Nitrogen Flow Model As a Tool for Evaluation of Environmental Performance of Agriculture and Effects on Water Environment

Journal of Developments in Sustainable Agriculture 8: 1-12 ( 2013)

Nitrogen Flow Model as a Tool for Evaluation of Environmental Performance of Agriculture and Effects on Water Environment

Junko Shindo*

Graduate School of Medicine and Engineering, University of Yamanashi, Takeda 4-4-37, Kofu, Yamanashi 400-8510, Japan

Based on the changes in food demand and production in East Asian counties, the temporal trends from 1980 to 2007 and spatial variation of nitrogen outflow via three channels (nitrogen balance (NB) in agricultural areas, human waste and atmospheric deposition) were estimated with a simple nitrogen flow model. In many countries, food consumption has rapidly increased due to population growth and increase in per capita food demand, especially demand for animal protein. In contrast, per capita food demand in India was almost stable despite economic growth. Re- flecting these food demands, consumption of inorganic nitrogen fertilizer and resulting nitrogen balance (NB) per unit agricultural land have increased drastically in many countries. In some countries such as Cambodia, Laos and Myanmar, NB was very low or negative even in 2007 suggesting that soil fertility may become exhausted. NB varies significantly within countries and it was extremely high in cities and provinces in China facing the East China Sea and the western Plain of Hindustan in India. Total nitrogen outflow from river basins, which is the sum of the outflows from the three channels in each 0.5° ×0.5°grid cell in the basins, also showed large variability. The Yangtze River basin accounted for the largest share of the total nitrogen outflow, about 20% of the total load of the study area. Comparing the average nitrogen outflow per unit land area in 2007, the Huai River basin had the largest value by a wide margin and may be severely polluted with nitrogen.

Key words: Nitrogen flow model, nitrogen balance, dietary transition, water pollution, East Asia

─────────────────────── was 8.9 t ha−1 in the Netherlands (FAO, 2012a). Ac- 1. Introduction cording to Smil (2001), the total amount of inorganic The industrial synthesis of ammonia by Haber and nitrogen fertilizer used in 1900 was estimated to be Bosch at the beginning of the twentieth century has about 340,000 t N, which was supplied from Chilean been said to be the most fundamentally important nitrates and recovery of ammonia from coke ovens. technical invention of that century. The world popu- Consumption increased to 11.0 million t N in 1961 and lation of 1.6 billion in 1900 could not have expanded to to 105.9 million t N in 2010 (FAO, 2012b). the current population without this invention (Smil, By the end of the 1980s, it was recognized, however, 2001). Application of inorganic nitrogen fertilizer that water quality had deteriorated severely over wide coupled with breeding of high-yield crop varieties has areas and that agriculture was a major contributor of enabled us to enhance crop yield markedly. Global the pollution (EEA, 2005; OECD, 2001). In Europe, average of wheat yield, for example, which was about various measures have been taken to reduce nitrogen 0.8 t ha−1 at the end of the nineteenth century (Smil, pollution due to agriculture including designation of 2001), increased to 1.1 t ha−1 by 1961 and to 2.8 t vulnerable zones, establishment of codes of good agri- ha−1 by 2007. The highest yield in the world in 2010 cultural practice for fertilizer application and manure

Received: September 11, 2012, Accepted: December 22, 2012 * Corresponding Author: Graduate School of Medicine and Engineering, University of Yamanashi, Takeda 4-4-37, Kofu, Yamanashi 400-8510, Japan. Tel: ＋81-55-220-8833, Fax: ＋81-55-220-8833, E-mail: [email protected] 2 J. Dev. Sus. Agr. 8 (1) management based on the nitrate directive imple- etal protein in 2007 was only 1.1 times the 1961 level mented in 1991 by the European Economic Commu- and has been decreasing recently. Changes in food nity (European Commission, 2002). The Organization consumption in Japan were similar at one time: per for Economic Cooperation and Development (OECD) capita food consumption increased until the mid- also developed an indicator ‘Nitrogen balance’ to 1980s and per capita animal food consumption until evaluate the environmental performance of agriculture the mid-1990s. Since then, however, consumption has of each member country (OECD, 2001). For the remained almost stable (Shindo et al., 2009). Dietary 2002-2004 period, South Korea and Japan had the transition to less vegetal staple food and more animal highest and the fourth-largest NB, respectively, per food along with increased total food consumption took hectare of agricultural land including pasture among place before the 1990s in Japan, whereas it has been the 30 OECD member countries (OECD, 2008). progressing since 1980s in China. Recent changes in In Asian countries other than Japan and South China’s food consumption seem to indicate, however, Korea, nitrogen status has also been drastically chang- that the dietary transition in China is coming to an end ing: increase of inorganic nitrogen fertilizer use has because the per capita total protein intake has shown been more rapid there than in other regions of the a slightly decreasing trend and per capita meat pro- world for the past several decades and more than 60% tein has been almost stable since 2000. In many other of global nitrogen fertilizer use is in Asia. This is due East Asian countries, increased meat consumption is to rapid economic growth and growing population that marked especially since 1980s (Table 1). Per capita have demanded more food. Therefore, concerns have meat consumption in Vietnam and Cambodia in 2007 been raised regarding the current situation and future was 4.4 times those in 1980; in South Korea, 4.2 times; trends for nitrogen flow and environmental effects and in Myanmar, 3.8 times. However, the absolute (Galloway, 2000; Zheng et al., 2002; Yamaji et al., values for meat consumption differed greatly from 2004; Chen et al., 2010). country to country. This study evaluates nitrogen flow in East Asia for In contrast for India, although total food consump- 13countries ranging from Japan to India (east to west). tion in 1980 had increased to 2.5 times the 1961 level Yearly changes in nitrogen balance were estimated to of 1.5 million t N to 4.0 million t N (Fig. 1b), this in- evaluate the environmental performance in agriculture crease was mainly due to population growth. Per of individual countries with a simple nitrogen flow capita total protein consumption has remained low model. The model also estimated nitrogen outflow despite recent economic growth (Table 1). Per capita from the grid cells within large river basins and ni- animal protein intake, dominated by milk, was still low trogen concentration in river water caused by nitrogen (10 g day−1) in 2007 although it has been gradually in- outflow via nitrogen balance in agricultural land, creasing. Per capita meat consumption has remained human waste and atmospheric deposition to evaluate quite low (1.2 g day−1 in 2007). Mahendra et al. (2004) potential environmental effects from nitrogen assoc- reported that per capita calorie and protein con- iated with food production and consumption. sumption decreased during the economic reform period from 1990 to 2000 in India based on consumption data 2. Food consumption and dietary from the National Sample Survey Organization in transition in East Asia India. Bangladesh and Indonesia also show low rates Food consumption has dramatically increased in of increase for animal protein consumption. many East Asian countries according to statistics 3. Nitrogen flow model (FAO, 2012c). In the case of China, food consumption has increased more rapidly than population growth The food consumption rate and its yearly change (Fig. 1a) and per capita protein intake has become 1.7 vary from country to country and hence the environ- times the 47 g capita−1 day−1 in 1961 at 80 g capita−1 mental effect due to agricultural nitrogen should differ day−1 in 2007 as shown in Table 1. As shown in Fig. spatially and temporally. In order to evaluate the po- 1a, the increase in meat consumption is marked since tential nitrogen effects associated with agriculture and 1980s: per capita intake of meat protein was 1.2 g food consumption based on these temporal trends for day−1 in 1961, 4.8 g day−1 in 1980 and 17.2 day−1 in food demand along with food production, trade, 2007. In contrast, the per capita consumption of veg- fertilizer use, etc. by using the statistical data for each Shindo: Nitrogen Flow Model for Evaluating Environmental Performance of Agriculture 3

Fig. 1. Annual human food consumption in nitrogen equivalent and population calculated based on the Food Balance Sheet data derived from FAO statistics (FAO, 2012c) and the change in population in (a) China and (b) India from 1961 to 2007.

country, a simple nitrogen flow model was created timated by using equations based on existing knowl- (Shindo et al., 2003, 2006, 2009). Such a model can edge and assumptions. The methods for estimating be used to evaluate future water quality trends or these flows were described in detail in our previous effects of measures taken to reduce nitrogen outflow paper (Shindo et al. 2009). Some portion of nitrogen from agricultural land based on assumed scenarios. was assumed to accumulate as soil organic matter in The scheme of the model is shown in Fig. 2. We the agricultural soil and the remaining nitrogen is assumed that nitrogen input as inorganic nitrogen exported to the sea through the groundwater layer and fertilizer (Nfert), biological fixation (Nfix), food import, river systems. During the export process from soil to etc. circulates in the food supply and consumption the sea, a portion of the nitrogen is assumed to be re- system and flows out to the environment via runoff and moved from the water by denitrification and by im- leaching from agricultural fields (NBG), human waste mobilization. These processes were assumed to disposal (NHW) and atmospheric deposition. In ad- proceed by a first-order reaction, depending only on dition to atmospheric ammonia deposition, NOx de- temperature and the residence time in the soil and position originating from fossil fuel combustion is also groundwater layers. included in the model. The magnitude of the flows indicated by arrows were directly derived from statistical data, calculated by budget calculations or es- 4 J. Dev. Sus. Agr. 8 (1)

Table 1. Changes in per capita protein intake in East Asian countries

Total Meat protein Change in consumption rate g protein capita−1 day−1 g protein capita−1 day−1 g protein capita−1 day−1 y−1 1961 1980 2007 1961 1980 2007 1961-1980 1980-2007 Bangladesh 40.8 40.9 49.0 1.0 0.8 1.2 −0.01 0.02 Cambodia 43.8 36.5 57.0 1.7 1.2 5.5 −0.02 0.16 China47.2 59.3 80.3 1.2 4.8 17.2 0.19 0.46 India 54.6 48.4 59.2 1.4 1.4 1.2 0.00 −0.01 Indonesia 37.6 48.2 56.3 1.2 1.6 3.8 0.02 0.08 Japan 74.7 85.7 91.1 2.4 9.7 14.9 0.38 0.19 South Korea 51.4 78.7 85.2 1.3 4.1 17.5 0.15 0.49 Laos 51.9 52.2 62.1 2.9 3.1 5.9 0.01 0.10 Malaysia 45.3 57.8 72.2 4.4 7.8 16.4 0.18 0.32 Myanmar 46.1 63.3 79.4 1.7 2.5 9.6 0.04 0.26 Philippines 39.3 49.0 57.4 4.0 5.4 10.4 0.07 0.18 Thailand 41.4 50.2 56.0 4.9 6.2 9.5 0.07 0.12 Vietnam 45.6 44.6 70.4 3.4 2.9 12.6 −0.03 0.36

Fig. 2. Schematic diagram of the nitrogen ﬂow model for food supply and consumption.

Nfert, nitrogenous chemical fertilizer; Nfix, biological nitrogen fixation; Ncrop, nitrogen uptake; Nbyproduct, nitrogen in crop byproduct; NNH3_VF,NNH3_VL, ammonia volatilization from fertilizer and livestock waste; NNH3_D,NNOx3_D, ammonia and NOx deposition; NLW,NHW, nitrogen in livestock waste and human waste; NBPmin,NLWmin; mineralized nitrogen from byproducts and livestock manure; NBG, nitrogen balance. Shindo: Nitrogen Flow Model for Evaluating Environmental Performance of Agriculture 5

half of nitrogen input in Bangladesh, China, India, 4. Nitrogen balance as an indicator of Indonesia, Malaysia and Vietnam. For the countries environmental performance other than Indonesia among these 6 countries and for of agriculture Thailand, NBA and NBF have become significantly Among the three channels of nitrogen outflow to the larger during the past 27 years. NBF is much larger environment dealt with by the nitrogen flow model, the than NBA for China (Fig. 3b) and this discrepancy is outflow from agricultural land is conceptually the same caused by the significantly larger NBF than NBA in as ‘nitrogen balance (NB)’ that has been evaluated as Hunan, Sichuan and other provinces located mainly an indicator of environmental performance of agricul- in the western region (Fig. 3c). NBA for China was ture for OECD member countries including Japan and about one third of that for Japan in 2007, whereas NBF South Korea. According to the OECD methodology, slightly exceeded the Japanese value. Spatial vari-

NB is defined as the difference between the total ability NBA and NBF were large among and within quantity of nitrogen input for the agricultural land due countries (Fig. 3b, c). In China, the NBA and NBF to inorganic fertilizers, livestock manure, biological values in three direct-controlled municipalities (Bei- nitrogen fixation and atmospheric deposition and the jing, Tianjin, and Shanghai) and all provinces facing total quantity of outputs due to uptake by harvested the East China Sea, such as Jiangsu and Fujian, were food, fodder crops and by pasture. OECD evaluated extremely high. This spatial variation was mainly the nitrogen balance for each country as gross NB caused by the uneven application rate of inorganic −1 (NBG;tNyear ) and as the NB per unit agricultural fertilizer to farmland, ranging in 2007 from more than −1 −1 −1 −1 −1 land area including pasture (NBA;kgNha year ) 500 kg ha y in Jiangsu Province to 70 kg N ha (OECD 2001, 2008). y−1 in Jilin Province. Similar large spatial variability I evaluated NB for countries in East Asia at the was observed also in India: application rates varied country-scale, at the region-scale within a country and from 343 kg N ha−1 y−1 in Punjab to 0.6 kg N ha−1 at grid-scale by using the nitrogen flow model men- y−1 in Nagaland, and states with high application rates tioned in the previous section (Shindo, 2012). Re- of inorganic nitrogen fertilizer were mainly in the garding the NB per unit agricultural land, NBA values plains of Hindustan. based on the definition by OECD methodology were 5. Potential nitrogen outflow too small for the regions such as China with vast from agricultural land unfertilized pasture because pasture was counted as agricultural area, and I considered it was not a sensitive Gross nitrogen balance (NBG) is considered to be an indicator of too much or too little nitrogen for such indicator of potential nitrogen load to the hydrosphere areas. I proposed an alternative indicator, NBF,which and atmosphere. Based on the NBG estimated for in- assumes that inorganic nitrogen fertilizer was applied dividual 0.5°×0.5°grid cells, potential nitrogen out- only to farmland (arable and permanent crop land, not flows to rivers were evaluated for large river basins. including pasture) and that nitrogen from livestock Fifteen river basins were identified from Total Runoff waste was added evenly to the agricultural land Integrating Pathways data (Oki and Sud, 1998), which (including pasture) within a region, i.e., by country, indicate outflow direction for each 0.5°×0.5°cell in district, or grid cell depending on the estimation scale continental East Asia (Fig. 4). Three river basins in (Fig. 3a). In Cambodia, Laos and Myanmar, applica- Japan (Ishikari, Shinano and Tone) were also evaluated tion rates of inorganic nitrogen fertilizer were low and for the comparison, though these basins contain only 6, as a result, nitrogen from biological nitrogen fixation 5 and 7 grid cells, respectively. The basic information was the largest input. For these countries NBA and for these basins is shown in Table 2. For the Amur NBF were very small or negative, which suggests that River basin, only the area within Chinese territory was crop cultivation is strongly dependent on soil fertility studied and NBG from Nepal and Bhutan were not and organic fertilizer and that soil fertility may become considered for the Ganges River and the Brahmaputra exhausted. The amount of nitrogen from livestock River basins. Of the global population, 27.6% live in wastes are comparable to those from inorganic fer- these river basins; there are 1554 million ha of farm- tilizer in Japan and South Korea, whereas inorganic land (14.7% of global farmland) and 34.8 million t N fertilizer is the largest input and accounts more than of inorganic nitrogen fertilizer (34.6% of global con- 6 J. Dev. Sus. Agr. 8 (1) sumption) was used in 2007. The amount of inorganic sin has the highest ratio of agricultural area including nitrogen fertilizer consumption tripled from 1980 to pasture (RA). Average application rate for fertilizer −1 2007. The proportion of farmland area (RF) is much was 137.2 kg N ha (per farmland) in 2007 and the higher in the Ganges, Godavari and Krishna River values were especially high in the Yellow, Huai, basins in and around India and in the Huai River basin Yangtze and Pearl River basins in China. in China than the other basins. The Yellow River ba- The total NBG in each river basin is shown in Table 3with average NB G per unit land area, average NBF and estimated nitrogen concentration in river water in

2007. NBG has increased drastically over a wide area (Fig. 5). Total NBG of these 18 river basins was 7.2 million t N in 1980, which increased to 27.1 million t N in 2007. The Yangtze River basin accounted for the

largest share of the total NBG both in 1980 and 2007, followed by the Huai River basin in 1980 and by the Ganges River and the Huai River basins in 2007. The Yangtze River basin in 2007 accounted for about 20% of the total load of the study area and about one third of the total load of the studied basins. The rate of in-

Fig. 3. Estimation of nitrogen balance (NB). (a) Proportions of nitrogen input from various sources in

2007, (b) NB per unit agricultural land area (NBA) and per unit farmland area (NBF) of each country in 1980 and 2007, (c) NBA and NBF of each province of China in 2007. Thick and thin Error bars represent the standard deviation of NBA and NBF, respectively, for the regions within Japan, China, and India. Nfert, nitrogenous chemical fertilizer; Nfix, biological nitrogen fixation; NNH3_D,NNOx3_D, ammonia and NOx deposition; NBPmin,NLWmin; mineralized nitrogen from byproducts and livestock manure; NBA, NBF, nitrogen balance per agricultural land and per farmland. Shindo: Nitrogen Flow Model for Evaluating Environmental Performance of Agriculture 7

Fig. 4. River basins modeled in this study. River basins: 1 Amur River, 2 Liao River, 3 Yellow River, 4 Huai River, 5Yangtze River, 6 Pearl River, 7 Red River, 8 Mekong River, 9 Chao Phraya River, 10 Salween River, 11 Irrawaddy River, 12 Brahmaputra River, 13 Ganges River, 14 Godavari River, 15Krishna River

Fig. 5. Spatial distribution of the estimated amount of potential nitrogen outflow into each grid cell due

to agriculture (NBG) for each 0.5°×0.5°grid cell in (a) 1980 and (b) 2007. The area of each grid cell is roughly 2,500 km2, depending on latitude. Gray color in the study area (for countries other than North

Korea, Nepal and Bhutan which are out of my study area) indicates negative NBG values. crease was high for Southeast Asian river basins such the Ishikari River basin in Hokkaido but decreased in as the Red and Mekong Rivers, where NBG was quite the other two basins. Average NBG in the Huai River low or even negative in 1980. NBG changes in the basin in 1980 was similar to the Tone River basin in basins in Japan were small: NBG increased a little in Japan in 1980 (Fig. 6). The average NBG for the Huai 8 J. Dev. Sus. Agr. 8 (1)

Fig. 6. Potential nitrogen outflow to the environment from each river basin from gross nitrogen balance

(NBG), human waste (HW) and atmospheric deposition in (a) 1980 and (b) 2007. NNH3_D,NNOx3_D, ammonia and NOx deposition; NHW, nitrogen in human waste; NBG, nitrogen balance.

River basin increased rapidly and reached 150 kg N river basin (Table 2) and human protein intake also ha−1 in 2007, significantly higher than the other river increased in many Asian countries. However, nitrogen basins in 2007. The Huai River basin is rather small outflows from human waste were much smaller than and the main provinces located in the basin are NBG for almost all river basins except for the Tone and Shandong, Henan, Jiangsu, and Anhui provinces that the Shinano River basins in Japan (Fig. 6). Thus, have high population densities (Table 2). Very inten- changes in nitrogen outflow to the environment at large sive agriculture is practiced in the Huai River basin and spatial scales were dominated mostly by agricultural the value of NBF is also the largest among the basins practices in East Asian river basins. Nevertheless, (Table 3). Although NBF in the Tone River basin is human waste can cause severe municipal water pollu- the second highest, average NBG is not very large tion due to inappropriate treatment and disposal. because of the low ratio of farmland in the basin. In 6. Estimation of water pollution contrast, the river basins in India showed relatively in Asian river basins high NBG values that I attribute to the large farmland area and moderate NBF. In the river basins in South- Based on the total nitrogen outflow from each grid east Asia, negative NBG were observed in considerable cell, nitrogen concentration in river water was esti- areas in 1980 and the Irrawaddy River basin still had a mated by a simple nitrogen export model. Areas with negative NBG in 2007. high river water nitrogen concentrations include the From 1980 to 2007, population increased in every Huai River basin and the lower reaches of the Yangtze Shindo: Nitrogen Flow Model for Evaluating Environmental Performance of Agriculture 9 2 76 20 95 80 88 152 630 128 258 219 per 216 258 649 239 179 208 348 567 km 2007 6.6 2.4 1.2 61.9 23.2 93.2 38.6 29.4 35.1 1841 79.2 10.4 75.8 69.5 2007 118.0 168.2 460.3 148.3 419.9 6 Population 10 4.7 1.1 2.2 9.5 87.1 39.0 25.9 16.4 68.7 26.4 19.1 86.4 51.3 45.1 47.1 1243 1980 124.2 339.7 249.2 9.8 58.3 65.2 67.9 75.6 80.5 70.7 98.6 2007 175.5 390.7 226.2 315.2 255.3 149.9 100.1 103.5 123.6 103.5 kg N/ha 8.4 7.3 7.3 81.1 72.7 20.5 35.8 18.2 19.8 40.6 22.4 25.0 1980 154.8 158.1 168.7 125.6 104.2 105.2 73 87 33 15 25 Fertilizer 914 362 488 906 1058 2768 5123 2132 1829 5566 1208 1261 2007 10926 34769 1000 t 43 16 55 70 41 19 20 940 104 385 776 224 570 348 332 1879 4177 1384 1980 11384 3 3.2 19.6 19.3 30.6 24.0 30.6 15.7 49.1 21.6 19.4 19.4 21.4 75.1 20.1 2007 65.9 10.7 60.7 17.1 F R 2.4 13.1 24.2 14.5 32.7 13.7 15.3 44.5 18.3 15.1 19.8 16.2 59.8 13.6 74.8 24.7 1980 15.6 65.1 A % Agricultural area* R 26.0 39.2 25.8 86.7 34.8 55.7 49.4 31.3 23.4 51.9 62.2 78.5 20.9 27.5 2007 67.1 12.1 61.0 21.9 Basic information for the studied river basins 2 1 − 1.5 0.5 1.4 0.4 1.5 0.9 1.3 1.8 1.0 0.8 1.5 0.5 1.1 1.1 1.6 0.8 1.2 1.5 my Annual tation* Precipi- Table 2. 2 5.8 6.7 7.7 ℃ 0.5 9.7 3.8 18.4 19.7 20.4 13.7 25.2 11.2 10.1 21.4 22.7 11.9 26.4 26.1 temp* Yearly average 1 12 16 14 426 197 136 778 810 267 181 335 868 408 740 294 228 619 2 1782 8107 This study 1000 km 17 11 14 409 229 171 945 806 189 179 272 414 226 320 651 Basin Area* ture 1722 9521 1930 1016 Litera- 5 indicate the portion of agricultural area with and without pasture, respectively. F 3 4 and R A World Resources Institute, 2012 Calculated from grid data providedR by Climatic Research UnitData of for University Russia of is East notData Anglia. included for except Nepal the and basin Bhutan area are from not literature. included except the basin area from literature. 1 2 3 4 5 * * * * * Liao Red Yellow Mekong Huai Chao Phraya Yangtze Salween Amur* Pearl Godavari Irrawaddy Shinano Total Tone Ishikari Krishna Ganges* Brahmaputra* 10 J. Dev. Sus. Agr. 8 (1)

Table 3. NB and nitrogen concentration in river water for the studied river basins in 2007

NBG N concentration NBF Total (1000 tN) average average at river mouth

1980 2007 kgNha−1 kgNha−1 mgNL−1 mgNL−1 Amur447 1331 15 67.0 1.5 Liao 280 840 43 128.6 2.1 2.9 Yellow 514 2598 33 140.5 1.2 1.9 Huai 1224 4081 153 310.7 5.2 4.8 Yangtze 2811 8999 50 202.4 1.8 2.4 Pearl 590 1508 35 165.5 1.4 1.7 Red 16 330 24 99.0 1.0 1.0 Mekong −29 586 7 28.0 0.3 0.3 Chao Phraya −43 191 11 34.0 0.3 0.4 Salween 5 56 2 17.4 0.3 0.2 Irrawaddy −59 −182 −4 −21.7 0.2 0.1 Brahmaputra 78 663 11 34.8 0.5 0.8 Ganges 823 4085 55 72.0 1.2 1.4 Godavari 233 945 32 52.9 0.8 0.7 Krishna 188 946 42 62.6 1.7 1.4 Ishikari 24 40 29 148.7 1.1 1.0 Shinano 32 25 21 185.7 2.2 1.7 Tone 91 7949 238.7 3.0 3.2

Total 7227 27121 33

and the Yellow Rivers (Fig. 7 and Table 3). There are some other high concentration areas including north- eastern China around the Liao River basin and the northwestern plain of Hindustan in the Ganges River basin. It is difficult, however, to validate estimates for such wide areas. The model aims to estimate the average property of nitrogen flow in areas of considerable extent, while actual observations of concentration have large spatial variability in rather small areas. Hence, we need measurements at a considerable number of sites for validation. In the case of Japan, there are more than 2000 monitoring sites for surface water quality and the prefectural average of total nitrogen concentration corresponded well with the average of estimates for each 8×8 km grid cell calculated by this model (Shindo et al., 2009). For other countries, I compared the estimated concentration with the measurements in GEMS/Water Fig. 7. Spatial distribution of estimated river water database (UN GEMS/Water, 2003). Unfortunately there total nitrogen concentration in 2007. were no measured nitrogen concentration data in GEMS/Water database for the Huai River and the Ganges River basins for which high nitrogen concent- Shindo: Nitrogen Flow Model for Evaluating Environmental Performance of Agriculture 11

that must be reduced by improved agricultural practices in each region and country. The approach can be used to predict future environmental loads caused by food production and supply based on various scenarios including socio-economic situation, food demand, changes in land use and climate and to evaluate the effectiveness of countermeasures to improve the nitrogen use efficiency in crop and livestock farming. Acknowledgments This work was partly supported by the Global En- vironment Research Fund (B-0801) of Japan’s Min- istry of the Environment. I thank the Research In- stitute for Humanity and Nature and the Frontier Research Center for Global Change for providing the estimated NOx emission data for Asian countries and Fig. 8. Comparison of estimated average nitrogen the National Center of GEMS/Water in Japan for concentration of river basins with measurement data providing the water quality data for Asian countries. provided by the GEMS/WATER project (2003). References

Chen, M., Chen, J., Sun, F., 2010. Estimating nutrient releases rations were estimated by the model. Although the from agriculture in China: An extended substance flow data were not sufficient for evaluating the average analysis framework and a modeling tool. Science of the total environment 408, 5123-5136. water quality of the river basins, the measurement data European Commission, 2002. Implementation of Council Di- did support the general trends of the estimates (Fig. 8). rective 91/676/EEC concerning the protection of waters against pollution caused by nitrates from agricultural Conclusions sources. Luxembourg: Office for Official Publications of the European Communities, 44 pp., ISBN 92-894-4103-8. My nitrogen flow model has shown that nitrogen [EEA] European Environment Agency, 2005. Source appor- outflow to the environment has increased greatly over tionment of nitrogen and phosphorus inputs into the aquatic the last three decades in wide areas of East Asia due to environment. EEA Report No. 7/2005, ISSN 1725-9177. the rapid increase of food demand in many East Asian [FAO] Food and Agriculture Organization, 2012a. Production countries and these increased nitrogen outflows have database. http://faostat.fao.org/site/339/default.aspx, accessed August 2012. had a negative influence on water quality. Intensive [FAO] Food and Agriculture Organization, 2012b. Resources / agricultural areas draining into the East China Sea, Fertilizer database. http://faostat.fao.org/site/575/default.aspx especially in the Huai River basin and around the lower #ancor, accessed August 2012. reaches of the Yellow and Yangtze River basins, were [FAO] Food and Agriculture Organization, 2012c. Resources / estimated by model results to be severely polluted by Fertilizer database. http://faostat.fao.org/site/354/default.aspx, accessed August 2012. agricultural nitrogen. I attribute the severe water pol- Galloway J. N., 2000. Nitrogen mobilization in Asia. Nutrient lution in these areas to the excessive use of inorganic Cycle in Agroecosystems 57: 1-12. nitrogen fertilizer. Mahendra, Dev., S., Ravi, C., Viswanathan, B., Gulati A., These results have large uncertainties caused by the Ramachander, A., 2004. Economic Liberalisation, targeted programmes and household food security: A Case Study of limited data in each country and by the overly sim- India, MTID Discussion paper, No. 68, International Food plified model structure. Thus, there is much room for Policy Research Institute, Washington, D.C. further improvement. Nevertheless, this kind of esti- [OECD] Organisation for Economic Co-operation and Develop- mation is necessary, in order to evaluate the historical ment, 2001. Environmental Indicators for Agriculture, vol changes in food demand and food production and how 3. Methods and Results, OECD, Paris, p 409. [OECD] Organisation for Economic Co-operation and Devel- they have caused environmental stress from an in- opment, 2008. Environmental performance of agriculture tegrated viewpoint in terms of nitrogen flow. This in OECD Countries since 1990. OECD, Paris, p 575. ISBN approach can also identify the major nitrogen sources 978-92-64-04092-2 12 J. Dev. Sus. Agr. 8 (1)

Oki, T., Sud, Y. C., 1998. Design of total runoff integrating land in Japan and 12 other Asian Countries based on a pathways (TRIP)―a global river channel network. Earth nitrogen-flow model. Nutrient Cycling in Agroecosystems interactions, 2: paper 1, 1-36. 94, 47-61. Shindo J, Okamoto K, Kawashima H., 2003. A model-based Smil, V., 2001. Feeding the world: a challenge for the twenty- estimation of nitrogen flow in the food production-supply first century. (Japanese trans: Henmi K, Yanagisawa K) system and its environmental effects in East Asia. Ecol MIT Press, Cambridge. Model 169: 197-212. UNGEMS/Water, 2003.???????????? Shindo, J., Okamoto, K., Kawashima, H., 2006. Prediction of World Resources Institute. Watersheds of the world: an earth- the environmental effects of excess nitrogen caused by in- trends special collection. http://www.wri.org/publication/ creasing food demand with rapid economic growth in watersheds-of-the-world, accessed August 2012. eastern Asian countries, 1961-2020. Ecol Model 193: Yamaji, K., Ohara, T., Akimoto, H., 2004. Regional-specific

703-720. emission inventory for NH3,N2 O, and CH4 via animal Shindo, J., Okamoto, K., Kawashima, H., Konohira, E., 2009. farming in South, Southeast, and East Asia. Atmospheric Nitrogen flow associated with food production and con- Environment 38, 7111-7121. sumption and its effect on water quality in Japan from 1961 Zheng, X., Fu, C., Xu, X., Yan, X., Huang, Y., Han, S., Hu, F., to 2005. Soil Sci Plant Nutr 55: 532-545. Chen, G., 2002. The Asian Nitrogen Cycle Case Study. Shindo, J., 2012. Changes in the nitrogen balance in agricultural Ambio 31, 79-87.