The role of nitrogen in world food production and environmental sustainability

Agriculture, Ecosystems and Environment 116 (2006) 4–14 www.elsevier.com/locate/agee

The role of nitrogen in world food production and environmental sustainability B. Eickhout *, A.F. Bouwman, H. van Zeijts Netherlands Environmental Assessment Agency, National Institute for Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands Available online 18 April 2006

Abstract On the basis of the FAO projection ‘World Agriculture: Towards 2015/2030’ we direct our discussion to food production, the consequences for land use, efficiency of nitrogen (N) and losses of reactive N to the environment during 1995–2030. According the FAO, global food production can keep pace with the increase in food demand in the coming three decades. However, according the projection used here, there will be a major global increase (8%) in arable land, most of it in developing countries and with a major impact on the extent of tropical forests. Further forest clearing may occur to compensate for declining soil productivity due to land degradation. Despite improvements in the N use efficiency, total reactive N loss will grow strongly in the world’s increasingly intensive agricultural systems. In the 1995–2030 period emissions of reactive N from intensive agricultural systems will continue to rise, particularly in developing countries. Therefore, the increase of N use efficiency and further improvement of agronomic management must remain high on the priority list of policy makers. # 2006 Elsevier B.V. All rights reserved. Keywords: Agriculture; Food security; Land use; Crop production; Environment; Reactive nitrogen; Sustainable development

1. Introduction For many decades, food security and hunger have been among the most important issues playing a role in the global political arena. At the first World Food Summit in 1974, political leaders from around the world set a goal to eradicate hunger in the world within 10 years. This ambitious goal was not met, leading to new goals at the second World Food Summit in 1996. The world leaders committed themselves to reduce the number of chronically undernourished by half by the year 2015. This target has been endorsed at many other meetings since then and is now known as one of the eight millennium development goals (MDGs) of the United Nations (United Nations, 2001). The MDGs commit the international community to adopt an extended view on development and recognize the importance of creating a global partnership to achieve sustainable economic growth. * Corresponding author. Tel.: +31 30 274 2924; fax: +31 30 274 4464. E-mail address: [email protected] (B. Eickhout). 0167-8809/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.agee.2006.03.009

The term ‘sustainable economic growth’ embraces the marriage of economy and ecology and has been introduced by Brundtland (1987) as sustainable development. The Brundtland report emphasizes the importance of development as a prerequisite for peace, security and protection of the environment. Hence, hunger eradication needs to be addressed within ecological constraints. The seventh MDG (ensuring environmental sustainability) is poorly elaborated in terms of measurable indicators. The United Nations General Assembly concluded that the selection of indicators for the environmental MDG would need further refinement (United Nations, 2001). So far, however, only greenhouse gas emissions, the extent of forest area and the access to improved water sources and sanitation have been defined as indicators. These are insufficient when it comes to assessing the potential for food security within environmental constraints (MNP, 2005). This paper focuses on the relationships between food production, the area of agricultural land and emissions of reactive nitrogen (N). Nitrogen is an essential element for plant growth and a key element of agricultural input. Rapid

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

increases of crop yields became possible when synthetic N fertilizer became available after the discovery of the Haber– Bosch process in the early 20th century (Smil, 2001). However, increased use of N fertilizers has also led to increased N losses from agro ecosystems, especially since the 1950s. Distribution of fertilizer N across the globe is very uneven. In some areas N is used excessively and leads to N pollution, causing a host of problems for human and ecological health. Other parts of the world suffer from reduced soil fertility, diminished crop production, and other consequences of inadequate N supply (Mosier et al., 2004). The recovery of fertilizer N in global crop production is about 50% (Krupnik et al., 2004; Smil, 1999). The surplus may accumulate in soils, or be lost to air, groundwater and surface water via various pathways. Losses from the soilplant system are due to denitrification in the form of gaseous dinitrogen (N2), nitrous oxide (N2O) and nitric oxide (NO), volatilization of ammonia (NH3), leaching of nitrate (NO3 ), runoff and erosion (Bouwman et al., 2002a; FAO/IFA, 2002). The environmental consequences of the different compounds are diverse. Essentially, all emitted NH3 is returned to the surface by deposition, one of the causes of soil acidification since the early 1980s (Van Breemen et al., 1982). Moreover, there is growing concern about the eutrophication of natural ecosystems and loss of biodiversity due to N deposition (Bouwman et al., 2002c). Nitrous oxide is one of the so-called greenhouse gases, contributing 6% of the anthropogenic greenhouse effect; it also contributes to the depletion of stratospheric ozone (IPCC, 2001). Finally, NO3 is an important pollutant of groundwater and surface water (Heathwaite, 1993; Johnes and Burt, 1993). Increasing N inputs to freshwater systems can, if sufficient P is present, cause eutrophication, generally accompanied by decreased diversity of both plant and animal species (Schindler, 1977; Vollenweider, 1992). The aim of this paper is to provide insight into the impact of historical and future changes of food production on the environmental losses of reactive N from agriculture. After introducing the data and methods (Section 2), we will discuss the changes in food consumption (Section 3) and production (Section 4), land use (Section 5), nitrogen use efficiencies (Section 6), and losses of reactive N (Section 7) in different regions of the world for the period 1970–2030.

2. Data and methods used For our analysis we used historic data for food demand and production from FAOSTAT (for the period 1970–1995; FAO, 2001) and a projection to 2030 from the FAO study ‘World Agriculture: Towards 2015/2030’ (Bruinsma, 2003). This projection has been transformed into spatial land use distributions with the Integrated Model to Assess the Global Environment (IMAGE-team, 2001). The IMAGE model generates 0.5 by 0.5 degree global land cover maps providing the grid cells covered by either agriculture (crops

5

and grassland) or natural ecosystems. Four broad groups of crops were distinguished: grassland, wetland rice, leguminous crops (pulses, soybeans) and upland crops. Within the category of grassland, we distinguished between grassland in intensive (landless and mixed) and pastoral livestock production systems and (semi)-natural and marginal grassland according to Bouwman et al. (2005a), who used estimates from Sere´ and Steinfeld (1996). Here, we use the term intensive agricultural systems for all mixed and industrial (landless) livestock production systems and all arable agriculture. The methodology for N balances calculation was taken primarily from Bouwman et al. (2005b). Many countries have vast areas of extensively used pastoral (semi)-natural and marginal grassland, typically with small N inputs. In contrast, other countries have primarily mixed and landless systems characterized by much larger N inputs. To make a good comparison between different countries, we excluded the areas of pastoral (semi)-natural and marginal grassland, and consider the N balances for crop and mixed and landless livestock production systems together. Including pastoral, natural and marginal grassland would give an unrealistic impression of the intensity of agricultural systems. N inputs in the surface N balance include biological N fixation, atmospheric N deposition, application of synthetic N fertilizer, and animal manure and animal N excreted during grazing. Outputs include N removal by crop harvesting and grazing. The N balance calculation is a static approach, i.e., soil N changes were neglected. The N balance surplus consists of NH3 volatilization, denitrification, leaching and surface runoff. All surface N balance input and output terms were allocated to the 0.5 by 0.5 degree resolution according to the fractions of the grid cells covered by wetland rice, leguminous crops, upland crops and fertilized grassland. Mean N application rates via chemical fertilizer for the mid-1990s were taken directly from IFA/IFDC/FAO (2003), with a few exceptions. For 1970 the application rates were multiplied with the change in crop yield compared to 1995 (Bruinsma, 2003) thus ignoring changes in crop N recovery. For N excretion rates per animal, data from Van der Hoek (1998) were used. All manure produced by pigs and poultry was assumed to be collected. For ruminants, the fraction of the manure that is excreted in pastures was calculated from the fraction grass in the ration according to Bouwman et al. (2005a) and the complement was assumed to be stored in animal houses. The animal manure produced within intensive and pastoral systems is distributed over different animal waste management systems according to Bouwman et al. (2005b). Animal manure available for application to cropland and grassland equals all collected manure corrected for manure used as fuel and building purposes. Within the intensive and pastoral systems in most developed countries, we assumed that 50% of the collected animal manure was applied to arable land and 50% to grassland. In most developing countries 95% of the collected manure is

6

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

assumed to be applied to crops and 5% to grassland, thus, accounting for the lower importance of grass compared to crops in developing countries (Sere´ and Steinfeld, 1996). Ammonia volatilization from the application of animal manure and N fertilizers was calculated according to Bouwman et al. (2002a). The direct emissions of N2O and NO from agricultural land were calculated according to Bouwman et al. (2002b) and nitrate leaching was calculated according to by Van Drecht et al. (2003). All calculations were done for each 0.5 by 0.5 degree grid cell, taking into account a variety of factors, related to agricultural management and climate and soil conditions. Ammonia volatilization from animal manure stored in animal houses and from grazing was calculated according to Bouwman et al. (1997). The removal of N from arable fields in the harvested product was calculated from crop production data (FAO, 2001) and projections (Bruinsma, 2003) and N content for 34 crops (Bouwman et al., 2005b). N removal by grazing and cutting of grass and hay were estimated as 0.6 times the sum of the N inputs minus the NH3 emission (Bouwman et al., 2005b). Here, we defined two indices for analyzing differences in the efficiency of N use and N loss to the environment for intensive agricultural systems, i.e. the ‘overall system N recovery’ (OSR) and the ‘nitrogen uptake ratio’ (NUR). The OSR is defined as the crop N export (sum of the N in harvested crop and grass parts and grass consumption by grazing) expressed as a percentage of the sum of all N inputs. The mixed systems considered here have integrated crop and livestock production, in which livestock production relies on a mix of food crops, crop by-products and roughage, consisting of grass, fodder crops, crop residues, and other sources of feedstuffs. In these mixed systems the byproducts of one activity (crop by-products, crop residues and manure) often serve as inputs for another. Therefore, the OSR is an indicator for the overall N use efficiency for the integrated crop-livestock production in intensive mixed systems, and for N losses. All N that is not taken up by the crop is assumed to be lost by NH3 volatilization, denitrification, leaching and surface runoff. However, since in the static surface N balance approach we do not account for changes in soil N, the calculated efficiencies may be high in agricultural systems where soil N depletion is a major source of plant N. In contrast, the N losses to the environment may be overestimated in a situation with soil N accumulation. The NUR is defined as the N in harvested crop parts (for all crops excluding legumes) divided by the N inputs from fertilizers and animal manure. Legumes are excluded because these crops can fix atmospheric nitrogen and generally receive no or small amounts of N fertilizer. This definition is comparable to the nitrogen use efficiency (NUE) as used by, for example, Dobermann and Cassman (2005) for cereals, which is defined as the ratio of cereal grain yield divided by the total amount of N fertilizer applied

to cereals. For our purpose the NUR has the advantage that is an indicator for the loss of N for situations of N surplus and for soil N depletion in situations of N deficit (NUR > 100%). Both the NUR and NUE ignore changes in soil N and the N inputs from atmospheric deposition and non-symbiotic biological N fixation by free-living organisms. Therefore, we recognize that N losses may be underestimated and N depletion overestimated.

3. Food consumption Food demand is driven by population growth and per capita consumption. According to the latest United Nations assessment of world population prospects, growth will gradually slow down from the present 1.5–1.2% yr 1 in 2015 and 0.9% yr 1 between 2015 and 2030, reaching a world total of 8270 million inhabitants in 2030 (Table 1). Although the growth rate of populations varies widely, most of the increases will be in developing countries. The projected growth in transition countries (Eastern Europe and the former USSR) will be negative in the coming three decades (between 0.2 and 0.3% yr 1). The growth of income is the other major determinant of increasing food demand. The development of per capita consumption is strongly related to economic development. The projected growth of per capita gross domestic product (GDP) is particularly high in developing and transition countries (about 4% yr 1 for the coming three decades, respectively), and somewhat lower for industrialized countries (2.7% yr 1) as seen in Table 1. World total food consumption (kcal) may rise by more than 50% between 1998 and 2030. The projections of food consumption for the different commodities suggest that the per capita food consumption will grow significantly. The Table 1 Total population, annual per capita GDP growth, total food consumption and meat consumption for developing, industrial and transition countries in the years 1965, 1998 and 2030 Year

Developing

Industrialized

Transition

World

6

Population (10 inhabitants) 1965 2657 1998 4572 2030 6869

706 892 979

Per capita GDP growth (% yr 1) 1998–2030 4.0

351 413 381

2.7

4.1

3714 5900 8270 2.6

1

Per capita total food consumption (kcal day ) 1965 2054 2947 1998 2681 3380 2030 2980 3500 Meat consumption 1965 1998 2030

(kg person 10 26 37

1

yr 1) 62 88 100

3223 2906 3180

2358 2803 3050

43 46 61

24 36 45

Source: Bruinsma (2003) and FAO (2001). Please note that in the rest of the study we used the years 1970, 1995 and 2030.

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

world average will exceed 3000 kcal per person per day in 2030 in the projection used. There is a sharp contrast between regions with relatively high levels of poverty and food insecurity. In South Asia, the long-term high GDP growth of more than 4% yr 1 holds the promise of having a positive impact on food security. In subSaharan Africa the limited growth of about 2.0% yr 1 will result in a much slower eradication of undernourishment. At low income levels, economic growth is generally accompanied by a significant structural change in consumption patterns. Diets shift to include more livestock products (Table 1) and less staples such as roots and tubers (Bruinsma, 2003). The per capita meat consumption is showing a strong global increase (25% between 1998 and 2030), with very fast growth in developing and transition countries (more than 40 and 30%, respectively). Although the annual meat consumption of 88 kg of meat per person is already much higher in industrialized countries than in developed countries, a further increase of about 15% is projected by Bruinsma (2003). This will lead to increased use of land for growing feedstuffs (Section 4) and to a rise in the use of fertilizers (Section 5). These trends in resource use are in line with foresights from Cassman et al. (2003) that the global grain demand will increase at a faster rate than population growth because of a greater per capita consumption of livestock products in developing countries. Analyses by Tilman et al. (2001) indicate that future land use and fertilizer use depend much more on per capita GDP growth (70%) than on population growth (30%). Currently, 2800 kcal per person per day of food energy are produced per capita on average. Since the average minimum food intake is between 1700 and 2000 kcal per person per day to prevent malnutrition, sufficient food should be available worldwide to feed everyone. However, at the global level, 776 million people were undernourished in 1997/1999; this number is expected to decline to 610 million in 2015 and to 440 million in 2030 (Bruinsma, 2003). As a result, the MDG to halve the proportion of the world population suffering from hunger by the year 2015 compared with 1990–1992 will not be achieved. Armed conflicts, bad governance, unclear property rights, lack of education and poor infrastructure will likely remain the major underlying causes of undernourishment (Bruinsma, 2003).

Table 2 Total yield of upland crops, leguminous crops and wetland rice for 1970, 1995 and 2030 for different world regions Regions

Year

Upland cropsa (Tg yr 1)

Legumi-nous crops (Tg yr 1)

Wetland rice (Tg yr 1)

Developing

1970 1995 2030

1928 3974 6641

42 102 251

225 412 603

Industrialized

1970 1995 2030

867 1196 1559

35 79 147

13 15 13

Transition

1970 1995 2030

559 478 563

9 6 5

1 1 1

World

1970 1995 2030

3355 5647 8763

86 186 403

239 427 618

Based on Bouwman et al. (2005a). a Include all crops except wetland rice and leguminous crops.

sometimes triple cropping. Genetic engineering may become more important in the near future; more rapidly in developed than in developing countries. Major benefits are expected from savings on external inputs like pesticides and herbicides (Bruinsma, 2003). Global crop production increased by 70% in the 1970– 1995 period and is expected to grow by another 60% in the coming three decades (Table 2; Bouwman et al., 2005b). Close to 90% of the production increase in the 1970–1995 period occurred in developing countries; for the period up to 2030 this is expected to be 85%. While the increase in rice production in the developing countries was 85% in the 1970–1995 period, the growth will slow down to less than 50% between 1995 and 2030. The global production of meat more than doubled between 1970 and 1995, and this increase will slow down to about 160% in the coming three decades (Table 3). Most of this increase (around 75% for 1970–1995, Table 3 Total production of meat for 1970, 1995 and 2030 for pastoral and intensive livestock production systems in different regions for the years 1970, 1995 and 2030 Regions

Year

Pastoral (Tg yr 1)

Mixed/industrial (Tg yr 1)

Developing

1970 1995 2030

8 11 14

28 120 249

Industrialized

1970 1995 2030

2 3 3

58 89 100

Transition

1970 1995 2030

0 0 0

22 24 29

World

1970 1995 2030

10 14 17

108 232 379

4. Food production From 1970 to date, food production has kept pace with increasing demand. New high-yielding varieties combined with improved crop management have been important for increasing yields in industrialized countries and Asia (green revolution). The use of tractors and farm machinery improved productivity per agricultural worker. Increased use of mineral fertilizers, pesticides and irrigation has led to substantial yield increases. In addition, cropping intensity has increased in many parts of the world, with double and

7

Based on Bouwman et al. (2005a).

8

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

and around 90% in the coming three decades) will occur in the developing countries (Bruinsma, 2003). The contribution of yield increase to the total growth of production for the period 1995–2030 is expected to be about 70% (Bruinsma, 2003), which is larger than the expected growth in world total food consumption (Section 3). This implies that crop yield increase alone would be more than sufficient to produce the extra food for direct human consumption needed in the coming three decades. The remaining production increase (300–400 Tg dry matter yr 1) equals the increase in production of food crops used to feed animals, implying that most of the projected arable land expansion in the next three decades is needed for increasing the production of animal feedstuffs (Bouwman et al., 2005a).

5. Land use Between 1970 and 1995 the global arable land area expanded from 1405 to 1494 Mha (Table 4; Bouwman et al., 2005a). Since the arable land areas decreased slightly in the industrialized and transition countries during this period, the expansion in the developing countries has been considerable (15%). Total grassland areas have increased by 4% in the 1970–1995 period, which adds another 150 Mha to the total expansion of the agricultural area (Table 4; Bouwman et al., 2005a). The global arable land area is expected to increase further from 1494 Mha in 1995 to 1609 Mha between 1995 and Table 4 Areas of grassland in pastoral, intensive (semi)-natural and marginal systems and of arable land in different regions in the world for the years 1970, 1995 and 2030 Systems 1970 Pastoral Mixed/industrial (Semi)-natural and marginal Total grassland Arable land 1995 Pastoral Mixed/industrial (Semi)-natural and marginal Total grassland Arable land 2030 Pastoral Mixed/industrial (Semi)-natural and marginal Total grassland Arable land

Developing (Mha)

Industrialized (Mha)

Transition (Mha)

World (Mha)

1281 209 616

175 242 371

0 100 273

1456 551 1261

2106 742

788 381

373 282

3268 1405

1451 239 618

147 236 372

0 90 261

1599 565 1251

2308 861

756 367

351 266

3415 1494

1437 297 618

157 190 373

0 96 248

1594 584 1239

2353 963

720 372

344 273

3416 1609

Based on Bouwman et al. (2005a).

2030 (Table 4). While the arable land area in industrialized and transition countries will increase only slightly during this period, there will be a major expansion in the developing countries (more than 10%). The global area under grassland is expected to change only slightly in the coming three decades (Table 4). However, in developing countries the area under intensively managed grassland may increase further as a result of a continued growth of intensive ruminant meat and milk production (Bouwman et al., 2005a). The agricultural land expansion will have a massive impact on the extent of tropical forests, being one of the indicators of the environmental MDG. However, the abovementioned numbers do not reflect the land transformations that occur as a consequence of loss of the land’s productivity and abandonment of eroded and degraded land. On the basis of a compilation of data, Bouwman and Leemans (1995) estimated an annual global loss of 4 Mha of degraded arable land that would need to be compensated for by forest clearing. Many low input systems have been able to sustain a low per capita food production volume at low levels of crop yields, through expansion of agricultural land and at the cost of soil fertility (FAO, 2001). In fact, crop yields have not increased substantially in many, primarily sub-Saharan, countries (FAO, 2001), where a high risk of land degradation prevails as a result of soil erosion, soil nutrient depletion and overgrazing (Delgado et al., 1999; Sere´ and Steinfeld, 1996). It is uncertain how much forested land has been permanently converted to agricultural land to compensate for the productivity decline, and how important the role of shifting cultivation is in maintaining the soil productivity in these regions.

6. Nitrogen use efficiencies The potential for cropland expansion is limited in most countries. Crop production increases will therefore come primarily from higher yields and thus higher nutrient application rates (Roy et al., 2002). In the 1970–1995 period, the global total N inputs from fertilizers, animal manure and N fixation in intensive agricultural systems almost doubled, from about 92 to 165 Tg yr 1 (Table 5). By contrast, the N inputs in extensive, pastoral livestock production systems increased only slowly. The aggregated NUR did not change in the industrialized countries (48–49%) between 1970 and 1995 (Table 6). In the transition countries, NUR values for 1970 are comparable to those in industrialized countries. However, after 1990, N fertilizer use strongly decreased in transition countries, with the NUR increasing to 67%. NUR exceeded 100% in 1995 in Russia, indicating an increasing risk of soil N depletion. Soil depletion is probably also the cause of high NUR values for many developing regions. The projected increase of global total crop production during 1995–2030 is almost 60% (Table 2) while the growth

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

9

Table 5 Inputs of N from animal manure in pastoral and intensive agricultural production systems, and of N fertilizer and biological N2 fixation in intensive systems for the years 1970, 1995 and 2030 Regions

Year

Animal manure Na Pastoral (N in Tg yr 1)

Intensive (N in Tg yr 1)

N fertilizer (N in Tg yr 1)

N2 fixation (N in Tg yr 1)

Total (N in Tg yr 1)

Developing

1970 1995 2030

22 24 30

19 30 40

9 53 73

10 16 25

59 123 168

Industrialized

1970 1995 2030

2 3 3

15 16 16

15 26 31

6 9 13

39 53 62

Transition

1970 1995 2030

0 0 0

9 8 9

7 5 6

2 2 2

19 15 17

Global

1970 1995 2030

24 27 33

43 55 64

31 83 110

18 27 40

116 192 247

Calculated from Bouwman et al. (2005b). a Total manure N not corrected for NH3 volatilization from collected manure. It excludes animal manure that is not part of the agricultural system, such as manure excreted in urban areas and stored but unused manure.

of the N inputs in the same time period is 30% (Table 5). The projected growth of N inputs is much lower than the increase during 1970–1995. The NUR in industrialized countries is expected to increase to about 60% in the coming three decades, and for the developing countries as a whole NUR is estimated to increase to close to 60% (Table 6). However, there are large differences between regions and countries. Soil N depletion will still occur in many countries in subSaharan Africa (NUR > 100%) and Southeast Asia (NUR 90%). In other developing countries the NURs for 2030 will be comparable to those of the industrialized countries. Although our results (Table 6) are aggregated efficiencies for a large number of crops, the trends in NUR are basically the same as those presented by Cassman et al. (2003) for

cereals alone, with decreasing efficiencies in the 1960–2000 period for most developing regions. The change in time of the NUR for Africa (a decrease between 1970 and 1995 and an increase in the coming three decades) and the transition countries (a sharp increase in the early 1990s) is also reported by Cassman et al. (2003) and Dobermann and Cassman (2005). In industrialized countries, the OSR was more-or-less stable (44–45%) between 1970 and 1995 and will improve to 52% in the coming decades (Table 6). There was a decrease in the OSR in the developing countries between 1970 (52%) and 1995 (42%), with 1995 values ranging from 31% (South Asia) to 55% (Southeast Asia). This decrease is related to the fast increase in N fertilizer consumption in many developing

Table 6 The overall system N recovery (OSR) and nitrogen uptake ratio (NUR) in different regions in the years 1970, 1995 and 2030 Regions

OSRa (%)

NURb (%)

1970

1995

2030

1970

1995

2030

North America Western Europe Transition countries Latin America Middle East + North Africa Sub-Saharan Africa South Asia East Asia Southeast Asia

42 44 38 51 54 48 42 61 58

43 49 46 43 47 51 31 43 55

51 58 55 49 50 58 38 39 61

49 44 43 82 85 139 99 108 135

48 54 67 49 58 108 41 48 78

63 68 83 66 63 131 58 42 90

World Developing Industrialized

46 52 44

43 42 45

47 45 52

67 103 48

52 51 49

61 58 62

Source: Bouwman et al. (2005b). a The overall system N recovery (OSR) is calculated as the N in all harvested crops and grass cut and consumed by grazing expressed as % of total inputs from N deposition, N2 fixation, N fertilizer and animal manure. b The nitrogen uptake ratio (NUR) is calculated as the N in harvested parts in upland crops and wetland rice (excluding leguminous crops) as % of N input from fertilizers and manure.

10

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

countries such as China (Table 6). For the developing countries the OSR will slowly increase in the coming decades to 45% in 2030. The cultivation of leguminous crops (pulses, soybeans) has a major influence on the OSR. The global N inputs from N fixation by legumes have strongly increased from 18 Tg yr 1 in 1970 to 27 Tg yr 1 in 1995, and will continue to grow in all world regions in the coming three decades (Table 5). This development is primarily related to the increasing demand for soybeans as an animal feed resource. The regions with the most prominent growth in soybean production are North and South America, and sub-Saharan Africa. In the projection used, the NUR is assumed to increase as a result of better management and techniques, such as integrated plant nutrition systems, and use of efficient fertilizers, matching application rates with plant demand, precision management, sophisticated schemes for timing and mode of fertilizer application, and crop residue and animal manure management, as discussed elsewhere (Bruinsma, 2003; Roy et al., 2002). Apart from management and crop yields, the N recovery in agricultural systems will depend on many other factors, such as climate and soil conditions and the mix of crops. The potential efficiency, therefore, is not the same for all countries. Results obtained at experimental stations (Balasubramanian et al., 2004) indicate that major increases in efficiency are possible with existing knowledge on management strategies. Further possibilities may arise from crop improvement (through plant selection, and breeding and genetic modification), in which yield stability is increased and yield losses are reduced. This will contribute to increasing the efficiency of fertilizer use. Changing socioeconomic circumstances or policies will have an important influence on whether new and emerging technologies for increasing NUR will gain acceptance by farmers (Giller et al., 2004).

7. Loss of reactive N: consequences for ecosystems and water quality The environmental consequences of N use in food production are strongly related to the OSR and NUR. The difference between the N inputs and the plant N uptake equals the amount of N lost from the system, if changes in soil N are ignored. As stated before, this assumption may lead to underestimations of N losses in situations of soil depletion. Despite the anticipated improvements in the N use efficiency, the total N loss is expected to increase from 109 to 132 Tg N yr 1 (+21%) between 1995 and 2030. It is expected that the total N loss will not change much in the industrialized and transition countries, but it will increase strongly in the developing countries from 67 to 93 Tg N yr 1 (+39%).

The total N loss for different countries and world regions shows contrasting developments. In North America and Western Europe we see a gradual decrease in total N loss as a result of an increasing OSR over the whole 1970–2030 period. In South and Southeast Asia there is an increase in N loss between 1970 and 1995, and a decrease between 1995 and 2030. Whether the loss of reactive N will increase or decrease is related to different developments that simultaneously influence the N recovery (Table 6). It will depend on the importance of intensification, on the one hand, and increasing efficiency on the other. The amount of N increases much faster in intensive production systems than in pastoral systems (Table 5). This development leads to more animal manure storage and animal manure spreading on arable land. The NH3 loss rates associated with such systems are much higher than for pastoral systems (Bouwman et al., 1997). As a result, the global NH3 volatilization from animal housing, fertilizer and animal manure application and grazing increased from 18 Tg N yr 1 in 1970 to 34 Tg N yr 1 in 1995 and will increase further to 44 Tg N yr 1 in 2030 (Table 7). The contribution of developing countries to the global agricultural NH3 source increased from 63% in 1970 to 76% in 1995 and to 80% in 2030. Ammonia emissions from industrialized countries increased slowly between 1970 and 1995, and a further slow increase (North America) or decrease (Western Europe) is projected for 2030 (Table 7). The contribution of industrialized countries to the global emission decreased from 33% to 22% in the 1970–1995 period, with a continued decrease to 18% in 2030. Table 7 Losses of reactive N from intensive agricultural production systems in developing, industrialized and transition countries in the years 1970, 1995 and 2030 Regions

Year

Emissions (Tg N yr 1) NH3 a

N2 O

NO

NO3

Developing

1970 1995 2030

8.6 23.8 33.0

1.2 1.9 2.5

0.5 0.8 1.2

4.8 14.7 21.9

Industrialized

1970 1995 2030

6.0 7.3 7.9

0.5 0.6 0.7

0.4 0.5 0.6

9.0 10.5 10.6

Transition

1970 1995 2030

3.5 3.0 3.1

0.3 0.2 0.3

0.2 0.2 0.2

4.5 3.3 2.8

World

1970 1995 2030

18.1 34.2 44.0

2.0 2.7 3.5

1.1 1.5 2.0

18.2 28.5 35.3

b

Calculated from data presented by Bouwman et al. (2005b). a The projections for 2030 are based on increasing efficiency in livestock production, leading to decreasing N excretion rates per unit of product. The projection does not account for increasing use of additional techniques for reduction of NH3 loss from manure stored in animal housings or from spreading of manure. b The projection for 2030 for countries of the European Union (EU) is based on the implementation of existing EU policies with maximum application rates for 2030 of 170–250 kg N ha 1 yr 1.

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

Fig. 1. Atmospheric N deposition rates for the year 1995 (a) and the change in kg N ha Bouwman et al. (2005b).

Geographically explicit results for atmospheric N deposition are shown in Fig. 1. The consequences of the high rates of N deposition on nature can be devastating. Currently, mainly ecosystems in the northern part of Europe, India and East China are affected by N deposition. However, within the coming 30 years, the pressure on ecosystems through atmospheric N deposition is expected to increase further, especially in India and Eastern China. These foresights raise the question whether the expected increase of food supply meets sustainable standards for (Asian) ecosystems. Mean nitrate leaching and runoff losses for 1995 (Fig. 2) also vary from low values in dry climates (e.g. South Asia, 10% of total N inputs) to close to 20% in more humid, temperate climates (e.g. Western Europe, 22%). World regions dominated by warm and wet climates (Asian regions) also show relatively low leaching rates (<16% of total N inputs) due to high denitrification losses (for example, 24% in East Asia). Decreasing trends in both denitrification and leaching are shown in North America and

1

yr

1

11

between 1995 and 2030 (b). Based on data presented by

Western Europe for the period 1970–2030, and in most of the world regions between 1995 and 2030 (Table 6). Again, the differences between 1995 and 2030 indicate that most of the increase will occur in India and China, returning major challenges to comply with the MDG of access to clean drinking water (Fig. 2). Growing river loads of reactive N will lead to increased incidence of problems associated with eutrophication in coastal seas. Bouwman et al. (2005c) estimated that the global river N flux to coastal marine systems will grow by about 10% in North America and Oceania, will decrease in Europe, and increase by 27% in developing countries. On the global scale and particularly in developing countries, agriculture is the major source of reactive N in surface water. However, in many river basins in developing countries human waste is the dominant contributor to the total N input to surface water, particularly in densely populated river basins or in basins where large population centers are located near the river mouth (Bouwman et al., 2005c).

12

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

Fig. 2. Nitrate concentration of water percolating from the soil to shallow groundwater for the year 1995 (a) and change in nitrate concentrations between 1995 and 2030 (b). Based on the model presented by Van Drecht et al. (2003) and data from Bouwman et al. (2005b,c).

8. Concluding remarks Global food production can meet the food demand towards 2030, but only with the use of improved technologies. However, because of other factors influencing the balance between food supply and demand (Lucas and Hilderink, 2004), it is highly uncertain if the MDG of halving the number of undernourished people will be met. The environmental consequences of the developments in agricultural production for the coming decades, as portrayed here, indicate that it will be difficult to meet the MDG ensuring environmental sustainability; this is due to deforestation (mainly in Africa and Latin America), nitrous oxide and ammonia emissions and deterioration of water quality. However, there is a lack of clearly defined environmental goals and as a result it remains uncertain whether the MDG of eradicating hunger will be achieved within environmental constraints. We recognize that scenarios and projections for future agricultural production and land use are fraught with

uncertainty. For example, the projection that we use for 2030 for the harvested area of cereals (796 Mha) differ by 44 Mha from the projection for 2025 by the International Food Policy Research Institute (Rosegrant et al., 2002). This illustrates there is disagreement between the two approaches. The expected changes in arable land are highly dependent on the expected crop yields and the amounts of N required to support these yields. Cassman et al. (2003) show that we may have difficulties in reaching sufficient increases in yield. Hence, there may be an increased demand for land, which is further aggravated by land degradation and the expected expansion of agriculture to marginal lands in some countries, especially when prime agricultural land is used for urbanization and industrialization. These observed trends also pose the question whether future increases in arable land can be considered sustainable. Our estimates suggest that emissions of NH3, N2O and NO to air, nitrate leaching to groundwater, N inputs to surface water and river N export to coastal aquatic systems, will all increase in the coming three decades, particularly in

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

the developing countries. These increases are mainly the result of the rapid increase in livestock and crop production. Apart from management, the N use efficiencies depend on many factors, such as climate, crop type and crop yield and soil conditions. There are ample management practices that contribute to improving the N use efficiencies. Further, promising strategies for increasing the N use efficiencies include the cultivar selection, plant breeding and genetic modification. The loss of soil productivity by depletion of the soil N (and P) will remain a severe environmental problem, especially in sub-Saharan countries. It may also lead to further forest clearing to compensate for the production loss. Re-cycling of human waste, a major N source in many river basins of developing countries is not a common practice but could be an option for reducing ongoing nutrient depletion of the arable land in many developing countries. Apart from the overall increases in fertilizer use and livestock production, there is also a global trend towards concentration of certain agricultural activities in peri-urban areas (FAO, 1997). This will lead to large local surpluses of N and P from animal manure and associated losses to aquatic systems and the atmosphere, hence, turning non-point source pollution into a more local problem.

Acknowledgements The work described in this paper formed part of the projects Integrated Terrestrial Modeling (S/550005/01/DD) and Sustainability Outlook (M/500013/01/DV) of the Netherlands Environmental Assessment Agency (MNP), associated with the National Institute for Public Health and the Environment (RIVM). Thanks are due to Klaas Van Egmond (MNP) and especially to Ken Giller and Oene Oenema (Wageningen University and Research Center) and three anonymous reviewers for providing extensive comments which led to major improvements of this paper.

References Balasubramanian, V., Alves, B., Aulakh, M., Bekunda, M., Cai, Z., Drinkwater, L., Mugendi, D., Van Kessel, C., Oenema, O., 2004. Crop, environmental, and management factors affecting nitrogen use efficiency. In: Mosier, A.R., Syers, J.K., Freney, J.R. (Eds.), Agriculture and the Nitrogen Cycle. Island Press, Washington, DC, pp. 19–33. Bouwman, A.F., Leemans, R., 1995. The role of forest soils in the global carbon cycle. In: McFee, W.W., Kelly, J.M. (Eds.), Carbon Forms and Functions in Forest Soils. Soil Science Society of America (SSSA), Madison, pp. 503–525. Bouwman, A.F., Lee, D.S., Asman, W.A.H., Dentener, F.J., Van der Hoek, K.W., Olivier, J.G.J., 1997. A global high-resolution emission inventory for ammonia. Global Biogeochem. Cycl. 11, 561–587. Bouwman, A.F., Boumans, L.J.M., Batjes, N.H., 2002a. Estimation of global NH3 volatilization loss from synthetic fertilizers and animal manure applied to arable lands and grasslands. Global Biogeochem. Cycl. 16 (2), 1024, doi:10.1029/2000GB001389.

13

Bouwman, A.F., Boumans, L.J.M., Batjes, N.H., 2002b. Modeling global annual N2O and NO emissions from fertilized fields. Global Biogeochem. Cycl. 16 (4), 1080, doi:10.1029/2001GB001812. Bouwman, A.F., Van Vuuren, D.P., Derwent, R.G., Posch, M., 2002c. A global analysis of acidification and eutrophication of terrestrial ecosystems. Water Air Soil Pollut. 141, 349–382. Bouwman, A.F., Van der Hoek, K.W., Eickhout, B., Soenario, I., 2005a. Exploring changes in world ruminant production systems. Agric. Syst. 84, 121–153. Bouwman, A.F., Van Drecht, G., Van der Hoek, K.W., 2005b. Nitrogen surface balances in intensive agricultural production systems in different world regions for the period 1970–2030. Pedosphere 15 (2), 137–155. Bouwman, A.F., Van Drecht, G., Knoop, J.M., Beusen, A.H.W., Meinardi, C.R., 2005c. Exploring changes in river nitrogen export to the world’s oceans. Global Biogeochem. Cycl. 19 , doi:10.1029/2004GB002314 GB1002. Bruinsma, J.E., 2003. World Agriculture: Towards 2015/2030. In: An FAO Perspective, Earthscan, London, 432 pp. Brundtland, G., 1987. Our Common Future: The World Commission on Environment and Development. Oxford University Press, Oxford, 318 pp. Cassman, K.G., Dobermann, A., Walters, D.T., Yang, H., 2003. Meeting cereal demand while protecting natural resources and improving environmental quality. Annu. Rev. Environ. Resour. 28 10.1–10.44. Delgado, C., Rosegrant, M., Steinfeld, H., Ehui, S., Courbois, V., 1999. Livestock to 2020. The Next Food Revolution. Food, Agriculture and the Environment Discussion Paper 28, International Food Policy Research Institute, Food and Agriculture Organization of the United Nations, International Livestock Research Institute, Washington, D.C., 72 pp. Dobermann, A., Cassman, K.G., 2005. Cereal area and nitrogen use efficiency are drivers of future nitrogen fertilizer consumption. Sci. China Ser. C Life Sci. 48 (Supp.), 1–14. FAO, 1997. Long-term Scenarios of Livestock-Crop-Land Use Interactions in Developing Countries. FAO Land and Water Bulletin No. 6, Food and Agriculture Organization of the United Nations, Rome, Italy, 144 pp. FAO, 2001. FAOSTAT Database Collections. Food and Agriculture Organization of the United Nations, Rome. , http://www.apps.fao.org. FAO/IFA, 2002. Global Estimates of Gaseous Emissions of NH3 NO and N2O from Agriculture. Food and Agriculture Organization of the United Nations/International Fertilizer Industry Association, Rome, 106 pp. Giller, K.E., Chalk, P., Dobermann, A., Hammond, L., Heffer, P., Ladha, J.K., Nyamudeza, P., Maene, L., Ssali, H., Freney, J., 2004. Emerging technologies to increase the efficiency of use of fertilizer nitrogen. In: Mosier, A.R., Syers, J.K., Freney, J.R. (Eds.), Agriculture and the Nitrogen Cycle. Island Press, Washington, D.C., pp. 35–51. Heathwaite, A.L., 1993. Nitrogen cycling in surface waters and lakes. In: Burt, T.P., Trudgill, S.T. (Eds.), Nitrate: Processes, Patterns and Management. Wiley and Sons, Chichester, pp. 99–140. IFA/IFDC/FAO, 2003. Fertilizer Use by Crop, 5th ed. Food and Agriculture Organization of the United Nations, Rome. IMAGE-team, 2001. The IMAGE 2.2 Implementation of the SRES Scenarios. A Comprehensive Analysis of Emissions, Climate Change and Impacts in the 21st Century. CD-ROM publication 481508018, National Institute for Public Health and the Environment, Bilthoven. IPCC, 2001. Climate Change 2001. The Scientific Basis. Contribution of Working Group I to the Third Assessment of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, 881 pp. Johnes, P.J., Burt, T.P., 1993. Nitrate in surface waters. In: Burt, T.P., Trudgill, S.T. (Eds.), Nitrate: Processes, Patterns and Management. Wiley and Sons, Chichester, pp. 269–317. Krupnik, T.J., Six, J., Ladha, K., Paine, M.J., Van Kessel, C., 2004. An assessment of fertilizer nitrogen recovery efficiency by grain crops. In: Mosier, A.R., Syers, J.K., Freney, J.R. (Eds.), Agriculture and the Nitrogen Cycle. Island Press, Washington, D.C., pp. 193–207. Lucas, P.L., Hilderink, H.B.M., 2004. The Vulnerability Concept and its Application to Food Security. Report 550015004/2004, Netherlands

14

B. Eickhout et al. / Agriculture, Ecosystems and Environment 116 (2006) 4–14

Environmental Assessment Agency, National Institute for Public Health and the Environment, Bilthoven, 40 pp. MNP, 2005. Outstanding Environmental Issues for Human Development. Netherlands Environmental Assessment Agency, National Institute for Public Health and the Environment, Bilthoven, 62 pp. Mosier, A.R., Syers, J.K., Freney, J.R. (Eds.), 2004. Agriculture and the Nitrogen Cycle. Island Press, Washington, D.C., p. 296. Roy, R.N., Misra, R.V., Montanez, A., 2002. Decreasing reliance on mineral nitrogen—yet more food. Ambio 31, 177–183. Rosegrant, M.W., Cai, X., Cline, S.A., 2002. World Water and Food to 2025: Dealing with Scarcity. International Food Policy Research Institute, Washington, D.C.. Schindler, D.W., 1977. Evolution of phosphorous limitation in lakes. Science 195, 260–262. Sere´, C., Steinfeld, H., 1996. World Livestock Production Systems. Current Status, Issues and Trends. Animal Production and Health Paper 127, Food and Agriculture Organization of the United Nations, Rome, 82 pp. Smil, V., 1999. Nitrogen in crop production: An account of global flows. Global Biogeochem. Cycl. 13, 647–662. Smil, V., 2001. Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of World Food. MIT Press, Cambridge, 339 pp.

Tilman, D., Fargione, J., Wolff, B., D’Antonio, C., Dobson, A., Howarth, R., Schindler, D., Schlesinger, W.H., Simberloff, D., Swackhamer, D., 2001. Forecasting agriculturally driven global environmental change. Science 292, 281–284. United Nations, 2001. Road Map towards the Implementation of the United Nations Millennium Declaration. Report to the secretary general. A/56/ 326, United Nations General Assembly, New York. Van Breemen, N., Burrough, P.A., Velthorst, E.J., Van Dobben, H.F., de Wit, T., Ridder, T.B., Reijnders, H.F.R., 1982. Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature 299, 548–550. Van der Hoek, K.W., 1998. Nitrogen efficiency in global animal production. In: Van der Hoek, K.W., Erisman, J.W., Smeulders, S., Wisniewski, J.R., Wisniewski, J. (Eds.), Nitrogen, the Confer-N-s. Elsevier, Amsterdam, pp. 127–132. Van Drecht, G., Bouwman, A.F., Knoop, J.M., Beusen, A.H.W., Meinardi, C.R., 2003. Global modeling of the fate of nitrogen from point and nonpoint sources in soils, groundwater and surface water. Global Biogeochem. Cycl. 17 (4), 1115, doi:10.1029/2003GB002060. Vollenweider, R.A., 1992. Coastal marine eutrophication: principles and control. Sci. Total Environ. Suppl. 1–20.