a FAO, Animal Production and Health Division, Rome, Italy
b Environmental Research Group Oxford (ERGO), Oxford, UK
Whilst still in its infancy, the development of geographic information systems (GIS) and remote sensing (RS) tools is paving the way for global land use monitoring. This paper provides a first, tentative description of livestock related land use.
Preliminary results indicate that, in most of the developing world, the distributions of man and his livestock are closely related. This association is particularly prominent in India for large ruminants, and in South East Asia for pigs and poultry. For example, it is estimated that in China some 95 percent of the human population is concentrated in 50 percent of the land area, the humid plus subhumid ecozones. Here, the average density of 260 persons per square km coincides with a livestock biomass of approximately 25 metric tonnes, out of which 36 percent, or 9 metric tonnes, is made up of live weight pigs plus poultry. The densely populated, higher rainfall areas of China thus form the global epicentre of monogastric animals. It is in these areas also, that the highest animal protein production increase occurs. These local developments strongly affect the global statistics on animal agriculture; it is estimated that in the year 2010 poultry meat and pork will constitute about 70 percent of the developing world's total meat production.
With the progressive prominence of so-called land-detached, monogastric animal protein production in peri-urban environments in South East Asia, the question arises how to accommodate this development within the global agricultural landscape. While crop productivity gains may assist in overcoming the increased demand for cereals for livestock feed, the dynamic global animal production and associated change in land utilisation pattern interferes with a whole series of complex development issues, including poverty, environmental safety and emerging diseases' risk.
The progressive aggregation of people, crops and livestock in the developing countries requires clarification in order to come to grips with the complexity of interactions among economic, social and ecological processes. GIS and RS provide the tools for such undertaking.
Key words. Livestock geography, land use, animal agriculture, geographic information systems, monogastric animals, animal protein, Asia, China.
Livestock species are grouped according to their digestion system into ‘ruminants’ and ‘non-ruminants’ or ‘monogastrics’. This division reflects the differences in the animals' feed requirements and, thereby, the livestock related land use. Whilst most ruminants are confined to grazing areas, monogastric animals may be kept anywhere. Because monogastric and dairy animals produce perishable consumables there is a tendency to keep these animals in the proximity of the market. In developing countries, where cooling, storage and transport problems complicate the distribution of meat, milk and eggs to urban populations, there is a strong tendency to establish production units in areas of high population and land pressure, and to bring in the feed from where-ever is the most economic and convenient.
While ruminants may be kept in sparsely inhabitated areas and harsh environments unsuitable for crop production, these species are increasingly kept near human settlement and associated cropping areas (Bourn and Wint, 1994). Despite this progressive trend towards intensification, integrated livestock-crop (and fishery/forestry) production is believed to be still the most common farming system across the globe. These assumptions can, however, only be confirmed by analysing the spatial distribution of livestock in relation to agro-ecological zonation, and the distribution of people and crop production, much of which is lacking.
This paper argues the need for a quantitative geographic framework to characterise global livestock production. The work presented is drawn from a series of preliminary FAO studies carried out in 1996 (FAO 1996a, and b), which aimed to establish the data processing and analytical techniques and protocols required to collate and interpret the information available in a systematic fashion.
The primary data sources held by FAO, and potentially suitable for mapping, include:
The WAICENT global data set on Country Agricultural, Land and Demographic variables for the years 1962, 1972, 1982, 1992 and 1994, in spreadsheet format, available through the Internet site (http://www. fao. org/lim500/agri_db. pl).
Raster images of agro-ecological zones (AEZ) and national boundaries, converted to vector polygon format;
Raster images of global human population numbers at a resolution of 5 × 5 minutes.
The animal population data are available only at national level, and not for within country agro-ecological zones. In contrast, global human population is available, in image format, at a resolution of approximately 10 × 10 kilometres. By dint of relatively simple, if labour intensive, image processing and GIS techniques, these can be used to produce human population numbers for each agro-ecological zone within each country.
A number of studies on Sub-Saharan Africa have shown that livestock and human population levels are closely associated (Bourn and Wint, 1994). Where there are more people there are more animals. This link is sufficiently robust to be evident at a global and continental level. This is shown in Table 1, which gives the best fits from a choice of linear, logarithmic, power, compound and growth regression equations for each category, together with their significance statistics. These relationships were applied to the AEZ human population data to predict total animal biomass in each AEZ.
It is, however, unlikely that a single equation would satisfactorily describe the relationship between the numbers of people and individual animal species at a global level - there is too much regional variability in livestock management practices and the range of species kept. Therefore, the numbers of individual species were estimated by calculating the number (or weight) of each livestock species per person for each country, and then applying these ratios to the individual AEZ human population data to give the weight of each species in each AEZ in each country. Full details of the methods and analysis protocols are given in (FAO, 1996a and b).
Table 1: Relationships between livestock and human populations
|Africa||All||Power||51||543.28||. 000||1.9669||. 914|
|Asia||All||Power||40||887.52||. 000||1.9105||. 957|
|South Am.||All||Power||14||54.82||. 000||2.5727||. 797|
|North Am.||All||Linear||23||66.55||. 000||61.2857||. 743|
|Europe||All||Power||41||209.91||. 000||1.7860||. 837|
|Australia||All||Linear||1||177.05||. 048||779.06||. 994|
|Oceania||All||Linear||4||117.94||. 000||135.598||. 967|
|World||All||Power||177||1026.64||. 000||1.9336||. 853|
Linear equations of form y =ax + b1.Power equation of form: y = xb1.
y= Livestock Density (kg/km2). x= Human Population (no/km2). R2 is redefined, as it was assumed
that there are no livestock where there are no people so the equations were forced through Zero
* All Species; SR = Small Ruminants; LR=Large Ruminants; NR = Non-Ruminants (Pigs andPoultry)
The distribution of different categories of livestock between continents shows Asia to be prominent in ruminants, poultry and pigs: it supports approximately 64 percent of the global livestock biomass. It should be stressed that the biomass distributions shown here reflect the weight and, thereby, the number of animals present, but does not provide information on the density of the animals, the productivity level or the volume of production. The latter two parameters indicate a different geographical focus with the emphasis on developed countries, particularly Western Europe (FAO 1996b).
Figure 1 also provides an indication of temporal changes in the livestock biomass levels. Pigs have increased most substantially in Asia, since 1962, though in recent years their biomass has stabilised. This suggests that any increases in production have been derived from rises in productivity. Poultry populations have risen in all continents, though by far the most strongly in Asia, both in absolute and relative terms. The number of ruminants, by contrast, have increased in the developing world, but have either remained static, or fallen, in the developed nations.
Figure 1: Livestock Biomass by Continent
Map 1: Livestock Biomass Density in Asia, 1994
The relevance of this geographic breakdown becomes even more apparent when Map 1 is considered, depicting the distribution pattern of livestock within Asia, with the major foci in India and China Further interrogation of this data base in a geographical information system (GIS) shows these two concentrations are comprised of ruminant and monogastric animals respectively.
The evolution of the ruminant versus monogastric livestock biomass in Asia according to ecozones is substantiated in Figure 2. There is a clear tendency towards a replacement of ruminants by monogastric livestock in wetter areas (see Map 2). However, some caution is required when interpreting this trend because the progressive aggregation of monogastric livestock in humid and moist subhumid areas in Asia is not universal, and appears not so much influenced by the climatic conditions in which the animals are kept but, rather, by human preferences or anthropogenic factors.
Map 2: Agro-ecological Zone in Asia
Figure 2: Non-Ruminants as a % of Total Biomass by Agro-Ecological Zone
The importance of the anthropogenic influence in determining non-ruminant distributions can be illustrated by geographical analysis of biomass distributions in relation to human population density. Both the poultry and pig biomass is shown in relation to human population in Asia (Figure 3) and compared to that of the remainder of the world (Figure 4). The accompanying pie charts detail the land area proportions corresponding with the respective human population density classes. In Asia, there is a very pronounced aggregation of both types of monogastric animals in the densely population areas. In the rest of the world, in contrast, this aggregation is less obvious, in that the non-ruminants are more evenly distributed in relation to human population density.
Interrogation of geographical information system data base shows that in China in 1994 approximately 95 percent of the total population (1.19 billion people) was distributed over 50 percent of the land area (4.57 million square kilometres), coinciding with the humid and subhumid ecozones. Here the vast majority of all the monogastric animals are kept. Thus, the wetter area of China forms the global epicentre of monogastric animals, with 9 tonnes live weight of monogastrics per square kilometre, in addition to an average of 260 people.
Total Poultry Weight (Tonnes) in relation to Human Population Density (No/Km2)
Total Pig Weight (Tonnes) in relation to Human Population Density (No/Km2)
Total Poultry Weight (Tonnes) in relation to Human Population Density
Total Pig Weight (Tonnes) in relation to Human Population Density (No/Km2)
Non-Ruminants as a Percentage of Total Livestock Biomass
As shown in Figure 5, monogastric animals constitute an increasing percentage of global livestock biomass. This trend is more marked for the continent of Asia as a whole. If, however, China is examined separately, this proportion is substantially higher than elsewhere, but has been static for the last 10 to 15 years. It should be noted that this apparent stability results from a rise in total livestock biomass levels in China, and obscures the absolute rise in non-ruminant biomass.
The results suggest that the distribution of livestock biomass is strongly influenced by human factors rather than climatic features or ecozone (LGP). The relative increase in the numbers of monogastric animals in Asia is closely associated with the areas of highest human population pressures. An extrapolation of current trends suggest that in the year 2010 poultry and pig meat would comprise 70 percent of the total meat production in the developing world (FAO, 1995). In order to understand the implications of the animal protein production upsurge in Asia it is necessary to consider the changes in (i) the demand for the individual animal protein products such as eggs, milk and meat, (ii) the animal productivity levels and (iii) the related grain and land requirements.
Demand for animal protein
FAO (FAO, 1996c) has grouped the world's nations into six different classes according to major food commodities consumed per capita:
|1. Rice;||2. Maize;|
|3. Wheat;||4. Milk/Meat/Wheat;|
|5. Millet/Sorghum;||6. Cassava/Yams/Taro/Plantain.|
Class 1 is mainly made up of countries from the Asia region with a high rice consumption and a relatively high proportion of animal protein in the diet that is increasing faster than anywhere else in the developing world. Despite this rapid growth, consumption levels of pork and ruminant meat is still only 29 and 10 percent respectively of those in developed countries. (FAO, 1995) Class 2 and 3 embrace diverse developing countries; class 4 includes mainly developed countries whilst the last two classes largely comprises the countries of inter-tropical Africa.
When these dietary habits and trends are considered in conjunction with the ongoing economic growth and rising income levels, and when taking into account the income and price elasticities of the demand for livestock products, it is not surprising that Asia shows the highest growth in animal protein production of all world regions, for poultry meat, eggs, pork, red meat and milk alike.
This increase is unlikely to proceed indefinitely. Consumption per capita is likely to eventually stabilize or become balanced with more fibers, to form what is perceived as a healthier diet. However, because of the projected absolute increases in number of people in Asia, the total consumption of animal protein, once this has attained its highest level, may not diminish much for a considerable period of time.
Efficiency of production
Equally dynamic are the levels of livestock productivity. With the continued advance of genetics, feeding, health care and management, it may be assumed that the level of productivity, measured in terms of efficiency of feed conversion, is likely to rise quite spectacularly in most developing countries in the short to medium term. The present level of productivity, estimated as the ratio of product to ‘standing crop’ of animals, suggests a predominance of low to medium input level production systems. This is corroborated by the figures on the sales of concentrate feed and veterinary products in South-East Asia which show the highest growth rates in the world, though they are still relatively low compared to intensive systems in developed countries (Yeo, 1996).
For the industrial systems it would appear that the end of the intensification trajectory has almost been attained. The conversion of 1.75 kg of concentrate feed into one kg of live weight chicken, which is nowadays feasible in modern production units, has approached physiological limits.
Feed grain and arable land requirements
In the developed world, about 65 percent of the total agricultural land area is allocated to the production of cereals for livestock feed (FAO, 1995). Thus, with emerging economies and middle-income countries in Asia following suit, animal agriculture would become the most important farming activity world-wide. However, not all the required grain can be grown in Asia. A recent analysis of different grain supply and demand projections models on China suggested most models to be plainly wrong but, interestingly, the better models showed similar results with grain production in the year 2010 ranging from 389 to 486 and the demand from 468 to 513 million metric tonnes. These findings suggest that China will depend on substantial although not excessive grain imports. (Shenggen Fan and Mercedita Agcaoili-Sombilla, 1997). It would appear that, globally, there is sufficient arable land to facilitate such development. Before 1960, most increases in grain production resulted from expanding acreage under cultivation, whilst between 1960–1990, at least 80 percent of the 110-percent increase in cereal production came from raising yields per acre. Improved cereal yields have thus kept the global area actually harvested for grain at around 600 million hectares. These increases in yields have largely derived from the use of improved varieties of cereals, a 70-percent expansion of irrigation (from 100 to 170 million hectares), and a more than tripling use of chemical fertilizer (Action Group on Food Security, 1994).
While there may globally be sufficient land resources to allow for the dynamic expansion of animal agriculture world wide there will be a series of both positive and negative effects on a wide range of factors including food and income security, environmental preservation and, last but not least, health. Regarding the latter, many diseases have developed varying degrees of resistance to anti-microbial drugs and the WHO has reported (WHO, 1996) at least 30 new diseases that have emerged in the last 20 years. Apart from increases in world trade and traffic it is believed that the quickening pace of emergence and re-emergence of infectious diseases may also be associated with the growing densities of man and livestock. A recent review (Plotkin, 1997) suggested that the multi-sectoral factors causing disease emergence means that the present policy and legal framework to address them is largely inadequate. This is illustrated by recent experiences in Europe with BSE-CJD, as well as infectious livestock plagues such as foot and mouth disease and classical swine fever.
The global village
A progressive rise of global animal agriculture will have mixed effects on the socio-economic and biophysical environments. Though medium-high input levels in peri-urban animal protein production will become more grain costly, this may be an inescapable transition stage towards the prevalence of higher resource use efficiencies, before economic and environmental goals can coincide.
The expected evolution of the global agricultural landscape is likely to be accelerated by recent international trade agreements. Globalization demands increased efforts to come to grips with the complexity of interactions among economic, social and ecological processes. Where agriculture is concerned, there is a clear geographic dimension to this type of analysis. As the present work shows, the dynamics in agriculture are highly relevant in the global resource management context.
Clarifying land pressures geographically; informed decision taking
Understanding global agriculture in turn demands a comprehension of the distributions of livestock and associated land use patterns that can allow policy makers and planners to make informed, data driven, decisions. This implies analyzing the available data in its geographical context, as illustrated by the preliminary work presented here. This approach can, however, only be effective if it is based on reliable, geo-referenced data, but will allow all available information - both remotely sensed, and that derived from more traditional types of agricultural statistics, to be integrated and evaluated effectively.
Action Group on Food Security (1994). “Feeding 10 billion people in 2050. The Key Role of the CGIAR's International Agricultural Research Centers”, A Report by the Action Group on Food Security, pp. 14.
Ausubel, J. H. (1996). Can Technology Spare the Earth?, American Scientist, Volume 84, 166–177.
Bourn, D. And Wint, W. (1994). Livestock, Land Use and Agricultural Intensification in Sub-Saharan Africa, ODI Network Paper 37a, London: Overseas Development Institute, pp. 22.
FAO (1995). World Agriculture: Towards 2010; An FAO study, Edited by Nikos Alexandratos, Food and Agriculture Organization of the United Nations, and John Wiley & Sons, Chichester, England. pp. 434–435.
FAO (1996a). Livestock Geography; A Demonstration of GIS techniques applied to Global Livestock Systems and Populations; Consultant report by William Wint. Food and Agriculture Organization, Rome, Italy. pp. 77.
FAO (1996b). Livestock Geography II; A Further Demonstration of GIS techniques applied to Global Livestock Systems, Populations and Productivity; Consultant Report by William Wint. Food and Agriculture Organization. Rome, Italy. pp. 90.
FAO (1996c). Food requirements and population growth, FAO, Technical Paper WFS 96/TECH/12 Provisional version May 1996, Food and Agriculture Organization, Rome, Italy. pp. 59.
FAO (1997). Food Outlook, global information and early warning system on food and agriculture. No. 1/2 1997, Food and Agriculture Organization, Rome, Italy. pp. 22.
Plotkin, B. J. and Kimball, A. M. (1997). “Designing an International Policy and Legal Framework for the Control of Emerging Infectious Diseases: First Steps”, Emerging Infectious Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA, USA, URL: http://www. cdc. gov/ncidod/EID/vol3no1/plotkin. htm, Posted: February 5, 1997.pp. 11.
Shenggen Fan and Mercedita Agcaoili-Sombilla (1997). “Why do projections on China's future food supply and demand differ?”, EPTD Discussion Paper No. 22, Environment and Production Technology Division, International Food Policy Research Institute, Washington, D. C., U.S.A.
WHO (1996). World Health Organization Division of Emerging and Other Communicable Diseases Surveillance and Control, Emerging and other communicable diseases strategic plan 1996–2000, WHO/EMC/96. 1., World Health Organization, Geneva.
Yeo, A. (1996). Vaccines & pet products take worldwide honours, Animal Pharm Review of 1996, PJB Publications Ltd 1996.