Box 4-3. Genetically modified soybeans in Latin America.

At the global scale, soybean is one of the fastest expanding crops; in the past 30 years planted area more than doubled (FAO, 2002b). Of the world's approximately 80 million ha, more than 70% are planted in the USA, Brazil and Argentina (Grau et al., 2005). Argentina's planted area increased from less than a million ha in 1970 to more than 13 million ha in 2003 (Grau et al., 2005). Soybean cultivation is seen to represent a new and powerful force among multiple threats to biodiversity in Brazil (Fearnside, 2001). Deforestation for soybean expan­sion has, e.g., been identified as a major environmental threat in Argentina, Brazil, Bolivia and Paraguay (Fearnside, 2001; Kaimowitz and Smith, 2001). In part, area expansion has oc­curred in locations previously used for other agricultural or grazing activities, but additional transformation of native veg­etation plays a major role. New varieties of soybean, includ­ing glyphosate-resistant transgenic cultivars, are increasing yields and overriding the environmental constraints, making this a very profitable endeavor for some farmers (Kaimowitz and Smith, 2001). Although until recently, Brazil was a key glo­bal supplier of non-GM soya, planting of GM soy has been le­galized in both Brazil and Bolivia. Soybean expansion in Brazil increased; as did research on soybean agronomy, infrastruc­ture development, and policies aimed at risk-reduction during years of low production or profitability (Fearnside, 2001). In Brazil alone, about 100 million ha are considered to be suit­able for soy production. If projected acreage in Argentina, Bra­zil and Paraguay are realized, an overproduction of 150 million tonnes will be reached in 2020 (AIDE, 2005).

to consider for land use dynamics are: the perceptions and values of local stakeholders land resources, its goods and services; land tenure and property rights and regulations; the development and adoption of new sources of AKST; and urban-rural connections.

4.4.4 Climate variability and climate change
Agricultural systems are already adapting to changes in cli­mate and climate variability in many part of the world. This is in particular the case in arid areas. The IPCC concluded in its latest assessment that it is very likely that humans caused most of the warming observed during the twentieth century (IPCC, 2007a). The report also indicates that future climate change is to be expected, as a function of continuing and increasing emissions of fossil fuel combustion products, changes in land use (deforestation, change in agricultural practices), and other factors (for example, variations in so­lar radiation). Changes in climate will not only manifest themselves in changes in annual means (precipitation, tem­perature) but also in changes in variability and extremes.

4.4.1 Driving forces of climate change
A set of scenarios (IPCC Special Report on Emissions Sce­narios—SRES) was used to depict possible emission trends

under a wide range of assumptions in order to assess the potential global impact of climate change (IPCC, 2000). Subsequent calculation showed that these scenarios result­ed in atmospheric concentrations of CO2 of 540-970 parts per million in 2100 compared with around 370 parts per million in 2000. This range of projected concentrations is primarily due to differences among the emissions scenarios. Model projections of the emissions of other greenhouse gas-ses (primarily CH4 and N2O) also vary considerably by 2100 across the IPCC-SRES emissions scenarios. The IPCC sce­narios are roughly consistent with current literature—with the majority of the scenarios leading to 2100 emissions of around 10-22 Gt C (Van Vuuren and O'Neill, 2006) (Figure 4-21) with projections by the IEA-2006 World Energy Out­look in the middle of this range. The IPCC-SRES scenarios do not explicitly include climate policies. Stabilization sce­narios explore the type of action required to stabilize atmo­spheric greenhouse gas concentrations (alternative climate policy scenarios may look into the impact of a particular set of measures; or choose to peak concentrations). Ranges of stabilization scenarios giving rise to different stabilization levels are compared to development without climate policy (Figure 4-22) (IPCC, 2007c). The ranges in emission path­ways result from uncertainty in land use emissions, other baseline emissions and timing in reduction rates.

4.4.4.2 Projections of climate change IPCC calculations show that different scenarios without climate policy are expected to lead to considerable climate change: the global mean surface air temperature is expected to increase from 1990 to 2100 for the range of IPCC-SRES scenarios by 1.4 to 6.4 C° (IPCC, 2007a) (Figure 4-23). The total range given above is partly a consequence of differences in emissions, but also partly an impact of uncertainty in cli­mate sensitivity, i.e., the relationship between greenhouse gas concentration and the increase in global mean temperature (after equilibrium is reached).  Over the last few years, there has been a shift towards expressing the temperature consequences of stabilization scenarios more in terms of probabilistic expressions than single values and/or ranges. A 50% probability level for staying below 2°C corresponds approximately to 450 ppm CO -eq, while for 2.5°C the cor­responding concentration is around 525 ppm CO2-eq. Simi­larly, a scenario that would lead to 2°C warming as the most likely outcome could also lead to a 0.9 to 3.9°C warming (95% certainty). Handling uncertainty therefore represents an important aspect of future climate change policy. Costs of stabilization increase for lower concentration levels, and very low concentration levels, such as 450 ppm CO2-eq may be difficult to reach (IPCC, 2007c).
     Combining the current scenarios with climate policy and the expected temperature increase for different green­house gas levels shows that the former would decrease the lower bound of the expected temperature increase to about 0.5-1.0 °C above the 1990 level (i.e., based on an insensitive climate system and using a strong climate policy scenario) (Van Vuuren et al., 2006). This implies that although these values may be uncertain, climate change is very likely. High rates of temperature change are in fact most likely to occur in the first half of the century as a result of climate sys­tem inertia, the limited impacts of climate policy and lower