ANNEX 2. IDENTIFY THE ENVIRONMENTAL CONSEQUENCES OF FARMING SYSTEMS OVER THE PAST 50 YEARS AND THEIR LONG-TERM SUSTAINABILITY
In this report we begin by summarising the broad impacts of agriculture on the wider environment. We then provide more detail on the priority areas: greenhouse gas (GHG) emissions, air and water quality, water resources and soil degradation, all of which need to be assessed when evaluating the sustainability of agricultural systems. We also summarize the interaction between agriculture and land use change (LUC), including ecosystem services and biodiversity. Finally we summarize the main environmental impacts of the farming systems identified in Task 1.
Summary of findings
The environmental impacts of farming systems over the past 50 years and their long-term sustainability have been evaluated with respect primarily to their impact on a range of environmental factors.
All farming systems may have adverse impacts on the environment; while it might be assumed that intensive systems have the greatest adverse environmental impact this is not always the case. Intensive production tends to emit less GHG per t of produce. This is mainly because intensive systems may use inputs more efficiently but also because intensive systems are not usually encroaching upon natural ecosystems and do not lead to indirect emissions from LUC. Intensive farming is also less likely to lead to land degradation since greater crop yields return more organic matter to land, hence better maintaining soil organic matter (SOM) and soil structure. Inputs of manures and fertilizers maintain soil fertility. However, intensive farming has greater impacts on local water and air quality and also makes greater use of water resources, principally via the need for irrigation water and also, to a much lesser extent, as a consequence of draining wetlands. The adverse impacts on air and water quality tend to result from intensive agriculture emissions being concentrated and causing point source pollution.
The key concerns for each farming system are:
Intensive arable
The major environmental impacts of intensive arable farming are a reduction in water quality and emissions of N2O arising from applications of fertilizer-N. However, in many EU countries inputs of fertilizer-N, the major source of emissions of N2O and NO3, have stabilised while yields of the major arable crops have continued to increase.
Intensive dairy farming
Intensive dairy production is a significant source of GHG but productivity can increase without corresponding increases in emissions. Large herd sizes and concentrations of livestock within small areas pose serious concerns for their impacts on water and air quality. In addition, indirect LUC may be stimulated by demand for ingredients such as soya or cassava for concentrated feeds.
Intensive livestock farming
Intensive livestock farming poses similar problems to those of dairy farming.
Intensive horticulture.
As horticulture, particularly vegetable production, becomes more specialised and intensive the large fertilizer applications required and intensive cultivation are likely to increase local water pollution.
Extensive beef and sheep grazing
Extensive livestock production is a significant source of the GHG CH4 and a driver to LUC. In addition, soil nutrient supplies can be depleted leading to a cycle of further extensification and demand for LUC to create new pastures.
Wetland rice cultivation
Irrigated rice cultivation is a major source of CH4 but changes to production methods are forecast to lead to no net increase in CH4 emissions to 2030 despite forecasts of substantial increases in production. However, rice cultivation is likely to continue to be a major source of N2O and NH3.
Irrigated
Without significant improvements in water use efficiency it will be difficult to greatly increase production from this farming system. In some regions maintaining current production may be difficult.
Smallholder rain-fed humid
Climate change is forecast to reduce rainfall in several of the regions which depend upon rain-fed smallholder production. Substantial areas are also prone to land degradation. Unless inputs can be made available to improve productivity and management to avoid degradation, there will be pressure for LUC.
Smallholder rain-fed highland
This farming type is particularly vulnerable to soil erosion. Dixon et al. (2001) suggest that this system has great potential for intensification, by using soil restoration methods and improved water management techniques. However, this may compromise the sustainability of the system itself.
Smallholder dry and cold
These systems often have low productivity due to environmental constraints including nutrient poor soils, low temperatures and lack of rainfall. This low input of precipitation explains the need for additional water abstraction in these areas, leading to water stress.
Dualistic mixed
Heavy demand for water in some areas is diminishing water resources while a combination of increased agrochemical inputs and decreased water flows is reducing water quality. In some areas soil reserves of P and K are not being maintained.
Fisheries
The loss of mangrove swamps to shrimp farming has had major adverse effects on coastal environments. The majority of sea fisheries are over-exploited and it may prove difficult to maintain current production.
1.Impacts of agriculture on the environment
Agriculture has changed the environment since the first farmers modified the existing habitat to cultivate crops and suppress competition from other plants and grazing by animals. As human populations have increased so has the replacement of natural ecosystems by agriculture. Of the world’s fourteen biomes, more than half have had 20-50% of their surface converted to croplands (Olson et al. 2001). Tropical dry broadleaf forest, temperate grasslands and flooded grasslands and savannas have had the greatest conversion to agriculture over the past 50 years. More than 50% of the global wetlands and 60% of major world rivers have been transformed or modified by humans over the past 100 years (Millennium Ecosystem Assessment 2005), reducing biodiversity through habitat flooding, disruption of flow patterns, and fragmentation of animal populations and travel corridors. Water abstraction from rivers in many regions of the world has significantly reduced flows and, in some cases, left some major rivers nearly dry.
As well as reducing biodiversity and modifying water supplies, the removal of plant cover and tillage of soil break down soil organic matter (SOM) and lead to substantial emissions of the GHG carbon dioxide (CO2). Breakdown of SOM, and release of CO2, may continue for decades after initial LUC and the process also releases nitrate (NO3) which can pollute watercourses. In recent years agricultural production has increased primarily through increased production per ha, reducing the pressure for LUC and consequent emissions of CO2 from soil. However, this has not eliminated the impact of increased agricultural production on GHG emissions. Increased production of arable and forage crops in many parts of the world has been made possible by increased use of mineral fertilizers, in particular N, and in the soil a proportion of that N is converted to the GHG nitrous oxide (N2O). The growing demand for livestock products has led to increases in emissions of the GHG methane (CH4) which arises primarily from enteric fermentation.
As agriculture intensifies soil quality can be reduced, in some cases leading to erosion, and further contamination of watercourses, and in extreme cases degradation such that the land can no longer be farmed. The input of agrochemicals and mineral fertilizers, especially nitrogen (N) can lead to further reductions in water quality due to increased NO3 leaching, phosphate (P2O5) enrichment and pesticides. Emissions to air may also be increased not only as N2O but also as ammonia (NH3). To be able to improve the sustainability of global production systems, it is necessary to first identify the environmental impacts of the systems.
1.1Greenhouse gases
Table 1 below presents current estimates of CH4 emissions expressed per litre (L) of milk produced in different world regions. There are very large differences among regions and that greater productivity tends to emit less CH4 per L.
Table 1. Default CH4 in different regions, per L of milk yield, from IPCC 2006 table 10.11, dairy cattle
Enteric CH4kg head-1yr-1 / Milk L head-1 yr-1 / g CH4 L-1NAFTA / 121 / 8400 / 14.4
EU / 109 / 6000 / 18.2
CIT / 89 / 2550 / 34.9
Oceania / 81 / 2200 / 36.8
Latin America / 63 / 800 / 78.8
N Africa, mid-East / 40 / 475 / 84.2
S Asia / 47 / 900 / 52.2
E and SE Asia / 61 / 1650 / 37.0
SS Africa / 40 / 475 / 84.2
The above picture is not quite so straightforward as it appears. The output is per lactating dairy cow. The greater yielding cows tend to have a shorter lifetime, going through fewer lactations within that lifetime, and hence the ratio of replacement animals (followers) per dairy cow is greater. Since these followers produce CH4 as they mature (but before they start producing milk) they will indirectly contribute to the GHG burden of milk production. In addition, the move toward breeds suitable only for milk production lessens the opportunity for unwanted male dairy calves to be sold to beef producers for fattening, hence indirectly increasing the GHG emissions of beef production (Webb et al., 2009). Nevertheless, when the contribution of followers is taken into account the ranking of the above regions is unlikely to change significantly; production in the NAFTA and EU regions will emit less GHG per L milk produced than production in other regions. But the difference will not be as great as that produced by the simple estimate above.
Table 2 below presents default emissions for beef production in different regions.
Table 2. Default CH4 emissions from beef production in different regions, from IPCC 2006, beef cattle
Enteric kg CH4head-1yr-1 / **meat kg head-1 yr-1 / kg CH4 kg-1 meatNAFTA / *53 / 65 - 69 / 0.8
EU / 57 / 65 - 69 / 0.9
CIT / 58 / 45 - 67 / 1.3
Oceania / 60 / 20 - 23 / 3.0
Latin America / 56 / 13 - 19 / 4.3
N Africa, mid-East / 31 / 13 - 19 / 2.4
S Asia / 27 / 15 - 22 / 1.8
E and SE Asia / 47 / 15 - 22 / 3.1
SS Africa / 31 / 13 - 19 / 2.4
*Includes beef cows, bulls, calves, growing steers/heifers, and feedlot cattle.
**Estimated from Bouwman et al. (2006)
These data also indicate that increased production of livestock products does not inevitably come at the price of increased emissions, albeit there is a general trend in that direction.
Table 3 presents average estimated N excretion for the main types of livestock in each region of the world. Emissions of N2O from manure management and following application of manures to land will be broadly in proportion to these estimates of daily N excretion. There is less variation in these estimates, which is partly due to their being expressed per day, and thus not taking account of the longer time to maturity in some regions. Such differences are found mainly with extensively-raised ruminants. Hence emissions of N2O per kg extensively-raised beef or lamb will be greater for produce raised in Africa, the Middle East and Latin America.
Table 3. Default N excretion, kg N (1000 kg animal)-1 mass day-1, from IPCC 2006
Dairy / Beef / Finishing pigs / Sheep / Buffalo / Horses / Layers / BroilersNAFTA / 0.44 / 0.31 / 0.42 / 0.42 / 0.32 / 0.30 / 0.83 / 1.10
EU / 0.48 / 0.33 / 0.51 / 0.85 / 0.32 / 0.26 / 0.96 / 1.10
CIT / 0.35 / 0.35 / 0.55 / 0.90 / 0.32 / 0.30 / 0.82 / 1.10
Oceania / 0.44 / 0.50 / 0.53 / 1.13 / 0.32 / 0.30 / 0.82 / 1.10
Latin America / 0.48 / 0.36 / 1.57 / 1.17 / 0.32 / 0.46 / 0.82 / 1.10
N Africa, mid-East / 0.70 / 0.79 / 1.57 / 1.17 / 0.32 / 0.46 / 0.82 / 1.10
Sub-saharan Africa / 0.60 / 0.63 / 1.57 / 1.17 / 0.32 / 0.46 / 0.82 / 1.10
S Asia / 0.47 / 0.47 / 0.42 / 1.17 / 0.32 / 0.46 / 0.82 / 1.10
E and SE Asia / 0.47 / 0.47 / 0.42 / 1.17 / 0.32 / 0.46 / 0.82 / 1.10
The other major source of GHG emissions from agriculture is N2O arising following application of mineral-N fertilizers. The current IPCC default emission factor (EF) is 1.0% of applied N. Although there is evidence of emissions being more or less than this default depending upon soils and subsequent weather conditions, several factors control N2O emissions and these will broadly correlate with applications of N fertilizer. Hence productions systems such as the high-input systems used in the EU, and increasingly China and parts of India, will emit much more than extensive systems in other regions. However, since yields in regions such as NW Europe are larger than in many other areas N2O emissions per kg of output will not be in proportion to overall N2O emissions.
As indicated above, GHG emissions, primarily in the form of CO2, arise when land is converted to agriculture. While soils under well-managed agricultural grassland may contain amounts of carbon similar to some natural ecosystems, soils under tillage will contain less (Guo and Gifford, 2002). However, when evaluating fluxes of CO2 to the atmosphere, it needs to be remembered that only considering changes to soil carbon takes no account of changes in above ground carbon stocks which may be greatly reduced by land use change. The current estimate of total carbon storage, both SOC and above ground, is c. 360 t/ha for Brazilian rainforest (IPCC, 2006) and hence conversion to either grassland or arable will lead to significant emissions of CO2.
1.2.Water and Air Quality
With respect to broader environmental impacts of N, in particular NO3 leaching and NH3 volatilization, Bouwman et al. (2006) report some differences among regions in the efficiency with which applied N is used, as reflected in the recovery of N as a % of inputs (nitrogen use efficiency: NUE). For most regions NUE for 1995 was reported as c. 50%, but only c. 40% for E Asia and 78 and 108% respectively for SE Asia and SS-Africa. Forecasts for 2030 were between c. 60 and 70%, but are rather less for E Asia (42%) and more for CIT (83%), SE Asia (90%) and SS-Africa (131%). The latter figure is a cause for concern. While a relatively large NUE indicates that losses of N to the environment are relatively small, some losses, especially of NO3 leaching and denitrification in humid areas, are unavoidable and even in a sustainably fertilized system NUE would be expected to be < 100%. If N removal in crops is exceeding inputs it means there is a net loss of N from the soil-crop system potentially leading to a reduction in soil fertility. The forecast increase in NUE suggests these problem of nutrient depletion will increase. In turn declining soil fertility is likely to increase pressure for clearance of forests and savannahs for agricultural production.
The small NUE for E Asia is due to the prevalence of paddy rice cultivation from which emissions of N are particularly large. Emissions of NH3 being increased by the high pH of these systems while waterlogging leads to intense denitrification, albeit the predominant loss is as molecular N (N2) which does not harm the environment. Emissions of NH3 are particularly large in China, and some CIT due to the use of the fertilizer ammonium bicarbonate (AB) (Cai et al., 1998). In their projections of fertilizer-N use to 2030, Bouwman et al. (2005) assumed 90% of the AB used in China would be substituted by urea, hence greatly reducing, but by no means eliminating, emissions of NH3.
1.3.Water resources
Views have been put forward that current patterns of water extraction are close to what can be sustained, or may be exceeding it. Some major rivers no longer reach the sea including the Indus, Rio Grande, Colorado, Murray-Darling and Yellow rivers. This is at least partly due to extraction for cereal production.
Freshwater fish populations are in decline. According to the World Wide Fund for Nature (WWF, 2003), fish stocks in lakes and rivers have fallen roughly 30% since 1970. This is a bigger population decline than that suffered by most forms of wildlife. Half the world’s wetlands, in one estimate, were drained, damaged or destroyed in the 20th century, mainly because, as the volume of fresh water in rivers falls, salt water invades the delta, changing the balance between fresh and salt water. On this evidence, there may be systemic water problems, as well as local disruptions.
1.4.Soil degradation
A UNEP survey(cited in Smeets et al., 2004) on soil degradation reported that the rate of soil erosion is 10 to 20 times the renewal rate in temperate regions and 20 to 40 times the same rate in the tropics. This results in an annual worldwide loss of cropland of between 5 and 12 million ha per year. Deforestation is thought to be responsible for 43% of the total erosion and overgrazing and mismanagement for 29% and 24% respectively (Smeets et al., 2004).
Despite this, in many parts of the world such as northwest Europe, soils have been cultivated for over two millennia and are more productive than ever. This production is largely dependent on inputs of mineral fertilizers and other agrochemicals, but the production of high yielding crops may lead to substantial returns of organic matter to soils, helping to maintain soil structure. This has led to the view being put forward that regions such as Europe, with fertile and stable soils, equitable climate and in consequence large yield potentials, have a duty to optimize food production to reduce the burden on pristine ecosystems and less resilient soils in other parts of the world.
1.5.Land use change, ecosystem services and biodiversity
Ecosystem services (ES) may be defined as 'the benefits of nature to households, communities, and economies.' (Boyd and Banzhaf, 2006). Clearly the continued provision of ES requires the natural ecosystem to remain intact, and hence changes in land use from natural ecosystems to agriculture reduces the provision of ES. It is presumed that increased food production is better achieved by increased production on existing agricultural land than by converting natural systems. The Millennium Ecosystem Assessment has categorised ecosystem services into the four general areas of support, regulation, provision, and cultural services, shown in Table 4 below. Each general area is sub-divided into increasingly detailed roles that support human society.
Table 4. Categories of ecosystem services (Millennium Ecosystem Assessment)
Support / Provision / Regulation / CulturalNutrient Cycling / Food / Climate Regulation / Aesthetic
Soil Formation / Fresh Water / Flood Regulation / Spiritual
Primary Production / Wood and Fibre / Disease Regulation / Educational
Fuel / Water Purification / Recreational
These ecosystem services depend on all of the component ecosystem species within: plants provide primary production and food for herbivores, soil invertebrates aerate the soils and recycle nutrients, bacteria and fungi decompose plant and animal litter, birds distribute seeds of plants and so on. All of these roles are interlinked among the thousands or millions of species that inhabit an ecosystem.
There is increasing evidence that high levels of biodiversity may act as “insurance” that buffer ecosystem services from environmental change. Therefore maintenance of species composition and abundance is essential to maintaining the ecosystem services upon which humans depend. Loss of biodiversity is directly associated with habitat change, climate change, invasive alien species, overexploitation and pollution (MEA 2005). The effects of this loss are particularly pronounced in areas that have been deforested or where wetlands have been drained. Key impacts here include loss of ecosystem services that act as natural breaks on flooding and erosion, as well as release of carbon stored in soils. Some of these changes are irreversible, particularly deforestation in areas that are vulnerable to desertification. These are crucial differences to the impact of deforestation in temperate and tropical regions. Unlike the deciduous forests of temperate regions, tropical rain forests are not easy to regenerate once lost.