severe shortage of arable land must find ways to increase the productivity of their land (as must Japan and the Netherlands). However, because of their low average standard of living they manage to operate their agricultural sector with a low labor productivity and therefore achieve a low income for farmers. Finally, several African countries, such as Burundi, Uganda, and Ghana, have such a low economic pressure (low material standard of living) that they do not require the boosting of their endowment of natural capital. Until now, agricultural production in these countries has not relied on heavy subsidies of technology and fossil energy. However, if demographic and/or economic pressure increases, as indeed is occurring all over the African continent, their situation could rapidly deteriorate further.

Relationships between GNP and economically active population in agriculture

The relationships between per capita GNP and the percentage of the economically active population in agriculture for the selected 20 countries is shown in Figure 6. The graph shows a gap in the per capita GNP between developed industrialized countries (GNP over $15,000/year per capita) and the less developed countries (GNP below $2,500/year per capita). The percentage of the labor force in agriculture is lower than 8% in developed countries and higher than 50% in countries with a GNP below $1,000. Figure 6 clearly shows that the lower the GNP per capita, the higher the percentage of farmers in the labor force.

Decreasing the percentage of farmers in the labor force has the effect of increasing the arable land available per farmer within a defined society. For instance, the availability of arable land per capita is about the same for the United States (0.76 ha) and Argentina (0.81 ha), however, because of the smaller percentage of farmers in the labor force (2.0% in the United States versus 10.6 % in Argentina), the arable land per agricultural worker is larger in the United States (64 ha) than in Argentina (21 ha) (Figure 7). [Note that the value of 0.76 ha of arable land per capita in the United States refers to the year 1989 ­the year used for the comparison. Due to the increase in the US population this value is today 0.72 as indicated in Fig. 3]. Similarly, European countries with a high population density, such as Germany FR and the U.K. have only 0.12 ha of arable land per capita, which is less than the arable land available per capita in India or Burundi. However, by keeping the number of farmers down (3.8% in Germany FR and 2.1% in UK compared to 66.8% in India and 91.5% in Burundi) the arable land available per farmer in Germany FR and the UK (6.4 ha and 11.2 ha, respectively) is much higher than in India and Burundi (0.8 ha and 0.4 ha, respectively) (Figure 7).

Availability of arable land and productivity of labor

The high economic pressure, that is the need to attain a high standard of living for farmers, indicates that the optimization of labor productivity is a major goal in agricultural development in developed countries. For this reason the availability of arable land per farmer becomes a key factor in achieving a high labor productivity.

The labor productivity of farmers in relation to the arable land per farmer for the selected countries is shown in Figure 8. Several interesting points emerge from this graph. First, there is a relation between availability of land (natural capital) and labor productivity. Australia, Canada and the United States, the three countries with the highest labor productivity (a crop output of more than 200 million kcal/worker/year) are also the three countries with the highest availability of arable land per worker, more than 60 hectares per agricultural worker. Second, the non-linearity of the relation for these three countries suggests that biophysical, al constraints, other than land availability, also affect agricultural production. For instance, cold weather in Canada and shortage of water in Australia affect productive output. The United States manages to obtain an output per worker that is almost double that achieved in Australia, with an amount of arable land per worker of only half that in Australia. Also this difference can be explained by the different level of technological subsidies related to the comparative advantage (more water) enjoyed by U.S. agriculture.

Third, European agriculture is evidently subject to severe biophysical constraints in terms of shortage of arable land per farmer, which is a consequence of excessive demographic pressure. Because of problems in maintaining farmers' income at acceptable levels, the number of farmers has decreased over the last few decades in the European Community to a mere 5% of the economically active population. In addition, farmers' incomes are still heavily subsidized by the European governments and further reductions in the number of farmers are expected. Although such a change will increase the arable land per farmer in Europe, land shortage due to demographic pressure will remain a problem and will make it impossible for European agriculture to reach the amount of arable land per agricultural worker currently available in the United States, Canada or Australia.

Australia, Canada, the United States, and Argentina combined currently produce about 81% of the net surplus of cereal traded on the international market (WRI, 1991). The relative abundance of land available for agriculture in these countries makes this possible.

Availability of arable land and productivity of land

The need to augment the productivity of land is dependent on the severity of biophysical constraints that are affecting a nation's agriculture. Two major ways to increase the productivity of land (yield/hectare) are fertilization and irrigation. The relationships between land availability and use of irrigation for the 20 selected countries is shown in Figure 9.

Irrigation is an expensive way to augment the yield per hectare. Besides water, irrigation requires expensive fixed investments and large energy inputs for operation. Farmers generally irrigate only when no alternatives are available or if subsidized. The data reported in Figure 9 confirm this. The more a country is faced with land constraints, the more its agriculture relies on irrigation. This is especially true if the percentage of farmers in the labor force is large. Exceptions are the African countries, Burundi, Ghana, Uganda, and Zimbabwe, that are mainly located in the humid tropics or sub-tropical areas and therefore enjoy an ample supply of rainfall.

The relationship between land and labor productivity in agriculture is reported in Figure 10 and reveals interesting data. For instance, although U.S. agriculture has a low performance in terms of yield per hectare, lower than Bangladesh, China, Costa Rica, Ghana, Egypt and the European countries, it has the best performance in terms of labor productivity. China, Bangladesh, and Egypt, which are densely populated, have such a severe land shortage that all technological and fossil energy inputs go into the boosting of land productivity, leaving little room for improvements in labor productivity. Note that China's land productivity is more than double that in the United States, 13 million and 6 million kcal of crop output/ha/year respectively, whereas the labor productivity in China is less than 1/lOOth that in the United States, 2.6 versus 386.1 million kcal/worker/year, respectively.

Substitution of technology for natural capital

Generally, the biophysical output per hour of labor in a society is affected by the availability of natural capital and that of human-made, technological capital. The relative severity of the internal and external pressure experienced by a particular country determines the room for substituting technological capital for natural capital. In economic sectors that heavily depend on natural processes, such as agriculture, forestry and fisheries, the availability of natural capital generally becomes the dominant factor in determining labor productivity. This is supported by the comparison of average farmers' productivity in the 20 countries studied.

In spite of the fact that Chinese farmers use more fossil energy per ha than U.S. farmers (5.7 versus 4.4 million kcal/ha/year), they have a productivity per hour of labor that is less than 1/100th that in the United States. This difference in labor productivity is explained both by the larger natural capital available in the United States (64 ha of arable land/worker in the United States versus 0.2 ha/worker in China) and the better technology available to U.S. farmers. For example, the production of nitrogen fertilizer in the United States by using natural gas is more efficient than the production of nitrogen fertilizer in China by using coal (Smil, 1991).

However, when the performance of European and U.S. agriculture is compared, we deal with two agricultures that have access to similar technologies. Still, the use of energy input per hectare by European farmers is more than twice that in the United States, that is the average in the European Community is greater than 10 million kcal/ha versus 4.4 million kcal/ha in the United States, whereas the productivity of EC farmers is only about one fourth that of U.S. farmers, 100 million kcal/worker/year in the EC versus 386 million/worker/year in the United States. This difference is due to land shortages in the European Community.

The effect of shortages of land on agricultural performance is even more evident when the United States is compared with Japan. In Japan, 80 pieces of machinery (tractors + harvesters) per 100 ha of arable land are used versus 3 pieces of machineryper 100 ha in the United States [FAO, 1991a], although American tractors and harvesters are larger than those used in Japan. In 1989, use of nitrogen fertilizer was 154 kg/ha in Japan versus 53 kg/ha in the United States, while 69% of Japanese arable land was irrigated compared with 10% in the United States (FAO, l991a,b). Yet the Japanese investment in technological capital does not make up for the shortage of natural capital. The productivity of Japanese farmers is only 14.8 million kcal/worker/year compared with 386.1 million kcal/worker/year for U.S. farmers. Again, shortage of arable land per agricultural worker (0.99 ha in Japan versus 64 ha in the United States) accounts for the significant difference in agricultural performance.

Energy efficiency of agriculture

The internal and external pressure on society also affect the energy efficiency of agriculture. An increase in GNP and/or higher demographic pressure increase the need for commercial energy in food production. The consumption of fossil fuel energy per hectare and per agricultural worker is shown in Figure 11.

Countries such as Japan and the Netherlands, with a high GNP per capita and high demographic pressure, have a high consumption of fossil energy both per hectare and per worker. Countries having a high GNP per capita but a relatively low demographic pressure, such as the United States, Canada, and Australia, have a high consumption of fossil energy per farmer (to achieve a high labor productivity) but relatively low energy consumption per hectare of arable land. The opposite is true for countries with a high population density and low per capita incomes, such as China and Egypt, that basically use fossil energy inputs to boost the productivity of land.

The output/input energy ratio, calculated as the quantity of food energy produced (only crops, no livestock) per unit of fossil energy consumed in its production, is plotted against the productivity of agricultural labor for the 20 selected countries in Figure 12. This graph shows that the 20 countries are roughly divided into four groups. (i) In the United States, Canada, and Australia, where the demographic pressure is low but the economic pressure is high, the energy output/input ratio is low (between 1 and 2) because of fossil energy subsidies required to keep the labor productivity high. (ii) In Bangladesh, China, and India, where the demographic pressure is high and the economic pressure low, the energy output/input ratio is greater than 1 (around 3). The more severe is the arable land shortage (e.g. Egypt) the lower the energy output/input ratio; (iii) In Japan, the Netherlands, Italy, and Germany, where the demographic pressure and the economic pressure are both high, the energy output/input energy ratio is less than 1. In these countries, more fossil energy is expended in agriculture than is produced in the form of crops. The need to obtain a high density of agricultural flows per unit of land and per unit of labor dramatically increases the required technological inputs. (iv) Furthermore, where the economic pressure is low and the exploited ecosystem rich, there is no need to boost the agricultural productivity of either labor or land. This situation is experienced in some African countries located in the humid tropics which have extremely high energy output/input ratios. However, their energy output/input ratios of about 30/1 or higher are not a sign of exceptional efficiency in agricultural production, but a sign of poor economic performance, that is, the energy ratios are high because the crop output per unit of labor is low. Put another way, a high energy efficiency in terms of output/input in agricultural production spells poverty for the farmers.

In conclusion, in countries with a high standard of living, an increase in population density results in a dramatic decrease of the output/input energy ratio in agriculture (Fig. 13). The less arable land available per capita in these countries, the higher the consumption of fossil energy to maintain food production. Thus, the denser developed countries are populated, the more they will depend on fossil energy inputs to bolster their food security.

Increased pressure on society forces modern agricultural techniques to increasingly rely on fossil energy inputs in order to obtain a high labor productivity and/or high yields. Thus, modern agriculture faces a continuous decline in the energy output/input ratio and an increased stress on the environment. Both negatively affect the sustainability of agricultural production in the future.

The international market: implications for agriculture

The same quantity of agricultural product traded on the international market, such as 1 kg of rice, 1 kg of shrimp or 1 cubic meter of timber, can have a widely different labor cost depending on where it was produced or harvested. Indeed, when farmers from the United States interact, via international trade, with farmers of less developed countries they face a difficult situation. Because of the need to obtain a high income that matches with the average standard of living in their society, farmers in the United States and other developed countries that operate on the international market must either (i) be able to generate a much higher biophysical output per hour of labor than farmers in developing countries; or (ii) obtain a higher added value for the same product. That is, the price must be higher than that on the international market through a system of tariffs and/or subsidies from their government.

For example, in 1990 the average economic productivity of labor was $25/h in the United States and $0.15/h in Burundi. In that same year, with the price of rice approximately $300/metric ton, 100 kg of rice corresponded to about 1.2 hours of labor in the United States and 200 hours of labor in Burundi. Hence, to be competitive U.S. farmers have to produce 167 times more rice per hour of labor then farmers in Burundi, or receive a much higher price per kg of produced rice. In practice, farmers in developed countries do indeed have a very high labor productivity, but additional measures in the form of government interventions are needed to protect farmers' income. Those measures become more important when the endowment of arable land and technology per farmer is small.

The ability to achieve a labor productivity that enables farmers in developed countries to compete on the international market depends on: (i) the availability of natural capital per farmer (such as arable land and water supply) that is continually shrinking because of population growth; (ii) the cost of technological capital used in agricultural production; and (iii) the gradient in standard of living among trading countries. When shortages in natural capital exist, especially in situations where the difference in standard of living (GNP) between competing countries is large, farmers in developed countries generally need government intervention in the form of tariffs and subsidies. For instance, the European Community, that suffers severe land shortages and therefore has limited potential to increase farmers' labor productivity, protects its farmers against international competition by guaranteeing prices that are higher than those on the international market. In recent years, this cost the European Community 37.3 billion ECU to taxpayers and 57 billion ECU from higher prices to consumers (average for 1987-1989). Whereas the same protection, in the same period, cost the United States 42.4 billion ECU to taxpayers and 22.7 billion ECU from higher prices to consumers (Koester, 1991). Without protection through tariffs, subsidies and international trade agreements, the number of farmers in developed countries would steadily decline, as would the number of workers employed in other economic sectors dealing with the direct exploitation of natural resources and processes. A cross-country comparison of developed countries highlights the fact that when economic pressure is high, income disparity between agricultural and nonagricultural labor is not related to price support, but is related to the availability of natural capital and the GNP of a country (Figure 14, after Koester, 1991).

Environmental costs of high labor productivity to agriculture

The high labor productivity of farmers in developed countries, such as the United States, has three major costs:

(i) The average area cropped per farmer is high, such that few farmers can be supported per unit of area. For instance, the same amount of arable land that supports 1 American farmer would support 320 Chinese farmers.

(ii) The investment of resources per farmer is high. For instance, the energy intensity of labor in U.S. corn production falls in the same category as that of U.S. manufacturing industries with 1,500 kWh per work-hour or more (Stanhill, 1984, p. 124). Nevertheless, agriculture is among the economic sectors with the lowest per capita income. The earnings for full time male and female workers in the category 'farmers, forestry and fishing' average only 75% and 80% respectively of the earnings of 'machine operators, assemblers and inspectors' in the manufacturing sector (USBC, 1991, p. 415);

(iii) The heavy use of technological inputs causes major environmental damage, such as soil erosion, mining of groundwater, pollution from fertilizers and pesticides and loss of biodiversity (Pimentel et al., 1989; Brown, 1993). In order to keep the labor productivity of farmers high, crop production is boosted by intensive use of machinery, fertilizers, pesticides, and irrigation, even if this pushes energy flows through the agroecosystem to unsustainable levels (Giampietro et al., 1992a,b).

The high standard of living in developed countries depends on their high labor productivity. For agriculture to be economically viable, farmers must have a labor productivity that matches the average of society. On the other hand, sustainable management of natural resources, that is limiting the rate of resource extraction to reduce environmental problems, almost invariably implies low labor productivity (yield/hour) and the need to subsidize farmers' income. To have both sustainable agriculture and high farmers' income creates a heavy economic burden for the government. For instance, in the United States the Acre Reduction Program (ARP), also called "set-aside", reached record levels of over 30 million hectares protected in this way during 1983, 1987 and 1988. Under this program, the government pays farmers several billion of dollars not to farm (Ervin, 1992). In a similar program of set-aside being implemented in the European Community, the area in set-aside is smaller (less than 4 million hectares), but could rapidly increase (Buckwell, 1992). Because of the larger percentage of farmers in the labor force and the limited amount of land, the amount of money paid to the farmer per set-aside hectare in Europe is generally higher than in the United States. For example, in Germany, such an amount can be as high as $600/ha per year (Fasterding et al., 1992).

Following the same principle, in 1992, Canadian fishermen of Newfoundland were paid not to fish at an estimated yearly cost to the Canadian government of US$400 million (Brown, 1993, p. 8). Many similar policies to save the fishery sector are implemented in other developed countries.

Social implications for rural communities

A high population density, such as in the EC and Japan, and the fact that the yield of crops is limited by biophysical constraints, means that the productivity of farmers in these countries can be raised only by decreasing their number to increase the arable land per worker. Further increases in technological capital will not help European or Japanese farmers to increase their productivity.

However, the process of reducing the number of farmers to increase the availability of arable land per farmer can not continue indefinitely. For as there numbers become thin (e.g., 2% of the labor force in the United States) further reductions not only become increasingly difficult because of the problem of managing more hectares per farmer, but not desirable because: (i) farmers in the United States and Canada are already experiencing major social stress due to their extremely low density that leads to high costs of social services and transportation; (ii) farmers are needed because of their role as stewards of the rural landscape; (iii) we need to preserve the culture and values characteristic of this social group that are fundamental for the survival of every civilized country.

Conclusion: The United States, a special case

Technological capital is increased by economic growth, that is more people with a higher income per capita. Economic growth, however, has the opposite effect on natural capital. In fact, when natural capital is measured on a per capita basis, an increase in population decreases the availability of natural resources. When the world was "empty", strengthening the economy was a desirable strategy because natural resources were abundant and importing natural resources at low cost was possible. Times have changed however, and the world is now full. Environmental resources are being used at an unprecedented rate. Heavy dependence on imports of natural resources (either oil or food) has become risky in an increasingly unstable world full of people competing for the shrinking endowment of natural capital.

The United States, along with Canada and Australia, are among the few countries in the world that still enjoy a fair amount of arable land per farmer and natural capital per citizen. Therefore, the United States can still afford to make rational choices regarding the use of their natural resources in relation to food supply. The future development of agriculture must be directed toward eco-compatible, yet economically viable solutions. This would include set-aside of marginal land and the use of low input agriculture based on rotation and fallow. It is important, however, to keep the productivity of farmers' labor high enough to avoid a heavy burden on the economy in the form of government subsidies and unstable agriculture. Furthermore, heavy dependence on imported oil should be avoided because of national security risk.

The feasibility of achieving self-sufficiency in food production depends on the trends in population growth. The current progression of the suicidal path of rapid population growth foretells a disaster. To avoid finding ourselves in a "no return" situation regarding the problem of food production as in China and Western Europe, the American public should debate and understand the issues of immigration and rapid population growth and their implications for future food security in their country now while there is still time to make adjustments. Self-sufficiency in food production and other basic resources should be viewed as a strategy to guarantee a continued high standard of living and national security to U.S. citizens in the face of turbulence that can be expected around the world in the next decades. There is no time for delay, choosing not to change the current pattern of high immigration and rapid population growth, means moving into the Malthusian trap in the United States.

III. Ecological constraints to food production in the United States

Introduction

In a recent article, John Bongaarts (1994) reports that it is difficult to obtain an accurate indication about the number of people that can be fed from a definite amount of land, since scientists working in different fields provide widely different estimates. Some agronomists and many economists generally see no problem in feeding 10 billion people on our planet (Waggoner/CAST, 1994), whereas ecologists argue that the current population is already too numerous given the environmental resources available to support human life (Ehrlich et al., 1993; Pimentel et al., 1994b; Vitousek et al., 1986).

Regarding these different outlooks about the possibility to feed the future human population, economists and ecologists are simply saying different things. What is considered technologically feasible by agronomists and economists, that is maintaining or improving current yields per hectare in the coming few decades by relying more on technology, fossil energy, soil degradation, and depletion of underground water reservoirs, is not sustainable in ecological terms on a long time scale of centuries. Ecologists do not deny that it is possible to transfer agricultural technologies that are presently in use in developed countries, to developing countries in Africa and Latin America to temporarily increase land productivity. What ecologists do say is that these, so called, technological fixes are not sustainable in the long run because they are (i) not ecologically compatible with the earth's resources; and (ii) are based on the depletion of fossil energy stocks, which are finite.

For an accurate picture of the ecological consequences already induced by intensive agriculture, we refer the reader to a report by the Dutch National Institute of Public Health and Environmental Protection (RIVM, 1992), that documents the environmental impact of agriculture in the Netherlands (e.g. pollution of the water table, spread of pesticides, destruction of natural habitats, loss of biodiversity). The Netherlands currently is desperately trying to reduce the serious pollution problem associated with agriculture.

To the best of our knowledge none of the technological optimists has dealt realistically with the ecological sustainability of food production. In general, the biophysical constraints, such as soil quality, factors responsible for the stability of the water cycle, role of specific biota and biodiversity in maintaining the long-term productivity of agroecosystems, stability of nutrient cycles, and off-site effects of pollution are simply ignored (Waggoner/CAST, 1994).

In fact, several biophysical parameters limit the productivity of land. These include soil structure and composition, that determine the ability of soil to retain water and nutrients, water supply, nutrient supply, biodiversity at the regional scale, land slope, and solar irradiation. For instance, the Australian continent appears virtually 'empty' when its population density is compared with that of other countries (see Table 2), yet several Australian scientists consider it to be already 'full' when measures of water resources, agricultural yields and ecological sustainability are taken into account (Stone, 1994).

Current levels of agricultural productivity are non-sustainable if they do not prevent environmental degradation and/or if they heavily depend on depletion of stocks of fossil energy and/or underground water. In this section, the current situation of U.S. agriculture in relation to its ecological sustainability is examined.

Food supply in the United States

Current estimates of U.S. food supply

According to FAO Food Balance Sheets (199lc) the per capita food energy available for consumption in the United States in 1989 was 3,595 kcal/day of which 66% from plant products and 33% from animal products. Direct per capita consumption of several food groups is listed in Table 6 (USBC, 1992).

The direct consumption of food by item tells only part of the story. For instance, according to FAO (199lc) the cereal grains consumed directly per capita are just a small fraction of the total per capita cereal grains consumption (directly and indirectly) in the United States. In fact, of the total domestic consumption of cereal grains 72% are used to feed livestock, 11% are for direct human consumption, and the remaining 17% are used by the food industry to produce different food products and alcoholic beverages. Therefore, almost 90% of the cereal grains are consumed indirectly by Americans. A similar pattern occurs for soybeans and oil seeds. A large fraction of soybeans is used for feeding livestock, either directly or in the form of by-products (bean meal) of soy oil production, and in the food industry to produce soy oil for human consumption.

To obtain an assessment of the total U.S. food demand the total internal demand was considered, including feed for livestock and raw food materials for the food industry, for the 239 million American citizens in 1989, as reported by FAO (199lc). On the basis of the assessment of the consumption per capita, the aggregated demand was calculated for a U.S. population of 260 million, that will probably be reached at the end of 1994. In this way, we obtained the following data for the total food demand: 222,600 million kg of cereal; 19,400 million kg of starchy roots; 51,200 million kg of sugar crops; 884 million kg of pulses; 30,000 million kg of oil crops; 28,600 million kg of vegetables; 34,560 million kg of fruit; 29,150 million kg of meat; 71,600 million kg of milk; and 4,400 million kg of eggs.

Land requirement of current food supply

In order to assess the amount of land needed to supply the food directly and indirectly consumed by the U.S. population, the land use in the United States must be assessed.

According to the U.S. Department of Agriculture, in 1989 the United States had available for agriculture: 239.5 million ha of grassland and pasture and 188 million ha of arable land. Of this arable land, almost 15% (27.5 million ha) was left idle, 22% (41.5 million ha) was used for export food crops (basically cereal grains and soybeans), and 3% (5 million ha) was used for growing non-edible crops such as cotton and tobacco, either for export or domestic use. The remaining 60% of the arable land (144 million ha) was used for the production of domestic food crops.

To assess the requirement of land for food supply we need to take into account the arable land used to produce food crops for export (to be subtracted) and the land used elsewhere to produce imported food crops (to be added). The trade balance for animal products is not considered here because the value of animal products exported ($6.5 billion) is roughly equal to the value of animal products imported ($5 billion). The major food crops involved in U.S. international trade are cereal grains, with a net export of 106 million metric tons or 37% of the total U.S. production, and soybeans, with a net export of 20.4 million metric tons or 38% of total U.S. production. The area of arable land used to produce this flow of food exports (corrected for imports) is about 28 million ha for cereal grains and 12 million ha for soybeans. We further reduced this area of 40 million ha by 25% to compensate for the arable land required to produce imported vegetables ($10 billion import vs $33 billion export). Therefore, considering the difference between export and import of food crops, an estimated 30 million ha of arable land is cultivated for the production of a net flow of export food crops, this is equivalent to about 16% of the total arable land in the United States.

Thus, with 188 million ha of arable land available and subtracting 15% for the hectarage left idle, 5% for the cultivation of non-food crops and 16% for the production of net food exports, we arrive at an estimate of about 120 million ha of arable land used in producing the domestic demand of food and feed crops in 1989 for 239 million Americans. Almost 85% of these 120 million ha of arable land was used to cultivate only 4 different crop groups, including cereal grains (40%), oil seed crops (12%), hop (9%), and hay, such as alfalfa (22%).

Note these figures refer to arable land that is planted as reported by USDA agricultural statistics and therefore are slightly higher than figures referring to arable land harvested, such as reported by most FAO statistics. For instance, the area planted to cereal grains in the United States in 1989 was 76 million ha, whereas the area of cereal grains harvested in that year was only 63.5 million ha. Climatic conditions, such as local shortage or excess of water supply, economic reasons, and/or other factors account for sensible differences between planted and harvested area, especially for cereal grains.

Adding the 120 million ha of arable land used for food and feed production for domestic use to the 239.5 million ha of grassland and pasture, totals 360 million ha of land used to supply the domestic demand for plant and animal food products in 1989. Based on 1989 production and processing technology, the per capita land requirement in 1989 was 0.5 ha of arable land and 1 ha of grassland and pasture for a high quality diet. Such a diet includes a variety of vegetables and fruits, a large supply of animal protein, and in general a diversity of other foods. This finding agrees with the estimate of Lal (1989) that about 0.5 ha of cropland per capita is needed to provide a nutritious diet of plant and animal products.

Land for future food supply

Different scenarios for future food supply in the United States are discussed in detail in Part IV of this report. We note here, however, that according to the estimate above, when the U.S. population reaches 520 million in the year 2050, as projected by USBC (1992), the need for land to support food production will reach 210 million ha of arable land and 520 million ha of grassland and pasture. Since only 188 million ha of arable land and 239 million ha of grassland and pasture are currently available in the United States, a doubling of the U.S. population would certainly lead to a biophysical constraint to the food supply in terms of arable land. This constraint will remain even if: (i) all available arable land were to be used for domestic food supply, eliminating crop exports and the production of non-food crops such as cotton and tobacco; (ii) in the year 2050, farmers would continue to have the same access to fossil energy inputs as are available today; and (iii) no further land degradation would occur during the next 50 years.

In the future, shortage of land for food production in the United States has negative implications not only for domestic food security but also for the depletion of environmental resources. First, to further increase the productivity of land in agriculture (yields per hectare) where possible, more technological inputs, such as fertilizers, pesticides, and irrigation will be used. This will, however, further increase adverse ecological impacts and dependence on oil imports, and actually may speed the loss of arable land due to soil erosion and other forms of degradation. Second, deforestation and use of marginal lands to expand the area of arable land and pasture lands for food production will increase loss of biodiversity through the destruction of habitats. Note marginal lands are more susceptible to negative environmental effects of intensive agriculture. Third, the export volume of agricultural products will decrease because of the increased domestic needs for these products. This may be particularly critical for the U.S. economic balance, since this trend is likely to occur at a time in which food products will become a valuable commodity because of worldwide scarcity.

On the positive side, there is some room to reduce the current per capita food demand while maintaining a varied diet. Indeed, the large amount of animal products and alcoholic beverages currently consumed in the United States reflects the indirect consumption of large amounts of grains in the form of livestock feed and raw materials, such as hop, for alcoholic beverages and provides some buffer for the projected increased food demand of 520 million Americans. Each American citizen consumes per year, on average, 30 kg of bovine meat, 20 kg of pork, 30 kg of poultry, 260 kg of milk, 16 kg of eggs, and 134 liters of beers. Comparing this level of consumption with that of other developed countries (e.g. Europe or Japan, see Table 5), it is reasonable to assume that the consumption of animal products and alcohol can be significantly reduced without adversely affecting the nutritional quality of the American diet. This change in consumption pattern would encourage the direct consumption of grains and legumes.

In fact, replacing animal products and alcohol in the diet with grains and legumes would free some arable land now needed for livestock feed and hop production (14 million ha were planted to hop in 1989, 75% of which for domestic consumption) and reduce the need for pasture and grass lands. This would reduce the amount of land needed for the per capita food supply. Such a change is important as reduced per capita availability of grassland and pastures due to population growth (from l ha per capita in 1989 to less than 0.5 ha per capita for a doubled population in 2050) will increase the dependence of livestock on arable land for production of feed grains. Equivalent feeding values of corn per animal product, adapted from data provided by the U.S. Department of Agriculture (1990, Table. 76, p. 57), are listed in Table 7. :These figures refer to the quantity of animal products consumed and do not include the maintenance of stocks required to guarantee a continuous flow of slaughtered animals.

The more the U.S. population grows, the more the option of a low-intensity use of land, such as beef production on range lands, will be reduced for U.S. agriculture. Beef lots coupled to intensive farming for livestock feed will become a forced choice to continue to provide animal products to 520 million Americans. In this way, the per capita supply of animal protein in the United States will be strongly dependent on the supply of oil, in the same way that today the supply of plant protein in China is linked to that of coal to produce nitrogen fertilizer.

Ecological compatibility of current U.S. agricultural production

Regarding the possibility of expanding and intensifying agricultural production in the United States to increase its future food supply, one should first answer the fundamental question of whether current agricultural production, the baseline for future projections, is even now sustainable according to ecological constraints. Put another way, can the intensity and acreage of U.S. agriculture be increased or should it be decreased to guarantee the long-term sustainability of food production?

In this section we present a list of checks on the eco-compatibility of current U.S. agriculture.

Impact of agriculture on soil and land

Over the last 200 years of farming, the United States has abandoned an estimated 100 million hectares (about 30%) of farmland because of erosion, salinization, and waterlogging, and the soil degradation problem appears to be worsening (USDA, 1971; 1989). Croplands lose an average of 17 t/ha/yr of soil to water and wind erosion combined, but in some states such as Iowa, erosion rates average 30/t/ha/yr. Pastures lose on average 6 t/ha/yr. About 90% of U.S. cropland is losing soil above the sustainable rate of 1/t/ha/year, and about 54% of U.S. pasture land (including federal land) is overgrazed and subject to accelerated erosion (Pimentel et al., 1994c)

In the United States, soil erosion losses, compounded by degradation caused by salinization and waterlogging, cause the abandonment of nearly 1 million ha of cropland each year. Even the most valuable soils are rapidly being degraded. For instance, Iowa, which has some of the best soils in the world, has lost one-half of its top soil after little more than 150 years of farming and continues to lose topsoil at an alarming rate of about 30 t/ha/yr (30 times faster than the rate of soil formation). A similar situation exists in the rich Palouse soils of the Northwest where about 40% of the soil has been lost in the past century. The majority of soil erosion on U.S. cropland, about 60%, is due to rainfall and water run-off, but for arid states wind erosion is the major cause (Pimentel et al. 1994c).

About 21 million hectares of U.S. cropland are considered highly erodible and the only way to halt erosion would be to convert the land to a. use which allows for permanent vegetative cover, like managed pasture.

Effects of erosion on agricultural productivity.

Crop yields on severely eroded soil are lower that those on protected soils because erosion reduces soil fertility and water availability. Corn yields on severely eroded soils are reduced by 12 to 21% in Kentucky, up to 24% in Illinois, 25 to 65% in the Southern Piedmont, and 21% in Michigan (Pimentel et al., 1994c). Erosion by water and wind adversely affects soil quality and productivity by reducing infiltration, water-holding capacity, nutrients, organic matter, soil biota, and soil depth.

Onsite economic costs. When erosion by water and wind occurs at a rate of 17 t/ha/yr, then each hectare annually loses approximately 100 mm/ha of water and 462 kg of nutrients. If water were to be replaced, it would cost the United States about $40/ha/yr to pump groundwater for irrigation, assuming water were available. An additional $100 per hectare would be needed to replace lost nutrients with fertilizers (Troeh et al. 1991). The billions of tons of soil and water lost annually from U.S. cropland translate into an on site economic loss of approximately $28 billion each year (Pimentel et al., 1994c).

Off site economic costs. Agricultural soil erosion also leads to extensive damage throughout the surrounding environment. These Off site costs include: roadway, sewer, and basement siltation; drainage disruption; foundation and pavement undermining; gullying of roads; earth dam failures; eutrophication of waterways; siltation of harbors and channels, loss of reservoir storage; loss of wildlife habitat and disruption of stream ecology; flooding; damage to public health; and increased water treatment costs. The total Off site costs of soil erosion have been estimated to total about $17 billion per year, in 1992 dollars (Pimentel et al., 1994c)

Total on site and Off site costs. The combined on site and Off site costs of erosion from agriculture in the United States are estimated at $45 billion per year or about $100/ha of cropland and pastureland (Pimentel et al., 1994c). This erosion cost increases production costs in U.S. agriculture by about 25% per year.

Energy costs of soil erosion. To compensate for the on site and Off site damage inflicted by soil erosion and associated rapid water runoff on agricultural production, an additional 1.8 million kcal of fossil energy per hectare are expended each year (Pimentel et al., 1994c), assuming an average erosion rate of 17 t/ha/yr. This means that approximately 10% of all the energy used in U.S. agriculture today is expended just to offset losses of soil nutrients, water and crop productivity caused by erosion (Pimentel et al., 1994c).

Effect of urbanization on arable land

Owing to the rapid spread of urbanization and highway networks an area equivalent to Ohio and Pennsylvania was black-topped over in the 30 year period between 1945 and 1978. The build-up area in the United States in 1989 was 99 million hectare (WRI, 1994). Almost half of the land that has been taken for housing and highways was the most productive U.S. agricultural land. Clearly, this is a serious loss of natural resources.

Constraints to water supply

Few plants send roots deeper than a meter and therefore can not reach groundwater resources and must depend on the water held in the top layers of soil. This moisture is provided either by rainfall or by pumping water from groundwater supplies. Groundwater is referred to as fossil water because it accumulates in aquifers deep below the surface and is replenished only very slowly. Less than 0.1% of the stored groundwater that is mined annually by pumping is replaced by rainfall. In the United States, surface water supplies about 60 % of the water used in irrigation and the remainder is from groundwater supplies (Gleick, 1993).

Each individual requires nearly 3 liter of fresh water per day for drinking, but uses at least 90 1/day for cooking, washing, and other domestic energy-related needs (Brewster, 1987). On average, each American uses about 400 liters/day for all domestic needs (Kendall and Pimentel, 1994). However, when the water consumption related to agricultural production and industrial uses is also included, each American uses a total of more 5,500 1/day of fresh water (Kendall and Pimentel, 1994).

The water used by agriculture accounts for as much as 85 % of all fresh water 'consumed' in the United States (Poster, 1989), because all crops require large amounts of water. For example, a corn crop that produces about 7,000 kg/ha of grain will take up and transpire about 4.2 million liters/ha of water during the growing season (Leyton, 1983). To supply this much water to the crop, not only must 10 million liters of rain fall per hectare, but a significant portion must fall during the growing season. Thus, the production of 1 kg of corn can require over 1,400 liters of water.

Overall water supply of the United States is better than that of many other countries, but is being stressed in southern and western areas (Pimentel et al., 1994b). For instance, the average annual decline of water tables is assessed in the range of 0.6-0.9 m in Arizona; 0.15-1.05 m in California, 0.75 m in Florida, and 0.3-1.2 m in Texas (U.S. Geological Survey, 1987). By the time the Colorado River enters Mexico it has literally disappeared because of the excessive removal of its water by the States of California, Arizona, and Colorado. Countless disputes between farmers and local communities over the appropriation of fresh water now are being reported in Texas. The annual overdraft of the great Ogallala aquifer located in the Great Plains is 130% to 160% above replacement. If this rate of overdraft continues, this vast aquifer is expected to become non-productive in about 40 years. The Southwest has only 6% of the country's available water as rainfall, but its large irrigated farms and growing metropolitan areas account for 36% of the nation's water use. California, for example, actually consumes more water than falls on the state in the form of precipitation (USGS National Water Summary, 1987; U.S. Census, 1990).

Another major threat to maintaining ample fresh water resources is pollution. The U.S. Environmental Protection Agency has found 98 different pesticides, including DDT, in groundwater in 40 states, contaminating the drinking water of over 10 million residents. Other types of pollution, such as nitrate nitrogen, also are reported in which agriculture is a major contributor (USBC, 1991).

What is a sustainable yield?

Current agricultural productivity worldwide is dependent on the depletion of stocks of resources, that is geological deposits of fossil energy and minerals, for the supply of technological inputs such as fertilizers, machinery, pesticides, and irrigation. Furthermore, intensive agricultural production has adverse ecological impacts that translate into large dollar costs and reduced long-term productivity.

It is therefore pertinent to determine what can be considered a sustainable yield if agricultural production is to conserve the long-term productivity of its soils and fare without relying on added fertilizer nutrients. In a review of available data on agricultural production that does not rely on fertilizer nutrients, Bender (1993) argues that such agriculture would result in lower yields than those commonly associated with organic farming. In fact, almost all organic farms add to the system one or more external nutrient sources, such as starter fertilizer, animal feed, or manure.

Bender (1993) provides the following yields: (i) for continuous grain cropping with no-rotation, 500 kg/ha in semiarid regions and 1000 kg/ha in humid regions; (ii) for grain cropping with rotation and using only green manure,.about 1,400 kg/ha; and (iii) for grain cropping with rotation and using animal manure about 2,000 kg/ha. To put these figures in perspective, for the United States in 1989, the average yield of cereal grains was 4,470 kg/ha and the average yield of corn under intensive cultivation (heavy use of fertilizer, irrigation, and pesticides) was 7,000 kg/ha.

Bender reports similar low yields for dry land production of wheat (900 and 1,000 kg/ha) and rice (1,200 and 1,400 kg/ha) in respectively, Japan and China. The only imported nutrients sources used to maintain productivity were urbanites' excrete, street refuse, canal mud, wood ash and wild plants.

If agricultural communities export the larger part of their harvested crops to urban areas elsewhere, it will be impossible to make up the resulting nutrient deficit in the soil, no matter how much care is taken in recycling nutrients in the rural community.

Annually 62 billion kg of nitrogen is taken from the atmosphere and converted mainly by biological fixation into forms of nitrogen that can be used by plants. In addition, the amount of nitrogen made available to plants in the form of fertilizer produced by fossil energy-based technology, is nearly equal to the total amount fixed annually by natural means, that is an additional 60 billion kg. Thus, human technology is able to generate a flow of nitrogen that would require an area nearly the size of the entire earth if it were fixed by natural processes. In this way, the heavy use of commercial fertilizers is boosting the productivity of land well beyond the constraints imposed by the biogeochemical cycle of nitrogen fixation.

However, by adopting a system of legume-fallow production, from 60-130 kg/ha of nitrogen can be added to the land during one year of fallow (Pimentel et al., 1989). This can replace the use of commercial fertilizer but doubles the land required for production (half of which in fallow). In addition, when crops produced are consumed elsewhere, alternative agricultural practices can not completely make up for a massive withdrawal of nutrients from the soil unless a large supply of arable land is available to leave the land fallow enabling natural process to fix nitrogen from the atmosphere.

Besides the availability of land, labor constraints also affect the feasibility of recycling nutrients at the farm level. For instance, most developed countries are faced with the excessive accumulation of manure resulting from intensive husbandry. Obviously, processing, transporting and spreading manure on cropland would be the best solution in ecological terms to reduce the economic cost of fertilizer and reduce pollution from the accumulation of manure. However, labor requirements for affluent manure management are high and do not make this solution in developed countries economically viable. Indeed, Chinese farmers, masters in recycling nutrients on their cropland, pay for this achievement in the form of a very low productivity per hour of farm labor.

Furthermore, in the United States and most other developed countries, the overwhelming effects of the economic pressure on the agricultural sector makes rural communities small and scattered with the bulk of the population concentrated in urban areas. Under these conditions, agricultural production is exported from rural to urban areas elsewhere and cropland will therefore always need a certain amount of imported nutrients to make up for those withdrawn by the crops, no matter what alternative techniques of farming are used. Recycling as much manure and other agricultural byproducts as possible has to be considered a major goal of agricultural techniques. However, as agricultural production becomes more specialized and separated from the areas of consumption, nutrient recycling becomes more difficult and expensive in terms of energy and labor.

Dependence of food supply on fossil energy

At present, the food supply of all developed countries is dependent on fossil energy. For instance, the output/input energy ratio of U.S. agricultural crops is 1.4 (Table 4). Thus, 0.7 kcal of fossil energy are consumed in the U.S. agricultural sector to produce 1 kcal of crop. However, the inputs used for the calculation of the energy output/input ratios listed in Table 4 are based on FAO statistics and include only fertilizers, irrigation, pesticides, machinery and fuel for field operations. This totals 850,000 billion kcal spent in U.S. agriculture in 1989. Other energy inputs are required in the agricultural sector, such as energy and machinery for drying crops, transportation of inputs and outputs to and from the farm, electricity, and construction and maintenance of buildings and infrastructures. When these inputs are also included, the total commercial energy used in U.S. agriculture well exceeds 1,000,000 billion kcal (=10¹⁵ kcal or about 5% of the total consumption of fossil energy in the United States) and brings the output/input energy ratio close to 1.

Based on this estimate, the food supply in the United States requires the use of 0.4 metric ton of oil equivalent per capita/year, or about 530 liters of oil equivalent, just for the agricultural sector. However, this is just a fraction of the total fossil energy expended in the entire food system. Including food processing, packaging, and distribution increases the commercial energy consumption by more than 3 times. Finally, also energy use related to shopping and home preparation of food, that in official statistics appears under the heading "residential energy use", contributes to fossil energy use in the food system. Taking into account all these energy uses related to the food system, the percentage of total U.S. energy consumption expended in the food system amounts to 17%.

IV. An outlook on food security in the 21st century: the United States and world

Grain supplies

Of the 183 nations in the world only few are net exporters of grains In 1989, the United States, Canada, Australia and Argentina supplied more than 81% of the net cereal grains export worldwide (WRI, 1994). The United States alone provided 54% of this. Export crops, including non-food crops, use about 25% of the current arable land in the United States. The dollar value of the total export of U.S. agriculture, in 1989, was nearly $40 billion, of which $3.3 billion were obtained from non food crops. The dollar value of total agricultural imports was $21.5 billion, leaving a net positive difference of $18 billion in 1989. Of the imported crops, $15.2 billion were spent on competitive crops, that is crops that could be produced in the United States, whereas $6.2 billion were for non-competitive imports, such as coffee, tropical fruits, etc..

The outlook for world food security is gloomy at best. Green revolution technologies have not been able to keep up improvements in yields with population growth. Per capita food production in Africa is down 12% since 1981 and down 22% since 1967 (Kendall and Pimentel, 1994). A large number of developing countries depend for their food security on imports of food (mainly grains) and/or fossil energy for fertilizers and other inputs in agriculture. This dependence on the international market makes developing countries vulnerable for both internal and external fluctuations in the food supply.

For instance, countries with a large population, such as P.R. China or India, can not rely on the international market to supply unexpected grain shortfalls. If, for example, the domestic supply of Chinese rice (173 million metric tons) had dropped by 10% in 1989 due to climatological factors or internal social unrest, the resulting need for 17 million metric tons of rice could not be met on the international market. In fact, in that year, the major net rice exporters were Asia, with a net export of 2.5 million tons, and Canada and the United States, with a net export of 2 million metric tons (USDA, 1992).

Indeed, communist China, under Mao Tze T'ung, experienced a serious famine which ran from 1958 through much of the 1960's. In previous decades, Marxists have often blamed the capitalist way of production for famines occurring throughout the world and denied the role of biophysical constraints in limiting food production. However, the worst famine recorded in recent history was experienced in Marxist China at a time of complete isolation from external influences. There, with a situation of total self-reliance chosen for ideological reasons that eliminated the option of import, the significance of land shortages and the need to control the population size was made crystal clear to the Chinese government (Abernethy, 1993).

Fossil energy

In the last decades, technological progress has been able to boost the productivity of agriculture due to the injections of fossil energy subsidies (Pimentel et al., 1990). Odum E.P. (1971, p. 46-47) effectively described these technological changes in agriculture as follows: "Those who think that we can upgrade the agricultural production of the so-called 'undeveloped countries' simply by sending seeds and a few 'agricultural advisors' are tragically naive! Crops highly selected for industrialized agriculture must be accompanied by the fuel subsidies to which they are adapted!" These fuel subsidies are for fertilizers, pesticides, and irrigation.

What are the implications of this huge dependence of many developing countries on oil for their food supply and, hence, survival? The International Institute for Applied Systems Analysis (IIASA, 1981) has predicted that if the world population doubles and developing countries increase their use of fossil energy, then fossil energy use worldwide will increase 2 to 3 times above the 1980 level by the year 2030. Starr et al. (1992) provide a similar estimate when they project that the global energy demand will increase to about 4 times the present level by the middle of the next century. Assuming that these estimates are accurate, the depletion of the world's non-renewable fossil energy resources will occur much more rapidly than is currently projected on the basis of the highly unlikely assumption that the current rate of consumption will remain constant in the future.

U.S. oil production has declined 400,000 to 500,000 barrels per year for the past two decades (USC, 1990). Currently the United States imports 58% of its oil and this percentage is projected to rise to 60-70% by the turn of the century (Gibbson and Blair, 1991). With U.S. oil reserves of only 15 to 20 years, the United States will enter the 21st century captive to a growing oil import bill. With an increasingly volatile oil market, due to instability in the Persian Gulf and the former Soviet Union combined with an 170% increase in the demand for oil in developing countries by the year 2010, U.S. dependence on oil becomes an ever-worsening strategy with regard to national security. The United States currently spends about $65 billion per year on oil imports plus an additional $50 billion to protect U.S. interests in the oil-rich Gulf.

Other developed countries are not better off. For instance, the food supply of the European Community and Japan is totally dependent not only on imports of food and livestock feed but large amounts of fossil energy. For instance, through fodder imports the Netherlands indirectly uses an area of agricultural land that is five to seven times larger than that within the country itself (RIVM, 1991, p. 108). Dutch agriculture, as well as the agriculture of Japan and other EC countries, can not provide a sufficient food supply for its population without importing both feed and fossil energy. The import of which is likely to be subject to large fluctuations on the international market due to political instability expected in the next century. Also the Chinese food supply has become entirely dependent on oil; "Half of all peasants in Southern China are alive because of the urea cast or ladled onto tiny fields-and very few of their children could be born and survive without spreading more of it in the years and decades ahead" (Smil, 1991, p. 593).

The increased instability of local governments due to population pressure and resource scarcity is beginning to destabilize social order. Civil war in Somalia, Burundi, ex-Yugoslavia, and the widespread unrest in many ex-Soviet Union Republics are examples of this. Western civilization is raising expectations and building up the internal pressure in many developing societies at the very same moment in which the external pressure, in the form of shortage of natural resources per capita, on society is increasing. The final result of which spells little good for the international order.

Another formidable risk in the global struggle for survival is that ecological limits as they affect the long-term sustainability of agroecosystems will be easily forgotten to cope with the emergencies of the moment. This is the reason why the United States, that still is in the privileged situation of having a little more time to reflect on its future development must face the implications of population growth and take action.

Agriculture is an activity in which technological capital has little room for substituting for shrinking natural resources (land, water, soil, biota). Moreover, if the price of oil rises as oil supplies become scarce, what will be the reaction of the poorest countries which depend on cheap oil for their survival? If food exporter countries will stop exporting because of increased internal demand, what will be the reaction of the many countries heavily dependent on the import of food and livestock feed for their survival? Will the environment be the big looser of the simultaneous increase of internal and external pressure in societies all over the world? In that case, what about the quality of life for future generations?

Different scenarios for land, population and food supply in the United States

Extrapolating the current situation into the future

The per capita requirement for land to supply U.S. food, based on the current pattern of crop production and technology, is 0.5 ha of arable land and I ha of grass and pastureland. The current amount of land in use for food production is 188 million ha of arable land and 240 million ha of grass and pastureland.

By extrapolating the present situation into the future, without considering the possibility of reductions in arable land and/or yields due to land degradation, shortage of fresh water supplies, shortage of oil supplies, or increase in oil prices, we find the maximum population size that could be fed is 350 million. This assumes, current technology and land supply, current diet composition and no major changes in structure and function of the agricultural sector. Yet, with this population size, the need for grassland will exceed the supply and therefore more grain will be required to maintain the current levels of beef production.

When the spread of urbanization, roads and other commercial uses of land is taken into consideration, arable land is further reduced. Right now, 0.4 ha of U.S. land per capita are built-up, of which 0.2 ha are taken from arable land. A further increase in population and associated urban development will reduce arable land per capita by at least 0.1 ha. Based on this assumption, the previous figure of 350 million people will be lowered to 315 million.

Clearly, the scenario of a maximum population of 350 or 315 million based on an agriculture using the current level of technological subsidies, implies that no food or feed crops will be exported and that the cultivation of non-food crops, such as cotton and tobacco, will be replaced by food and feed crops. Indeed, a significant increase in domestic food needs due to population growth will absorb the food surplus currently exported.

Regarding the reliability of this scenario we have to note that progress in agricultural technologies provides some room for improvement in yields in the future. However, we are skeptical about the possibility of achieving dramatic changes in agricultural performance in the near future from technological fixes such as biotechnology. In the open field, it is the entire network of energy and matter flows that stabilizes a certain level of productivity. If hundreds of kilograms of nitrogen are taken from the soil and millions of liters of water are transpired during the growing season, those inputs must somehow be made available to the crops. Technological improvements can not continue to counteract most of the major decreases in productivity that are occurring due to soil erosion and salinization. Technology, for example, can not speed reformation of topsoil or make more fresh water.

U.S. agriculture for a population of 520 million

A doubling of the U.S. population will result in a drastic reduction in the per capita availability of land to 0.25 ha of arable land and 0.5 ha of grassland and pasture. Such a reduction will be augmented by the increased use of arable land for economic development, such as urbanization and roads, thereby lowering the supply of arable land to about 0.15 ha per capita. At present, such a problem seems less relevant on grassland and pastureland. Assuming that the United States will maintain its high GNP per capita, which implies heavy economic pressure on agricultural performance, then U.S. agriculture will experience a situation similar to that of present-day Italian agriculture, which has only 0.16 ha of arable land per capita but a high GNP (Tables 2 and 3).

In other words, a doubling of the population in the United States will reduce the supply of arable land per capita to present European levels (Table 3). As noted in the international comparison, this will bring about a dramatic decrease of the output/input energy ratio in agriculture from the current 1.4 to about 0.6 as in Italy today. This suggests nearly a 3-fold increase in the consumption of fossil energy over current levels in order to increase the productivity per hectare. In fact, the 2.6 increase in the demand of fossil energy for technological inputs will be coupled with increased energy costs of transportation because of the spatial distribution of the population in a country where huge distances separate areas of production from those of consumption.

As a result of a population doubling, agricultural production can be expected to consume at least three times as much fossil energy as at present. This and other changes will intensify adverse environmental impacts such as soil degradation, deforestation, and expansion of production on marginal lands. Moreover, it is uncertain whether such an adjustment toward a doubling of the food supply would be possible without a dramatic rearrangement of the pattern of land use in U.S. agriculture. Because of expected shortages in water supply in many areas now in production in the western and southern states, future agricultural production may have to be concentrated in areas with greater availability of water. An assessment of the environmental and economic costs of such a rearrangement are beyond the scope of this report.

Sisler (1988) has estimated on the basis of food elasticity data that, in the United States, a 1% increase in demand or 1% decrease in supply will generate a 4.5% increase in food price at the farm gate, and that if the US population will double in 2050, depending on technology, diet and land degradation, prices of food could increase 3 to 5 times compared to today.

In the year 2050 we can expect increases in the price of food all over the world. If the difference between the increase in oil price and the increase in food prices on the world market will remunerate the technological boosting of land productivity, most probably some economists and politicians will push for it. We can only hope that, at that time, shortages of water, oil, and other environmental resources both in the United States and worldwide will have convinced the general public of the importance of preserving environmental equilibria, in spite of economic pressure for increasing the return of economic investments in the short term.

Since any assessment is heavily dependent on the assumptions made by the authors, it is difficult to objectively predict the consequences of a population size of 520 million on the socio-economic structure of the United States. For instance, if future generations of American citizens are not concerned about ecological compatibility, they could decide to continue to boost agricultural yields per hectare even if they will have at that time a limited amount of arable land available (0.15 ha per capita). As a matter of fact, this is what all European countries are doing at present. Also Japan, China and Egypt, with diminishing supplies of land per capita, are degrading their soils and environments and becoming increasingly dependent for their food security on vanishing fossil energy stocks and imports.

Perhaps it will be possible to feed 520 million U.S. citizen with 0.15 ha of arable land available per capita, but only for a short period of time since such a solution would not be sustainable because of dependency on fossil energy stock depletion and lack of ecological compatibility. Clearly, the costs of that choice would be immense in ecological and energetic terms and would represent a suicidal choice for the future of the country.

As an alternative, it is also plausible that the option of intensifying agricultural production will simply not be there for Americans in 2050. This is a distinct possibility because fossil energy may have become too scarce and expensive, water resources at peril, and soils too seriously degraded to support the required level of agricultural production.

In summary, 520 million Americans represent a number that does not fit with the concepts of sustainability, ecocompatibility, and long-term self-reliance. However, if the U.S. population reaches that size, nothing will be left to be done in 2050. The vital resources that enable agriculture to provide food security will be depleted and there is no way to replenish them. It is only by acting now that the United States can prevent a future disaster.

U.S. agriculture based on ecocompatibility

In this scenario we assume that, in order to reduce dependence on imported oil, restore the fertility of soil and reduce the environmental impacts of agriculture, present U.S. agriculture will undergo sweeping changes in techniques of production that rely less heavily on fossil energy and other agricultural inputs and rely more on rotation and other fanning techniques requiring low technological inputs. The feasibility of such a change will require several major adjustments in the structure and function of the agricultural sector in the United States, especially in terms of new government policies and legislative interventions.

Based on such modifications of production techniques, crop yields would be reduced. Assuming a reduction to 65% of the current level of production (Bender, 1993), the demand of land per capita will amount to 0.8 ha of arable land and 1.5 ha of grassland. This scenario further assumes that 15% of the arable land will remain cultivated to export crops and non-food crops such as cotton, which will present an additional limit to the supply of land for food production.

According to this scenario, the maximum size of U.S. population that could be fed by such an ecocompatible agriculture would be only about 210 million. This number could be increased to reach 240 million if cultivation of export crops and other non-food crops were to be eliminated. Note, fossil energy inputs are still a required input in this system.

References

Abernethy, V. 1993. The demographic transition revisited: lessons for foreign aid and U.S. immigration policy. Ecological Economics 8 (3): 235-252.
Bender, M. 1993. Agricultural productivity in a post-fossil fuel era. The Land Report, No. 47. The Land Institute, Salina, Kansas.
Berio, A-J. 1984. The analysis of time allocation and activity patterns in nutrition and rural development planning. Food and Nutrition Bulletin 6(1): 53-68.
Bongaarts, J. 1994. Can the growing human population feed itself? Scientific American, March, 36-42.
Brewster, J.A. 1987. World Resources 1987. A Report by the International Institute for Environment and Development and World Resources Institute. Basic Books Inc., New York.
Brown, L. 1993. A new era unfolds. In: L. Brown et al. (Eds.), State of the World 1993, pp. 3-21. W.W. Norton & Company, New York.
Buckwell, A.E. 1992. Should we set aside set-aside? In: Set-aside. BCPC Monograph No. 50. Proceedings of a Symposium organized by The British Crop Protection Council held at Cambridge University l5th-18th September 1992, pp. 275-283. Farnham, UK, British Crop Protection Council.
Daly, H.E. 1992. From empty-world economics to full-world economics: Recognizing an historical turning point in economic development. In: R. Goodland, H.E. Daly and S. E1 Serafy (Eds), Population, Technology, and Lifestyle, pp. 23-37. Washington, DC: Island Press.
DOE. 1990. Annual Energy Outlook. Washington, DC: U.S. Department of Energy.
DOE. 1991. Annual Energy Outlook with Projections to 2010. Washington, DC: U.S. Department of Energy.
Ehrlich, P.R., A.H. Ehrlich, and G.C. Daily. 1993. Food security, population, and environment. Population and Development Review 19 (1): 1-32.
Ervin, D.E. 1992. Some lessons about the political-economic effects of set-aside: the United States' experience. In: Set-aside. BCPC Monograph No. 50. Proceedings of a Symposium organized by The British Crop Protection Council held at Cambridge University l5th-18th September 1992, pp. 3-12. Farnham, UK, British Crop Protection Council.
FAO. 1991a. Production Yearbook 1990 (Vol. 44). Rome: Food and Agriculture Organization of the United Nations.
FAO. 1991b. Fertilizer Yearbook 1990. Rome: Food and Agriculture Organization of the United Nations.
FAO. l991c. Food Balance Sheets. Rome: Food and Agriculture Organization of the United Nations.
Fasterding, F., Plankl, R. and Jones, A. 1992. The impact of the European Community' setaside policy in Germany. In: Set-aside. BCPC Monograph No. 50. Proceedings of a Symposium organized by The British Crop Protection Council held at Cambridge University l5th-18th September 1992, pp. 269-274. Farnham, UK, British Crop Protection Council.
Georgescu-Roegen, N. 1971. The Entropy Law and the Economic Process. Cambridge, MA: Harvard University Press.
Giampietro, M. 1994. Using hierarchy theory to explore the concept of sustainable development. Futures, 26 (6): 616-625.
Giampietro, M. and Pimentel, D. 1993. The tightening conflict: population, energy use, and the ecology of agriculture. The NPG Forum, October. Negative Population Growth Inc., Teaneck, NJ.
Giampietro, M., Cerretelli, G. and Pimentel, D. 1992a. Energy analysis of agricultural ecosystem management: human return and sustainability. Agriculture, Ecosystems and Environment 38: 219-244.
Giampietro, M., Cerretelli, G. and Pimentel, D. 1992b. Assessment of different agricultural production practices. Ambio 21(7): 451-459.
Giampietro, M., Bukkens, S.G.F. and Pimentel, D. 1993. Labor productivity: A biophysical definition and assessment. Human Ecology, 21 (3): 229-260.
Gibbons, J.H. and Blair, P.D. 1991. U.S. energy transition: on getting from here to there. Physics Today 44: 22-30.
Gleick, P.H. 1993. Water in Crisis: A Guide to the World's Fresh Water Resources. Oxford University Press, New York.
Hall, C.A.S., and Cleveland, C.J. and Kaufmann, R. 1986. Energy and Resource Quality. John Wiley ~ Sons, Inc., New York.
Hall, C.A.S., Pontius, R.G., Jr., Coleman, L. and Ko, J-Y. 1994. The environmental consequences of having a baby in the United States. Population and Environment, 15 (6): 505-524.
International Institute for Applied Systems Analysis (IIASA). 1981. Energy in a Finite World-A Global Systems Analysis. Edited by Energy Systems Program Group of IIASA. Ballinger, Cambridge, MA.
Kaplan, R.D. 1994. The coming anarchy. The Atlantic Monthly 273: 44-76.
Kendall, H.W. and D. Pimentel. 1994. Constraints on the expansion of the global food supply. AMBIO 23 (3): 198-205.
Koester, U. 1991. Economy-wide costs of farm support policies in the major industrial countries. In: K. Burger, M. de Groot, J. Post and V. Zachariasse (Editors), Agricultural Economics and Policy: International Challenges for the Nineties. No. 7 in the series Developments in Agricultural Economics, pp. 5-17. Elsevier, Amsterdam.
Lal, R. 1989. Land degradation and its impact on food and other resources. In: D. Pimentel (Ed.), Food and Natural Resources, pp. 85-140. Academic Press, San Diego.
Lawson, R.L. 1991. The U.S. should increase its use of coal. In: C. P. Cozic and M. Polesetsky (Eds.), Energy Alternatives (pp. 41-45). San Diego: Greenhaven Press.
Leyton, L. 1983. Crop water use: principles and some considerations for agroforestry. In: P. A. Huxley (Ed.), Plant Research and Agroforestry, pp. 379-400. International Council for Research in Agroforestry, Nairobi, Kenya.
McSweeney, B.G. 1979. Collection and analysis of data on rural women's time use. Studies in Family Planning 10(11/12): 379-383.
Mueller, E. 1984. The value and allocation of time in rural Botswana. Journal of Development Economics 15: 329-360.
Odum, E.P., 1971. Fundamentals of Ecology. W.B. Saunders Company, Philadelphia.
Pimentel, D., Culliney, T.W., Buttler, I.W., Reinemann, D.J. and Beckman, K.B. 1989. Low-input sustainable agriculture using ecological management practices. Agriculture, Ecosystems & Environment 27: 3-24.
Pimentel, D., Wen Dazhong and Giampietro, M. 1990. Technological changes in energy use in U.S. agricultural production. In: S.R. Gliessman (Ed.), Agroecology, pp. 305-321. Springer-Verlag, New York.
Pimentel, D., U. Stachow, D.A. Takacs, H.W. Brubaker, A.R. Dumas, J.J. Meaney, J.A.S. O'Neil, D.E. Onsi, and D.B. Corzilius. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience 42: 354-362.
Pimentel, D., M. Herdendorf, S. Eisenfeld, L. Olander, M. Carroquin. Corson, J. McDade, Y. Chung, W. Cannon, J. Roberts, L. Bluman and J. Gregg. 1994a. Achieving a secure energy future: Environmental and economic issues. Ecological Economics, 9 (3): 201-219.
Pimentel, D., R. Harman, M. Pacenza, J. Pecarsky, and M. Pimentel. 1994b. Natural resources and an optimum human population. Population and Environment 15 (5): 347-369.
Pimentel, D., Harvey, C., Resosudarmo, P., Sinclair, K., Kurz, D., McNair, M., Crist, S., Shpritz, L., Fitton, L., Saffouri, R. and Blair, R. 1994c Environmental and economic costs of soil erosion and conservation benefits. Unpublished manuscript.
Postel, S. 1989. Water for Agriculture: Facing the Limits. Worldwatch Paper 93. Worldwatch Institute, Washington, DC.
Rijksinstituut voor Volksgezondheid en Milieuhygiene (RIVM). 1992. National Environmental Outlook 2, 1990-2010. RIVM, Bilthoven, The Netherlands.
Silvestri, G. and Lukasiewicz, J. 1991. Occupational employment projections. Monthly Labor Review, November: 64-80.
Simon, J. L. 1981. The Ultimate Resource. Princeton University Press, Princeton.
Sisler, D. 1988. Personal communication. Department of Agricultural Economics, Cornell University, Ithaca, NY.
Smil, V. 1991. Population growth and nitrogen: An exploration of a critical existential link. Population and Development Review 17 (4): 569-601.
Stanhill, G. 1984. Agricultural labour: from energy source to sink. In: G. Stanhill (Ed.), Energy and Agriculture, pp. 113-130. Springer-Verlag,Berlin.
Starr, C., Searl, M.F. and Alpert, S. 1992. Energy sources: a realistic outlook. Science, 256: 981-987.
Stone, J. 1994. Empty or full? The debate over the population of Australia Paper presented at The Optimum Population Conference, Cambridge, UK, August, 1993. Abstract published in Better World (Magazine of the Optimum Population Trust), No. 3, January 1994, p. 22-23.
Stout, B.A. 1991. Handbook of Energy for World Agriculture. New York: Elsevier.
Troeh, F.R., Hobbs, J.A. and Donahue, R.L. 1991. Soil and Water Conservation. Prentice Hall, Englewood Cliffs, NJ.
U.S. Bureau of the Census (USBC). 1975. Historical Statistics of the United States Colonial Times to 1970, Bicentennial Edition, Part 1/2. Washington, DC.
U.S. Bureau of the Census (USBC). 1991. Statistical Abstract of the United States 1991. Washington, D.C.: U. S. Department of Commerce.
U.S. Bureau of the Census (USBC). 1992. Statistical Abstract of the United States 1992. Washington, D.C.: U. S. Department of Commerce.
U.S. Bureau of the Census (USBC). 1992. Current Population Reports. USBC, Washington, DC.
U.S. Congress (USC). 1990. National Energy Strategy: Hearing Before the Subcommittee on Oversight and Investigations of the Committee on Energy and Commerce. U.S. Congress, House of Representatives, Washington, DC.
U.S. Department of Agriculture (USDA). 1971. Agriculture and the Environment.
USDA, Economic Research Service (ERS), Washington, DC.
U.S. Department of Agriculture (USDA). 1989. The Second RCA Appraisal. Soil, Water, and Related Resources on Nonfederal Land in the United States. Analysis of Conditions and Trends. USDA, Washington, DC.
U.S. Department of Agriculture (USDA). 1990. Agricultural Statistics. Government Printing Office, Washington, DC.
U.S. Department of Agriculture (USDA). 1992. Agricultural Statistics. Government Printing Office, Washington, DC.
Vitousek, P.M., Ehrlich, P.R., Ehrlich, A.H. and Matson P.A. 1986. Human appropriation of the products of photosynthesis. BioScience 36, 368-373.
Waggoner/CAST (Council for Agricultural Science and Technology). 1994. How Much Land Can Ten Billion People Spare for Nature. Report prepared by Paul E.
Waggoner. Task Force Report, No. 121, February 1994. CAST, Ames, IA.
WDF 1989. World Development Forum 7 (10): 1.
World Resources Institute (WRI). 1991. World Resources 1990-91. New York: Oxford University Press.
World Resources Institute (WRI). 1992. World Resources 1992-93. New York: Oxford University Press.
World Resources Institute (WRI). 1994. World Resources 1994-95. New York: Oxford University Press.

Source: Silvestri and Lukasiewicz (1991).
Occupational Groups percentage of total occupation in the year) (% change in employment)

1 Agricultural production in terms of energy.
2 Food supply (per capita). Actual per capita food intake is lower due to post-harvest losses.

Appendix

Sources and specifications for the comparison of the agricultural performances of 20 countries

I Calculation of energy output in agriculture

Source of data: FAO Food Balance Sheet (199lc).

Total output is the sum of plant and animal products (including export but not imports). The conversion factor to obtain energy output from kg output has been estimated for each country separately on the basis of the local agricultural production pattern using food group-specific energetic conversion factors provided by FAO. Plant products were divided into the following food groups: cereals, starchy roots, sweeteners, pulses, oil crops, vegetables, fruit, stimulants, spices, and alcoholic beverages. Animal products included meat, milk, eggs, animal fats (fish and seafood are not included).

II Calculation of energy inputs in agriculture

Sources of data: FAO (1991a) for machinery (number of tractors and harvesters) and irrigation (percentage of arable land irrigated); FAO (199lb) for fertilizers (N as N; P as P2Os; K as K2O) expressed in metric tons; WRI (1994) for pesticides (metric tons).

Conversion factors:

Machinery: Average weight per piece of machinery 15 tons for USA, Canada, and Australia; 8 tons for Argentina and European countries; 6 tons for Africa and Asia (after Stout, 1991, p. 75)

Fuel consumption per tractor or harvester: 5 tons/year for USA, Canada, Australia; 3.5 tons/year for Argentina and European countries; 3 tons/year Asia and Africa (based on Stout, 1991, p. 75)

Energy equivalent for machinery weight: 34,230 kcal/kg of machinery (this includes manufacturing, maintenance, repair, and transportation) divided by 10 in order to discount on a life span of 10 years (based on Stout, 1991, p. 75).

Energy equivalent of fuel: 10,081 kcal/kg (42.2 GJ/metric ton oil).

Irrigation: 2,000,000 kcal/ha/year for Argentina, USA, European countries, China, and Asia; 2,300,000 kcal/ha/year for Africa and Australia (based on Stout, 1991, p. 88).

Pesticides: 70,000 kcal/kg for pesticides in developing countries; 100,000 kcal/kg for developed countries (after Helsel, 1992, p. 194-196).

Fertilizers: the energy conversion includes the production, packaging, and transportation of fertilizers; Nitrogen (as N): 18,657 kcal/kg; Phosphate (as P2O5): 4,157 kcal/kg; Potash (as K2O): 3,273 kcal/kg (after Helsel, 1992, p. 184).

NOTES: (i) Recent data on pesticide consumption are not available. (WRI data are extrapolated from the last available FAO data before FAO discontinued its pesticide recording in 1985/86). Recent trends in pesticide consumption in agriculture show that the number of treatments has increased but that lower dosages per treatment are applied (more toxic products are available). Furthermore, more effective pesticides (requiring lower dosage) are more expensive in energetic terms.

(ii) In general, it is difficult to assess input/ output energy ratios for agriculture since different authors have different ideas of what should be considered inputs of agriculture (e.g. energy spent for drying of crops, transportation of outputs and inputs to, from and on the farm, construction of buildings and infrastructures). To have comparable data, we relied on FAO statistics and applied homogeneous conversion factors on these data. Therefore, the data presented are perhaps not as accurate as some country specific assessments available for selected countries (from specific case studies). On the other hand, in this way we have a set of data that can be used for international comparison.

FAO. 1991a Production Yearbook 1990 (Vol. 44). Rome: Food and Agriculture Organization of the

United Nations (FAO).

FAO. 1991b. Fertilizer Yearbook 1990. Rome: FAO.

FAO. l991c. Food Balance Sheets. Rome: FAO.

Helsel, Z.R 1992. Energy and alternatives for fertilizer and pesticide use. In: R.C. Fluck (Ed.), Energy in Farm Production (Vol. 6 of Energy in World Agriculture), pp. 177-201. Elsevier, Amsterdam.

Stout, B.A. 1991. Handbook of Energy for World Agriculture. New York: Elsevier. World Resources Institute (WRI). 1994. World Resources 1994-95. New York: Oxford University Press.

Table of Contents  Please send me your comments.